CEN/TC 250/SC 2 N2087

CEN/TC 250/SC 2/WG 1 N 1314

Claus V Nielsen, DTI, 22 Feb. 2020
CEN/TC 250/SC2/WG1/TG10

When comparing the carbonation rates observed after 25 years for the outdoor conditions (XC4 in
Fig. 2.4.4) it is seen that they are significantly lower than the reference specimens stored indoor.
They are in fact much lower than the ratios applied in the model calculations in this report, cf. Table
2.4.3.

Fig. 2.4.4: Carbonation rates (mm/year½) observed for concretes stored indoor for 10, 20 and 25
years. Furthermore, the carbonation rates measured for outdoor exposure for 25 years are given for
the same concretes. The w/c ratios and the 28-days strengths are indicated in the diagram.

Fig. 2.4.5 contains the results for the higher quality concretes. Mixes 6-11 are used for many outdoor
applications where carbonation takes place and chlorides are often present (XD1, XS1). Mixes 12-16
are considered to be high-quality concretes for civil structures with harsh exposure conditions (e.g.
bridges). None of the specimens in Fig. 2.4.5 experienced full depth carbonation during the 25-years
period.

One observation from the diagram is that the carbonation rates calculated at 10, 20 and 25 years are
fairly identical (about 10% variation) for most of the concretes, which confirms the square-root time
law. The observations of outdoor (XC4) exposed specimens for 25 years shows to be significantly
lower than the reference (indoor) specimens. For many of the high-quality concretes, only 1 mm car-
bonation depths were measured after 25 years of outdoor exposure, which is in contrast to say 25
mm carbonation depth measured on the same specimens stored inside.

It is also obvious that the differences between the various concretes’ carbonation resistance become
rather small for w/c ratios around 0.40, despite different strengths, w/c and binder compositions.
The average carbonation rates measured for the best concrete qualities lie from 4.5 to 5.5
mm/year½, which is probably corresponding to resistance class XRC7.

Section 2-4_Simplified calculation_carbonation_cln_2021-02-26.docx

For citations, copying and referencing see Note on Front page! page 157
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, 22 Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Fig. 2.4.5: Carbonation rates (mm/year½) observed for concretes stored indoor for 10, 20 and 25
years. Furthermore, the carbonation rates measured for outdoor exposure for 25 years are given for
the same concretes. The w/c ratios and the 28-days strengths are indicated in the diagrams for each
concrete.

Norwegian field data

A large Norwegian field investigation [Helland 2016] has been reported where existing concrete have
been sampled from different locations and the carbonation depth has been measured. Fig. 2.4.6
gives the results in terms of the carbonation rate (carbonation depth divided with the exposure age
at the time of sampling). In 2000 samples were taken from concrete structures based on CEM I and
in some cases with added silica fume (open symbols in Fig. 2.4.6). In 2013 a new investigation was
made on concretes made with blended fly ash cement (CEM II/A-V) with about 19% FA.

Helland (2016) converts the measured carbonation depths from the actual w/c ratio of the actual
concrete to a reference w/c = 0.60 by means of a Norwegian conversion factor [Skjoldsvold 1986].
However, such a conversion is not applied in Fig. 2.4.6 as it does not provide any increase in the pre-
cision because of the large scatter experienced on these measurements. The results in Fig. 2.4.6
show a slight tendency towards a higher carbonation rate for exposure class XC3 (sheltered) but the
number of observations is too few to show statistical significance. The coefficient of variations range
between 30 to 40% on the 2000 investigation and approximately 58% on the 2013 investigation.

In the 2013 investigation a mean value of 2.7 mm/year½ is found and in the 2000 investigation the
mean value for concretes containing pure CEM I is 1.6 mm/year½ and 2.1 mm/year½ for the con-
cretes with CEM I and SF.

Section 2-4_Simplified calculation_carbonation_cln_2021-02-26.docx

For citations, copying and referencing see Note on Front page! page 158
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, 22 Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Fig. 2.4.6: Carbonation rates (mm/year½) measured on cores sampled from different outdoor struc-
tures having w/c ratio around 0.60. The typical age oat the time of sampling was 10 years.

It is expected that the concretes represented in Fig. 2.4.6 are similar in terms of performance with
those depicted in Fig. 2.4.4 even though it is recognised that the cement types are not exactly identi-
cal. If the Norwegian data are used as input for the stochastic modelling, then the cumulative distri-
bution of the initiation phase will look like Fig. 2.4.7. The Norwegian requirements call for minimum
cover thickness of 25 mm for exposure class XC4, which is applied, together with an execution toler-
ance of cdev = 10 mm.

If we consider only the initiation phase it is seen that concrete based on fly ash cement only reaches
tini = 40 years if the safety index β=1.5 is to be fulfilled. For the pure Portland cement (CEM I) the ini-
tiation phase easily reaches several hundred years under the same assumptions and when silica
fume is added tini = 85 years is reached. Moreover, since the Norwegian field observations are based
on the square-root time law, which is generally more conservative than the models applied in the
present document (Table 2.4.3) the model predictions in Fig. 2.4.7 will also be conservative.

However, if the execution tolerance is reduced from cdev = 10 mm to say 5 mm (higher precision on
the building site) then the predictions will be less conservative. This is illustrated by comparing Figs.
2.4.7 and 2.4.8. The initiation phase is reduced significantly when the execution tolerance on the re-
inforcement is improved.

Section 2-4_Simplified calculation_carbonation_cln_2021-02-26.docx

For citations, copying and referencing see Note on Front page! page 159
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, 22 Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Fig. 2.4.7: Cumulative probabilities for the initiation phase calculated on basis on the measured data
from Fig. 2.4.6 and minimum cover thickness 25 mm. It is assumed that the square-root time law is
applicable (b = 0.5). The red circles indicate the failure probability of 7% (β=1.5) which defines the
limit state.

Fig. 2.4.8: Similar to Fig. 2.4.7. Only difference is the assumed execution tolerance on the rebars.

Section 2-4_Simplified calculation_carbonation_cln_2021-02-26.docx

For citations, copying and referencing see Note on Front page! page 160
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, 22 Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Other Nordic data on the carbonation rate

During the period from 2003 to 2005 a Nordic collaboration was carried out investigating the CO2
uptake potential for concrete both during its service life and afterwards when the concrete is demol-
ished and crushed. A state-of-the-art report was prepared [Lagerblad 2005] gathering information
on the carbonation rates to expect at different exposure conditions and different concrete qualities
(Table 2.4.6). The results from this investigation were mainly focusing on Nordic outdoor exposure
conditions (wet, humid, cold). The main results are now included in the standard EN 16757 which
constitutes the rules for including the CO2 uptake into environmental product declarations to be
used for life-cycle-assessments. Table 2.4.6 contains the carbonation rates for CEM I concretes. In
case of blended cements or additions of limestone filler, fly ash, silica fume or GGBS the rates are
increased by 5 to 10 % for substitution up to 30% and for slag cement the carbonation rate is in-
creased up to 30%.

The figures in Table 2.4.6 for outdoor exposed to rain (XC4) are in line with the Norwegian findings
described previously. However, for the XC3 exposure conditions the numbers seem to be higher
than what is observed in the field.

Table 2.4.6: Carbonation rates for different exposure conditions and concrete qualities according to
Lagerblad (2005). Based on CEM I concrete types. The values included in EN 16757 is shown in paren-
theses.
Carbonation rates, mm/year½
Strength class,
< C15 C15 – C20 C25 – C35 > C35
cylinder MPa
Wet/submerged
2.0 (-) 1.0 (-) 0.75 (-) 0.5 (-)
(XC2)
Buried in ground 3.0 (-) 1.5 (1.1) 1.0 (0.8) 0.75 (0.5)
Outdoor exposed
5.0 (5.5) 2.5 (2.7) 1.5 (1.6) 1.0 (1.1)
to rain (XC4)
Outdoor shel-
10 (11) 6 (6.6) 4 (4.6) 2.5 (2.7)
tered (XC3)
Indoors dry (XC1) 15 (16.5) 9 (9.9) 6 (6.6) 3.5 (3.8)

An extensive Finnish investigation on existing concrete facades and balconies made from precast
concrete elements is reported in Lahdensivu (2012). The investigation links the degradation and the
reinforcement corrosion with investigations of the concrete quality, the cover thicknesses and the
carbonation and chloride content. The results from this investigation have also been utilised in a re-
cent article on the role of carbonation on corrosion [Angst et al. 2020].

Carbonation rates were measured from cored samples by means of the square-root time law. Sev-
eral thousand samples were taken from buildings from the 1960’ies up to the 1990ies. Most of the
samples were unprotected from the rain (XC4) and the carbonation rates are of the same order of
magnitude as described Helland (2016), ranging from 2 to 3 mm/year½ (average) having coefficients
of variation from 40 to 60% and in some cases up to almost 100%. Hence, very large variation in the
results exist.

Section 2-4_Simplified calculation_carbonation_cln_2021-02-26.docx

For citations, copying and referencing see Note on Front page! page 161
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, 22 Feb. 2020
CEN/TC 250/SC2/WG1/TG10

The various field observations seem to confirm the carbonation rates that is typically applied in mod-
elling. Furthermore, they seem to confirm the validity of the simple square-root time law. However,
it is acknowledged that the observations are mainly related to the Northern European wet and hu-
mid climate conditions.

2.4.5 References
Angst, U., et al.: “Corrosion of steel in carbonated concrete: mechanisms, practical experience and
research priorities – a critical review by RILEM TC 281-CCC”, RILEM Technical Letters, Vol. 5, pp. 85-
100, 2020

Bager, D.: “Aalborg Portland durability project – 25 years judgement”, Nordic Concrete Research
Mini-Seminar on Exposure Sites, 2008.

EN 16757, Sustainability of construction products – Environmental product declarations – Product

category rules for concrete and concrete elements, CEN, 2017.

fib: Model Code for Service Life Design, Bulletin no. 34, International Federation of Structural Con-
crete, Lausanne, Switzerland, 2006.

Greve-Dierfeld, S.v., et al.: ”Understanding the carbonation of concrete with supplementary ce-
mentitious materials: a critical review by RILEM TC 281-CCC”, RILEM Materials and Structures, Vol.
53, No. 136, 2020.

Helland, S.: “Assessment of carbonation ingress in-field as a tool to calibrate code requirements”,
Proceedings fib Symposium, Cape Town, SA, Nov. 2016.

Lagerblad, B. Carbon dioxide uptake during concrete life cycle, State of the art, CBI Report 2, 2005

Lahdensivu, J.: Durability properties and actual deterioration of Finnish concrete facades and balco-
nies, Tampere University of Technology, Report 1028, 2012

Skjolsvold, O.: “Carbonation depths of concrete with and without condensed silica fume”, Proceed-
ings: Fly ash, silica fume and natural pozzolans in concrete, ACI SP-91, pp. 1031-1048, 1986.

Section 2-4_Simplified calculation_carbonation_cln_2021-02-26.docx

For citations, copying and referencing see Note on Front page! page 162
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

XRC Model comparison and analyses

S. v. Greve-Dierfeld, F. Toutlemonde, C.V. Nielsson Draft 2 18.10.2020
update 24.11.2020 update 18.2.2021, 28.2.2021, 30.04.2021, 18.05.2021, 28.05.2021, 31.05.2021

2.5 Model comparison

The models applied for the carbonation of concrete are different but give similar results. All models
describe the progress of the carbonation front as a time-power law multiplied by a term describing
the effect of the environment (relative humidity). Moreover, the (admissible) duration of the
propagation phase is deduced from the design service life when considering the carbonation front
progress for deriving the cover depth. The following numbers and differences in the model input
parameters are shown in table 2.5.1.

In Table 2.5.1 fenv and ke are different abbreviations for the effect of environmental condition (relative
humidity) on carbonation rate. The time exponent b describes the progress of carbonation depth with
time. The time exponent can be expresse by b=0.5-w. CO2 is the ambient CO2 concentration and tprop
is the propagation phase.

Table 2.5.1 Summary of model input. These numbers are values taken from subsection 2.1-2.4 and for
more details it is referred to section 2.1-2.4
parameter exposure Model 2.1 Model 2.2 Model 2.3 Model 2.4
class and
related
mean RH
fenv, ke XC2≥ 85% 0.30 0.40 0.50 0.5
RH
XC3: 75% RH 0.80 0.75 0.87 0.87
XC3: 65% RH 1 1 - -
XC4: 75% RH 0.70 0.75 0.83 0.80

XC4: 65% RH 0.85 1 - -

time XC2 ≥ 85% b=0.25 w=0.4 b=0.40 b=0.40

exponent b, RH
b=0.5-w XC3: 75% RH b=0.40 w=0 b=0.45 b=0.45
XC3: 65% RH b=0.45 w=0 - -
XC4: 75% RH b=0.35 w=0.24 b=0.40 b=0.45
XC4: 65% RH b=0.40 w=0.15 - -
CO2 XC2: ≥ 85% 480 From XRC definition (400)
concentration RH
[ppm] XC3: 75% RH 470
XC3: 65% RH 425
XC4: 75% RH 470
XC4: 65% RH 425
tprop, (if XC2: ≥ 85% 5 9 (10) 5
relevant, RH*
mean values XC3: 75% RH 25 13 (20) 10
for xcorr = 50 XC3: 65% RH 45.5 - -
mm) XC4: 75% RH 5 4 (5) 5
XC4: 65% RH 12.5 - -

For citations, copying and referencing see Note on Front page! page 163
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

* In Chapter 2.1 XC2 is calculated with carbonation at RH = 90% (for the case no leaching is considered), for the case leaching
is considered it is referred to chapter 2.1; in chapter 2.2 XC2 is calculated with carbonation at RH = 85%

Figure 2.5.1a) and b) show the fraction by which each result of model reduces the concrete cover in
the exposure classes compared to XRC reference conditions. In Figure 1 a) and b) 1 means no influence
on carbonation depth and 0.5 means the carbonation depth is half of the initial value). All influences
combined are shown in figure 1 a) and b) right. No safety margins nor safety factors have been applied
on the model parameters in the figures 2.5.1.

For calibration of the input values, model 2.1 used data from exposure tests exposed to a characteristic
average humidity area in Switzerland and a dry region in Switzerland. In the chapter 2.2 the model of
[Izquerdo 2002] was applied with data from Spain and fib MC for SLD 2006]. Model 2.3 used data from
experience of site structures in France and Model 2.4 used boundaries derived from the model
calculations of Model 2.1, 2.2 and 2.3.

Figure 2.5.1: Influence of each model parameter on the concrete cover (carbonation depth at initiation
time)

Figure 2.5.2 shows average values of carbonation depths in dependency of average values of
carbonation rates considering the input parameters summarized in table 2.5.1. Note: Relation to the
XRC definition depends on the probabilistic assumptions, as detailed in subsections 1.2 and in chapter
2.6.

For citations, copying and referencing see Note on Front page! page 164
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Figure 2.5.2: comparison of mean carbonation rate and mean carbonation depth at initiation time

The differences in the models, especially the time exponent, explain some differences in the relation
between XRC designation and mean carbonation rates, due to different allocations of average
carbonation rate after 50 year and average carbonation rate after 1 year. For example, chapter 2.3
the mean carbonation rate after 50 year is translated to the mean value of carbonation rate at 1 year
assuming a 0.45 times power law.

For citations, copying and referencing see Note on Front page! page 165
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

2.6 XRC comparison

Table 2.6.1 displays the recommended minimum cover depth values which are selected for the
Eurocode 2 revision. The values depend on the different exposure classes with respect to carbonation
and on the required design service life (50 or 100 years).

In table 2.6.1 the given concrete cover depth refers to cmin,dur. The correlation between the mean value
of concrete cover and cmin,dur is different for each model. This is explained in chapter 1.2. with further
details in chapters 2.1-2.4. The XRC designation is the carbonation rate derived from the 90% quantile
of carbonation depth under XRC reference conditions. The CoV of carbonation rate expected under
the XRC reference condition is taken different in each model and includes effects of the variability of
the material and of the curing and execution. CoV is in the lower range and was used in the range of
CoV = 25-50%). As detailed in Subsections 2.2 and 2.4 each model has used different CoV of placed
concrete and different definitions of cmin,dur.

As an example, for model 2.1 the carbonation rate is normal distributed. As an example, the mean
value of carbonation rate k of a placed concrete in XRC reference conditions after 50 year of exposure
is km = 3.0 mm/year0.5 and assuming carbonation rate is normal distributed and the CoV of the placed
concrete in XRC reference conditions is CoV = 50% (upper value) the XRC designation is XRC5 and
calculated as follows:
Ø 3.0 mm/year0.5 + 1.282*0.5*3.0= 5 mm/year0.5 → km=3.0 = XRC5
Note: the CoV is assumed under constant relative humidity and moisture conditions including the
effect of execution and production

In table 2.6.1 XRC is allocated to the minimum concrete cover cmin,dur

As an example, in Model 2.1 and 2.4 cmin,dur corresponds to the 5% quantile of the average value of
concrete cover (carbonation depth after 50 year). It is calculated as follows:
Ø Cmin,dur = xc(tSL) – 1.645 * standard deviation of concrete cover c = xc(tSL) – 10 mm

As another example in model 2.3 the XRC designation associated to the characteristic carbonation
depth not exceeded with a 90% probability in the reference situation at 50 years, helps deriving the
mean value of carbonation rate at 1 year assuming a 0.45 times power law for the carbonation front
in the reference conditions and a 25% CoV (lower value). With these assumptions:
- XRC5 corresponds to kµ = 4.6 mm/year0,5

Derivation of cmin,dur to ensure meeting a target reliability index of 1.5 is detailed in Subsection 2.3

To compare the calculated (and rounded) minimum concrete cover depths from sections 2.1-2.4, the
diagrams in Figs. 2.6.1 and 2.6.2 show the recommended minimum cover depths and the different
model approaches taken from sections 2.1-2.4.

The comparison in Figs. 2.6.1 and 2.6.2 are for service lives of 50 and 100 years, respectively. Only XC2
to XC4 are depicted because in general, the durability modelling for XC1 only causes very limited
provisions for the minimum cover depth as seen in Table 2.6.1.

For citations, copying and referencing see Note on Front page! page 166
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Table 2.6.1: Minimum cover depths cmin,dur recommended in prEN 1992-1-1 (D7), clause 6.5.2.2 (Table
6.3(NDP).

Exposure class XC1 XC2 XC3 XC4

Design service life (years) 50 100 50 100 50 100 50 100
XRC 0,5 10 10 10 10 10 10 10 10
XRC 1 10 10 10 10 10 15 10 15
XRC 2 10 15 10 15 15 25 15 25
XRC 3 10 15 15 20 20 30 20 30
XRC 4 10 20 15 25 25 35 25 40
XRC 5 15 25 20 30 25 45 30 45
XRC 6 15 25 25 35 35 55 40 55
XRC 7 15 30 25 40 40 60 45 60
Accounting for the different possible interpretations in the target reliability it was agreed to select
for the Eurocode revision the higher recommended values, which are displayed in Table 2.6.1. This is
clearly seen from the data in Figs. 2.6.1.

Figure 2.6.1: Comparison of calculated minimum concrete cover depths (rounded values) as a
function of the XRC classes for b = 1.5. The dashed lines are the recommended values given in Table
2.6.1.

For citations, copying and referencing see Note on Front page! page 167
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Chapter 3.1: Chloride

Stefanie v. Greve-Dierfeld Draft 4 29.02.2021

3.1 Full probabilistic approach I (SvGD)

3.1.1 Introduction
3.1.2 Limit state function and models
3.1.3 Model parameter
3.1.4 Parameter study
3.1.5 Minimum cover calibration
3.1.6 Results
3.1.7 Verification with tested concretes
3.1.8 References

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 168
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

3.1.1 Introduction
The transport of chloride ions is for simplicity modelled with the ERFC solution to Fick`s second law of
unidirectional diffusion under the assumption of infinity by delivering a total chloride profile at time t.
Permeation, capillary suction in partially saturated concretes may considerably increase the ingress
rate at the surface.
Because some of the chlorides are bound only part of the chlorides are mobile and affect passivation
performance and corrosion kinetic. Part of the bound and unbound chlorides originate from the raw
materials. The type of compound affects binding capacity (CaCl, MgCl > NaCl). Chloride binding capacity
in turn affects the maximum and total chloride content, which is often higher than what can be
expected from equilibrium and higher than during tests where NaCl is the main source of chlorides.
Carbonation or leaching can liberate bound chlorides that can be washed off or redistribute. This can
results in a peak or increases of total amount of chlorides behind the concrete surface and / or
decreases the amount of total chlorides at the concrete surface.
In addition to that, ongoing hydration, and other sources of alteration of the pore structure, dry out
and temperature variation in turn affects the effective chloride diffusion coefficient with time.
The calculations are performed with an apparent diffusion coefficient at time t determined according
to Figs. 1 considering a chloride peak concentration.

10
CEM I + 4% SF w/z=0.35
XRDS reference conditions
XS1
Dapp(t) [10-12 m2/s]

XS3

1 fenv
fenv a
aXRDS
a

0.1
0.1 1 10
time [year]

Figure 3.1.1 apparent diffusion coefficient and a (a) procedure to design with XRDS based on the
apparent diffusion coefficient after 50 years under XRDS reference conditions and (b) apparent
diffusion coefficients determined in [Costa et al. 2006]

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 169
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

3.1.2 Limit state function and models

3.1.2.1 XRDS-Classes

The designation of XRDS classes for resistance against corrosion induced by chloride ingress is derived
from the depth of chlorides (XRDS) penetration [mm] (characteristic value 90 % fractile),
corresponding to a reference chlorides concentration (0,6 % by mass of binder cement + type II
additions), assumed to be obtained after 50 years on a concrete exposed to one-sided penetration of
reference seawater (30 g/l NaCl=3% NaCl-solution) at 20 °C. XRDS has the dimension of a diffusion
coefficient [10-13 m²/s].

3.1.2.2 Limit state

The limit state function used to calculate the concrete cover required for a target reliability level of b
= 1.5 is given with Eq. (1).

{ ( , )− <0= − < 0} < ℎ = − Eq. (1)

p is the failure probability [-]

a is the concrete cover [m]
ccrit is the corrosion inducing chloride threshold level [wt.%/binder]
c(a,tini) is the chloride content at the depth of reinforcement at initiation time [wt.%/binder]
ptarget is the target failure probability 7% related to the reliability index b = 1.5 [-]
tini is the initiation time according to Eq. (2) [year]
tSL design service life [year]
xcor is the uniform diameter reduction [mm]
vcorr is the uniform corrosion rate [mm/year]
a concrete cover [mm]
xcrit depth with critical chloride content [mm]

The semi probabilistic limit state equation is given with Eq. (2).
≥ ∙ , , Eq. (2)
gf partial factor for the material variable under curing, execution, and moisture conditions [-]
xcrit,k Characteristic value of the depth with corrosion inducing chloride content [mm]
cmin is the design value of concrete cover [mm]
The design value of concrete cover is calculated with Eq. (3).
= −∆ Eq. (3)
Dcdev deviation from nominal cover [mm] and simultaneously the partial factor for the geometry
variable [mm] it depends on the execution class EC
cnom characteristic value of concrete cover [mm]
tini,d initiation time [year]
The design value of initiation time is calculated with Eq. (4)
,
, = − , = − ∙
Eq. (4)
,

gcor is the partial factor for the corrosion rate

tcor,d design value of corrosion propagation time [year] prior to crack initiation
tSL end of design working life or target design working life (Decision N140: tSL = 50 and tSL = 100
year) [year]
Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 170
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

xcor,k characteristic value of corrosion loss [mm] prior corrosion induced cracking
vcor,k characteristic value of corrosion rate [µm/year]

Table 3.1.1 summarizes possible partial factors and the reliability level expected if (i) the design
equation is used without corrosion phase bini and (ii) if the design equation is used for corrosion
initiation and propagation phase bini+corr.
Table 3.1.1: Selected partial factors for the additional parameter study

D c [mm] gf [-] (i) gcor [-] (i) (ii) bini [-] (i) bini+cor [-] (i) (ii)
acc. to 1.1 (1.0-1.4) 1.0 ≥ 0.5
Tolerance 1.6 (1.2-1.9) 1.0 (< 1.1) ≥ 1.5
class
For the calibration of partial factors the following parameter study has been applied as summarized in
Table 3.1.2.
Table 3.1.2: Parameter study applied to derive partial factors

Parameter Unit Mean value Standard deviation Distribution function

DXRDS 10-13 m2/s 0.5-10 CoV = 35-70 Normal
a - 0.0-0.65 CoV = 30 - 50 Normal and beta a=0, b=0.5
fenv - 0.5-1.0 CoV = 5-15 Normal
fT - 0.56-0.80 CoV = 5-20 Normal
Cs Wt.- 1.5-4 CoV=25-75 Lognormal
Ccrit %/binder 0.4-1.0 CoV=20 Beta a=0.2, b=2.0
C0 0.04-0.2 - constant
Dx mm 0-15 CoV=50 Beta a=0, b=50
c mm 5 - 100 3-9 Normal and Lognormal
xcorr mm 50 CoV=20-40 constant and normal
vcorr mm/year 5 - 80 CoV = 60-130 lognormal and normal

Figure 3.1.2 shows a (left) summary of sensitivity factors and figure 3.1.2 (right) shows ranges of
required partial factors in dependency of target reliability.

Figure 3.1.2: sensitivity (left. DAfStb623 and additional according to table 2) and ranges of partial
factors gf (right) for b = 1.5

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 171
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

3.1.2.3 fib / ISO 16204

The fib model used to calculate the chloride content at the depth of reinforcement at initiation time
is shown in Eq. (7-8).
⎡ ⎤
∙ ⎢⎢1 − ⎥
∆
( , )= ,∆ ⎥ Eq. (5)
⎢ ∙ ∙ ( ) ∙ ⎥
⎣ ⎦
erf error function [-]
Cs,Dx maximum chloride content after 50 year of exposure [wt.%/binder]
a concrete cover [m]
t0 reference time t0=0.0767 [year]
Dx depth with corresponding (maximum) chloride concentration Cs,Dx [m]
MRCM is the chloride migration coefficient [m2/s] at time t0 = 0.0767
ke is the effect of temperature [-]
a is the age exponent [-]

= − Eq. (6)

be temperature coefficient [K] ND(4800/700)

Tref reference Temperature [K]
Treal temperature on the structural element or the ambient air [K]

3.1.2.3 XRDS-model for placed concrete

Because, the designation of XRDS classes for resistance against corrosion induced by chloride ingress
is derived from the depth of chlorides (XRDS) penetration [mm] (characteristic value 90 % fractile),
corresponding to a reference chlorides concentration (0,6 % by mass of binder cement + type II
additions), assumed to be obtained after 50 years on a concrete exposed to one-sided penetration of
reference seawater (30 g/l NaCl=3% NaCl-solution) at 20 °C. XRDS has the dimension of a diffusion
coefficient [10-13 m²/s], the model has to be modified for placed concrete. The following modifications
have been made:

Ø The model is based on the diffusion coefficient under XRDS conditions (i.e. curing class 2)
DXRDS. DXRDS is related to the apparent diffusion coefficient at time tini by the ageing factor a
which depends on the exposure classes and the conditions like time of first exposure to
chlorides, moisture conditions during exposure, possibility of further hydration and maturing.
Ø An additional factor is included, that accounts for e.g. the diffusion in partially saturated
conditions (fenv) different from XRDS definition.
Ø An initial chloride content is included.
Ø The effect of temperature on the apparent diffusion coefficient is considered from
experimental results.

The model used to calculate the chloride content at the depth of reinforcement at initiation time is
shown in Eq. (7-9).

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 172
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

∆
( , )= + ,∆ − ∙ Eq. (7)
∙ ∙ ∙ , ∙

( ) = ∆ +2∙ ∙ ∙ ( ) ∙ ∙ Eq. (8)

,∆

( )= , · 50 ∙ Eq. (9)

erfc complementary error function [-]

Cs,Dx maximum chloride content after 50 year of exposure [wt.%/binder]
a concrete cover [m]
Dx depth with corresponding (maximum) chloride concentration [m]
Ci is the initial chloride content i.e. from raw material [wt.%/binder]
Dapp,tini is the apparent chloride diffusion coefficient at reference time t0 equals 1 year [m2/s]
DXRDS is the mean value of XRDS designation in [m2/s]
fT is the effect of temperature (different from 20°C) on ionic transport [-]
fenv is the effect of partially saturated on apparent chloride diffusion coefficient [-]
a age exponent [-] under exposure
aXRDS age exponent [-] under XRDS reference conditions

3.1.3 Model parameter

In the full probabilistic design, each parameter is applied as random variable. The random variable is
described with distribution function and distribution parameter.
In the semi probabilistic design described in this chapter, each parameter is applied by its mean values
(= characteristic values). The characteristic carbonation depth and the mean value of concrete cover
are increased by the partial factors. The partial factors for chloride depth gf, the partial factor for
concrete cover Dc and the partial factor for the corrosion rate gcor are given in chapter 3.1.2.2.
concrete cover a:
Usually, the nominal value of concrete cover is smaller or in the range of the mean value of concrete
cover [DuraCrete 2000, JCSS 2000, Greve-Dierfeld 2015]. The concrete cover may be described normal
distributed (conservative) [JCSS2000, fib bulletin 34]. Considering, that negative values are almost
impossible the concrete cover can be described with a Lognormal, a student or a Neville distribution
[RILEM, DBV, JCSS] (less conservative). The variability of this is affected by the evenness the structural
element is cast against, surface treatment measures and the precision of rebar placing as well as the
maximum grain size of concrete. To ensure the possibility of quality control, the minimum concrete
cover is defined as the 5%-quantile of the concrete cover. In turn a lower execution quality should lead
to a higher specified nominal value of concrete cover and a higher execution quality will lead to a lower
specified nominal value of the concrete cover. While the minimum concrete cover (5%-quantile as
characteristic value) can be ensured. Table two shows the selected ranges of differences between
mean value of concrete cover and 5% quantile of concrete cover (minimum concrete cover). Higher
standard deviations will lead to higher mean values of concrete cover, cp. Table 3.1.3.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 173
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Table 3.1.3 Selected ranges of 5 % quantiles of concrete cover (cmin), selected mean values of concrete
cover cm and corresponding standard deviations and partial factors Dc
5% quantile of concrete Mean value of Distribution Dc [mm] Standard
cover = cmin [mm] concrete cover cm function deviation s [mm]
[mm]
15 - 65 5 3
10 - 60 20 - 70 Normal 10 6
5 - 55 15 9

Corrosion rate vcorr

The corrosion rate of ordinary black steel in concrete induced by chlorides usually starts (regarding
significant corrosion rates) with pitting corrosion and is mainly governed by macro-cell activity, which
affects corrosion rate until the anode to cathode ratio reaches about 15 to 20 [Alonso et al. 1995].
Primary factors controlling corrosion are the conductivity of the electrolyte, the porosity and the
chemical composition of the liquid surrounding the steel and increases with chloride content. Other
influences are e.g. the O2 supply and the temperature as it controls solubility and mobility and the steel
quality [Tuutti 1980, Nürnberger 1996, Müller et al. 2010, Schiegg et al. 2017 and 2018, Contectvet
2000]. Table 4 summarizes corrosion rates determined from different sources.
Table 3.1.4 corrosion rate vcor from [Andrade et al. 2018, Contenctvet 2000, DuraCrete 2000, Schiegg
et al. 2017&2018, Müller et al. 2010, Tuutti 1980, Nürnberger 1996]
Exposure Distribution Mean value [source] single values [source] CoV [%]
class function [mm/year]
XD1 4 1-8 [Contectvet2000, Hunkeler et al. 30
2006]
XD2 Lognormal 30 1-15 [Contectvet2000, Hunkeler et 90-150
or Weibull al. 2006]
XD3 max 30 5 – 60 [Schiegg et al. 2018, Greve- 50-130
Dierfeld et al. 2010, Contectvet2000,
Hunkeler et al. 2006]
XS1 30 [DuraCrete2000] 5-60 [Contectvet2000, Tuutti1985] 50-130
XS2 0-4 [DuraCrete2000] 0-12 [Contectvet2000, 30-50
DuraCrete2000; Tuutti1982]
XS3 70 [DuraCrete2000] 10-120 [Contectvet2000, Greve- 80-100
Dierfeld et al. 2010]

Tolerable corrosion loss xcorr

The end of design service life is defined as the time until a certain number of chlorides have been
reached the reinforcement (initiation time) plus a certain time of corrosion (propagation period). The
tolerable rebar diameter reduction xcrit was found to increase with increasing concrete cover,
decreasing rebar diameter and increasing porosity. [Tuutti1992] found zero crack width in the case of
average corrosion induced diameter reduction in the anodic area in the range of 50-100 mm for
concrete cover 10 to 30 mm respectively. This corrosion loss xcrit equals a pit depth of 0.5 – 1.0 mm
for a pitting factor of ap=9 - 10 [DuraCrete2000] or it equals a pit depth of 0.15 mm to 0.6 mm for
pitting factors found in 3 – 6 [Tuutti1982]). Table 3.1.5 summarizes the selected tolerable corrosion
loss.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 174
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Table 3.1.5 average tolerable diameter reduction in the anodic area

Exposure class Distribution Xcorr Mean value m [mm] (=characteristic value) CoV [%]
function
all constant 50 - 100 -

Chloride threshold value initiating reinforcement corrosion Ccrit

Figure 3 summarizes influences on the critical chloride threshold level [DuraCrete2000, Angst et al.
2019, Pettersen 1995]. Additionally, the total critical threshold level increases with chloride binding
capacity, c.p. chapter 2.1.1 and in turn with higher initial chloride content. Additionally, the threshold
level is affected by the rebar quality and production process (e.g. prickled, geometry, hot/cold rolled)
[e.g. Bisschop et al. 2017]. Generally, the chloride inducing threshold level (Ccrit) decreases with
decreasing OH- concentration [e.g. Breit 2006] and therefore with increasing degree of carbonation or
Ca-leaching. Other species are likely to affect the passivation performance.

Figure 3.1.3: Chloride threshold values in dependency of relative humidity according to

DuraCrete2000

Table 3.1.6 summarizes critical chloride contents determined from literature allocated to specific
exposure classes.

Table 3.1.6 chloride inducing threshold levels (Ccrit in wt.%/cement), the proposed values from fib 2006
and DuraCrete2000 are not distinguished with respect to the exposure class
Expo- Distribution Mean value m Single values from different sources CoV
sure function [wt.%/cement] [%]
class
XD1 -
XD2 Normal, - submerged samples 0.8-2.3 [Bisschop et al. 2016]
XD3 Lognormal, 0.60 [fib2006], 0.68 [Schiegg et al. 2017],
Beta a=0.2 0.48 [DuraCrete 0.4-1.0 [Greve-Dierfeld et al. 2017] 20 -
XS1 and b=2.0 2000] 1.5 - 2.2 [Frederiksen 2000] 30
XS2 [fib 2006] 0.8 - 2.8 [Frederiksen 2000, Sörensen et al. 2018]
XS3 0.3 - 2.0 [Frederiksen 2000, Sörensen et al. 2018]
0.77 [Markeset 2007]

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 175
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Maximum chloride concentration Cs

The maximum chloride concentration depends on for example on the distance to the chloride source,
chloride concentration of the ambient solution, the chloride binding capacity (increases with the
cement content and is dependent on the type of binder), the concrete age / exposure time, the time
of investigation (moisture state). Table 3.1.7 summarizes maximum chloride concentrations (acid
soluble concentrations) in structural elements exposed to chlorides in Switzerland and Germany
[Greve-Dierfeld et al. 2017, Schiegg et al. 2017&2018, Bisschop et al. 2016, Greve-Dierfeld et al. 2017,
Kapteina 2011, Müller et al. 2010, Ungricht et al. 2008, fib Bulletin 76, DAfStb issue 623 2016] and in
XS exposed structures summarized in [fib bulletin 76, Müller et al. 2010 and Nürnberger 1996, Greve-
Dierfeld et al. 2014, Markeset et al. 2010, Helland et al. 2010, Wall 2007, Ghods et al. 2005, Fluge et
al. 2001]. Brackish water or in baltic sea lower chloride concentrations are apparent. For XD exposed
structural elements it was distinguished between XD1 and XD3 by the structure. For example, tunnel,
gallery roofs, bridge bottom views where related to XD1 and the lower part of the tunnel walls, bridge
columns, car park floors, retaining walls to XD3. It should be noted, that especially for XD exposed
structural elements the time of inspection is of interest [Kapteina 2011, Ungricht et al. 2008]. XS
exposed structures were allocated into XS3 until a distance to wave crest of 10 m otherwise the
structural element was allocated to XS2 (submerged) respectively XS1 (> 10 m).

Table 3.1.7 maximum chloride concentrations

Exposure Distribution Mean value m [mm/year] (=characteristic value) CoV [%]
class function
XD1 0.5 – 2.0
XD2 1.5 – 4.0 50-90 (75)
XD3 Normal 1.5 – 4.0
XS1 1.0 - 2.0 40 – 90 (45)
XS2 2.0 – 4.0 20 – 50 (25)
XS3 2.0 – 4.0 40 – 70 (45)

Figure 3.1.4 exemplifies chloride profiles determined from XD-exposed structures and figure 3.1.5
exemplifies chloride profiles from XS-exposed structural elements.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 176
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Figure 3.1.4: Chloride profiles from XD exposed structures and exposure tests

Figure 3.1.5: Chloride profiles from XS exposed structures and exposure tests

Initial chloride concentration Ci=C0

At existing structures in Switzerland C0 of 0.04 wt.%/cement are frequently found. The tolerable
chloride content in raw materials gives raise, that the initial chloride content is around 0.1 wt.-%. The
initial chloride content depends on the mix design. For design of new structures, the assumed initial
chloride concentration must be the same as the chloride content class specified. Current EN 206
allow specifications up to 0.4%.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 177
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Depth with maximum chloride content Dx

The depth with maximum chloride content may deviate from concrete surface. There are several
reasons for this. These are for example capillary suction [Gang et al. 2014] and water vapour transport
(wick-actions) [Olsson 2019, Baroghel-Bouny 2007], carbonation [Geng et al. 2016, Li et al. 2018, Wang
et al. 2017, Liu et al. 2017, Malheiro et al. 2018] and leaching [De Weerdt et al. 2016] and wash-off or.
Therefore, the actual depth Dx depends on factors like environmental conditions that favour capillary
suction, water vapour transport, carbonation or leaching. Therefore, at initiation time, the concrete
quality (w/c ratio) and the carbonation resistance have a huge influence on the depth Dx. Table 3.1.8
summarizes suggested values for Dx in dependency of the exposure class (transport mechanism
involved) assessed and verified from inspected structural elements. It must be considered that the
application of Dx after 50 year of exposure (XRDS) inherently affects the age exponent a due to the
time dependency of Dx. Therefore, the age exponent increases with increasing progress of Dx (i.e.
carbonation rate). Generally, it can be assumed, that concretes with a higher Chloride resistance may
(lower Ca-content and higher replacement levels by SCM) exhibit higher Dx and with lower chloride
resistance lower Dx.

Table 3.1.8 Depth Dx

Exposure Distribution Mean value Single values derived from investigations in CoV [%]
class function m [mm] CH
XD1 Beta a=0, 0 [fib 2006] 5 – 25 [Expert report CH] 50%
XD2 b=50 0 – 10 [expert report; sewage building CH] [fib2006]
XD3 [fib2006]; 10 [fib 2006] 5 – 25 [expert reports CH] (s=7-15
XS1 Normal 0 [fib 2006] 5 – 25 [derived from carbonation depth] mm
XS2 (s=7-15) 0 – 10 [from leaching depth in eg. Eckström] for ND)
XS3 10 [fib 2006] 0 – 10 [derived from carbonation depth]

Effect of temperature on apparent diffusion coefficient fT

Ionic diffusion is affected by the temperature as well as dissolution and chloride binding which is
evaluated for example in [Dousti et al. 2015]. The effect on ionic diffusion or migration is often
modelled with the Arrhenius Law. Ionic diffusion is increased with temperature and with this the
apparent diffusion coefficient and on migration coefficient [Care et al. 2008, Touil et al. 2017, Dousti
et al. 2013, fib 2006]. Some researches found that increased porosity may be one reason for increase
in migration coefficient at increased temperature [Care et al. 2008]. Table 8 summarizes temperature
conditions in dependency of exposure class (XD1-XD3 CH and D average temperature ranges; XS1 –
XS3 ranges of temperature North Sea NL, BE, D). Additionally, table 3.1.9 summarized proposed,
calculated, and determined values for fT from the above-named sources.

Table 3.1.9 Effect of temperature on the apparent diffusion coefficient

Exposure Distribution Average annual Mean values m [mm/year] CoV [%]
class function temperature m/s
[°C] normal D
XD1 10 (8-14)/8 0.56 [fib2006],
XD2 10/5 0.6 [Touil et al. 2017],
XD3 Normal 10 (8-14)/8 0.7-0.8 [Yuan et al. 2009] 30 – 36 [fib2006]
XS1 10 (9-11)/6 0.5-0.9 [Dousti et al.
XS2 10 (9-11)/5 2013&2015, interpolated]
XS3 10 (9-11)/6

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 178
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Environmental effect fenv, age exponent under site conditions a and apparent diffusion coefficients
Dapp(t) at exposed structures:
As described initially, the age exponent and the environmental effect fenv are correlated and depend
on the degree of saturation and degree of hydration at the first time of chloride exposure and their
development. The concrete starts with a concrete quality expected after curing protocol curing class
2. If the concrete is directly exposed to water contact with chlorides, ongoing hydration will affect
ageing exponent. In addition to that, precipitates, moving depth with maximum chloride content affect
the age exponent binder dependent.

Ø ongoing hydration, dry out and early chloride exposure (XS1, XD1, XD3)
Ø partially saturated concrete and late chloride exposure (XD1, XD3)
Ø Moving maximum chloride concentration (all exposures)
Ø Ongoing hydration under saturated conditions early exposure (XS2, XD2, XS3)
Ø Ongoing hydration under saturated conditions late exposure (XD2)

In the exposure class XS2 one difference between XRDS model and fib model exist, because of the
exposure time prior to corrosion. The XRDS model starts with an apparent diffusion coefficient
expected after curing protocol according to curing class 2. The fib model starts with a diffusion
coefficient of a concrete cured for 28 day. Therefore, the age exponents in the XRDS model is slightly
higher than in the fib model due to the initially lower apparent diffusion coefficient.

Figure 3.1.6 shows apparent diffusion coefficients of existing structures from Bamforth 2004. The
apparent diffusion coefficients are in the range of < 0.05 up to app. 10 · 10-13 m2/s for FA and BFSC
concretes and 5 up to > 50 · 10-13 m2/s for ordinary Portland cement concrete.

Figure 3.1.6: Apparent diffusion coefficients for

CEM I (left) slag cement concrete (middle) and
concretes with FA (right) from [Bamforth 2004]

Figure 3.1.7 shows collected apparent diffusion coefficients of single structures with specific
combinations of cement + addition.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 179
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Figure 3.1.7: Apparent diffusion coefficients from existing structures from site investigations and
data analyses provided by members of fib TG8.5/5.6 [Greve-Dierfeld et al. 2014]

Table 3.1.10 factors for fenv derived from literature review and age exponents under different exposure
conditions and according to figures 3.1.6-3.1.7.
Table 3.1.10 factors fenv and a from existing structures and from experimental investigations of chloride
transports under partially saturated conditions [Olsson2019, Zhang2020, Shah2019, Honglei et al.
2020, Savoye et al. 2018, Shah et al. 2019, Kapteina 2011]
Expo- fenv [-] a [-] Comment and source
sure Distri- mean Co Distri- mean CoV
class bution value V bution value [%]
func- [%] functi
tion on
XD1, Const. 1 0 0.6-0.8 XD1 [Kapteina 2011]
XD3, 1 Beta 0.60 CEM I [Polder2012]
XS1, 1 a=0, 0.65 40 CEM II/B-S, III/A [Polder2012]
tests 1 b=1 0.70 CEM I > 50% BFS [Polder2012]
1 0.80 CEM I < 30% FA [Polder2012]
ND 0.2-0.3 10 0.5-0.6 Wet dry cycles [Gang et al. 2015]
0.1-0.3 - Sd = 80% [Zhang et al. 2020]
0.1-0.5 - Sd = 80% [Savoye et al. 2018]
0.6-1.0 0.6-0.4 55-85%RH [Vera et al. 2007]
0.2-0.6 0.15-0.25 Atmospheric [Costa et al. 2001]
XD2, Const. 1 0 0.27 CEM I [Bamforth 2004]
XS2, 1 DRCM 0.30 CEM I [Gehlen 2000]
XS3 1 DRCM 0.45 CEM I + slag [Gehlen 2000]
1 0.62 CEM I + slag [Bamforth 2004]
1 0.70 CEM I + FA [Bamforth 2004]
1 0.46-0.64 CEM I +5-8%SF [Helland et al.]
1 0.61-0.84 CEM I+5-8%SF+46-48%BFS
1 0.4 CEM I [Polder2012]
1 0.45 CEM II/B-S, III/A [Polder2012]
1 0.5 CEM I > 50% BFS [Polder2012]
1 0.7 CEM I < 30% FA [Polder2012]
1 0.2 CEM I XS3 [Markeset et al. 2010]
1 0.38-0.43 CEM I +8-20%SF
1 0.46-0.52 CEM I+10-35%FA

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 180
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Average app. diffusion coefficient at Initiation time Dapp(tini) in dependency of DXRDS,m [10-13 m2/s]
(average Diffusion coefficient under XRDS reference conditions:
Especially in XD1 conditions, where the initiation time is much lower than the design service life and
the age exponent may differ strongly from age exponents expected under unidirectional diffusion in
saturated conditions, the apparent diffusion coefficient at initiation time differs from DXRDS. Equation
(10) repeats the calculation procedure and table 3.1.11 compares apparent diffusion coefficients at
initiation time and apparent diffusion coefficients under XRDS reference conditions.

( )= , · 50 ∙ Eq. (10)

Table 3.1.11: Example for Diffusion coefficient at initiation time in the exposure class XD1 in
dependency of DXRDS,m
DXRDS,m a [-] aXRDS [-]* tini = 40 year tini = 90 year
[10-13 Dapp(tini) [10-13 m2/s] respectively factor for DXRDS,m
m2/s]
0.8 0.3 0.17 0.09
0.5 0.37 0.19
0.7 0.81 0.42
1 0.7 0.3 0.24 0.14
0.5 0.53 0.30
0.7 1.17 0.66
0.6 0.3 0.35 0.22
0.5 0.77 0.48
* Some single values can be found in section 3.1.7

Determination of XRDS designation mean value DXRDS,m [10-13 m2/s]:

Because XRDS are defined under specific, constant reference conditions: 20°C, 30g/L NaCl, the
D(XRDS) consider the following variabilities:
Ø Variations from the concrete plant
Ø Variation in compaction on site
Ø Variation in curing
Ø Variation in the ageing under XRDS conditions (aXRDS)
Ø (Some unavoidable variations in testing)
Variations of the climate conditions within one design situation are included in the factor fenv, fT, a and
in turn in the requirements for the concrete cover.
Some data on variability regarding the factory production process including placed have been
evaluated in [Ait Mokhdar et al. 2013 and Jacobs et al. 2018]. In table 3.1.12 the average CoV and
average standard deviation is given determined from the above listed data and used in the calculations.
Table 3.1.12: variations of DXRDS,m
Distribution function Average CoV [%] Source
Normal 33-40 EN 12390-11
Normal 49-116 Gehlen 2000
Normal 66-106 Bamforth 2004

XRDS is the 90% quantile of the apparent diffusion coefficient expected under XRDS conditions
(DXRDS). XRDS is calculated with Eq. (11).

= + 1.282 ∙ ∙ Eq. (11)

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 181
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

DXRDS is the mean value of XRDS designation in [m2/s]; apparent diffusion coefficient under XRDS
conditions (curing class 2, 20°C, 30g/L NaCl)
XRDS is the 90% quantile of DXRDS for the case of a normal distribution [m2/s]

Table 3.1.13 summarizes DXRDS,m in dependency of CoV within the lower ranges given in table 3.1.12.
XRDS [10-13 0.5 1 1.5 2 3 4 5 6 8 10 15 20 25
m2/s]
CoV DXRDS,m [10-13 m2/s]
40 0.3 0.7 1.0 1.3 2.0 2.6 3.3 4.0 5.3 6.6 9.9 13.2 16.5
60 0.3 0.6 0.8 1.1 1.7 2.3 2.8 3.4 4.5 5.7 8.5 11.3 14.1
100 0.2 0.4 0.7 0.9 1.3 1.8 2.2 2.6 3.5 4.4 6.6 8.8 11.0

3.1.4 Parameter study

Table 3.1.13 and Table 3.1.14 show the parameter study applied to determine minimum concrete
cover in dependency of XRDS classes.
Table 3.1.13 Parameter study for XD exposures, mean values (distribution functions are shown in the
previous tables
Para- Unit Distribution XD1_1 XD1_2 XD2 XD3
meter function m CoV % m CoV % m CoV % m CoV %
DXRDS 10-13 m2/s Normal 0.1< m < 25 and CoV = 40-100%
aXRDS - Constant* See section 3.1.7 (0.3-0.7)
a - 0.65 - (0.3-0.7)=aXRDS
Dapp(tini) 10-13 m2/s Normal See Table 3.1.11
Cs wt.%/binder Lognormal 1.5 75 1.5 75 3 45 3 45
Ccrit wt.%/binder Lognormal 0.6 20 0.6 20 0.9 20 0.7 20
Ci=C0 wt.%/binder Constant 0.1-0.2
fenv - Constant 1.0 - 0.4 - 0.8 - 0.5 -
fT - Constant 0.7 - 0.7 - 0.6 - 0.7 -
Dx mm Beta 10 50 10 50 5 50 10 50
xcorr mm Constant 50
vcorr mm/year normal 5 30 5 30 10 130 30 130
* Variability is included in the apparent diffusion coefficient.
condition is accounted for with Eq. (9) respectively Eq. (
Table 3.1.14 Parameter study for XS exposures, mean values (distribution functions are shown in the
previous tables
Parameter Unit Distribution XS1 XD2 XD3
function m CoV % m CoV % m CoV %
XRDS 10-13 m2/s Normal 0.1< m < 20 and CoV = 60%
aXRDS - Constant* See section 3.1.7 (0.3-0.7)
a - 0.65 - (0.3-0.7)=aXRDS
Dapp(tini) 10-13 m2/s Normal Table 3.1.11
Cs wt.%/binder Lognormal 2 65 4 25 4 45
Ccrit wt.%/binder Lognormal 0.6, 20%
Ci=C0 wt.%/binder Constant 0.1-0.2
fenv - Constant 1.0 - 1.0 - 1.0 -
fT - Constant 0.56 - 0.56 - 0.56 -

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 182
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Dx mm Beta 10 50 0 50 10 50
xcorr mm Constant 50
vcorr mm/year normal 30 120 20 50 70 120
* Variability is included in the apparent diffusion coefficient.

3.1.5 Depth with critical chloride content versus DXRDS,m

Cover calibration requires to allocate the cnom to cmin, which was performed with Table 2. D(tini) was
allocated to XRDS with Eq. (8) and Eq. (9). Figure 5 shows the calculated concrete cover.

Figure 3.1.8: Minimum concrete cover in dependency of XRDS for b=1.5 (left) tSL = 50 year and (right)
tSL=100 year

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 183
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

3.1.6 Results
Figure 3.1.6 shows the approximate curves to the median between maximum and minimum depth
xcrit for the exposure classes of XD1/XD2a and the exposure classes XD3/XD2b.

Figure 3.1.9 XRDS and cmin

Table 3.1.15 summarizes the curves approximated to the calculated minimum cover.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 184
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Table 3.1.15 summarizes minimum concrete cover versus approximate values of figure 3.1.9
cmin,dur [mm]
XD1/XD2a XD3/XD2b
b=1.5 b=1.5 b=0.5
XRDS 50 years 100 years 50 years 100 years 50 years 100 years
0.5 20 20 20 30 25 20
1 20 25 25 35 25 25
1.5 20 30 30 40 30 30
2 20 30 35 45 30 35
3 25 35 40 55 35 45
4 25 35 45 65 35 50
5 30 40 50 70 40 60
6 30 40 55 80 40 65
8 35 50 60 90 45 65
10 35 50 65 90 45 70
15 40 55 70 100 55 75

3.1.7 Verification
3.1.7.1 Link between XRDS and durability testing standards
XRDS is the apparent diffusion coefficient Dapp(50year) in reference conditions:
Ø At 20°C
Ø Saturation in 30g/l NaCl
Ø Curing class 2

Actual testing conditions (CEN Test Standards) and national testing standards are listed in table
3.1.16.

Table 3.1.16: National and European testing standards

- XRDS Options
A B C
EN 1992-1-2, EN 12390-11:2015 (all
CEN - FprEN 12390-18:
EN 206 procedures shall ensure SIA 262-1/B
standards 2020
unidirectional diffusion)
30g/L NaCl, 20°C, 28d water curing at 20°C, Migration Migration
Reference Depth with 0.6 by mass Vacuum water saturation, 28d water curing 28d water curing at
conditions of binder-cement + type 2 d preparation, age at at 20°c 20°c
(XRDS) and II additions, Dapp(50year) exposure 30d, 3% NaCl, Chloride Chloride transport
test 10-13 m2/s chloride transport at 20°C transport at 20°C at 20°C
conditions Curing Class 2 Dnss(90day) in 10-12 m2/s, Mnss(28d) in 10-12 Mnss(28d) in 10-12
Cs(90d), Ci m2/s m2/s
Correlation
factors to fexe, aXRDS aXRDS, fexe, fMnss aXRDS, fexe, fMnss
XRDS

From Table 3.1.17 it can be seen that the curing and testing conditions are similar in between XRDS
reference conditions and testing conditions. That means the samples and XRDS means water

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 185
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

saturation over the entire exposure time at constant 20°C. It can be assumed, that the degree of
hydration is similar and the temperature and environmental factors fsd and fT (respectively ke) equals
1. Finally, the age exponent is the one expected in a 30% NaCl solution (or in see water conditions
submerged).

Table 3.1.17: factors to link XRDS to test condition.

XRDS Temperatur Curing type / duration migration vs. CoV of placed
ageing
Definition / moisture prior to chloride load diffusion concrete
Dapp(50year)
10-13 m2/s
fT = fenv = 1 aXRDS
(90% fexe fMnss CoV
quantile)

3.1.7.2 Current deemed-to-satisfy provisions in Switzerland

Table 3.1.18 shows current performance requirements in the exposure classes XD1-3. The exposure
class XD2 is subdivided into exposure class XD2a with < 0.5 g/l Cl and XD2b with > 0.5 g/l Cl.
Table 3.1.18: Performance requirements
Exposure class XD1 XD2a XD2b XD3
Service life 50y 100y 50y 100y 50y 100y 50y 100y
Value DCl [10-12 m2/s] according to SN EN - - - - 10 10 10 10
206:2019
Guidance value k [mm/year0.5] according - - - - 13 13 13 13
to SN EN 13670:2009
Req. cmin [mm] according to SIA262:2019 35 45
Tolerance Dcdev [mm] according to SIA 10
262:2019 (smoothed elements)

3.1.7.3 Corelations
Figure 3.1.10 shows correlations between curing duration and curing temperature on chloride
migration coefficient from data investigated in [Audenaert et al. 2010, Jacobs 2016, Jung et al. 2018,
Gjorv O. 2014].

Figure 3.1.10: Influence of curing duration dc and curing temperature prior to testing on chloride
migration coefficient
Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 186
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Figure 3.1.11 shows (a) apparent diffusion coefficients determined in ponding tests and (b) chloride
migration coefficients. All coefficients have been determined at concretes cured for 28-day prior
vacuum saturation. Curing times different from 28-day are given with empty symbols.

Figure 3.1.11: Chloride diffusion coefficients determined under conditions similar to XRDS definition
(but curing time 28day) (left) and chloride migration coefficients (right)

Figure 3.1.12 summarizes age exponents from literature determined at concretes exposed at an age
of 28 day to conditions similar than XRDS conditions.

Figure 3.1.12: age exponents determined under conditions similar than XRDS reference conditions
but exposed at an age of 28 day

The apparent diffusion coefficient of the placed concrete (Curing Class 2) in XRDS reference
conditions at reference time t = 1 year can be calculated with Eq. (12).
.
(t = 0.0767year) = , / Eq. (12)

D0 is the apparent diffusion coefficient at time t=1 year under XRDS reference conditions in
[m2/s]
aXRDS age exponent [-] under XRDS reference conditions

The XRDS classes and the corresponding diffusion coefficients DXRDS, and those at reference time t0
after placing and curing in CC2 are given D0 are given in Table 3.1.19 in dependency of age exponent
under reference conditions and standard deviation.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 187
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Table 3.1.19 XRC, carbonation rate at initiation time tini, carbonation rate at reference time t0 after
placing and curing in CC2 are given in Table 9 in dependency of age exponent and standard deviation.
XRDS [10-13 CoV aXRDS 0.5 1 1.5 2 3 4 5 6 8 10 20
mm2/s] [%] [-]
MRCM (t=28 0.0 0.0 0.1 0.1 0.1 0.2 0.3 0.3 0.4 0.5 0.7 1.3
day) [10-12 40% 0.3 0.2 0.5 0.7 0.9 1.4 1.8 2.3 2.8 3.7 4.6 9.2
m2/s] for 0.4 0.4 0.9 1.3 1.8 2.6 3.5 4.4 5.3 7.1 8.8 18
fMnss=1 0.5 0.8 1.7 2.5 3.4 5.1 6.8 8.4 10 14 17 34
0.6 1.6 3.2 4.8 6.5 9.7 13 16 19 26 32 65
MRCM (t=28 0.0 0.0 0.1 0.1 0.1 0.2 0.2 0.3 0.3 0.5 0.6 1.1
day) [10-12 60% 0.3 0.2 0.4 0.6 0.8 1.2 1.6 2.0 2.4 3.2 3.9 7.9
m2/s] for 0.4 0.4 0.8 1.1 1.5 2.3 3.0 3.8 4.5 6.0 7.5 15
fMnss=1 0.5 0.7 1.4 2.2 2.9 4.3 5.8 7.2 8.7 12 14 29
0.6 1.4 2.8 4.1 5.5 8.3 11 14 17 22 28 55
MRCM (t=28 0.0 0.0 0.0 0.1 0.1 0.1 0.2 0.2 0.3 0.4 0.4 0.9
day) [10-12 100% 0.3 0.2 0.3 0.5 0.6 0.9 1.2 1.5 1.8 2.4 3.1 6.1
m2/s] for 0.4 0.3 0.6 0.9 1.2 1.8 2.3 2.9 3.5 4.7 5.9 12
fMnss=1 0.5 0.6 1.1 1.7 2.2 3.4 4.5 5.6 6.7 9.0 11 22
0.6 1.1 2.1 3.2 4.3 6.4 8.6 11 13 17 21 43

3.1.8 References
Audenaert A, Yuan Q, De Schutter G (2010) On the time dependency of the chloride migration
coefficient in concrete. Construction and Building Materials 24(3): 396-402.

Sang-Hwa Jung, Hwa-Sung Ryu, Subbiah Karthick, Seung-Jun Kwon (2018) Time and Crack Effect on
Chloride Diffusion for Concrete with Fly Ash. International Journal of Concrete Structures and
Materials 12, Article number: 14.

Jacobs F (2016): Nachbehandlung von Beton. Beton 1+2:24-28.

Gjorv O (2014) Durability design of concrete structures in severe einvorinments.

Kapteina G (2011) Modell zur Beschreibung des Eindringens von Chloriden in Beton von
Verkehrsbauwerken. Dissertation Technische Universität München.
DBV 2015 Concrete Cover and Reinforcement. Securing of Concrete Cover at Design, Production and
Installation of Reinforcement and Concrete according to Eurocode 2
JCSS (2000), JCSS probabilistic model code Part 1: Basis of design. Joint Committee on Structural Safety.
JCSS (1999), JCSS probabilistic model code Part 3: Resistance models. Joint committee on structural
safety.
DAfStb Issue 622 (2016) Durability design rules for XC-exposed structural elements (in German) Greve-
Dierfeld (2016).
Kenel A, Jacobs F (2018) Measurement and evaluation of the cover depths. Final Report Astra AGB
2014/007 (in German).

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 188
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

DuraCrete 2000 Statistical quantification of the variables in the limit state function. BriteEuram final
Report
Alonso C, Andrade C, Farina J, Lopez F, Merino P, Novoa R (1995): Galvanic corrosion of steel in
concrete. Materials Science Forum 1992-1995:899-906
Contectvet (2000) A validated users manual.
Tuutti, K. 'Service life of structures with regard to corrosion of embedded steel' in 'Performance of
concrete in marine environment', ACI SP-65 (1980) pp.223-236.
Hunkeler F, Mühlan B, Ungricht H (2006) Risk to spalling due to reinfocement corrosion (Risiko von
Betonabplatzungen infolge Bewehrungskorrosion). Report No. 603 (in German)
Nürnberger U (1996) Corrosion and corrosion protection in the building industry (Korrosion und
Korrosionsschutz im Bauwesen) ISBN-13 : 978-3000085536
Y. Schiegg, F. Hunkeler, H. Ungricht, D. Keller, Measures to increase durability – Results
to the field study in the Naxberg tunnel, VSS report no. 683 (2017)

Schiegg Y, Bisschop J, Greve-Dierfeld S (2018) Monitoring rebar corrosion propagation in concrete –

Results of the Naxberg field experiment after 12 years. Eurocorr 2018

Harald S. Müller, Edgar Bohner, Christian Fischer, Joško Ožbolt, Christoph Gehlen, Kai Osterminski,
Peter Schießl, Stefanie von Greve-Dierfeld (2010) Dauerhaftigkeitsbemessung von
Stahlbetonbauteilen auf Bewehrungskorrosion; Teil 2: Dauerhaftigkeitsbemessung. Rissbildung infolge
Bewehrungskorrosion; Verbundfestigkeit von korrodierter Bewehrung in Beton; Voll-probabilistische
Modellierung von Bewehrungskorrosion: Ein Beitrag zur Dauerhaftigkeitsbemessung;
Lebensdauerbemessung-Anwendung und Validierung. DAfStb Issue 602

Tuutti, K. 'Service life of structures with regard to corrosion of embedded steel' in 'Performance of
concrete in marine environment', ACI SP-65 (1980) pp.223-236.

Schiegg Y, Bisschop J, Greve-Dierfeld S (2018) Monitoring rebar corrosion propagation in concrete –

Results of the Naxberg field experiment after 12 years. Eurocorr 2018

K. Pettersson. Chloride threshold value and the corrosion rate in reinforced concrete. In: Proc. of
the Nordic Seminar. Lund, 1995, 257-266

Angst U, Elsener B, Larsen C, Vennesland Ø (2009) Critical chloride content in reinforced concrete —
A review. Cement and Concrete Research 39(12):1122-1138.

Angst U, Geiker M, Alonso M, Polder R, Isgor B, Elsener B, Wong H, Michel A, Hornbostel K, Gehlen C,
Francois R, Sanchez M, Criado M, Sörensen H, Hansson C, Pillai R, Mundra S, Gulikers J, Raupach M,
Pacheco J, Sagues A (2019) The effect of steel concrete interface on chloride-induced corrosion
initiation in concrete: a critical review by RILEM TC 262-SCI.

Breit W (2006) Critical corrosion inducing chloride content – state of the art and new investigation
results. Beton 145-168.

Boschmann Käthler C, Angst U, Aguilar AM, Elsener B (2019) A novel approach to systematically
collect critical chloride contents in concrete in an open access data base. Data in brief 27:104675.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 189
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Bischop J, Schiegg Y (2016) Effect of rebar and cement type on the critical chloride content of
cementitious materials. EuroCorr 2016.

Greve-Dierfeld S v, Bischop J, Schiegg Y (2017) Nichtrostende Bewehrungsstähle zur Verlängerung der

korrosionsfreien Lebensdauer von Stahlbetonbauteilen. Beton und Stahlbetonbau 112(9):601-610.

Frederiksen, J. M. (2000). On the need for more precise threshold values for chloride initiated
corrosion. Materials and corrosion, 60 (8), 597-601.

Markeset G, Vennesland Ø (2008) Critical chloride content in reinforced concrete. COIN workshop, 5-
6 June 2008, Trondheim, Norway

Sörensen HE, Poulsen SL, Jönsson U (2012) Chloride threshold values from concrete blocks exposed
at rödbyhavn marine field exposure site.

Poulsen SL, Sörensen HE (2012) Chloride threshold values – state of the art. Danish Expert Centre for
Infrastructure Constructions.

Ungricht H, Hunkeler F (2008) Massnahmen gegen chloridinduzierte Korrosion und zur Erhöhung der
Dauerhaftigkeit. Final Report ASTRA 628.

Bisschop J, Schiegg Y, Hunkeler F (2016) Modelling the corrosion initiation of reinforced concrete
exposed to deicing salts. Final Report Astra 676.

fib Bulletin 76 (2016) Benchmark for deemed-to-satisfy provisions in Standards.

DAfStb issue 623 (2016) Design for durability with partial safety factors including a benchmark for
deemed-to-satisfy rules. Greve-Dierfeld S.v., Gehlen C.
Baroghel-Bouny V (2007) Water vapour sorption experiments on hardened cementitious materials.
Part II: essential tool for assessment of transport properties and for durability prediction, Cem.
Concr. Res. 37 (2007) 438–454.

Olsson N (2018) Experimental studies of ion transport in cementitious materials under partially
saturated conditions. Dissertation Universite Paris-Est and Lund University

De Weerdt K, Orsakova D, Müller A, Larsen C, Pedersen B, Geiker M (2016) Towards the

understanding of chloride profiles in marine exposed concrete, impact of leaching and moisture
content. Construction and Building Materials 120:418-431.

Geng J, Li, LY, Easterbrook D (2016) Effect of carbonation on release of bound chlorides in chloride
contaminated concrete. Magazine of concrete Research 68(7):353-363

Gang X, Yun-Pan L, Yi-Biao S, Ke X (2014) Chloride ion transport mechanism in concrete due to
wetting and drying cycles. Fib Structural Concrete DOI: 10.1002/suco.201400035.

Li K, Zhang Y, Wang S, Zeng J (2018) Impact of carbonation on the chloride diffusivity in concrete:
experiment, analysis and application. Materials and Structures 51-164

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 190
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Wang Y, Nanukuttan S, Bai Y, Basheer PAM (2017) Influence of combined carbonation and chloride
ingress regimes on rate of ingress and redistribution of chlorides in concrete. Construction and
Building Materials 140:173-183.

Liu J, Qiu Q, Chen X, Nigxu H, He Y, Ma Y (2017) Understanding the interacted mechanism between
carbonation and chloride aerosol attack in ordinary Portland cement concrete. Cement and Concrete
Research 95:217-225.

Malheiro R, Camoes A, Meira G, Amorim MT, Castro-Gomes J, Ferreira R (2018) Behaviour of

cementitious matrices subjected to the combined action of chloride ions and carbonation. Revista
Materia ISSN 1517-7076 artigo e-12162.

Touil B, Ghomari F, Bezzar A, Khelidj A, Bonnet S (2017) Effect of temperature on chloride diffusion in
saturated concrete.

Dousti A, Rashetnia R, Ahmadi B, Shekarchi M (2013) Influence of exposure temperature on chloride

diffusion in concretes incorporating silica fume or natural zeolite. Construction and Building
Materials 49:393-399.

Dousti A, Shekarchi M (2015) Effect of exposure temperature on chloride-binding capacity of cement

materials. Magazine of Concrete Research 67(15):821-832.

Care S (2008) Effect of temperature on porosity and on chloride diffusion in cement pastes.
Construction and Building Materials 22(7):1560-1573.

M.N. Balakrishna, Fouad Mohamad, Robert Evans, M.M. Rahman (2020) Chloride Diffusion
coefficient in partially saturated conditioned concrete cubes. International Research Journal of
Engineering and Technology (IRJET) 07(04):4732-4739.
Honglei C, Zuquan J, Tiejun Z, Benzhen W, Zhe L, Jian L (2020) Capillary suction induced water
absorption and chloride transport in non-saturated concrete: The influence of relative humidity,
mineral admixtures and sulfate ions. Construction and Building Materials 236:117581.
https://doi.org/10.1016/j.conbuildmat.2019.117581

Y Zhang, Z Yang, G Ye (2020) Dependency of unsaturated chloride diffusion on the pore structure in
cementitious materials. Cement and Concrete Research 127:106919

SHA Shah, R Liu (2019) Penetration of chloride ion in cement concrete structure exposed to drying
and wetting cycles Journal of Mechanical and Civil Engineering 16(2): 13-21

Fraj AB, Bonnet S, Leklou N, Khelidj A (2019) Investigating the early-age diffusion of chloride ions in
haredening slag-blended mortars on the light of their hydration progress. Construction and Building
Materials 225:485-495

Savoye S, Mace N, Lefecre S, Spir G, Robinet JC (2018) Mobility of chloride through cement-based
materials under partially saturated conditions. Applied Geochemistry 96:78-86.

De Weerdt K, Orsakova D, Müller A, Larsen C, Pedersen B, Geiker M (2016) Towards the

understanding of chloride profiles in marine exposed concrete, impact of leaching and moisture
content. Construction and Building Materials 120:418-431.

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 191
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Q Yuan, C shi, G De Schutter, K Audenaert (2009) Effect of temperature on transport of chloride ions
in concrete. Proceeding of second conference of Concrete Repair, Rehbilitation and Retrofitting ,
Dresden, Germany.

Gang X, Yun-Pan L, Yi-Biao S, Ke X (2015) Chloride ion transport mechanism in concrete due to
wetting and drying cycles. Structural Concrete 2:289-295

Minimum cover calibration and verification for XRDS

For citations, copying and referencing see Note on Front page! page 192
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

CHAPTER 3.2

METHODS BASED IN fib MODELS FOR DURABILITY DESIGN (MC2010) and in

JCSS PROBABILISTIC MODEL CODE.
David IZQUIERDO and Carmen ANDRADE

Chloride induced corrosion

3.2.1 Objective
3.2.2 Chloride induced corrosion
3.2.2.1 Model of the initiation period in marine environments
3.2.2.1.1 Input Parameters of the chloride model and their statistical
characterization
3.2.2.1.1.1 Surface chloride concentration and their CoV
3.2.2.1.1.2 Critical chloride content (Ccr) and its CoV
3.2.2.1.1.3. Ageing factor n and its CoV
3.2.2.1.1.4 Skin zone (x)
3.2.2.1.1.5 Chloride velocity VCl (t) and its CoV
3.2.2.1.1.6 Concrete cover and its CoV
3.2.2.1.1.7 Summary of input parameters of initiation of corrosion
due to chloride ingress
3.2.2.2 Model of the propagation period
3.2.3.2.1 Input Parameters of the propagation model and their
statistical characterization
3.2.2.3 Service life model
3.2.3 Formulation of Limit State Function
3.2.4.1 Sensitivity factors
3.2.4 Design values
3.2.4.1 Design values for the propagation period
3.2.4.2 Cover depths for chloride induced corrosion
3.2.5 References

For citations, copying and referencing see Note on Front page! page 193
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

3.2.1 Objective
The objective of present document consists in establishing the cover depths for 50 and 100
years that fulfil the definition of the Exposure Resistance Classes (ERC) given in chapter 1 in the
Introduction. For that objective the principles given in the Probabilistic Model Code of the JCSS
and the carbonation model of the fib MC2010 have been used.
For achieving that objective, the steps followed are:
 Description of the Time-explicit mathematical model used for calculating the service
life to fulfil the definition of ERC
 Phases of the model and selected input parameters
 Statistical characterization of the input parameters.
 Formulation of the Limit state function (LSF). Reliability level of compliance of the LSF.
In present document the reliability factor =1.5 has been adopted.
 Probabilistic calculations of the cover depths complying with the ranking of ERC
defined in the chapter 1 of Introduction.

3.2.2 Chloride induced corrosion

3.2.1.1 Model the initiation period in marine environments
The time explicit chloride model selected is that of fib MC2010. It is based on the classical 2nd
Fick’s law with time variant diffusion coefficient and a skin zone x. Hence, the chloride
concentration at a depth x can be calculated through:
𝑥−∆𝑥
𝐶 (𝑥, 𝑡) = 𝐶0 + (𝐶𝑠 − 𝐶0 ) ∙ 1 − erf 2 𝐷𝑎𝑝(𝑡)∙𝑡
Eq 3.2.1

Where:
C0 is the initial chloride concentration of chloride in concrete in %
CS is the concentration at the surface (a fitted value not a real one)
erf is the error function
Dapp (t) is the apparent diffusion coefficient for chlorides at time t, which usually is
estimated with Eq. 3.2.2:
𝑡0 𝑛
𝐷𝑎𝑝 (𝑡) = 𝐷0 ∙ 𝑡
Eq. 3.2.2

Where: t0 is the reference time for Dapp evaluation and,

n is the so-called ageing factor, that accounts for the apparent decrement of Dapp with
time.

For citations, copying and referencing see Note on Front page! page 194
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

Probabilistic evaluation of all the input parameters in Eq. 3.2.1 is complex, since a total number
of 5 variables has to be calibrated in a posterior analysis. The equation is then simplified as was
made that of the carbonation model by embodying several parameters in the Velocity of
chloride ingress, VCl. The rearranged equation supposes the following mathematical change of
variables:

Being the Ccr the critical chloride content (in %) it can be used to define the variable :
𝐶𝑐𝑟 −𝐶0
𝜉= Eq. 3.2.3
𝐶𝑠 −𝐶0

and then,

𝑉𝑐𝑙 (𝑡) = 𝑒𝑟𝑓 −1 (1 − 𝜉) 2 𝐷 (𝑡0 ) ∙ (𝑡0 )𝑛 Eq. 3.2 4

In consequence the time to depassivation can be calculated as

2
𝐶− Δ𝑥 1−𝑛
𝑡𝑑𝑒𝑝 = Eq. 3.2.5
𝐶𝐶𝑙 (𝑡)

For calculation, all scatter is merged into Vcl and n in order to make calibration easier and
feasible.

3.2.2.1.1 Input parameters of the chloride model and their statistical

characterization

It consists on calculating or adopting the Coefficient of Variation to be applicable to

each mean value of the input parameter. Calculation are made based in mean values.
3.2.2.1.1.1 Surface chloride concentration
For XS2 and XS3classes the chloride surface concentration is made to depend on the cement
type (Izquierdo, D. Andrade, C., 2011) and ((fib), 2015).
Table 3.2.1 Values of the surface concentration in function of cement type considered in the
calculations for XS2 and XS3
Cement type Cs [%Con]
CEM I 0.35
CEM III/B 0.35
CEM II/A-V 0.55
CEM II/A-D 0.50

For exposure case XS1, surface concentration is dependent on many parameters (seashore
distance, height of exposure, wind direction, wave height, etc.). In figure 3.2.1 are shown the
data used for the calculation of the scatter and due to it a simplified ranking approximation
was made (Izquierdo, D. Andrade, C., 2011).

For citations, copying and referencing see Note on Front page! page 195
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

1
Data
0,9
Cs [% by mass of concrete ]

0,8 fractile 95%

0,7
Average value
0,6
0,5
0,4
0,3
0,2
0,1
0
0,001 0,01 0,1 1

Inverse of diatance to the sea border [1/m]

Figure 3.2.1 Relation between distance to shoreline and superficial concentration of

chlorides (the X axis shows the inverse to the distance to the seashore)
The exposure class XS1 is not defined in detail in the EN206 and in reality, covers a wide range
of distances and locations with respect to the shoreline. For the sake of this exercise a value of
100 m is adopted and therefore average value of 0.2 % by weight of concrete of surface
chloride concentration with respect to the concrete mass is taken for the calculations.
A CoV = 50% was taken in all exposure classes.

3.2.2.1.1.2 Critical chloride content (Ccr)

Critical chloride content is widely characterized in the literature (Izquierdo et al. 2004). In
present calculations an averaged value of Ccr = 0.6% by cement weight and a CoV = 30% are
adopted, with a normal distribution.

3.2.2.1.1.3. Ageing factor n

Ageing factor is in many cases the most influencing variable in Eq. 3.2.1. Hence a proper
calibration of this variable is essential. For this exercise several data sources have been used in
order to account for the longest exposure periods because at short periods the aging factor n
may be still evolving and then, with a high uncertainty. In table 3.2.2 is provided the values
considered and their bibliographic source, together with the CoV recorded.

For citations, copying and referencing see Note on Front page! page 196
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

These data enabled to propose in all cases a CoV = 20% (upper boundary of recorded values)
for being adopted in the calculations.
Table 3.2.2 Values of aging factors used in the calculations and references
Cem type Source nµ nµ
(XS2/XS3) (XS1)
CEM I ((fib), 2015), (Izquierdo, D. 0.45 0.60
Andrade, C., 2011), (Polder, R.B.
Rooij, M.R., 2005)
CEM II/A-V (Izquierdo, D. Andrade, C., 2011), 0.80 0.60
(Polder, R.B. Rooij, M.R., 2005)
CEM III/B ((fib), 2015), (Polder, R.B. Rooij, 0.50 0.70
M.R., 2005)
CEM II/A-D ((fib), 2015) 0.40 0.65

3.2.2.1.1.4 Skin zone (x)

It is named “convection zone” in MC2010, however the mechanisms acting are not only
convention and then in preset exercise will be named “skin zone”. It is considered only in XS3
environment, where it has been shown that the combination of carbonation and chloride
ingress more often leads to a non-fickian diffusion profile (with a maximum in the interior of
the concrete). For the calculations, an average value of 10 mm ((fib), 2015) and CoV = 50%
were adopted.

3.2.2.1.1.5 Chloride velocity VCl (t)

The simplified Eq. 3.2.5 embodying several input parameters and resulting in a VCl is used for
the calculations.

2
1.8 y = 1.6878x-0.657
1.6 R² = 0.7926
Coefficient of Variation

1.4
1.2
1
0.8
0.6
0.4
0.2
0
0 5 10 15 20 25
Apparent Vcl [mm/year^0.5]

Figure 3.2.2 Average and CoV for apparent chloride ingress rate

For citations, copying and referencing see Note on Front page! page 197
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

Chloride ingress is dependent of cement type, exposure conditions and concrete quality and
then in a parallel manner than in the case of the carbonation rate, the VCl value has been found
to depend on its average values, as shown in figure 3.2.2. It is a hyperbolic function whose
formula will be used in the calculations for the CoV.
−0.657
𝐶𝑜𝑉𝑉𝐶𝑙 = 1.6878 ∙ 𝑉𝐶𝑙 Eq.3.2.6

3.2.2.1.1.6 Concrete cover

The same CoV = 30% than in chapter 2.2 for carbonation is adopted in present calculations
((fib), 2015) (Izquierdo, D. Andrade, C., 2011) (Izquierdo. D, 2001).
Table 3.2.3 Adopted Coefficient of variation of the concrete cover thickness
Type of execution Dist. Type Bias CoV
in situ – normal conditions Log-normal 1.0 30%
Precast – dedicated quality control Normal 1.0 10%

3.2.3.2 Summary of input parameters of initiation of corrosion due to chloride ingress

Table 3.2.4 Input parameters for the chloride modelling
Parameter Units Average value CoV (%) Statistical
distribution
C0 [wt.-%/cem] 0.01 20 Log-Nor
Cs,Dx [wt.-%/conc] See Table 3.2.1 for 50
XS2/XS3 Log-Nor
And 0.2% for XS1
Ccrit [wt.-%/conc] 0.1 30 Normal
x [mm] 10mm only in XS3 50 D
c [mm] several 30 See Table
3.2.3
VCl [m²/s] several See
equation Log-Nor
3.2.5
n [-] See table 3.2.2. 30
t0 [years] 28 days - -
t [years] 50 and 100 years - -

3.2.2.2 Model of the propagation period

The propagation model is the same than for carbonation described in (Andrade 1989):

𝑃
𝑡𝑝 = 𝑉𝑐𝑜𝑟𝑟 Eq.3.2.6
𝑐𝑜𝑟𝑟

where tp is the corrosion propagation time in years, Pcorr (µm) is the accumulated corrosion or
attack penetration after a certain period of time and Vcorr (µm/year) is the annually averaged

For citations, copying and referencing see Note on Front page! page 198
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

corrosion rate.

3.2.2.2.1. Input Parameters of the propagation period and their statistical

characterization
For propagation period, the values considered taken are those given in Table 3.2.5 (Andrade,
C., 1999). For scatter quantification, after (Izquierdo, D. Andrade, C., 2011), it was obtained
that 60% of CoV variation is shown in exposure cases with constant conditions (e.g.: XS1/XS2)
and 90% in XS3 with wet-dry cycles.
Table 3.2.5 Values of the corrosion rate adopted in exposure class and their corresponding CoV.
Propagation pedirods until Pcorr = 500 µm
CoV Vcorr,D tp[yr]
Exposure Vcorr [µm/y]
(%) =1,5 =1,5
XS1 30 60 56.3 1
XS2 10 60 13.1 4
XS3 70 90 105.0 0

3.2.2.3 Service life model

The service life is composed of an initiation (ti) period and a propagation (tp) one:

𝑡𝑆𝐿 = 𝑡𝑖 + 𝑡𝑝 Eq. 3.2.7

3.2.3 Formulation of Limit State Function

The probabilistic and partial factor methodology used next are those recommended in the
Probabilistic Model Code of the JCSS.
The method is the same than that described in chapter 2.2.3 for carbonation

3.2.3.1 Sensitivity factors

Sensitivity factors and target – reliability values (Izquierdo, 2019) have been calculated for the
input parameters of the model. The results obtained are the following:
a) For the case of chloride induced corrosion (seawater source), the sensitivity factors
are shown in figure 3.2.3 can be deduced that, on the resistance side Cover and
ageing factors are leading values, whereas chloride ingress rate (VCl) is the leading
variable on the action side.

For citations, copying and referencing see Note on Front page! page 199
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

Figure 3.2.3 Sensitivity factors for Chloride induced corrosion (max. 50µm loss of rebar section)

For calculation purposes the parameters given in Table 3.2 6 are adopted

Table 3.2.6 Sensitivity factors of the input service life parameters in the case of
carbonation
Variable Name  Type
Cover C 0.40 Resistance
Chloride Ingress rate VCl -0.80 Action
Corrosion rate VCorr ~0 Action
Ageing factor n 0.60 Resistance

In the same manner than in the case of carbonation, the summatory  > 1, that implies that
the values are slightly conservative. If a further refinement would be required reported values
for  could be divided by =1 in order to normalize the values. However, for this application
and in order to follow EN1990 procedure, no normalization to 1 was adopted.

Another conclusion from this sensitivity analysis is that the most sensitive parameter are the
chloride ingress rate and the aging factor. Then corrosion rate in this case is not predominant
as the values are very high resulting in relatively short propagation periods not impacting
significantly in the total service life except as will be justified in next paragraphs.

For citations, copying and referencing see Note on Front page! page 200
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

3.2.4 Design values

3.2.4.1 Design values for propagation period

Since the additive definition of service life: initiation + propagation period and given the fact
that concrete cover is only affecting the first one, in order to determine the required cover for
each exposure class and concrete property will obtained by subtraction of propagation period
from the total required service life.
𝑡𝑑𝑒𝑝 (𝑐𝑜𝑣𝑒𝑟) = 𝑆𝑒𝑟𝑣𝑖𝑐𝑒 𝐿𝑖𝑓𝑒 − 𝑡𝑝𝑟𝑜 Eq. 3.2.8

Applying design values to Eq. 3.2.8, yields:

𝑡𝑑𝑒𝑝 (𝑐𝑜𝑣𝑒𝑟), 𝑑 = 𝑆𝑒𝑟𝑣𝑖𝑐𝑒 𝐿𝑖𝑓𝑒 − 𝑡𝑝𝑟𝑜 , 𝑑 Eq. 3.2.9

Therefore, design values for several reliability levels are obtained considering a log-normal
distribution, 500 µm as maximum pitting attack giving:
50
𝑡𝑝𝑟𝑜,𝑑 = Eq. 3.2.10
𝑉𝑐𝑜𝑟𝑟,𝑑

Where Vcorr,d can be calculated as:

𝑉𝐶𝑜𝑟𝑟,𝑑 = 𝑉𝑐𝑜𝑟𝑟,𝜇 𝑒0.3·𝛽·𝐶𝑜𝑣 Eq. 3.2.11

Where  = -0.30 is adopted for XS cases. Calculated values for corrosion rate and propagation
period until Pcorr = 500 µm (pitting and end of service life) were given in Table 3.2.5. In view of
the short design propagation periods, no propagation has been discounted from the initiation
in the calculation of service life.

3.2.4.2 Cover depths for Chloride induced corrosion

The cover depth values are given in terms of cmin,dur (where 10 mm for tolerance is subtracted
from the design value of concrete cover). Table 3.2.7 shows the calculated minimum cover
depth values (in mm) for 50 and 100 years for each ERC (from 0.5 to 7) and exposure classes
XS1 to XS3). They should be rounded to the closest value ranked every 5 mm. The rounded
values are shown in Table 3.2.8.

Table 3.2.7 Calculated minimum values (10 mm were subtracted from the nominal cover used
in the calculations) for 50 and 100 years of service life)
Calculated mínimum cover depths
XS1 XS2 XS3
=1.50 50 years 100 years 50 years 100 years beta=1.50 100 years
XRD 0.5 16.0 18.1 19.0 24.7 32.0 37.4
XRS 1 22.0 24.8 30.0 36.7 42.0 51.0
XRS 2 30.0 34.1 45.0 54.5 56.0 69.5
XRS 3 35.0 41.0 56.0 68.7 68.0 83.4
XRS 4 40.0 46.8 66.0 81.0 78.0 94.9

For citations, copying and referencing see Note on Front page! page 201
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

XRS 5 44.0 51.8 75.0 92.0 85.5 104.8

XRS 6 48.0 56.3 83.0 103.0 92.5 113.8
XRS 7 51.2 60.4 90.0 111.5 98.5 121.9
XRS 8 55.0 64.2 97.0 120.3 104.5 129.4
XRS 8.5 56.3 66.0 100.0 124.5 107.0 133.0

Table3.2 8 Rounded minimum values of cover depths

Rounded mínimum cover depths
XS1 XS2 XS3
=1.50 50 years 100 years 50 years 100 years beta=1.50 100 years
XRD 0.5 20 25 20 25 35 40
XRS 1 25 30 30 40 45 55
XRS 2 30 35 45 55 60 70
XRS 3 35 40 55 70 70 N.R.*
XRS 4 40 50 65 N.R.* N.R.* N.R.*
XRS 5 45 55 75 N.R.* N.R.* N.R.*
XRS 6 50 60 N.R.* N.R.* N.R.* N.R.*
XRS 7 55 65 N.R.* N.R.* N.R.* N.R.*
XRS 8 60 N.R.* N.R.* N.R.* N.R.* N.R.*
XRS 8.5 65 N.R.* N.R.* N.R.* N.R.* N.R.*
*Not recommended

3.2.5 References
 Andrade, C., Alonso, M.C., y Gonzalez, J.A., (1990) “An initial effort to use the corrosion
rate measurements for estimating rebar durability”, in Corrosion Rates of Steel in
Concrete, ASTM STP 1065, eds. N.S. Berke, V. Chacker, y D. Whiting, Philadelphia, USA,
pp.29-37
 Andrade, C. (2014). Probabilistic Treatment of the Reinforcement Corrosion. 2013 W.R.
Whitney Award Lecture: Corrosion vol. 70, No. 6.
 Andrade C. (2017) -Reliability analysis of corrosion onset: initiation limit state- Journal of
Structural Integrity and Maintenance- Nov Vol. 2, no. 4, 200–208.
https://doi.org/10.1080/24705314.2017.1388693
 DAfStb Heft 601 Durability design of reinforce concrete structures for reinforcement
corrosion – Part 1: System relevant parameter of reinforcement corrosion.
(Dauerhaftigkeitsbemessung von Stahlbetonbauteilen auf Bewehrungskorrosion – Teil 1:
Systemparameter der Bewehrungskorrosion) Authors: M. Beck, A. Burkert, A. Faulhaber,
C. Gehlen, J. Harnisch, B. Isecke, J. Lehmann, K. Osterminski, M. Raupach, P. Schießl, W.
Tian, J. Warkus.
 DAfStb Heft 602 Durability design of reinfoced concrete structures for reinfocement
corrosion– Part 2: Durability design (Dauerhaftigkeitsbemessung von Stahlbetonbauteilen
auf Bewehrungskorrosion: Teil 2: Dauerhaftigkeitsbemessung) Authors: E. Bohner, C.

For citations, copying and referencing see Note on Front page! page 202
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Final draft (August 2022)

Fischer, C. Gehlen, H. S. Müller, K. Osterminski, J. Ožbolt, P. Schießl, S. von Greve-

Dierfeld
 CEN/TS 17440 2020, Technical Specification – Assessment and retrofitting of existing
structures, European Committee for Standardization, CEN, Brussels, July 2020.
 CONTECVET –(2001) A validated user’s manual for assessing the residual life of concrete
structures, DG Enterprise, CEC. (The manual can be downloaded from the web site of
http://www.ietcc.csic.es/index.php/es/publicaciones-2/manual-contecvet
 EN 1992-1-1:2004 Eurocode 2: Design of concrete structures - Part 1-1: General rules and
rules for buildings
 Gehlen C.(2000) Probabilistische Lebensdauerbemessung von Stahlbetonbauwerken-
Zuverlässigkeitsbetrachtungen zur wirksamen Vermeidung von Bewehrungskorrosion“,
PhD
 Izquierdo D. (2001): Basis of design for a probabilistic treatment of reinforcement corrosion
processes- PhD. Polytechnic University of Madrid. Faculty of Civil Engineering
 Izquierdo D., Alonso C., Andrade C., Castellote M. (2004) “Potentiostatic determination of
chloride threshold values for rebar depassivation. Experimental and statistical study”.
Electrochimica Acta 49 pp. 2731–2739.
 Izquierdo D. Andrade C. (2009) CALIBRATION of models for establishing the durability limit
state)- 7th colloquia of managers and technicians of cement plants November -MÁLAGA –
SPAIN
 Izquierdo D., Andrade C., Arteaga A., Tanner P. (2003) Economical optimization of
appraising concrete degradation processes3rd International IABMAS Workshop on Life –
Cycle Cost Analysis and Design of Civil Infrastructure Systems and the fib WP 5.3-1 ,TG 5.6
and the Joint Committee on Structural Safety Workshop on Probabilistic Modelling of
Deterioration Processes in Concrete Structures- Lausanne- Switzerland- March 24-26.
 Izquierdo D., Andrade C., Arteaga A. (2000) - Risk analysis of crack wit¡dth limit state due
to reinforcement corrosion- International RILEM Workshop on Life Prediction and Aging
Management of Concrete Structures. Paris. Dan Naus Editor. Rilem Publications
 JCSS (2000), JCSS probabilistic model code Part 1: Basis of design. Joint Committee on
Structural Safety.
 MC2010 fib Model Code for Concrete Structures 2010
 prEN 1990:2020 (E), Eurocode – Basis of structural and geotechnical design, CEN TC 250
Enquiry Draft, European Committee for Standardization, CEN, Brussels, September 2020.
 prEN 1990-2:2020, Eurocode – Basis of assessment and retrofitting of existing structures:
general rules and actions, CEN TC 250/WG2/WG2.T2 Project Team Version, European
Committee for Standardization, CEN, Brussels, September 2020.
 Tanner P, Lara C, Bellod JL, Sanz D. (2019) “The plastic cathedral”: Innovation to extend the
service life of a heritage structure. Structural Concrete.; 1–16.
https://doi.org/10.1002/suco.201900365
 Tuutti, K. (1982) “Corrosion of steel in concrete”, Swedish Cement and Concrete Institute
(CBI) nº 4-82. Stockholm.

For citations, copying and referencing see Note on Front page! page 203
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Background Document for EN1992 section 6

Final February 18, 2021

François Toutlemonde

Contribution to Part 3:

Models for chlorides-induced corrosion (XRDS)

3.3 Deterministic approach using margins and characteristic values

3.3.1 XRDS class designation and rationale of associated model assumptions

The class designation for resistance against corrosion induced by chloride ingress (XRDS) is derived
from the depth of chlorides penetration in mm (characteristic value 90 % fractile), corresponding to a
reference chlorides concentration (0,6 % by mass of cement + type II additions), assumed to be
obtained after 50 years on a concrete exposed to one-sided penetration of reference seawater (30 g/l
NaCl) at 20°C. Being based on a depth in mm, it can be related for calibration to objective measures of
chlorides profiles without further model interpretation. Yet for an easier interpretation in relation to
a material property, easier implementation of intermediate classes, and avoidance of numerical
confusion with another property, it has been decided that XRDS should have the dimension of a
diffusion coefficient (10-13 m²/s).
The dimension of a diffusion coefficient is the square of a length divided by a time. The length given in
the definition is the depth of chlorides penetration in mm (characteristic value 90 % fractile), noted dCl,
corresponding to the reference 0,6% chlorides concentration (by mass of cement + type II additions),
assumed to be obtained after 50 years in the reference conditions. The time t50 given in the definition
is 50 years. Thus, XRDS is to be proportional to dCl²/t50 and the only arbitrary choice is the coefficient
of proportionality.
The reason of having chosen a coefficient equal to ¼, which leads to XRDS = dCl²/631 since t50 is
expressed in seconds and dCl in mm, so that XRDS is in 10-13 m²/s, can be justified as detailed hereafter.

The mathematical description generally used as a reference for describing the distribution of the
chloride content in the cover concrete due to chloride diffusion, following e.g. fib Bulletin 34, fib Model
Code 2010 or ISO 16204, derives from the solution of Fick’s second law. The corresponding equation
reads:
C(x,t) = Ci + (Cs – Ci) [erfc {(x – Δx) / 2 (Dapp(t). t)1/2}]
In this equation, x represents the considered depth within the cover concrete and t the duration of
exposure to the chlorides ingress, erfc is the complementary Gauss error function, Dapp(t) is the
apparent chloride diffusion coefficient, Ci is the initial chloride content. In saturated conditions such as
referred to in the XRDS definition, the “convection depth” Δx (which is sometimes also interpreted as
a surface layer where carbonation can take place and disturb chlorides binding) is null. Cs is classically

For citations, copying and referencing see Note on Front page! page 204
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

designated as the “surface concentration”, although this peak chloride concentration is not directly
obtained as if the concrete porosity was filled by the surrounding seawater.

If we assume that the profile corresponding to the 90% fractile of chloride content follows the same
mathematical form, then we can write for t = t50:
0,6 % = Ci + (Cs – Ci) [erfc {dCl / 2 (Dapp(t50). t50)1/2}]
Having defined XRDS = ¼ dCl²/ t50 leads to 0,6 % = Ci + (Cs – Ci) [erfc {XRDS/Dapp(t50)}1/2]
We can then interpret XRDS as the apparent diffusion coefficient at 50 years (characteristic value,
namely corresponding to the 90% fractile profile), which reads XRDS/Dapp(t50) = 1, under certain
combinations of Cs and Ci values.

The reference combination corresponds to chloride concentrations (expressed as a weight ratio to the
cement + type II additions content) equal to Ci = 0,05 % (realistic situation for concrete mixes meeting
the minimum chloride class requirement) and Cs = 3,55 % (which is in line with typical values obtained
from field investigations, e.g. on structures exposed to Atlantic water).
Alternative examples of combinations compatible with interpretation of the XRDS designation as
representative of the apparent diffusion coefficient include:
Ci = 0 (only the propagation of chlorides from outside is considered) and Cs = 3,81 % (by mass of cement
+ type II additions) which is the solution of 0,6 % / Cs = erfc(1) = 0,157.
Ci = 0,1 % (limiting value for chloride class 0,1) and Cs = 3,28 % (by mass of cement + type II additions)
which is the solution of 0,5 % / (Cs – 0,1) = erfc(1) = 0,157.
Ci = 0,2 % (limiting value for chloride class 0,2) and Cs = 2,74 % (by mass of cement + type II additions)
which is the solution of 0,4 % / (Cs – 0,2) = erfc(1) = 0,157.

Moreover, if a surface concentration Cs = 3,3 % is assumed and Ci = 0 is considered (to focus on the
external chloride diffusion only), XRDS/Dapp(t50) is obtained as the solution of the following equation:
0,6 % / 3,3 % = erfc({XRDS/Dapp(t50)}1/2) = 0,1818, thus {XRDS/Dapp(t50)}1/2 = 0,945 leading to
XRDS/Dapp(t50) = 0,893 or Dapp(t50) / XRDS = 1,12
If a surface concentration Cs = 3,9 % is assumed and Ci = 0,1 % is considered, XRDS/Dapp(t50) is obtained
as the solution of the following equation: 0,5 % / 3,8 % = erfc({XRDS/Dapp(t50)}1/2) = 0,1316, thus
{XRDS/Dapp(t50)}1/2 = 1,065 leading to XRDS/Dapp(t50) = 1,134 or Dapp(t50) / XRDS = 0,882

In sum, XRDS = 1/4 dCl²/t50 (which reads XRDS = dCl²/631 with the proper units) is a direct
correspondence between the exposure resistance class designation and the chloride depth. It
corresponds to the characteristic value of the apparent chloride diffusion coefficient at 50 years under
a realistic reference set of assumptions, and in any case leads to XRDS values of the same order of
magnitude as Dapp(t50) for other realistic sets of assumptions. As such, it includes the progressive ageing
in the diffusion process over the 50 years-expected service life.

Presently used mixes generally correspond to depths in chloride penetration under the reference
conditions of XRDS definition, ranging from typically 15 mm (in case of very high-performance
concrete) to up to about 80 mm (in fact for concrete types to be used only in XD1 / XS1 environment),
depending on the severity of XD/XS exposure and on the design service life. This leads to XRDS between
0,5 and 10 which is deemed convenient for class designation and corresponds to recommended values
in Table 6.4. If relevant in the place of use, chloride diffusion in non-saturated conditions for XS1/XD1
might make sense to be considered, it would then turn out possible to include additional classes
corresponding to XRDS15, XRDS20, XRDS25 and XRDS30 corresponding to dCl values equal to 97,3;
112,4; 125,6 and 137,6 mm respectively – as detailed hereafter.

For citations, copying and referencing see Note on Front page! page 205
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

3.3.2 Initiation and propagation times

In the approach detailed hereafter, where computations are deterministic while input values include
characteristic values and margins, the limit state function considered is the random variable G(X):
G(X) = tini + tprop - tSL
With tSL = design service life (50 or 100 years)
tini = time until reinforcement depassivation / initiation of corrosion
tprop = time for (acceptable) propagation of corrosion
The design service life is conventional (fixed in reference to EN 1990) and used as deterministic in the
calibration process described hereafter.

The time for propagation of corrosion depends on availability of water and oxygen, which is deemed
to be described by the (environmental) exposure classes. It may also depend on the cracking of the
cover zone, which is limited depending on the (environmental) exposure class (SLS verification
according to EN 1992-1-1, section 9). In a first approach however, it has not been considered as
depending on the concrete itself. Thus, variance of the G(X) variable is obtained as the quadratic sum
of variances of tini and tprop variables because the sources of deviation are independent. In the following
t*prop will be used as “safe” estimate of acceptable tprop and the fractile estimation of tini– (tSL – t*prop)
will be considered for quantifying the reliability index (β) in relation to its target value.

In the case of chlorides ingress, pitting corrosion is the main harmful corrosion mechanism and its
quantitative development with time is not subject of well-established knowledge. Some authors have
attempted to build an equivalence between pitting depth development with time and an equivalent
loss of rebars diameter, so that tprop can be defined from laboratory measurements of corrosion rate.
This approach may be questioned due to the non-constant corrosion rate when chlorides reach the
reinforcement. Therefore, the following times for corrosion propagation t*prop have been considered
from experts’ opinion as characteristic values for the acceptable duration of reinforcement corrosion
propagation under XD and XS classes. These values are somewhat larger than what would derive from
established corrosion current intensity measurements since a time of corrosion “initiation”
corresponding to low-current is included. They however turn out safe with respect to estimated
propagation times for realistic concrete mixes studied in the French national PerfDuB project (El Farissi,
2020).

XS1 XS2 XS3 XD1 XD2 XD3

t*prop (years) 10 10 2 15 5 5

Under XS3 exposure (tidal and splash zone) the acceptable time for corrosion propagation is almost
negligible due to daily supply of alternatively chlorides, water and oxygen. This time is higher in
submerged conditions (XS2) due to reduced oxygen availability, or at a higher distance from seawater
droplets (XS1) due to reduced chlorides concentration. Moreover, under XS1 conditions, the cover
concrete may not be constantly saturated, which significantly slows down the diffusion process.
Average lower corrosion propagation rates have been considered for XD3 (resp. XD1) due to the
seasonal supply of de-icing salts, as compared to XS3 (resp. XS1). In the case of XD2, often associated
to pools / industrial waters and liquids, a safe low value has been considered due to difficult inspection

For citations, copying and referencing see Note on Front page! page 206
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

and possible concomitant chemical mechanisms accelerating the development of corrosion.

Moreover, it has been assumed that the corrosion rate does not primarily depend on the cover nor on
the concrete type; thus t*prop has been taken identical whatever the XRDS class under calibration. This
assumption may be penalizing concrete mixes with higher resistivity (e.g. depending on the cement /
binder type).

3.3.3 Diffusion of chlorides – influence of exposure conditions

The model used for describing the distribution of the chloride content in the cover concrete derives
from the simplified description of the Fick’s second law solution in a saturated porous material. It is
assumed to represent a stabilized profile after some years, following e.g. fib Bulletin 34. The
corresponding equation reads:

C(x,t) = Ci + (Cs – Ci) [erfc {(x – Δx) / 2 (Dapp(t). t)1/2}]

In this equation, x represents the considered depth within the cover concrete and t the duration of
exposure to the chlorides ingress, erfc is the complementary Gauss error function, Dapp(t) is the
apparent chloride diffusion coefficient, Ci is the initial chloride content.

Under XS1/XD1 and possibly also XS3/XD3 exposure classes, the chlorides ingress may take place in a
non-saturated material due to concomitant exposure to rainfall and drying. Therefore, the validity of
the equation is limited to x higher than the “convection depth” Δx (which is sometimes also interpreted
as a surface layer where carbonation can take place and disturb chlorides binding). Cs is then
interpreted not as the classically designated “surface concentration”, but as the peak chlorides
concentration which is observed for x = Δx. This description has been agreed in fib MC 2010 and used
e.g. by Rahimi (2020). From field observations, the following values have been assumed for Δx.

Δx (mm)
Design service life (years) XS1 XS2 XS3 XD1 XD2 XD3
50 5 0 8 5 0 8
100 5 0 10 5 0 10

In saturated conditions, the peak / surface concentration Cs can be computed as the product of the
chloride concentration of the surrounding liquid (e.g. seawater in XS2), supposed to be filling the whole
porosity of the material accessible to water, multiplied by an “enrichment factor” which accounts for
evidence of a higher concentration especially in high performance concrete (El Farissi, 2020). This
enrichment factor typically ranges from 2 to 3,5 and decreases with concrete porosity. Since a fix value
has been considered for Δx, this assumption is still safe even for very high-performance concrete
exhibiting an apparent very high enrichment factor, although a moderate modulation has been
considered for this enrichment factor and for concrete porosity depending on the XRDS class, so that
the peak Cs values variations are limited to less than 10% with a typical content of 3,2% of cement +
type II additions.

Due to uncertainty related to XD2 in terms of surrounding chlorides concentrations, in a first approach
the same Cs and Ci chlorides contents have been assumed as in the reference equation corresponding
to XS2. In tidal or splash zones (XS3), where the convection zone is submitted to drying and wetting

For citations, copying and referencing see Note on Front page! page 207
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

cycles, the peak of the profile, which is shifted at the edge of this zone, corresponds to an “enriched”
chlorides content typically equal to 1,2 times the value taken for the XS2 condition. Although fib
Bulletin 76 would advise not to consider it, this choice has been based on typical average field data
and is intended to compensate the deviations between effective chloride profiles, as measured, and
representation of the tail of the curve corresponding to an equivalent diffusion process.

In the case of XD3 exposure, with the typical situation of parts of civil structures exposed to deicing
salts in winter such as in Alpine / mountain regions of Central Europe, the cycles associated to deicing
salts are non-permanent, and the salts concentration should be lower than in XS3, provided water
runoff is correctly ensured at the surface of the member. It should however be noted from recent
chlorides content measurements on a 40 years-old Austrian retaining wall (Binder et al., 2019), that
modelling the chloride profile as stationary is far from observations: location of the peak of the
chlorides content varies from 1 cm-deep in winter to roughly 3 cm-deep in summer, with peaks of
about 10 kg/m3 (related to the concrete) in winter and 6 kg/m3 in summer. These concentrations are
to be compared to sea water (21 g/l for chloride ions) which at the concrete surface (porosity estimated
as 18%) would lead to 4 kg/m3. For the sake of simplicity, it has been considered a peak chloride
concentration in XD3 equal to 1,1 the one corresponding to XS2, and it has also been considered that
the same ratio relating the peak chlorides concentrations applies to x, the depth corresponding to the
reference chloride content. This is not mathematically exact but the order of magnitude is quite
correct.

Finally, in XD1 and XS1 conditions, the salt concentration at the member surface has been taken as half
of the one under XS2 conditions, which accounts for e.g. dilution by rainwater. For the sake of
simplicity, it has also been considered that the same ratio applies to x, the depth corresponding to the
reference chloride content, as the ratio associated to the chloride concentration. This is not
mathematically exact but the order of magnitude is quite correct.

This influence of the exposure conditions on the boundary conditions of the diffusion model is thus
summarized as follows:

XS1 XS2 XS3 XD1 XD2 XD3

Peak chloride Cs/2 Cs 1,2 Cs Cs/2 Cs 1,1 Cs
concentration

Noticeably enough, determination of the chloride profiles from real structures exhibits a high scatter
in the peak concentration, even in the same exposure class, which may depend on local variations of
exposure to rainfall, local humidity and wind exposure, local concrete porosity etc. It has not been
found realistic to use statistical parameters and characteristic values from these experimental
determinations of chloride profiles, which mainly cover XS3 situations and could be controversial.
Taking account of these variations has been addressed by taking conservative values for the ratios XS3
vs. XS2, XS1 vs. XS2, XD3 vs. XS2, and XD1 vs. XS2 peak concentrations, and for the enrichment factor
leading to Cs.

For citations, copying and referencing see Note on Front page! page 208
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

3.3.4 Complementary physical assumptions associated to XRDS classes

The designation of ERCs for resistance against corrosion induced by chloride ingress (XRDS) is derived
from the depth of chlorides penetration in mm (characteristic value 90 % fractile), corresponding to a
reference chlorides concentration (0,6 % by mass of cement + type II additions), assumed to be
obtained after 50 years on a concrete exposed to one-sided penetration of reference seawater (30 g/l
NaCl) at 20°C. Having the dimension of a diffusion coefficient (10-13 m²/s), it is calculated as the
rounded value of 1/631th of the square of this depth of chlorides penetration in mm, which
corresponds to verifying x (depth) / [2 (XRDS. t (50 years))1/2] = 1

Under these reference conditions, Δx = 0 thus Dapp(t) can be related to XRDS, Ci, the reference chloride
concentration (0,6 % by mass of cement + type II additions), and Cs. However, as indicated previously,
this relation is not unique. It is thus necessary to precise assumptions on porosity and binder content
that have been made in order to express Dapp(t) in relation to XRDS, for further application of the
model.

Presently used mixes in XD/XS environment correspond to depths of chloride penetration under the
reference conditions associated to XRDS definition, ranging from typically 15 mm (in case of very high-
performance concrete) to up to about 80 mm (in fact for concrete types to be used only in XD1 / XS1
environment), depending on the severity of XD/XS exposure and on the design service life. Therefore,
in order to establish the recommended provisions, computations have been done for XRDS classes
ranging from XRDS0,3 to XRDS11. Since resistance to chlorides ingress is positively correlated to
concrete compactness (namely, chlorides diffusivity is positively correlated to porosity), and due to
present deemed-to-satisfy provisions for concrete mixes employed in XD/ XS environments, it has been
considered that the cement + type II additions content vary from typically 280 to 450 kg/m3, and the
water to binder ratio also typically varies from 0,5 to below 0,3 for the range of concrete types
considered, leading to porosity ranging from 12 to 15 %. The resulting correlated material parameter
sets have been considered. Interestingly, the surface chlorides concentration in the reference XS2
conditions slightly varies from 3 to 3,3 % of the binder content, which is considered a safe estimate,
since values representative of the minimum binder contents according to present standards have been
accounted for. Namely, this corresponds to typical peak chlorides contents of 12 to 13 kg/m3 in
submerged concrete, which can be considered a reasonable design value.

Being fixed the depth corresponding to XRDS definition and the reference chloride concentration, the
reference 50 years duration of the diffusion process, and Cs, Dapp can be derived with an upper bound
value considering that the initial chlorides concentration is null, although this is never strictly true (the
diffusion process is maximized). Inversion of the erfc function thus leads to the following Dapp values,
which can be interpreted as a characteristic value (90 % fractile) of the apparent diffusion coefficient
of the concrete at 50 years (namely, chlorides have diffused more slowly or until a lower depth in 90%
of the cases). It can be noted that based on this set of assumptions, Dapp/XRDS varies between 1,11
and 1,22. This ratio exhibits low variability and is close to the 1,0-value referred to in the XRDS
definition, the reason for this being the set of assumptions corresponding to Cs, Ci and the influence
of the porosity and cement + type II addition content considered for the Dapp determination.

For citations, copying and referencing see Note on Front page! page 209
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

XRDS Chloride Binder (cement + Porosity Enrichment Cs (XS2) Cs (XS2) Dapp at 50

class depth type II addition) factor years
content
(mm) (kg/m3) (%) (kg/m3) (% 10-13 m²/s
binder)
0,3 13,8 450 12 3,75 13,5 3 0,365
0,5 17,8 435 12,5 3,5 13,125 3 0,606
0,8 22,5 420 13 3,25 12,675 3 0,97
1 25,1 410 13,3 3,1 12,369 3 1,213
1,5 30,8 400 13,5 3 12,15 3 1,811
2 35,5 390 13,7 2,9 11,919 3,1 2,394
2,5 39,7 380 13,8 2,85 11,799 3,1 2,954
3 43,5 370 13,9 2,8 11,676 3,2 3,484
3,5 47 360 14 2,75 11,55 3,2 4,012
4 50,2 350 14,2 2,7 11,502 3,3 4,508
5 56,2 340 14,3 2,6 11,154 3,3 5,635
5,5 58,9 330 14,4 2,5 10,8 3,3 6,198
6 61,5 320 14,5 2,4 10,44 3,3 6,805
7,5 68,8 310 14,7 2,3 10,143 3,3 8,452
8 71,1 300 14,75 2,25 9,95625 3,3 8,939
9 75,4 300 14,8 2,2 9,768 3,3 10,207
10 79,4 290 14,9 2,15 9,6105 3,3 11,174
11 83,3 280 15 2 9 3,2 12,61
15 97,3 280 15 2 9 3,2 17,195
20 112,4 280 15 2 9 3,2 22,926
25 125,6 280 15 2 9 3,2 28,658
30 137,6 280 15 2 9 3,2 34,39

In the XRDS definition, the reference chlorides concentration (0,6 % by mass of cement + type II
additions) has been selected as possibly related to the threshold for initiation of the rebars corrosion,
noted Cth, although this concept is highly debated and its quantitative determination quite variable,
including expressing this threshold as a ratio with respect to the binder content. With this definition,
discussions have arisen concerning the interpretation of the chlorides content threshold, the role of
the initial chlorides content, distinction between free and total chlorides content, etc. In a general
approach for calibration it is not possible to comply with all cases. Following (Angst et al., 2009) and
(Cao et al., 2019) a chloride content equal to 0,6 % by mass of cement + type II additions has been
deemed as a relatively low value for representing the threshold triggering the corrosion propagation,
noting the fact that the definition has been associated to total chlorides content.

Moreover, based on these data, it has been noted that this threshold tends to be lower for concretes
with a high porosity (high water to binder ratio) and higher for high performance concrete. It has also
been noted that the threshold would be lower in situations of salty water available (XS2, XD2, XS3,
XD3) and higher in non-constantly saturated situations representative of XD1 /XS1 (Angst et al., 2009).
Erring on the safe side, it has been considered that for the range of XRDS classes of concrete under
consideration, the thresholds would range from 0,5 to 0,7 % by mass of binder in XD1 / XS1 classes,
depending on the concrete porosity, and from 0,4 to 0,6 % by mass of binder in XS2, XD2, XS3, XD3
environments, also depending on the concrete porosity. The following chloride thresholds Cth were
thus considered as corresponding to the end of initiation period. Knowing Dapp and Cs in the reference

For citations, copying and referencing see Note on Front page! page 210
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

conditions, the depth xth corresponding to a probability of 90% not having reached this chloride
content threshold can be calculated by inverting erfc, as an intermediate indicator.

XRDS class Cth Cth xth Cth Cth xth

(XS1/XD1) (XS1/XD1) (XS2/XD2 (XS2/XD2
XS3/XD3) XS3/XD3)
(% binder) (kg/m3) (mm) (% binder) (kg/m3) (mm)
0,3 0,7 3,15 12,8 0,6 2,7 13,8
0,5 0,69 3 16,6 0,59 2,57 17,9
0,8 0,68 2,86 21,2 0,58 2,44 22,8
1 0,67 2,75 23,9 0,57 2,34 25,7
1,5 0,66 2,64 29,5 0,56 2,24 31,7
2 0,65 2,54 34,2 0,55 2,15 36,8
2,5 0,64 2,43 38,6 0,54 2,05 41,5
3 0,63 2,33 42,5 0,53 1,96 45,7
3,5 0,62 2,23 46,3 0,52 1,87 49,8
4 0,61 2,14 49,9 0,51 1,79 53,4
5 0,6 2,04 56,2 0,5 1,7 60,4
5,5 0,58 1,91 59,7 0,48 1,58 64,3
6 0,56 1,79 63,4 0,46 1,47 68,2
7,5 0,54 1,67 71,6 0,44 1,36 77,3
8 0,53 1,59 74,5 0,43 1,29 80,4
9 0,52 1,56 79,6 0,42 1,26 86,1
10 0,515 1,49 84,4 0,415 1,2 91
11 0,5 1,4 89,4 0,4 1,12 97
15 0,5 1,4 104,4 0,4 1,12 113,2
20 0,5 1,4 120,5 0,4 1,12 130,8
25 0,5 1,4 134,8 0,4 1,12 146,2
30 0,5 1,4 147,6 0,4 1,12 160,1

3.3.4 Minimum cover and reliability format

The equation describing the chloride profile is to be applied in the different exposure conditions in
order to determine the depth x corresponding to reaching the chloride threshold Cth at tini = tSL – t*prop:

C(x,t) = Ci + (Cs – Ci) [erfc {(x – Δx) / 2 (Dapp(t). t)1/2}]

In this application, Cs, Δx and t*prop are varied depending on the exposure class and design service life
as discussed before. While 20°C has been considered as a reference temperature in the XRDS
definition, it has been deemed that a more realistic calibration should consider an average 15°C
temperature. Based on e.g. (Jones et al., 1995) and further works, the resulting diffusion coefficient
has been multiplied by 0,75, leading to corresponding chloride depths multiplied by 0,866.

In the approach adopted, only chlorides transferred from the environment have been considered,
while the concrete mixes have been assumed to meet current requirements in terms of initial chlorides
content. In fact it is meant that the chloride class requirements are fulfilled and that aggregate shall

For citations, copying and referencing see Note on Front page! page 211
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

meet the standards requirements imposing to be washed, in case of marine origin. Only considering
chlorides brought by external transfer would formally correspond to Ci = 0 in the diffusion equation.
The underlying assumption is that the chlorides initially present in concrete (which are limited by the
chloride class) are mainly bound and do not contribute to the corrosion risk associated to the Cth
content, which has been taken from the lower bound estimate of the corrosion threshold ranges
expressed in terms of total chloride content.

As an equivalence to considering the limit sate function associated to time, the reliability associated to
the durability limit state (target Beta index) has been evaluated as the probability of the specified
minimum cover for durability cmin,dur to be exceeded by the progress of the chlorides ingress above the
considered threshold. Namely, this relies on the definition that cmin,dur = cnom - Δc,dev is a limiting
minimum value in the sense of the execution standard complemented by the provisions valid in the
place of use, so that the probability of having reinforcement at a distance to the surface lower than
cmin,dur is negligible. With this assumption, estimating the probability of reaching the durability limit
state is done without considering the real distribution of rebars location1 around cnom. Subsequently,
the target reliability is thus assumed to be reached when the probability of having a chloride depth
(depth xCl where the chlorides concentration exceed the corrosion threshold) in excess of cmin,dur at
time tSL - tprop is lower than the target reliability.

With a target β equal to 1,3 and an assumed normal distribution this could read:

P(cmin,dur < xCl(tSL – t*prop)) < 10%

Deriving xCl from the equation above and adopting it as the recommended cover depth means that this
depth is overpassed by the threshold value of the chlorides profile with a 10 % probability. In fact, the
cover concrete in the considered structure belongs to the ERC considered, namely given the “regular”
curing process (that shall be described in EN 13670) 90 % of it would have a chloride content lower
than the reference value in the reference environment. The distribution thus includes scatter due to
production and execution. The t*prop has been taken as a safe (low) estimate. And, as detailed before,
the other assumptions taken to represent the climate influence are assumed to introduce no unsafe
bias, although they can hardly be reliably attached to a precise fractile.

In order to extrapolate from 90 % quantile (β = 1.3) to a 93,6 % quantile (corresponding to β = 1,5), an

additional margin has to be introduced. However, field data corresponding to chlorides profiles are
not available equally in all exposure conditions to derive statistical figures. For a given concrete, it is
considered that the chlorides diffusion process is relatively homogeneous in saturated conditions (XS2
and XD2) and in case of regular periods of saturation (XS3 and XD3) and limited influence of
carbonation. For β = 1,5, it has thus been deemed that effects of scatter in the material properties
could be accounted for with an additional 5 mm margin in the minimum cover.

1
Noticeably enough, no consideration has been given to the “real” location of the reinforcement,
accounting of cnom instead of cmin,dur: especially, no assumption has been deemed as having to be
included concerning the margin for deviation in the rebars placement, except related to the definition
of cmin,dur = cnom – Δc,dev being the lower limiting value for rebars location, because the statistical
distribution of the margin in excess of cmin,dur is highly variable depending on the construction
conditions and production control, which would make a statistical modelling uncertain.

For citations, copying and referencing see Note on Front page! page 212
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

In non-saturated conditions however with probable concomitant carbonation (XS1 and XD1),
uncertainties in the chlorides transfer process motivate taking a higher margin. This margin should
range from 5 to 10 mm depending on the concrete quality, an empirical formula for this extra-depth
in mm has been taken as 0,04 times the square of the concrete porosity (%). Calibration of the
reliability enhancement associated to this extra cover could however deserve additional checking.

When applying the model describing the chloride concentration at the end of the initiation time,
namely tSL – t*prop, the apparent diffusion coefficient considered should be affected by the ageing
factor. Namely, it is empirically admitted that the chlorides profile follows the time and space
dependency according to :
[erfc {(x – Δx) / 2 (Dapp(t). t)1/2}],
with Dapp = D0 (t0/t)a
with the ageing factor α typically ranging from 0,3 to 0,7 depending on the binder type.

Assuming α = 0, namely considering the assumption of a “stabilized profile” without ageing, would
have resulted in extrapolating the definition as follows: for the times to initiation tSL – t*prop considered,
namely from 35 to 98 years, the depth x(tSL – t*prop) corresponding to the reference chloride content
evolves as the square root of time with respect to the reference x(50 years). This has been considered
as excessively severe, although effective evidence of ageing after about 30 years of age appears as
limited. Therefore, an α = 0,2 value has been considered for derivation of cmin,dur values in the different
exposure conditions and design service lives, which corresponds to a variation of x – Δx as a 0,4 time-
power law when extrapolating from the XRDS definition at 50 years to determine the xCl value.

As an illustration of the whole computational sequence for the determination of raw cmin,dur values, the
following examples are given:

XS2 – XRDS3 – 50 years

XRDS3 corresponds to dCl=43,5 mm. With a 0,53 % binder critical chloride concentration and a Cs value
equal to 3,2 % binder, xCl corresponding to reaching the threshold at 20°C in reference conditions is
45,7 mm. Considering the reduction of diffusion due to the average temperature and the 10 years for
propagation, this depth comes to 36,2 mm. The additional 5 mm-margin leads to 41 mm.

XD3 – XRDS1,5 – 100 years

XRDS1,5 corresponds to dCl=30,8 mm. With a 0,56 % binder critical chloride concentration and a Cs
value equal to 3 % binder, xCl corresponding to reaching the threshold at 20°C in reference conditions
is 31,7 mm. Considering the reduction of diffusion due to the average temperature it comes to
27,5 mm. Then a 10 mm-convection depth is considered, a 10 % increase in the peak concentration,
and a time extrapolation from 50 to 95 years (considering tprop = 5 years). This depth comes to 46,5 mm.
The additional 5 mm-margin leads to 52 mm.

A selection of obtained recommended cover thickness raw values is given hereafter and has been used
together with full-probabilistic modelling approaches to draft recommended provisions at the
European level.

For citations, copying and referencing see Note on Front page! page 213
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Resulting Cmin,dur recommended cover thickness raw values (mm)

versus the risk of corrosion induced by chlorides corresponding to β = 1,5 target reliability

Exposure class XS1 XS2 XS3

Design service life (years) 50 100 50 100 50 100
XRDS 0,5 17 20 19 25 31 37
XRDS 1 21 25 25 33 39 47
XRDS 1,5 24 28 30 40 45 56
XRDS 2 26 31 34 45 51 62
XRDS 3 29 36 41 55 60 75
XRDS 4 32 40 47 63 68 85
XRDS 5 35 44 53 71 75 95
XRDS 6 38 48 59 80 83 105
XRDS 8 43 55 69 93 95 122
XRDS 10 47 60 77 105 106 136

Exposure class XD1 XD2 XD3

Design service life (years) 50 100 50 100 50 100
XRDS 0,5 17 20 18 22 29 34
XRDS 1 20 25 24 30 37 44
XRDS 1,5 23 28 28 36 42 52
XRDS 2 25 31 31 41 47 58
XRDS 3 28 35 38 49 55 69
XRDS 4 31 40 43 57 62 78
XRDS 5 34 43 48 64 68 87
XRDS 6 37 47 54 71 75 96
XRDS 8 41 54 63 83 86 111
XRDS 10 45 59 70 93 96 124

3.3.5 Rounding, sensitivity and smooth continuity considerations

For practical implementation of recommended cover depths values that should be multiple of 5 mm,
a safe rounding process has been adopted, as follows:
- for recommended depths ending by 7 to 1, value rounded to the next number ending by 0.
- for recommended depths ending by 2 to 6, value rounded to the next number ending by 5.
Moreover, cmin,dur values higher than 80 mm have been considered as oversensitive to model
assumptions, with limited practical feedback. They would have led to unsatisfactory provisions for
thick unreinforced cover concrete. Therefore, it has been decided to indicate “N.A.” (not applicable) in
the corresponding cells.

In previous calculations, the diffusion coefficient has been taken identical in any of the exposure
classes, corresponding to chlorides diffusion in a saturated porous medium. This assumption is highly
questionable for XD1 and XS1 conditions, especially when interpretation of the limit between these
classes and XS3/XD3 classes (typically resulting from national interpretation in guidance for selecting
exposure classes) clarifies that splash of salty water droplets is excluded in the former ones. Additional
calculations have been done, considering a diffusion coefficient multiplied by 0,377 = 0,856, namely
corresponding to diffusion within concrete pores at a probable 85 % relative humidity (close to the
equilibrium associated to self-desiccation), following models detailed in (Buchwald, 2000) and (Mai-

For citations, copying and referencing see Note on Front page! page 214
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Nhu, 2013). Additional classes corresponding to higher apparent diffusion coefficients appear as
relevant in the continuity of presently accepted cover depths and mixes.
It has also been noticed that for many industrial facilities corresponding to XD2 classification, the
chlorides concentration of the brine is much lower than seawater. An alternative computation has thus
been performed assuming Cs/2.

Taking these assumptions and the rounding process into consideration, and ensuring that XD2
provisions keep stricter than XD1 (light grey cells), leads to the following table of recommended cover
depths. Light blue cells correspond to different provisions between XD1 and XS1, and between XD3
and XS3, respectively. Figures in red correspond to presently recommended cmin,dur values as from Table
4.4N of Eurocode 2 (structural classes S4 for 50 years-service life and S6 for 100 years-service life,
respectively).

Exposure class XS1 XS2 XS3

Design service life (years) 50 100 50 100 50 100
XRDS 0,5 15 20 20 25 30 40
XRDS 1 20 20 25 35 40 50
XRDS 1,5 20 25 30 40 45 55
XRDS 2 20 25 35 45 50 65
XRDS 3 25 30 40 50 60 75
XRDS 4 25 30 50 65 70 NA
XRDS 5 25 35 55 70 75 NA
XRDS 6 30 35 60 NA NA NA
XRDS 8 30 40 70 NA NA NA
XRDS 10 35 45 NA NA NA NA
XRDS 15 40 50 NA NA NA NA
XRDS 20 45 55 NA NA NA NA
XRDS 25 45 60 NA NA NA NA
XRDS 30 50 65 NA NA NA NA

Exposure class XD1 XD2 XD3

Design service life (years) 50 100 50 100 50 100
XRDS 0,5 15 20 15 20 30 35
XRDS 1 20 20 20 20 40 45
XRDS 1,5 20 25 20 25 45 55
XRDS 2 20 25 20 25 50 60
XRDS 3 25 30 25 30 55 70
XRDS 4 25 30 25 30 65 NA
XRDS 5 25 35 30 35 70 NA
XRDS 6 30 35 30 40 75 NA
XRDS 8 30 40 35 45 NA NA
XRDS 10 35 45 40 50 NA NA
XRDS 15 40 50 45 60 NA NA
XRDS 20 40 55 55 70 NA NA
XRDS 25 45 60 60 75 NA NA
XRDS 30 50 65 65 NA NA NA

Additional considerations related to the modelling sensitivity are given hereafter:

For citations, copying and referencing see Note on Front page! page 215
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

- Chlorides ingress model: the Fick’s second law associated to an ideal diffusion model is known to be
simplistic and reasonably valid only for the inner tail of the chlorides profile. The present approach
includes an enrichment factor and a convection depth, which is in line with up-to-date state-of the art.
However, calibration of the associated input is debatable and high variability of field data e.g. for peak
concentration is of clear evidence. Moreover, interaction of chlorides with the concrete depending on
the binder type will significantly affect the effective critical chloride content for corrosion, which can
hardly be considered at the design step when the concrete mix is not known.
- Since the propagation period is relatively short, the sensitivity to the assumption associated to the
critical pitting corrosion depth of the durability limit state is rather low, which limits a source of
divergence in appreciation of the models.
- Climatic conditions: chloride diffusion appears to be significantly affected by the temperature. Local
appreciation of this influence both for XS ad XD classes should be re-considered in National Annexes.
Moreover, appropriate consideration of this effect for environments associated to XD exposure classes
may be more difficult due to higher seasonal variations directly linked to chlorides supply.

- The deterministic computations in the present subsection only rely on a characteristic value of the
expected chloride ingress depth for estimating the material performance, which is consistent with the
XRDS definition. Conversely, the probabilistic models need to use an estimated mean value of the
performance, identified at early age and derived from the ERC definition, and an estimate of the ageing
factor, and have to extrapolate back to a low quantile value, which makes the computation rather
sensitive to assumptions associated to the scatter in material properties and boundary conditions. This
scatter may be real for site concrete, it may also be an artefact of durability parameters measurement.
In view of further comparison between the present approach and full-probabilistic models, the
following assumptions have been made: the coefficients of variation (CoV) associated to the chloride
depth corresponding to the XRDS class designation range from 70% to 100%, their value increases for
higher quality concretes. A direct determination of this coefficient has been associated to the
deterministically computed Dapp(t50) which is used for cover depth derivation. This CoV however
includes uncertainty in the diffusion coefficient associated to the concrete mix, variations in the peak
concentration (or enrichment factor) and scatter and uncertainty in the ageing factor determination.
If a characteristic value of the migration coefficient is used for concrete performance-based
specification, it should be derived from the mean value of the Dapp(t50) associated to the XRDS class,
considering discrete approximate values of the ageing factor, and include a coefficient of variation
associated to the migration test, typically 30%. A selection of corresponding possible provisions is given
hereafter, for the XRDS classes corresponding to the most frequently used combination of exposure /
cover depths.

Dapp(50) Dapp(50)
dCl charac µ M_RCM (90) µ M_RCM (90) specif = 1,6µ
-13
(mm) 10 10-13 10-12 m²/s 10-12 m²/s
m²/s m²/s
Alpha Alpha Alpha Alpha Alpha Alpha Alpha Alpha
class = 0,3 = 0,4 = 0,5 = 0,6 = 0,3 = 0,4 = 0,5 = 0,6
XRDS1 25,1 1,213 0,55 0,27 0,46 0,78 1,32 0,4 0,7 1,2 2,1
XRDS1,5 30,8 1,811 0,831 0,41 0,69 1,18 2 0,7 1,1 1,9 3,2
XRDS2 35,5 2,394 1,112 0,55 0,93 1,57 2,67 0,9 1,5 2,5 4,3
XRDS2,5 39,7 2,954 1,388 0,68 1,16 1,96 3,33 1,1 1,9 3,1 5,3
XRDS3 43,5 3,484 1,657 0,81 1,38 2,34 3,98 1,3 2,2 3,7 6,4
XRDS10 79,4 11,174 5,85 2,87 4,87 8,27 14,05 4,6 7,8 13,2 22,5
XRDS15 97,3 17,195 9,062 4,44 7,54 12,82 21,77 7,1 12,1 20,5 34,8
XRDS30 137,6 34,39 18,125 8,88 15,09 25,63 43,54 14,2 24,1 41 69,7

For citations, copying and referencing see Note on Front page! page 216
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

To establish this table the following equation has been considered:

Dapp = k . M_RCM(t0) A(t), with A(t) = (t0/t)a
And t0 = 90 days, t = 50 years and k=1 corresponding to the safe assumption consisting in neglecting
the chemical interaction of chlorides with the cement matrix.

A similar table could be established for t0 = 28 days, in this case M_RCM(t0) directly refers to the result
obtained from EN 12390-18 testing standard. Yet this option could be less reliable due to possibly
overestimated ageing factors of blended cements (Mai-Nhu et al. 2015; Linger et al. 2017; Denis et al.
2018).

For further use in the comparison with other modelling approaches the correspondence between XRDS
classes and the mean apparent diffusion coefficient is given extensively.

XRDS dCl Dapp(50)_charac CoV (%) Dapp(50)_µ

mm 10-13 m²/s 10-13 m²/s
0,3 13,8 0,365 100 0,16
0,5 17,8 0,606 98 0,269
0,8 22,5 0,97 96 0,435
1 25,1 1,213 94 0,55
1,5 30,8 1,811 92 0,831
2 35,5 2,394 90 1,112
2,5 39,7 2,954 88 1,388
3 43,5 3,484 86 1,657
3,5 47 4,012 84 1,932
4 50,2 4,508 82 2,198
5 56,2 5,635 80 2,782
5,5 58,9 6,198 78 3,099
6 61,5 6,805 76 3,447
7,5 68,8 8,452 74 4,337
8 71,1 8,939 73 4,618
9 75,4 10,207 72 5,308
10 79,4 11,174 71 5,85
11 83,3 12,61 70 6,646
15 97,3 17,195 70 9,062
20 112,4 22,926 70 12,083
25 125,6 28,658 70 15,104
30 137,6 34,39 70 18,125

With all above assumptions, it turns out difficult to provide a general straightforward determination
of the reliability index associated to meeting the recommended minimum cover depth. Importance of
scatter in the inputs and sensitivity of the models lead to hardly demonstrating achievement of a target
reliability higher than 1,5. It seems however rather satisfactory to have kept in the determination of
XRDS and XRC provisions this consistent and similar minimum target reliability whatever the type of
exposure and corrosion mechanism, namely for XD, XS and XC environments, since this expresses
equality in the durability objective for similar constructions, whatever their location.

For citations, copying and referencing see Note on Front page! page 217
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

References
Angst U., Elsener B., Larsen C., Vennesland Ø., “Critical chloride content in reinforced concrete – A
review”, Cement and Concrete Research 39 (2009), 1122-1138.

Binder F., Burtscher S., Strauss A., “Prediction of chloride profiles and discussion of time variant
alterations”, IABSE Symposium 2019 Guimarães, Portugal, March 25-27, pp. 1250-1260.

Buchwald A. (2000) “Determination of the ion diffusion coefficient in moisture and salt loaded
masonry materials by impedance spectroscopy”, in 3rd int. Symp., Vienna, pp. 475-482.

Cao Y., Gehlen C., Angst U., Wang L., Wang Z., Yao Y., “Critical chloride content in reinforced
concrete – An updated review considering Chinese experience”, Cement and Concrete Research
117 (2019), 58-68.

Denis P.E., Linger L., Aït Alaïwa A., Ben Attaya S., Hoarau G., Caro J., Cussigh F. (2018), Use of
resistivity as a concrete quality routine control tool. Outcomes of records gathered during 2-year
for the New Coastal Road on Reunion Island (France) offshore viaduct, fib Congress, Melbourne,
Australia.

El Farissi A. (2020) Prédiction de la durée d’utilisation des ouvrages en béton armé par une
approche performantielle dans le cas de la corrosion induite par la carbonatation ou l’attaque des
ions chlorure, Ph.D. Thesis, univ. La Rochelle, France

fib Bulletin No. 34: Model Code for Service Life Design - Model Code”, Schiessl P. et al., Technical
report, 2006.

fib Bulletin No. 76: Benchmarking of Deemed-to-Satisfy provisions in Standards, Gehlen C. et al.,
2015.

fib Model code for concrete structures (2010), Wiley

ISO 16204:2012, Durability – Service life design of concrete structures

Jones M.R., Dhir R.K., Gill J.P. (1995) “Concrete surface treatment: Effect of exposure temperature
on chloride diffusion resistance”, Cement and Concrete Research, 25, 197-208.

Linger L., Mai-Nhu J., Rougeau P., Torrenti J.M., Outteryck C., Denis P.E., Magné S., Dupuy P. (2017),
Implementation of a global durability approach in close cooperation between Owner/Engineer,
Designers and Contractor’s Joint-ventures for the New Coastal Road on Reunion Island (France),
Proceedings of the 2017 fib Symposium, Maastricht, The Netherlands, pp 2601-2612.

Linger L., collected durability indicators during production control of concrete for a large viaduct,
PerfDuB projet, personal communication, 2020.

Mai-Nhu J. (2013) « Corrosion des armatures du béton : couplage carbonatation/chlorures en

présence de cycles hydriques », Ph.D. Thesis, univ. Toulouse - CERIB.

Mai-Nhu, J., Rougeau, P., Linger, L., Denis, P.E., Magné, S. (2015) The New Coastal Road on Reunion
Island (France): Application of durability model to a real case, fib symposium, Copenhagen.

Rahimi A. (2020) Semi-probabilistic approach to durability design and assessment of reinforced

concrete members exposed to the action of chlorides, BAW, Karlsruhe.

For citations, copying and referencing see Note on Front page! page 218
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Claus V Nielsen, DTI, Feb. 2020

CEN/TC 250/SC2/WG1/TG10

Minimum cover for resistance against chloride ingress

(input to section 3.4 in the Background report for EC2 chapter 6) updated
from earlier version in document N315

3.4 Simplified calculations for chloride ingress

3.4.1 Introduction
This presentation is based on the principles for chloride ingress and the service life design model
that is described in a recent ACI article [Nielsen 2019]. The approach has been to make a simple intu-
itive method for assessing the expected design service life, considering the initiation period where
the chloride ions diffuse into the concrete towards the reinforcement. The model described in Niel-
sen (2019) is briefly described herein. It is based on the fib Model Code approach [fib 2006] which
has been applied in some of the previous models and the input parameters are following:

• The rate of chloride ingress is defined by the apparent diffusion coefficient Dapp [10-12 m2/s]
which is defined at the age 50 years, taking into account the ageing factor.
• The severity of exposure is defined by the chloride surface concentration Cs [%w-c.m.].
• The critical chloride concentration for the initiation of corrosion Ccr [%w-c.m.].
• Initial chloride content Ci [%w-c.m.] of the concrete mix.
• Cover thickness c to the reinforcement [mm].

These parameters are treated as stochastic variables, defined by distribution functions, mean values
and standard deviations, and Monte-Carlo simulations have been applied to obtain the distribution
of the time for the critical chloride concentration to reach the reinforcement.

3.4.2 Model and assumptions

Initiation phase

The calculations herein are based on the error function solution [fib 2006]. The time-dependent in-
gress of chlorides follows the following equation:

𝑥
𝐶(𝑥, 𝑡) = 𝐶𝑠 − (𝐶𝑠 − 𝐶𝑖 ) ∙ erf { } (1)
2√𝐷𝑎𝑝𝑝 (𝑡)∙𝑡

Where C denotes chloride concentration in weight-% of the cementitious material [%w-c.m.]. The
parameters are defined as:

x = coordinate from the surface [mm];

t = time from the beginning of the exposure [years];

Cs = surface concentration corresponding to x = 0;

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 219
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Ci = initial chloride concentration at t = 0;

Dapp = apparent chloride diffusion coefficient at time t [10-12 m2/s];

The field observations have shown that the diffusion coefficient decreases with age and the follow-
ing ageing effect model is used:
𝑡 𝛼
𝐷𝑎𝑝𝑝 = 𝑘𝑒 𝑘𝑡 𝐷0 ( 𝑡0 ) (2)

Where  = ageing factor, D0 = the reference diffusion coefficient, typically determined experimen-
tally at t0 = 28 days. The two factors ke and kt are taking the temperature influence into account and
the transfer between the field observations and the laboratory results. In the present calculations
both ke and kt are unity.

Another simplification that is applied herein, is the depth x that is used to describe that the peak
concentration of chlorides is located a few millimetres below the exposed surface rather than at x =
0. In the present calculations x is omitted [Nielsen 2019].

When the chloride content in the depth of the reinforcement bars (cover thickness c) reaches the
critical threshold value Ccr for corrosion initiation the initiation period tini is found by inserting (1) into
the following equation:

𝐶(𝑥 = 𝑐, 𝑡 = 𝑡𝑖𝑛𝑖 ) = 𝐶𝑐𝑟 (3)

Propagation phase
When the initiation phase is done, and the chlorides have reached the critical level at the dept of the
reinforcement, it may start to corrode during the so-called propagation phase. The definition of the
design service life is a sum of the initiation and the propagation phases (tSL = tini + tprop ).

In the present calculations the propagation phase has been included in a simplified manner:

• For the less severe exposure conditions XS1 and XD1 a constant propagation phase of tprop =
10 years has been applied.
• For the other exposure conditions where the concrete is directly exposed to chlorides from
deicing agents and/or seawater the propagation period is not taken into account. Thus, tprop
= 0 for all other exposure conditions than XD1 and XS1.

These assumptions are in line with the those given in the previous models.

3.4.3 Input parameters

The following parameters are treated as stochastic variables. See more details in [Nielsen 2019].

• Reference diffusion coefficient D0: Normal distributed with mean value and CoV = 20%.
• Ageing factor : Beta distributed between 0 and 1 with mean value and CoV = 20 to 30%.
• Surface concentration Cs: Log-normal distributed with mean value and CoV = 45%.
• Critical threshold level Ccr: Beta distributed between 0.2 and 2 with mean value 0.6 %w-c.m.
and CoV = 25%.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 220
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

• Cover thickness c: Normal distributed with mean value equal to the nominal cover and
standard deviation = cdev/1.64. Hence, if the tolerance is cdev = 10 mm the standard devia-
tion is 6 mm whis is applied in the examples given herein.

All other parameters are constants. In Nielsen (2019) a large number of combinations of the varia-
bles, corresponding to various binder types, ageing factors, etc., have been simulated by means of
Monte-Carlo and the variation of the ingress rate of chlorides after a certain number of years has
been quantified and collected in a single nomogram as shown the following examples.

The apparent diffusion coefficient Dapp

For the diffusion coefficient, the diagram in Fig. 3.4.1 shows the effect of the ageing factor (Eq. 2).
Typically, factor  = 0.3 is applied for CEM I and the higher values of  ranging from 0.5 to 0.6 will be
applied for slag cements and fly ash cements [fib 2006, Bamforth 2004]. However, it is also recog-
nised that large scatter is found on these values. The boxes named I – IV in Fig. 3.4.1 represents dif-
ferent concrete qualities in terms of increasing diffusion coefficients Dapp(50 yrs).

Fig. 3.4.1: Diagram illustrating the ageing effect in Eq. (2) up to the age of 50 years.
Note that the unit µm2/s is equal to 10-12 m2/s. the red circles indicate the range of values that are
applied in the next section examples.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 221
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Field observations show that for concretes with CEM I and water to binder ratios around 0.40, the
apparent diffusion coefficient would expectedly end up in class III or IV in Fig. 3.4.1. For fly ash ce-
ments the result would be class II and for slag cements we would typically end up in class I or better
[Nielsen, 2019]. Hence, the classes indicated in Fig. 3.4.1 illustrates the actual range of performance
which is often observed in the literature for concrete mixes exposed to chlorides from deicing agents
and seawater [fib 2006] [Helland et al. 2010].

Regarding the ageing effect it can be debated whether it shall be applied for the full duration of the
service life (say up to 100 years or more) or if it should be cut-off after a certain period. It is out of
the scope to go deeply into this discussion and there is hardly any long-term evidence to support it.
However, there has been reported observations from slag cements up to 60 years [Bamforth 2004]
but others seem to suggest a cut-off after 25 to 30 years [Nielsen 2019] which would be a conserva-
tive approach. The different models presented in sections 3.1-3.3 have applied different approaches,
regarding the ageing effect, which will result in differences in the minimum cover thicknesses, c.f.
section 3.5 and 3.6.

In the present calculations the ageing effect beyond 50 years has been neglected which means that
the lines of Fig. 3.4.1 will become horizontal after 50 years. This is considered to be a conservative
approach.

The Dapp(50 years) values that are applied herein are based on actual observations from field expo-
sure sites [Nielsen 2019]. Hence, they are obtained from temperature conditions that are about 8 to
10 ⁰C (annual average). If the correct modification of the diffusion coefficient according to the fib
Model code [fib 2006] were to be applied the reference diffusion coefficient (at 20 ⁰C reference tem-
perature) would be about twice as high as the observed field values. However, this is far from the
realistic experiences from the laboratory. Therefore, it has been chosen to use kt = ke = 1 in (2) which
again is a conservative assumption. More details of the background for this discussion are found in
Nielsen (2019).

The surface concentration Cs

The surface concentration depends strongly on the exposure conditions and the severity of the ex-
posure (e.g. the salinity of the seawater, the amount of deicing, the location of the structural ele-
ments in relation to the shore/road and the direction of dominating wind, etc.). Therefore, the varia-
tion found in the literature is immense [Nielsen 2019, Helland et al. 2010]. Table 3.4.1 give the val-
ues applied herein which are chosen on basis of the assumptions described in the previous model
calculations (Sections 3.1-3.3). Based on the literature referred to herein, these values seem to be
conservative – especially for the XD3 conditions which differ significantly between the different
countries.

The exposure classes XD2 and XS2 are not included in the calculation examples provided in the next
section. The reason for this is that XD1/XS1 typically represents the less severe chloride exposure
and XD3/XS3 represents very severe exposure conditions and XD2/XS2 is found somewhere in be-
tween.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 222
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Table 3.4.1: Surface concentration used for the calculations.

Exposure
XD1/XS1 XD2/XS2 XD3/XS3
classes
Typical expo- Air bourne chlorides. Submerged in sea- Direct spray from
sure condi- Vicinity of coastal ar- water, permanently roads. Splash zone on
tions eas and/or large roads wet marine structures.
Mean Cs
1.5 (CoV = 45%) Not included herein 4.0 (CoV = 45%)
[%w-c.m.]

3.4.4 Relationship between the XRDS classes and the diffusion coefficient
The exposure resistance classes against chlorides are denoted XRDSxx, where the designation xx is
the 90%-quantile of the diffusion coefficient under standardised exposure conditions for 50 years
[10-13 m2/s]. This has been explained in detail in the previous sections and it is not repeated here.

The apparent diffusion coefficient Dapp applied in the previous section (Eqs. (1) and (2)) is the mean
value obtained after 50 years. It relates to the XRDS classes and Fig. 3.4.2 shows the dependency
taken from section 3.1 and 3.2, which correspond approximately to the following simple relation-
ship:

Dapp(50 years) = ½*XRDS (4)

This is applied in the examples in the next section. For each of the Dapp values, depicted in Fig. 3.4.2
(red circles), that represent a wide range of realistic concrete qualities to be expected for chloride
exposure, the minimum cover thicknesses are calculated by means of the simplified method outlined
above and in Nielsen (2019).

Fig. 3.4.2: Mean Dapp(50 years) as a

function of the XRDS class designation.
Note that units are equal to 10-13 m2/s.
The boxes and the red circles represent
those depicted in Fig. 3.4.1. The dashed
red line represents slope 1:2.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 223
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

3.4.5 Examples for exposure conditions XD1/XS1

Based on the input parameters described in the previous sections, Fig. 3.4.3 shows the design ap-
proach in a single diagram. First Cs = 1.5 %w-c.m. is chosen (a) and then the critical threshold value is
chosen Ccr = 0.6%w-c.m. (b). Then the red dashed line is followed until the Dapp(50 years) value is
reached (c) and finally the necessary minimum cover thickness is found in the lower left part (e), cor-
responding to the allowable probability of failure Pf. The lines show 5 different levels of Pf (50, 40,
31, 20 and 9.7%) that correspond to β equal to 0, 0.25, 0.5, 0.85 and 1.3, respectively. Actually, in
order to reach β = 1.5 as is the basis for the models in this report the cover thicknesses were to be
increased slightly (say extra 5 mm). However, due to the general conservative assumptions de-
scribed above this is neglected here.

Fig. 3.4.3: XD1/XS1 minimum cover thickness for design service life 50 years (tini = 40 years + tprop = 10
years) with a probability of failure approximately 10% which corresponds to β ≈ 1.3. Diagram taken
from Nielsen (2019). Red lines and circles represent the input values used in the examples herein.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 224
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Fig. 3.4.4 show the similar picture for service life 100 years. Like it was described earlier, the Dapp(50
years) is maintained for the 100 years predictions, which corresponds to the assumption that the
ageing effect is negligible beyond 50 years.

For both Figs. 3.4.3 and 3.4.4 the execution tolerance is cdev = 10 mm and the initial chloride con-
centration is Ci = 0.05 %w-c.m.

Fig. 3.4.4: XD1/XS1 minimum cover thickness for design service life 100 years (tini = 90 years + tprop =
10 years) with a probability of failure approximately 10% which corresponds to β ≈ 1.3. Diagram
taken from Nielsen (2019). Red lines and circles represent the input values used in the examples
herein.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 225
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

3.4.6 Examples for exposure conditions XD3/XS3

Similarly, with the diagrams presented in the previous section for XD1 and XS1 this section includes
the diagrams for the more severe exposures XD3 and XS3. When the exposure conditions are in-
creased (Cs raised from 1.5 to 4 %w-c.m.), the corresponding chloride ingress increases accordingly
as it is seen by direct comparing of Figs. 3.4.3 and 3.4.5.

The exposure resistance class XRDS10 has been excluded since it results in excessive cover thickness
and instead XRSD2 has been inserted in Fig. 3.4.5 and XRSD4 has been inserted in Fig. 3.4.6.

Fig. 3.4.5: XD3/XS3 minimum cover thickness for design service life 50 years (tini = 50 years + tprop = 0
years) with a probability of failure approximately 10% which corresponds to β ≈ 1.3. Diagram taken
from Nielsen (2019). Red lines and circles represent the input values used in the examples herein.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 226
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Fig. 3.4.6: XD3/XS3 minimum cover thickness for design service life 100 years (tini = 100 years + tprop =
0 years) with a probability of failure approximately 10% which corresponds to β ≈ 1.3. Diagram taken
from Nielsen (2019). Red lines and circles represent the input values used in the examples herein.

3.4.7 Summary of example results

The results from the examples given in Figs. 3.4.3 – 3.4.6 are summarised in Table 3.4.2 below.

As it is seen in the next sections the results for 50 years of service life scenario are corresponding
well with the other models in sections 3.1-3.3. However, when it comes to 100 years of service life
the differences become significant and the model presented herein produces significantly higher val-
ues of the minimum cover thickness than what is seen from the other models. It is believed that
these discrepancies are related firstly to the ageing effect approach taken for the apparent diffusion
coefficient beyond the 50 years of service life. Secondly, the different model approaches contain dif-
ferent assumptions related to the effect of ambient temperature on the apparent diffusion coeffi-
cient.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 227
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Claus V Nielsen, DTI, Feb. 2020
CEN/TC 250/SC2/WG1/TG10

Table 3.4.2: Minimum cover thickness [mm] from the examples described above.
Exposure
XD1/XS1 XD2/XS2 XD3/XS3
classes
Not ap-
Service life 50 years 100 years 50 years 100 years
plied
XRSD1 25 40 - 40 65
XRDS2 - - - 50 75
XRDS3 35 60 - 55 85
XRDS4 - - - - 100
XRDS6 45 70 - 80 -
XRDS10 55 90 - - -

3.4.8 References

Bamforth, P.B.: Enhancing Reinforced Concrete Durability, Guidance on Selecting Measures for Mini-
mising the Risk of Corrosion of Reinforcement in Concrete, TR61, The Concrete Society, Camberley,
UK, 2004.

fib: Model Code for Service Life Design, Bulletin no. 34, International Federation of Structural Con-
crete, Lausanne, Switzerland, 2006.

Helland, S., Aarstein, R. & Maage, M.: “In-field performance of North Sea offshore platforms with re-
gard to chloride resistance”, Structural concrete, Vol. 11, No. 1, pp. 15-24, International Federation
of Structural Concrete, Lausanne, Switzerland, 2010.

Nielsen, C.V.: “Service Life Modeling of Chloride Ingress”, ACI Materials Journal, Sept. 2019, Vol. 116,
pp. 17-30, 2019.

Section 3-4_Simplified calculations_chloride_cln_2021-02-28.docx

For citations, copying and referencing see Note on Front page! page 228
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

XRDS Model comparison and proposal for concrete cover

S.v.Greve-Dierfeld, F. Toutlemonde, CV Nielsen, C. Andrade, D. Izquerdo Lopez
Draft 6

3.5 Model comparison

The models applied for the chloride induced corrosion are similar in all approaches for describing the
incubation and initiation period where the chloride ions progress towards the reinforcement. The
models are based on the erfc solution to Fick`s second law of diffusion.

The XRDS designation represents the apparent diffusion coefficient as a characteristic value (not
exceeded in 90% of the cases) with the units [10-13 m2/s] and refers to the depth where an average
concentration Cth equal to 0.6 wt.% chloride/binder-cement+type II addition is reached after 50 year in
30g/L NaCl of external solution at 20°C by unidirectional diffusion. The XRDS designation refers to a
concrete of at least execution class 2. Under actual exposure conditions temperature and moisture
condition, the chloride load or the chloride content at reinforcement that may lead to corrosion
initiation may differ from the XRDS-reference condition and XRDS-reference depth.

Table 3.5.1 summarises the parameters used in the erfc solution models. Furthermore, table 3.5.1
allocates the defined XRDS reference conditions to model parameters and model factors. The model
factors equal 1 if the condition in the exposure class is like the XRDS reference condition. The model
parameter has the same average value as the parameter of the reference condition if the exposure
condition equals (in average) the reference condition.

Table 3.5.1 summary of mean values of model parameter and transfer parameter in comparison with
XRDS-reference conditions
parameter XRDS-reference Exposure Model Model Model Model
condition class 3.1 3.2 3.3 3.42)
XS1 0.65 0.6 0.2 0.3-0.6
Time exponent Apparent XD1 0.50 - 0.2 (to 50 y)
a=n diffusion XS2 0.3-0.6 0.8 0.2 -
coefficient XD2 0.40 - 0.2
(Dapp at t=50 year) XS3 0.3 - 0.6 0.8 0.2 0.3-0.6
XD3 0.40 - 0.2 (to 50y)
XS1 0.56 1
XD1 0.70 -
Factor for the 20°C: XS2 0.56 1 0.75 1
effect of (ke=fT=1) XD2 0.60 -
temperature [-] XS3 0.56 1
k e = fT XD3 0.70 -
Factor for the XS1 1 1 (1)1)
effect of Fully saturated XD1 0.4 -
transport in condition XS2 1 1 1
partially satu- (fenv = 1) XD2 0.8 - 1 1
rated conditions XS3 1 4
fenv [-] XD3 0.5 -

For citations, copying and referencing see Note on Front page! page 229
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

parameter XRDS-reference Exposure Model Model Model Model

condition class 3.1 3.2 3.3 3.42)
Max. chloride 30g/L NaCl XS1 2 0.2 1.5- 1.5
concentration Cs XD1 1.5 - 1.65
[wt.-%/cem] for (Variable;
Model 3.2: Cs=3.8wt.%/cem. XS2 4 0.35-0.55 3.0-3.3 -
is taken as
[wt.%/concrete] XD2 3 -
reference
simplistic) XS3 4 0.35-0.55
3.6-4.0 4
XD3 3 - 3.3-3.6
Initial chloride No definition all 0.1-0.2 0.01in high 0.05
content Ci=C0 Cs and
[wt.-%/cem] (Ci=0 wt.%/cem.) low Cth
Corrosion depth with an XS1 0.6 0.1 0.5 to 0.6
inducing average chloride XD1 0.6 - 0.7
chloride content content 0.6 as of XS2 0.6 0.1 0.4 to -
Cth=Ccrit [wt.- cem + typeII ad. XD2 0.9 - 0.6
%/cem]; Model (Cth = 0.6 XS3 0.6 0.1 0.4 to 0.6
3.2 [wt.%/conc.] wt.%/cem) XD3 0.7 - 0.6
XS1 10 0 5 0
Depth with Unidirectional XD1 10 - 5
maximum diffusion XS2 0 0 0 -
chloride XD2 5 - 0
concentration (Dx = 0) XS3 10 10 8 (10)3) 0
Dx [mm] XD3 10 - 8 (10)3)
Depth with XS1 10 1 10 10
tprop, mean average 0.6 wt.% XD1 - 15
values [year] chloride after 50 XS2 5 4 10 -
years XD2 - 5
with tSL=tprop+tini XS3 1 0 2 0
(tprop=0) XD3 2 - 5
1)
a higher margin is considered in XD1/XS1 due to the uncertain transport process. A lower value is
considered admissible depending on local conditions 2) ageing was considered up to an age of 50
years.3) for tSL=100 year

Each factor and each average model parameter, that differs from the XRDS reference condition affects
the calculated depth with the specific chloride content. For example, if the temperature is lower than
in the temperature in XRDS reference conditions, the depth allocated to XRDS reference condition is
reduced in dependency of the factors proposed by each model in table 3.5.1. Another example is the
critical chloride content. If the critical chloride content is higher than the reference `fictive` threshold
given for XRDS reference depth, than the XRDS depth is reduced following the erfc-solution model.

For citations, copying and referencing see Note on Front page! page 230
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Fig. 3.5.1 average effect of each model parameter on depth with average chloride content of 0.6

Figure 3.5.2 shows the calculated depth with critical corrosion induction chloride content in each
exposure class for average apparent diffusion coefficients after 50 years of exposure to XRDS reference
conditions.

For calibration of the input values, model 3.1 used data from exposure tests and existing structures in
Switzerland and literature (i.e. from Denmark, Sweden, France, Denmark). In the chapter 3.2 the model
of [Izquerdo 2002] was applied with data from Spain and fib MC for SLD 2006]. Model 3.3 used data
from experience of site structures in France and Norway. Model 3.4 used boundaries derived from the
model calculations of Model 3.1, 3.2 and 3.3, from data from Denmark and fib MC for SLD 2006.

For citations, copying and referencing see Note on Front page! page 231
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

For citations, copying and referencing see Note on Front page! page 232
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Figure 3.5.2: Average depth with critical chloride content xcrit according to each model in
dependency of average apparent diffusion coefficient determined after 50 years in XRDS reference
conditions

The differences in the models, especially the critical chloride content, chloride surface concentration
and initial chloride content explain some of the differences. Others are explained by the different
environmental conditions. Furthermore, age exponents different from XRDS definition i.e. under dry
conditions (XD1, XS1) explain the differences in XD1/XS1 conditions.

For citations, copying and referencing see Note on Front page! page 233
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

3.6 XRDS comparison

Table 3.6.1 displays the recommended minimum cover depth values in dependency of XRDS classes
which have been selected for the Eurocode 2 revision. The values depend on the different exposure
classes and on the required design service life (50 or 100 years).

In table 3.6.1 the given concrete cover depth refers to cmin,dur. The correlation between the mean value
of concrete cover and cmin,dur is different for each model. This is explained in chapter 1.2. with further
details in chapters 3.1-3.4. Examples are given in the following.

The XRDS designation has the units of a diffusion coefficient in 10-13 m2/s derived from the 90% quantile
of depth with 0.6 wt.% chloride/binder-cement+type two addition after 50 year in XRDS reference
conditions. The CoV of apparent diffusion coefficient under XRDS reference conditions (respectively,
the CoV of the depth after 50 years under XRDS reference conditions) includes the following effects:

Ø variability of the raw material and batching

Ø variability of curing and execution
Ø variability in maturing and ageing
Ø Variability of maximum chloride concentration under reference conditions

But excludes variability of environmental conditions and the variability of critical chloride content.

CoV was used with CoV = 40-100 %. The lower the CoV used to calculate the XRDS values based on the
mean values from figure 3.5.1 the more conservative is the XRDS class designation but the lower the
CoV used to calculate the mean value from XRDS value, the less conservative is the concrete cover.

As an example, if the apparent diffusion coefficient (respectively, the depth with chloride content of
0.6 wt.%/binder after 50 years of exposure in XRDS reference condition) is normally distributed and if
the mean value is Dm(XRDS)= 3.97 10-13 m2/s and assuming the CoV of the placed concrete in XRDS
reference conditions is CoV = 40% (lower value of the range) the XRDS designation is consequently
XRDS 6 calculated as follows:
Ø 3.97 10-13 m2/s + 1.282*0.4*3.97 = 6 10-13 m2/s → Dm(XRDS)= 3.97 10-13 m2/s = XRDS 6
Subsections 3.1-3.4 give details on the CoV and calculation procedure used to determine the 90%
quantile of the apparent diffusion coefficient, in dependency of the selected distribution function.

Table 3.6.1: Minimum cover thickness recommended in prEN 1992-1-1 (D7), clause 6.5.2.2 (Table
6.3(NDP).

XS1 XS2 XS3 XD1 XD2 XD3

Design Service Life 50 100 50 100 50 100 50 100 50 100 50 100
XRDS0.5 20 20 20 30 30 40 20 20 20 30 30 40
XRDS1 20 25 25 35 35 45 20 25 25 35 35 45
XRDS1.5 25 30 30 40 40 50 25 30 30 40 40 50
XRDS2 25 30 35 45 45 55 25 30 35 45 45 55
XRDS3 30 35 40 50 55 65 30 35 40 50 55 65
XRDS4 30 40 50 60 60 80 30 40 50 60 60 80
XRDS5 35 45 60 70 70 NA 35 45 60 70 70 NA
XRDS6 40 50 65 80 NA NA 40 50 65 80 NA NA
XRDS8 45 55 75 NA NA NA 45 55 75 NA NA NA
XRDS10 50 65 80 NA NA NA 50 65 80 NA NA NA

For citations, copying and referencing see Note on Front page! page 234
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Accounting for the different possible interpretations in the target reliability it was agreed to select for
the Eurocode revision the average of recommended values from Figs 3.6.1, which are displayed in
Table 3.6.1. For values in excess of 80 mm deviations between models was considered excessive and
corresponding minimum cover depths values were decided not to be recommended (indication NA –
“not applicable”).

The XRDS classes are assumed safe for structural elements exposed to Atlantic-sea / Mediterranean
sea water. Because de-icing salt exposure is assumed similar or less severe, but highly variable through
Europe the provisions for XS and XD have been merged.

Figure 3.6.2: Comparison of calculated minimum concrete covers (rounded values) as a function of the
XRDS classes for b = 1.5. The dashed lines are the recommended values given in Table 3.6.2. Only
XD1/XS1 and XD3/XS3 exposure classes are shown as examples. Note that the x-axes differ

For citations, copying and referencing see Note on Front page! page 235
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG10 FprEN 1992-1-1:2023

Background to clause 6.5.2.2(4)

Minimum cover for durability of prestress-
ing tendons

Mikael Hallgren Tyréns AB

Carmen Andrade CIMNE UPC
Francois Toutlemonde Université Gustave Eiffel
Stefanie von Greve-Dierfeldt TFB AG

02.03.2021

For citations, copying and referencing see Note on Front page! page 236
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Section 6.5.2.2 Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared to explain the increase of minimum cover for durability for prestressing
tendons, as an addition to the minimum cover for durability recommended for reinforcing steel. The text in this
document is proposed to be implemented into the background document to clause 6.5 Concrete Cover.

6.5.2.2(4) Reason for the clause

Minimum cover for durability The clause states that for prestressing tendons, pre- or post-tensioned, the cover values in Table 6.3 and Table
of prestressing tendons
6.4 in clause 6.5.2.2(1) should be increased by +10 mm.
This 10 mm-increase is quantitatively in line with the difference between recommended cover depths as from
Tables 4.4N and Tables 4.5N of EN 1992-1-1:2004, which has been subject to very limited changes in national
provisions; it does not account for any new contradictory material, but it has been deemed as a simplification
in the sense of ease of use for the designer. When cover depths are determined from annex P in the revised
Eurocode, these tables will be used, thus a certain quantitative consistency is kept between this way and the
ERC approach of section 6.
The reason for this is due to the more severe consequences of a failure of a prestressing tendon compared to
the failure of a reinforcement bar. The more severe consequences require a lower probability for failure, i.e. a
higher β index.
These "more severe consequences" of prestressing steel corrosion could be detailed as follows: prestressing
tendons or bars would be susceptible to stress corrosion cracking, so that failure of wires or bars may suddenly
follow corrosion initiation, leading to immediate loss of capacity. This prevents from including an admissible
time for corrosion propagation in the design service life.
Therefore, in exposure conditions like XS3 or XC4 where the admitted time for corrosion propagation is low,
the additional 10 mm-cover depth can directly be interpreted as a safety margin to lower the occurrence of
depassivation and will increase the β index for corrosion propagation at the end of the design service life, as
compared to β = 1,5 assumed for carbon steel reinforcement in Tables 6.3 and 6.4 in clause 6.5.2.2(1). For
XC4, the reliability is typically increased from 1,5 to a value ranging from 2,0 to 2,3 (the increase is lower for
higher cover depths / higher XRC) and for XS3, from 1,5 to about 2,0.
For XC2, XC3, XS1/XD1 and XS2/XD2 exposure classes where a propagation time of typically 10 to 15 years
has been considered for reinforcing steel, the 10mm-cover depth increase for prestressing steel both compen-
sates this exclusion of the propagation period and reduces the probability of reaching depassivation (typically
from 10% to less than 3%).
An exception to this need of increase of minimum cover for durability is made for internal bonded posttensioning
systems that are provided with protection level 2 or 3 and internal unbonded prestressing tendons are encased
in corrosion resistant sheaths. For these cases, the minimum covers according to Tables 6.3 and 6.4 in clause
6.5.2.2(1) are considered to be sufficient and safe. The protection levels 2 and 3 presumes grouted polymer
ducts. Further guidance for details and the assessment of tendon protection levels and requirements on the
polymer ducts can be found in the fib Bulletin 75 [1].

References [1] Ganz H.R. et al., “Polymer-duct systems for internal bonded post-tensioning”, fib Bulletin 75, Lau-
sanne, 2014.

List of versions of this document

03.02.2021. pre-draft
02.03.2021 1st complete draft
17.02.2023 reviewed, only editorial update

Hallgren M. et.al. 02.03.2021

For citations, copying and referencing see Note on Front page! page 237
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

4.2 Background document to subsection

6.5.2.1(2)
∆cmin considering the increased uncertainty and
variability of concrete cast directly against soil

Carola Edvardsen COWI, Denmark

28.04.201

For citations, copying and referencing see Note on Front page! page 238
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Introduction
This document has been prepared in order to validate the proposed EC2-provisions to increase the minimum
cover by cmin considering the increased uncertainty and variability of concrete cast directly against soil.

6.5.2.1 (2) Reasons for introducing increased minimum cover

Increased minimum cover In recent years, an increasing number of retaining structures (e.g. diaphragm walls DWs and secant piles SPs)
have been built where the concrete is cast directly against soil und not mechanically compacted. Studies, e.g.
[1], [2], [3] and investigation at actual projects e.g. [4] have shown that there is a reduction in the concrete
quality compared to concrete cast against rigid formwork and compacted. This has especially been observed
at diaphragm walls. This reduced concrete quality leads to reduced ERCs and should be compensated by e.g.
an increased horizontal minimum cover. A cover increase of 5 mm is recommended in FprEN 1992-1-1.
Due to the absence of confinement when casting directly against soil in combination with the absence of a
mechanical compaction, the concrete at DWs and SPs will be more porous and more inhomogeneous
compared to concrete cast against a rigid formwork and compacted. In addition, the concrete can contain
other "hidden" defects such as soil inclusions, wash-outs, bleeding, segregation etc. as shown below. These
defects are unfortunately not always visible by the naked eye when looking on the concrete quality at the
external surfaces of such structures and thus not detectable unless cores are extracted which was however
seldomly done in the past. First recently some data from actual projects and field investigations turned up.
The following figures are taken from cores extracted at actual project in Europe [4] with retaining walls. The
higher porosity is visible at epoxy-impregnated thin sections, see Figure 1. The water to cement ratio of the
porous zones was found to be more than 0.70 and approximately 0.40 to 0.45 in unaffected parts of the core
which would represent the expected concrete quality for the given concrete mix when cast against formwork
and compacted.

Figure 1: Images of a concrete core extracted from a DW and impregnated with fluorescent epoxy. Areas with
high porosity under normal light (left) and ultraviolet light (right) appear green
Examples of soil inclusion are given in Figure 2-4 showing veins/voids with lighter colour and higher porosity in
the cover zone, i.e. interior "defects" not visible from the outside. For concrete a light discoloration of the
cement paste is an undisputable indicator of reduced concrete quality enabling aggressive ions to move much
more rapidly towards the reinforcement than unaffected concrete does.

C. Edvardsen 28.04.2021
For citations, copying and referencing see Note on Front page! page 239
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Figure 2: Inferior concrete quality at extracted Figure 3: Polished section of a core from DW
core from SP due internal soil inclusions (not showing veins/voids (light colour) in the cover
necessarily always visible at the concrete zone filled with trapped support fluid/bentonite
surface) mixture

[1] M.M.R. Boutz, B.J. Admiraal, R. van Berkel, P.R.B: van der Werff, “X10 Concept to improve quality
References
and durability of diaphragm walls", DFI-EFFC International Conference on Deep Foundations and
Ground Improvement, Rome, Italy, 2018.
[2] M. Czopowska-Lewandowicz, “Concrete properties in diaphragm wall embedded in non-cohesive
soils". MATEC Web of Conferences 174, 01007 (2018) ECCE 2018
[3] I.H. Wong, “Experience with waterproofness of basements constructed of concrete diaphragm walls
in Singapore”, Tunnelling and Unterground Space Technology, 12, 4, p. 491-495, 1997.
[4] Undisclosed major infrastructure project in Europe. Confidential.

List of versions of this document

28.04.2021 1st draft
20.02.2023 peer reviewed

C. Edvardsen 28.04.2021
For citations, copying and referencing see Note on Front page! page 240
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Background to subsection 6.5.1 (3)

Nominal cover for
bored piles and diaphragm walls

Frank Fingerloos German Society for Concrete and

Construction Technology

Berlin, 15.01.2021

For citations, copying and referencing see Note on Front page! page 241
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Introduction
This document has been prepared in order to explain the proposed EC2-provisions for the nominal cover of
bored piles and diaphragm walls according to concrete design with EC2 and execution requirements accord-
ing to EN 1536/EN 1538.

6.5.1(3) Problem
Nominal cover for bored piles
and diaphragm walls The geotechnical execution standards EN 1536 [1] and EN 1538 [2] contain design and execution rules for
concrete members according to Figure B.6.5.1 and B.6.5.2:

Figure B.6.5.1: Bored pile with circular and non Figure B.6.5.2: Diaphragm wall – Geometry of a panel
circular cross-sections (barrettes) [1] [2]
The definition of cover in EN 1536 [1] and EN 1538 [2] is:
Cover: Distance between the outside of the reinforcement cage and the nearest concrete surface
([1] 3.39 and [2] 3.13)
The requirements to concrete cover in the geotechnical standards are summarized in Table B.6.5.1.
The requirement of 7.7.1 in [1] and [2] seems clear. The design rules of EC2 for minimum cover for durability
and bond apply.
The requirements of 7.7.2 in [1] and [2] are not so clear. The meaning of cover in relation to execution as
minimum or nominal cover is open.
The NOTE to 7.7.2 should give clarity about the nominal cover. But there is an inconsistency: It is stated, that
the minimum cover of EC2 or the minimum cover in relation to execution is the nominal cover to be used
for bored piles or diaphragm walls. This is confusing and inconsistent.
From the EC2-designers view it is clear, that a minimum cover has to be increased by an allowance of
deviation considering the execution tolerances to specify the nominal cover using in design documents.
If the cover in relation to execution according to 7.7.2 in [1] and [2] is really the nominal cover including exe-
cution tolerances, it has to be clarified by the responsible standardisation committees of CEN/TC 288. An
improvement of the NOTE to 7.7.2 in next issues of [1] und [2] is necessary.
Table B.6.5.1: Requirements to concrete cover for bored piles and diaphragm walls in geotechnical standards
EN 1536 [1] 7.7 Minimum and nominal cover EN 1538 [2]: 7.7 …
7.7.1 The minimum cover in relation to environmental conditions and to adhesion shall comply with EN 1992 (all parts).
7.7.2 The minimum cover in relation to execution shall not be less than: 7.7.2 The minimum cover in relation to
– 75 mm for barrettes; execution shall not be less than 75 mm.
– 60 mm for piles with D > 0,6 m; or
– 50 mm for piles with D ≤ 0,6 m,
– unless otherwise agreed.
NOTE The minimum cover in relation to execution (in order to ensure that the concrete flows freely) is specified by
reference to target values and not by reference to execution tolerances. So the nominal cover is the greater of the
minimum cover in relation to environmental condition and to adhesion and the minimum cover in relation to execution.
7.7.3 The minimum cover in relation to execution should be increased to –
75 mm where:
– piles penetrate soft soil and are constructed without a casing;
– submerged placement of concrete with 32 mm max. aggregate is used;
– silica fume is used as cement replacement;
– reinforcement is installed subsequent to concrete placement; or
– the bore hole walls have uneven surfaces.
7.7.4 The minimum concrete cover in relation to execution may be re- –
duced to 40 mm to the external face of a permanent casing or lining,
where used.

F. Fingerloos 15.01.2021
For citations, copying and referencing see Note on Front page! page 242
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Due to the misunderstanding NOTE to 7.7.2 in [1] und [2] the clarifying clause 6.5.1 (3) is implemented in
FprEN 1992-1-1:2023:
(3) For bored piles and for diaphragm walls the nominal cover values for durability and bond of this Euro-
code apply. The cover values according EN 1536 and EN 1538 considering the type and deviations of
execution should be checked additionally. The largest value according to this Eurocode and EN 1536 or EN
1538 applies.
So it is clear, that the nominal cover of EC2 including the allowance of deviation cdev according to 6.5.3 is to
consider in eyery case (see extract of Table 6.7(NDP). Only if the cover of execution in EN 1536 or EN 1538
is greater, then these values apply. The designer shall to check both values.
Table 6.7(NDP) – Allowance for deviation cdev in EC2
Case cdev
8 Concrete cast against prepared ground (including uneven blinding layer) 40 mm a)
9 Concrete cast directly against unprepared soil 75 mm a)
a)
These allowances for deviation cdev apply for the nominal cover according to this Eurocode also for
bored piles and for diaphragm walls unless a National Annex gives other special values.
With two examples for bored piles the design of cover may be illustrated (from CEN-TC250-SC2-
Example 1
WG1_N1050).

Example 2

References [1] EN 1536:2010: Execution of special geotechnical work – Bored piles.

[2] EN 1538:2010: Execution of special geotechnical work – Diaphragm walls.

List of versions of this document

15.01.2021. 1st draft
17.02.2023 peer reviewed, only editorial update

F. Fingerloos 15.01.2021
For citations, copying and referencing see Note on Front page! page 243
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Background to
6.5.3 Allowance in design for deviation

Frank Fingerloos German Society for Concrete and Construction Technology

06.05.2021

For citations, copying and referencing see Note on Front page! page 244
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

Introduction
This document has been prepared in order to explain the proposed design values for the allowance in deviation.

6.5.3 Allowance in design The values for the allowance in deviation cdev are concentrated in one Table 6.7(NDP). The designer has to
for deviation choose an appropriate deviation value for nominal cover cnom under consideration of the expected execution
quality and the necessary reliability to reach the minimum cover cmin in the finished member.

EN 13670 [1] contains geometrical tolerances for execution in clause 10. The tolerance classes are consid-
ered as:
 Tolerance class 1: normal tolerances,
 Tolerance class 2: lower tolerances intended to be used with adjusted material factors in EN 1992-1-1,
Annex A (in prEN 1992-1-1-D7: Table A.1a(NDP), case (a)).

Explanation of Table 6.7(NDP) – Allowance for deviation cdev

Case cdev Illustrating figure
Commentary

10 mm
In general: for execution in To- value from
lerance class 1 according to EN 1992-1-1:
1 EN 13670 2004:
Standard cases in buildings by 4.4.1.3(1)P,
usual execution quality. accepted by
22 NSBs [2]

For execution in Tolerance 5 mm

class 2 according to EN 13670 new proposal
Where tolerances are specified acc. to per-
according to Tolerance class mitted sectio-
2 TC 2, the execution specification nal deviations
should request documentation that for TC 2 in
the assumptions made in the de- EN 13670,
sign are actually fulfilled in the fi- Fig. 3 b
nished member.

5 mm
Where fabrication is subjected value from
to a quality assurance system, EN 1992-1-1:
3 in which the systematic monito- 2004:
ring includes measurements of 4.4.1.3(3)
the cover 10 mm …
5 mm (4.3N)
0 mm
Where it can be assured that an
value from
accurate measurement device is
EN 1992-1-1:
used for systematic monitoring
4 2004:
and non conforming members
4.4.1.3(3)
are rejected (e.g. precast ele-
5 mm … 0 mm
ments)
(4.4N)
For concrete members in expo- 5 mm
sure class XC1, where the risk (but not in
5 of corrosion is insignificant combination See line 2
New proposal considering, that a with cmin,b for
shortfall of cmin,dur is without risk. bond)
For concrete cast against sur-
faces with exposed aggregate
(e.g. interfaces)
New proposal considering
- the lower corrosion risk inside a
concrete member without direct 5 mm
6
exposion due to environmental
condition,
- the lower bond failure risk inside
a concrete member due to redu-
ced cracking or spalling of « in-
ternal » cover.

Draft 2 06.05.2021
For citations, copying and referencing see Note on Front page! page 245
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

10 mm +
For concrete cast against une- dimension of
venness due to formwork or ex- unevenness
7
cavation sheeting (e.g. ribbed fi- similar to
nishes or architectural textures) EN 1992-1-1:
2004: 4.4.3(4)

40 mm a)
(increased mi-
nimum cover
EN 1992-1-1:
Concrete cast against prepared 2004: 4.4.3(4),
8 ground (including uneven blin- accepted by
ding layer) 20 NSBs [2],
here adopted
as increased
cdev on safe
side)

75 mm a)
(increased mi-
nimum cover
EN 1992-1-1:
2004: 4.4.3(4),
Concrete cast directly against
9 accepted by
unprepared soil
20 NSBs [2],
here adopted
as increased
cdev on safe
side)

a)
These allowances for deviation cdev apply for
the nominal cover according to this Eurocode
also for bored piles and for diaphragm walls
unless a National Annex gives other special va-
8/
lues.
9
See also BD to 6.5.1 (3).
Depending on real execution methods and national
experience adjusted values may be given.
Post-installed reinforcing bars
5 mm or
Ensuring the minimum concrete
according to
cover cmin for post-installed reinfor-
a project
cing bars in all directions
specification
- according to Table 11.2 for cmin,b
(reduced value
depending on the drilling
because the
10 method and the execution
real direction
procedure,
and length of
- according to Tables 6.3(NDP)
the drilling
and 6.4(NDP) or Q.3(NDP) for
hole can be
cmin,dur,
controlled and
- or according to Table P.2(NDP)
checked)
for cmin,dur.

[1] EN 13670:2009: Execution of concrete structures.

References
[2] CEN/TC 250/SC 2/WG 1 N 265 – Ignatiadis, A.: Analysis of NDPs in EN 1992-1-1 (August 2015)

List of versions of this document

16.02.2021 1st draft
06.05.2021 2nd draft
17.02.2023 Peer reviewed, only editorial update

Draft 2 06.05.2021
For citations, copying and referencing see Note on Front page! page 246
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG6 FprEN 1992-1-1:2023

Background document to subsection 7.2.1

Geometrical imperfections

Simon N.M. Wijte Eindhoven University of Technology,

the Netherlands

23-12-2022

For citations, copying and referencing see Note on Front page! page 247
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.2.1 Background to FprEN 1992-1-1

Introduction
This document has been prepared in order to validate the proposed EC2-provisions for geometric imperfections
that should be considered in the structural analysis to determine the internal forces in bracing structures and
vertical load bearing elements.
When imperfections are considered a distinction can be made between local and global imperfections. Local
imperfections are imperfections that will influence the internal forces in one element only. An example of a local
imperfection is a deviation from an assumed perfect straightness of a column. Global imperfections on the other
hand are imperfections that will influence the internal forces in the bracing elements of a structure like shear
walls and floor diaphragms.
In nowadays construction methods, hybrid structures are used more often. For instance for structures with
braced steel columns and concrete shear walls it is of importance that rules for global imperfections are
harmonized over the several Eurocodes. This to prevent that structural engineers have to make a choice
between the global imperfections which are prescribed in the several material related Eurocodes. For the future
it is worthwhile to consider these global imperfections in the Eurocode for loading, for instance in EN 1992-1-1.
Hereafter the relations between the several Eurocodes with respect to the global imperfections is discussed
based on the first generation of Eurocodes. When they are available a similar comparison will should made for
the second generation of the Eurocodes. Additional a relation between the geometrical imperfection in EN 1992-
1-1[2005] and the allowable deviations according to EN 13670 [2009] is discussed.

7.2.1 Summary of material related Eurocodes

Global imperfections Concrete structures
EN 1992-1-1 [2005] – Eurocode 2: Design of concrete structures – Part 1-1: General rules and rules for
buildings
Geometric imperfections are considered in clause 5.2. The global imperfections are described as follows:

Figure C7.2.1-1: Extracts from EN 1992-1-1 [2005]

Steel structures
EN 1993-1-1 [2011] – Eurocode 3: Design of steel structures – Part 1-1: General rules and rules for buildings
Geometric imperfections are considered in clause 5.3. The global imperfections are described in 5.3.2 and 5.3.3
as follows:

S.N.M. Wijte 23.12.2022

For citations, copying and referencing see Note on Front page! page 248
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.2.1 Background to FprEN 1992-1-1

Figure C7.2.1-2: Extracts from EN 1993-1-1 [2011]

Timber Structures
EN 1995-1-1 [2005] – Eurocode 5 – Design of timber structures – Part 1-1: General – Common rules and rules
for buildings
Imperfections are described in 5.4.4 which deals with the influence of imperfections on the internal forces in
plane frames and arches.

S.N.M. Wijte 23.12.2022

For citations, copying and referencing see Note on Front page! page 249
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.2.1 Background to FprEN 1992-1-1

Figure C7.2.1-3: Extracts from EN 1995-1-1 [2005]

Masonry structures
EN 1996-1-1 [2013] – Eurocode 6 – Design of masonry structures – Part 1-1: General rules for reinforced and
unreinforced masonry structures
Global imperfections are described in paragraph 5.3

Figure C7.2.1-4: Extracts from EN 1996-1-1 [2013]

Comparison
From these summaries from the several Eurocodes, table C7.2.1-1 is derived in which the inclinations that
should be taken into account in the design and a possible reduction of this inclination related to the height of the
structure are described.

Table C7.2.1-1: Inclinations according the different Eurocodes

Inclination reduction
concrete 1/200 height 2/h  2/3
1
number (0,5(1+1/m))
steel 1/200 height 2/h  2/3
1
number (0,5(1+1/m))
timber 1/200 height 2,23/h 1
masonry 1/(100htot)

S.N.M. Wijte 23.12.2022

For citations, copying and referencing see Note on Front page! page 250
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.2.1 Background to FprEN 1992-1-1

When both the inclination and the reduction for the height of the structure are considered, it can be concluded
that there is no significant difference between the several Eurocodes.

Both the EC2 and EC3 give an additional reduction that takes account for the number of elements, m. This
reduction is also identical.

Only EC2 and EC3 describe the effect of imperfections on floor diaphragm. In FprEN 1992-1-1 it is done in a
similar way.

Conclusion
From the comparison it shows that the global imperfections that influence the internal forces in the bracing
system are kind of similar in the 4 considered Eurocodes. In Eurocode 2 and Eurocode 3, additional also the
influence of imperfections on the internal forces in the diaphragms is described.
However some definitions of variables such as m and upper and lower limitations of these variables can be
different between the several Eurocodes. For that reason the definition of m is updated in FprEN 1992-1-1 to: ‘is
the number of load bearing members in one section that bear a significant part of the vertical load and that, due
to their inclination, contribute to the effect considered’ in the general description. Is is even described more
precise for the several situation to consider, such as a single columns, a bracing structure or a floor diaphragm.
In this way the definition of m is more in line with EC3, however, EC3 states that it should be about 50% of the
average load. This is more precise, but it is a discontinuous criterion and if a column for instance carries 48%, it
is questionable whether it should not be included. Besides, this percentage will change with every load
combination.

Due to the fact that structures are often not created from one material only for the ease of use it is of importance
that the global imperfections described in the Eurocodes are as identical as possible. In the revision process of
the Eurocodes this requires coordination. This coordination task can be simplified when global imperfections will
be described together with other load(effects) in Eurocode 1. l

S.N.M. Wijte 23.12.2022

For citations, copying and referencing see Note on Front page! page 251
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.2.1 Background to FprEN 1992-1-1

7.2.1 Geometrical imperfections and allowable deviations

Imperfections and EN 13670 [2009] – Execution of concrete structures
deviations
In clause 10.4 of this code the allowable deviations for columns and walls are presented in a table, which is
presented hereafter in table C7.2.1-2.

Table C7.2.1-2: Allowable deviations according figure 2 of EN 13670 [2009]

Inclination
The inclination of the structure considered in the design, should be compared to the total height of the structure.
There it should be noted that the execution requirements are valid for each bracing element (shear wall) while in
the design the effect of the inclination can be reduced when several walls are available. Hereafter in figure
C7.2.1-5 the rules for deviations during construction and imperfections for design are compared, for the situation
where only one wall is considered.

S.N.M. Wijte 23.12.2022

For citations, copying and referencing see Note on Front page! page 252
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.2.1 Background to FprEN 1992-1-1

Execution rule:
 = the smaller of 50 mm/h and /(200n)
where:
n is number of stories (assumed n = h/3m)

Design rule
= h 1/200
where
h = 2/h
 2/3
1

0.006

0.005

0.004

0.003


design
execution
0.002

0.001

0
0 10 20 30 40 50
total height [m]

Figure C7.2.1-5: Comparison of inclination q due to execution and design rules [n = h/(3
metres)]

Mid height eccentricity of a single member

The eccentricity at mid height is influenced by both the curvature of the single members as well as the horizontal
deviation of sequential walls. It is assumed that 50% of the moment due to the horizontal deviation is taken by
the floors and the other 50% is resisted by both the walls. So each wall is loaded by an eccentricity that equals
25% of the horizontal deviation.
The limit of the horizontal deviation is 15 mm or t/30. The last term is only governing when the wall thickness t >
450 mm. As this is not within practical limits, it is not considered here.
The total eccentricity at mid height due to the allowable deviations in the execution phase therefore leads to the
following eccentricity at mid height:

Execution rule:
e = 3,75 mm + h/300  18,75 mm
Which, in figure C7.2.1-6, is compared to the initial eccentricity in the design.

S.N.M. Wijte 23.12.2022

For citations, copying and referencing see Note on Front page! page 253
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.2.1 Background to FprEN 1992-1-1

Design rule:
e = h 1/400 l0
where
h = 2/h
 2/3 and  1

For this comparison its assumed that the effective length l0 equals the height of the wall h.

25
eccentricity at mid height [mm]

20

15

design
10
execution

0
0 1 2 3 4 5 6
single wall height [m]

Figure C7.2.1-6: Comparison of eccentricity due to execution and design rules

Conclusion
From figure C7.2.1-5 it can be concluded that the inclination of shear walls, considered in the design is large
enough. It even can be stated that specially for structures with a significant total height, the design rule in EN
1992-1-1 [2005] can be considered to be conservative. This among other reasons has led to the introduction of
7.2.1.1(3) in which, as an alternative method it is described that geometrical imperfections that should be taken
into account in the design also may be derived from specific execution specifications. In that situations the
geometrical imperfections to be considered should be at least 1,2 times the specified allowable deviations. This
is to account for deviations other than those that can be measured, such as due to differential shrinkage and
differential temperatures. This has also led to the lowering of the lower limit of variable n form 2/3 to 0,4.

From figure C7.2.1-6 it should be concluded that the initial eccentricities for single walls or columns considered
in the design is small in relation to the allowable eccentricities and horizontal deviations. This should be a
relevant item when the revision of EN 13670 is considered.

References

EN 1992-1-1 [2005] – Eurocode 2: Design of concrete structures – Part 1-1: General rules and rules for buildings
EN 1993-1-1 [2011] – Eurocode 3: Design of steel structures – Part 1-1: General rules and rules for buildings

S.N.M. Wijte 23.12.2022

For citations, copying and referencing see Note on Front page! page 254
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.2.1 Background to FprEN 1992-1-1

EN 1995-1-1 [2005] – Eurocode 5 – Design of timber structures – Part 1-1: General – Common rules and rules for buildings
EN 1996-1-1 [2013] – Eurocode 6 – Design of masonry structures – Part 1-1: General rules for reinforced and unreinforced masonry structures
EN 13670 [2009] – Execution of concrete structures

List of versions of this document

23.2.2021 1st draft
25.2.2021 2nd draft
23.12.2022
21.03.2023 peer reviewed and editorial update

S.N.M. Wijte 23.12.2022

For citations, copying and referencing see Note on Front page! page 255
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG6 FprEN 1992-1-1:2023

Background document to clause 7.3.2 (3)

Linear elastic analysis with redistribution

without explicit check on the rotation
capacity

Simon N.M. Wijte Eindhoven University of Technology,

the Netherlands

28-03-2023

For citations, copying and referencing see Note on Front page! page 256
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (3) Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared in order to validate the EC2-provisions for the rules to apply redistribution of
moments resulting from a linear elastic analysis without verification of the rotation capacity, see 7.3.2(3). Rules
for a further redistribution of moments, where verification of the rotation capacity is required, are addressed in
7.3.2(5) and are considered in the background document that has been written for that clause.

The proposal is based on the EN 1992-1-1:2004 and NEN 6720 (VBC 1995) [1] the latest Dutch code for the
design of concrete structures.

7.3.2(3) Reasons for change

Redistribution In 5.5 of EN 1992-1-1 a relation is described between the ratio of redistribution  (= Mredis/Melas) and the ratio of
the depth of the compression zone and the effective depth xu/d (kx). For concrete with a characteristic strength
not greater than 50 MPa. This relation is:

  k1 + k2 kx (eq 5.10a)

where k1 and k2 are NDP’s of which the recommended value are given as:
k1 = 0,44

 0,0014 
k2 = 1,25 0,6  
  cu2 

cu2 = 0,0035 for fck  50 MPa

Hence:

  0,44
  0,44 + 1,25 kx or kx  = 0,8  - 0,35
1,25

The relation between  en kx for fck  50 MPa is shown in figure C7.3.2(3)-1

δ k x ( δ) 
0.7 0.208
0.4 0.8 0.288
0.9 0.368
0.3
kx( δ) 1 0.448
0.2

0.1

0
0.7 0.75 0.8 0.85 0.9 0.95 1
δ
Figure C7.3.2(3)‐1: Relation between  and kx according 5.10a

This relation is valid for concrete strengths up to 50 MPa. For higher strengths equation 5.10b should be used.

  k3 + k4 kx (eq 5.10b)

where k3 and k4 are NDP’s of which the recommended value are given as:
k3 = 0,54

 0,0014 
k4 = k2 = 1,25 0,6  
  cu2 

The relation between  en kx for several concrete strengths is shown in figure C7.3.2(3)-2

For citations, copying and referencing see Note on Front page! page 257
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (3) Background to FprEN 1992-1-1:2023

Figure C7.3.2(3)‐2: Relation between  and kx according 5.10b for fck = 50, 60 and 80 MPa

In these relations a variation of the yield strength of the reinforcement, which may have a more significant
influence on the ductility than the ultimate strain of the concrete, is not considered. Furthermore, the use of
prestressing steel is not considered. Therefore, a proposal for a change is given hereafter.

Background
General
Rotation capacity is required to allow for redistribution. Rotation capacity is created by the yielding of
reinforcement. Yielding of reinforcement is also desired in cases of the determination of the moment resistance,
even without redistribution.

Besides the considered redistribution of the effect of the considered external forces, there is also a need to
make the none considered redistribution, caused by the effects of shrinkages, deformation due to temperature
effects and differential settlement of supports, possible.
xu
d

Figure C7.3.2(3)‐3: Stresses and strains at ultimate moment resistance

As shown from figure C7.3.2(3)-3, yielding of the tension reinforcement will be assured when the depth of the
compression zone is limited:

xu  cu2  cu2 E s
= kx  =
d fyk  cu2 E s  fyk
 cu2 
Es

To ensure yielding of the reinforcement a margin  can be applied to the strain that will occur in the
reinforcement before the concrete will fail. This margin is required because the considered material properties
are not the actual properties and it is those actual properties that will determine whether yielding of steel or
crushing of concrete will be the decisive criteria. Besides that the assumption that plane sections remain plane
is not always a valid assumption, hence:

For citations, copying and referencing see Note on Front page! page 258
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (3) Background to FprEN 1992-1-1:2023

 cu2 E s
kx 
 cu2 E s   fyk

This is valid for reinforcing steel. In structures in which prestressing steel or a combination of prestressing steel
and reinforcing steel is applied it is also desirable that yielding of steel occurs before the concrete will fail in
compression. Prestressing steel will start to behave non-linearly at a stress level of fpd. However in prestressing
steel already the working stress p,m is present. So non-linear behaviour will start in case of a prestressing steel
stress increase equal to fpd - p,m.

Application in the Netherlands

To consider the presence of prestressing steel, in the Netherlands the equation has been changed to:
 cu2 E s
kx.max =
 cu2 E s   f

Where:
f is the steel stress increase where nonlinear behaviour occur, taken as a weighted average of the
reinforcing steel and the prestressing steel
(fpd   p,m )Ap  fyd As
=
Ap  As

 is the margin taken as 1,4.

For situations without considered redistribution where cu2 equals 0,0035 and Es = 2,0ꞏ105 MPa this results in:

500
kx.max =
500  f

For fyk = 500 MPa the value of kx.max = 0,50, which is significantly greater than kx.max according equation 5.10
where kx.max = 0,45.

Beside this, a linear relation is assumed between the redistribution ratio  and the allowable depth of the
compression zone:

500 f
kx.max = - (1 - ) =  -
500  f 500  f

This equation can be rewritten as:

f f xu
kx.max =  -  = +
 cu2Es  cu2Es d
f f
 

Figure C7.3.2(3)-4: Relation between  and kx according proposal for fck = 50, 60 and 80 MPa, fyk = 500 MPa
and  = 1,4

For citations, copying and referencing see Note on Front page! page 259
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (3) Background to FprEN 1992-1-1:2023

Proposal for an alternative equation 5.10

Based on the previous the following equation applied as an alternative for equation 5.10 in EN 1992-1-1:2004

1 xu
= +
 E d
1  cu s
f

 0,7 where Class B and Class C reinforcing steel or prestressing steel is used (see Annex C)
 0,8 where Class A reinforcing steel is used (see Annex C)

where:
 is the ratio of the redistributed moment to the elastic bending moment

(fpd   p,m )Ap  fyd As

f =
Ap  As

xu is the depth of the neutral axis at the ultimate limit state

d is the effective depth of the reinforcing and prestressing steel
p,m is the long-term stress level in prestressing tendons at the state of zero (elastic) strain of the concrete
at the same level.

For 1/ a factor equal to 0,7 is chosen.

It should be noted that in FprEN 1992-1-1:2023 the ultimate strain cu is not varied for the different concrete
grades due to which the influence of the different concrete grades is lost. However, as can be seen in figure
C7.3.2(3)-4, the influence of the different values for cu is limited.

[1] NEN, TGB 1990 Regulations for concrete - Structural requirements and calculation methods, NEN
References
6720, Delft 1995.
.

List of versions of this document

18.2.2021 1st draft
23.12.2022
27.03.2023 Peer review
27.03.2023 Final

For citations, copying and referencing see Note on Front page! page 260
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to clause 7.3.2 (5)

Formula for Rotation Capacity

Alejandro Pérez Caldentey, Simon Wijte, Anssi Laaksonen 28.03.2023

For citations, copying and referencing see Note on Front page! page 261
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to paragraph 7.3.2 (5) (Rotation Capacity).

7.3.2 (5) Reasons for change

Rotation capacity EN 1992-1-1:2004 provides rotation capacity in graphical form in Figure 5.6N, reproduced below in Figure
C7.3.2(5)-1. There are several reasons why this presentation of rotation capacity should be revised:
- With the current proposal for factor cc, the ultimate strain of concrete is not a function of the
concrete strength anymore, so there is no compelling need for the curves in Figure 5.6N of EN
1992-1-1:2004 to be dependent on the concrete strength.
- The curves of Figure 5.6N do not seem to be totally correct, since for high values of the xu/d ratio,
the rotation capacity should be independent of the steel class since failure occurs by crushing of
concrete.
- It is not fully transparent how the curves of Figure 6.6N have been derived. There is very limited
background information which does not provide information on the how, for instance, tension
stiffening effects were considered what the refined curves look like or how the simplified bi-linear
law shown of Figure 5.6N was derived.
- There is no information available regarding the model uncertainty factor that was applied to
produce this graph and no information regarding how it fits with experimental data.
- There is no information regarding the limits of validity of this simplification.
- EN 1992-1-1:2004 does not account for the size effect which has been shown by tests to be a
major variable to predict rotation capacity (see, for instance, [C07-8]).

Figure C7.3.2(5)‐1Reproduction of Figure 5.6N of EN 1992-1-1:2004 which provides the rotation

capacity
The above points fully justify a revision of paragraph 5.6.3(4) of EN 1992-1-1:2004.

Enhancement of ease of use

With today’s modern computational methods, it is not practical to formulate rules and requirements in codes
in terms of graphs. A much more efficient way is to present the information in a closed-form equation, if
possible. The proposed change achieves this goal.

Problem statement
Significant rotation capacity and thereby structural ductility can be achieved by providing the structure with
ductile sections. In structures that can be assimilated to continuous beams, plastic rotation occurs mostly in
the zone around the negative bending moment. The area from the support to the point of zero moment can
be isolated and assimilated to a simply supported span subjected to a point load (the reaction) and the
distributed loads which have the opposite direction. The problem can further be simplified (and tested) by
disregarding the effect of the distributed loads. In this way the problem can be modelled as a simply
supported beam with a span equal to the distance between the points of contraflexure and a single
concentrated load where the slenderness corresponds to roughly half of the span of the continuous beam
subjected to a point load (see Figure C7.3.2(5)-2).

For citations, copying and referencing see Note on Front page! page 262
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

The plastic rotation capacity is the rotation due to the excess curvature with respect to the curvature
corresponding to the yielding moment that occurs between the section at which the yielding moment (Myd)
is reached and the section at which the ultimate moment (MRd) is reached (see Figure C7.3.2(5)-2). Due to
shear cracking, it can be assumed that the moment curve is shifted horizontally by a given magnitude. In
Figure C7.3.2(5)-2, the magnitude of the shift is assumed to be one effective height (d). Reality, of course
can be more complex.
The bottom part of Figure C7.3.2(5)-2 shows a graphical representation of the plastic rotation.

Figure C7.3.2(5)‐2 Top: Bending moment (shifted due to shear cracking near the point of application
of the load) and Bottom: curvatures and graphical representation of plastic rotation
Note that in Figure C7.3.2(5)-2 the curvatures are mean curvatures, since curvatures will, have maximum
values at the cracks and will reduce between cracks due to tension stiffening effects. This will result in a
reduction of the rotation capacity with respect to the assumption that all sections are fully cracked. So,
tension stiffening is an unfavourable effect when dealing with rotation capacity. A more detailed
representation of the curvature law is given in Figure C7.3.2(5)-3.

For citations, copying and referencing see Note on Front page! page 263
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Figure C7.3.2(5)‐3 A more detailed representation of the distribution of curvatures in the plastic
rotation zone, taken from reference[C07-6] (Eligehausen, 1987)

Deduction of the Expressions included in 7.3.2(5)

Expression (7.18) of FprEN 1992-1-1:2023 is based on the integration of a simplified plastic curvature law
as shown in Figure C7.3.2(5)-4. The variation in curvatures from the yielding to maximum is assumed to be
linear. The above simplifications and others that will be adopted in the next section are warranted due to the
large scatter that occurs when examining experimental data, as will be shown later. Based on this distribution
of forces, the plastic rotation can be evaluated as shown in Eq. (1)

  1  1  
       
 l   r u , m  r  y , m   lcf  1   1  1 
 pl , simpl   2  cf  aMy    2   aMy     2d       
 
  2  2  2   r  y,m   r u , m  r  y , m 

 
  1  1  
         (1)
  lcf    r u , m  r  y , m   1   1  
  2   aMy   2d         
2 2  r   r  y , m  
    u ,m

 
 
 lcf  1  1 
   aMy  2d        
 
 2    r u , m  r  y , m 

For citations, copying and referencing see Note on Front page! page 264
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Figure C7.3.2(5)‐4 Simplified plastic curvature law

The value of aMy, that is, the distance from the point of contraflexure to the point where the yielding moment
is reached can be derived as follows:

lcf 
M Rd  PRd 
4  M yd  2aMy  a  M yd lcf
 My (2)
PRd M lcf M Rd 2
M yd  aMy  Rd
2 

Myd and MRd can be determined assuming linear behaviour of concrete in compression for Myd and a
rectangular stress block for MRd, as shown in Figure C7.3.2(5)-5, x should be larger than xu, since in most
cases the increase in strain at the reinforcement will be much larger than the increase of the strain at the
compression fibre. At the limit they will be the same for a section that reaches, at the same time, yielding
and maximum concrete strain. Of course, it is possible that when the yielding moment is reached the
behaviour of concrete is not linear anymore. This will occur for sections with large steel quantities. In these
cases, however, the rotation capacity will be small, and plastic analysis would not be recommended. For
this reason, the assumption is maintained in the derivation that follows, without further limitation than that
the rotation capacity cannot be negative.

Figure C7.3.2(5)‐5 Sectional analysis to determine Myd and MRd

For the description of both the moment resistance and the corresponding curvature the stress-strain diagram
shown in Figure C7.3.2(5)-6 is used. In this diagram ud,ef is the strain in the reinforcement that is reached
when the ultimate strain in the concrete is determining the ultimate moment resistance.

For citations, copying and referencing see Note on Front page! page 265
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Figure C7.3.2(5)‐6 Stress-strain diagram for reinforcement steel

With the assumptions described in Figure C7.3.2(5)-5, the yielding and the ultimate bending moments can
be derived, and from these expressions, aMy can be determined as shown in Eq. (3):

 1 x   1 x
M yd  As f yd d 1    1  
 3d  f yd  3 d  lcf
  aMy  (3)
 x  f s , ef  xu  2
M ud  As f s , ef d 1  0.4 u  1  0.4 d 
 d   
The mean curvature at yielding (1/r)y,m can be derived from Eq. (4)

1  yd
   TS My
 r  y,m  x
d 1  
 d (4)
M 
TS My  1  0.6  cr   0.4
M 
 y 
In Eq. (4) TSMy is the tension stiffening factor for SLS and Mcr is the cracking moment. For tension stiffening
the upper bound is taken (factor of 0.6) because this is the conservative assumption.
The mean curvature at ULS (1/r)u,m can be derived from Eq. (5):

1   ud  
   TSMu min  ; cu 
. (5)
 r u , m   d  xu  xu 
The two limits of the above expression depend on whether the failure of the section is due to failure of steel
or failure of concrete respectively. In this case the effect of tension stiffening is very high, because when the
second branch of the constitutive law of steel is reached a small increase in steel stress will produce very
large strains. This means that the strain at the crack can be very large, while the strain between cracks can
be quite small. To model Tension Stiffening effects, the Tension Chord Model (Marti, 1998) [C07-7] will be
adopted. This model is relatively simple. It considers a constant bond stress whose value depends on the
concrete strength and on whether the steel has yielded or not (see Eq. (6)).

 2

 0.30 f 3
cm if  s  f yd
b   2
(6)
0.60 f 3 if  s  f yd
 cm

The tension stiffening factor TSMu represents, approximately, the ratio between the mean strain in steel
between the section of the crack and the section between cracks, and the maximum strain at the crack (see
Eq. (7)).

For citations, copying and referencing see Note on Front page! page 266
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

srm

   y  dy
s
(7)
TS Mu  0

 ud , ef
For the determination of TSMu, two cases can be distinguished, as shown in Figure C7.3.2(5)-7, where s1 is
the distance between the crack and the point where fyd may be found as the stress in the rebar. The first
case is where srm/2 is smaller than s1 and second case is the oposite. Hereafter s1 is described as a function
of srm and . Where  is the ratio between the difference in the force in the bar between the ultimate force
and the yielding force end the force that can be transmitted by bond between the section of the crack and
the section between cracck with the lowr bond stress.

srm
s1  
2
( f s , ef  f yd ) AØ 2 ( f s ,ef  f yd )Ø
 
 Ø 0,3 f cm 2/3
srm 0, 6 srm f cm 2/3

:
- In the first case where the difference between the ultimate force and the yielding force of the bar
is larger than the force that can be transmitted by bond between the section of the crack and the
section between cracks with the lower bond stress (>1.00), then the variation in force will be
linear and the minimum and maximum strains and TSMu factor can be determined as shown in Eq.
(8):

1
 min   su ,ef     yd 
 ud , ef
 ud ,ef   min
 mean  (8)
2
 mean 1   yd 
TS Mu   1 1  
 ud ,ef 2   ud ,ef 
- If, on the contrary, the difference between the ultimate force and the yielding force in the bar is
smaller than the force that can be transmitted by bond between the section of the crack and the
section between cracks with the lower bond stress (<1.00), then the variation in force will be bi-
linear and the minimum and maximum strains and TSMu factor can be determined as follows.
First, the force in the bar at the section in the middle between two cracks can be determined from
Eq. (9):
2 2
s 
Fmin  A f s , ef   0.3 f cm3 s1   0.6 f cm3  rm ,cal  s1  
  2 
A f f  s s ,ef yd

2
srm , cal
 A f s ,ef  A  f s ,ef  f yd    0.6 f cm3 1    
2 (9)
 2 
 A f s ,ef  A  f s ,ef  f yd  1  1     
  
2 
 A f s ,ef  A  f s ,ef  f yd    1
 
Since the bond stress that can be transferred between the section of the crack and the section
between cracks is larger than the maximum force in the bar minus the force at yielding, the stress
at the section between cracks will be within the elastic range and the minimum strain can therefore
be derived as shown in Eq. (10):

For citations, copying and referencing see Note on Front page! page 267
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

 yd  min  min A
 
f yd Fmin , 2 
A f s ,ef  A  f s , ef  f yd    1
A  
 2
f s ,ef   f s , ef  f yd   1  
  min   yd  
f yd (10)
 2  2
f s ,ef  2    f yd  1  
     
  yd
f yd
 f s ,ef  1  2 
  yd  2  1     1   
 f      
 yd

The mean strain can then be determined from Eq. (11):

 
 ud , ef   yd s1  yd   min  s1 
 s , mean   1  
2 srm 2  s rm 
2  2 
 ud ,ef   yd  yd   min  1 1   
  1      ud ,ef   yd   min 
2 2 2 2 2
 1  f s , ef  1  2   1   
  ud ,ef   yd   yd  2  1     1    
2 2  f       2
 yd

 1  f s , ef  1   1  1  1   
  ud ,ef   yd   yd  2 1    1     
2 2  f yd         2

  1  f s ,ef   1  1  1    
  ud ,ef   yd     2  1 1      (11)
2  2   f yd      2 

   
  
  
  1 1      f s , ef 1   1  
  ud ,ef   yd       1   1      
2 2 
  2   f yd 2    
  
 1
 
1
 
 2  2
 
  1  f s , ef 1   1 
  ud ,ef   yd     2    
2  2  f yd 2     
  
    f s ,ef  1 
  ud ,ef   yd 1    1  2     
2 
 2  f yd 
   
The tension stiffening factor referred to the steel strain can then be obtained from Eq. (12):

 mean   yd    f s , ef  1 
TS Mu , s    1     1  2      (12)
 ud ,ef 2  ud ,ef  2     
  f yd 

For citations, copying and referencing see Note on Front page! page 268
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Figure C7.3.2(5)‐7 Distribution of strains between cracks in the zone of maximum strain and
correspondence with the stress-strain diagram of steel, force variation in the bar and bond stresses
As a simplification it is assumed that the tension stiffening factor for strains can be assimilated to the tension
stiffening factor relating to curvatures. This is because it is not easy to determine how the compressive strain
will vary from the section of the crack to the section between cracks, since there may be slip at the section
between cracks. This proposal is conservative since it maximizes the tension stiffening factor and is also
not very unrealistic since generally the strain in the reinforcement will be much larger than the strain in
concrete (see Eq. (13)).

1  s , mean   c , mean

  
 r mean d
TS Mu    s , mean  TS Mu , s
1  ud ,ef   c ,max  ud ,ef (13)
  
r
 u ,ef   cu

d
By replacing the above formulations into Eq. (1), the final formulation is derived as follows:

 1   yd 
 pl , simpl   k1, pl lcf  2d      TS My  (14)
  u , m
r d  x 
where:

  1 x 
1  f yd 1   
k1, pl  1  3d 
2 f s , ef  xu  
  1  0.4 d  
  

M 
TSMy  1  0.6  cr   0.4
M
 y 

1   ud  
   TS Mu min  ; cu  .
 r u , m   d  xu  xu 

For citations, copying and referencing see Note on Front page! page 269
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

 1   yd 
1   1   if   1.00
 2   ud , ef 
TS Mu 
  yd    f s , ef  1 
2  1     1  2      if   1.00


ud , ef  2  f yd 
   
A  f s ,ef  f yd    f s , ef  f yd 
 2
 2
srm ,cal
 f 3
cm 2 f cm3 srm ,cal
2

Calibration of the model

The model as developed above, introduces some very simplified assumptions such as that the variation of
curvature between the section of maximum moment and the section of the yielding moment is linear, that
the shift in the bending moment law is equal to one effective depth on each side of the hinge and that the
maximum strain of concrete it 0.35%. These assumptions do not necessarily correspond to reality. For this
reason, in prEN 1992-1-1:2020-11 several calibration factors were introduced to adapt the model to the
experimental evidence. Additionally, as the variation of rotation capacity measured in similar specimens is
very high, a model uncertainty factor has also been introduced to provide a reliability factor, , equal to 3.8.
The formulation included in FprEN 1992-1-1:2023 is thus altered with respect to the above derivation as
shown in Eq. (15).

 pl , simpl 
k k l  2kd d    1 
y 1, pl cf  yd 
    TS My 
   u , m
r d  x 
(15)
1   ud  
 TS Mu min  ;  cu
 
 r u , m   d  x  x 
 u u 

where:
ky is a shape factor for the variation of mean curvature between the section of yielding to the section of
rupture;
kd is a factor to quantify the shift of the bending moment law due to diagonal shear cracking;
 is a factor to account for the experimentally demonstrated fact that the 0.35% limit to concrete strain
is an extremely conservative value for most design situations, as it will be shown below. This factor
is key to explain the size effect observed in rotation capacity tests ([C07-5], [C07-8], [C07-9], [C07-
11], [C07-12]);
 is a model uncertainty factor to guarantee an adequate level of reliability of the formulation.

Regarding the maximum compressive strain of concrete

Tests by Mattock [C07-11] and Corley [C07-12] will be used to illustrate the fact that limiting the concrete
strain to 0.35% is overly conservative. Furthermore, it will be seen that the maximum compressive strain is
dependent on the size of the element (which explains the clear size effect shown in the experimental series
considered herein) and on the amount of transversal reinforcement provided.

Table C7.3.2(5)-1 provides a general overview of the range of variables which are encompassed by these
tests. Note that the reinforcement ranges, especially for the Corley tests are very high, which would provide
high strains in the compressive fibres. Also, note that the value of the ratio between the ultimate steel stress
and the yield stress is very high for European Standards, as factor k is limited to a maximum of 1.35 for
class C steel, according to EN 1992-1-1:2004, and, also, to FprEN 1992-1-1:2023.

Table C7.3.2(5)‐1 – Range of parameters covered in the tests by Mattock and Corley
# tests Range Range Range Range Range
d [mm] xu/d k  w

Mattock 31 254‐508 0.27‐0.47 1.70‐1.82 0.74%‐2.95% 0.33%‐1.47%

Corley 40 152‐800 0.30‐0.60 1.40‐1.80 1.94%‐5.24% 0.15‐2.62%

Figure C7.3.2(5)-8 and -9 show the results pertaining to the maximum measured compression strain in the
tests of Mattock and Corley as a function of the effective depth. The tests are 3-point bending tests. The
compressive strain was measured using a Linear Differential Transformer Gage located at mid-span. The
measuring length of the gage is not reported, but from the picture included in the publication it is estimated
to be at least 20 cm long. It is certain that measuring the compressive strain is not an easy task and that,
for instance, the tests by Bigaj [C07-9], show a significant variation of the compressive strain along the

For citations, copying and referencing see Note on Front page! page 270
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

compressed fibre. Nonetheless, the measured values are so much above the 0.35% limit that it must be
conceded that this limit is just a very conservative lower bound. In fact, the conclusion by Mattock is as
follows:

“The maximum concrete compressive strain in a reinforced concrete beam at a section of

maximum moment under a single concentrated load or over a support can be very much in
excess of the usually assumed value of 0.003. Consequently, the curvature at ultimate
strength can also be very much greater than the value calculated on the assumption that
the maximum concrete compressive strain is limited to 0.003”

Corley arrives at a virtually identical conclusion.

Figure C7.3.2(5)-8 and -9 also clearly show that the maximum compressive strain is a function of the
effective depth and that this function has a hyperbolic shape.

Mattock ‐ measured ultimate concrete strain as a function of effective depth

0.070

0.060

0.050
cu,exp

0.040

0.030

0.020

0.0035
0.010

0.000
0 100 200 300 400 500 600
d[mm]

Figure C7.3.2(5)‐8 Ultimate compressive strain in concrete as measured by Mattock [C07-11] in

rotation capacity tests as a function of the effective height of the specimen

Corley ‐ measured ultimate concrete strain as a function of effective depth

0.070

0.060

0.050
cu,exp

0.040

0.030

0.020

0.0035
0.010

0.000
0 100 200 300 400 500 600 700 800 900
d[mm]

For citations, copying and referencing see Note on Front page! page 271
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Figure C7.3.2(5)‐9 Ultimate compressive strain in concrete as measured by Corley [C07-12] in

rotation capacity tests as a function of the effective height of the specimen
Figure C7.3.2(5)-10 and -11 also show that the maximum compressive strain increases as the transversal
reinforcement ratio, w, increases. This relationship will be assumed to be linear. Based on this evidence, it
is suggested to formulate the maximum concrete strain as shown in Eq. (16).

d ref
  cu   cu , d ,   k ,0  k ,1  k ,2  w   cu , max (16)
w
d

where k0, k1 ,k2 and cu,max are calibration factors.

Mattock ‐ measured ultimate concrete strain as a function of the transversal

0.070 reinforcement ratio

0.060

0.050
cu,exp

0.040

y = 0.6657x + 0.0112
0.030 R² = 0.0452

0.020

0.0035
0.010

0.000
0.000 0.005 0.010 0.015 0.020 0.025
w

Figure C7.3.2(5)‐10 Ultimate compressive strain in concrete as measured by Mattock [C07-11] in

rotation capacity tests as a function of the effective transversal reinforcement ratio

Corley ‐ measured ultimate concrete strain as a function of transversal

0.080 reinforcement ratio

0.070

0.060

0.050 y = 1.7383x + 0.0049

R² = 0.7285
cu,exp

0.040

0.030

0.020
0.0035
0.010

0.000
0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040
w

For citations, copying and referencing see Note on Front page! page 272
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Figure C7.3.2(5)‐11 Ultimate compressive strain in concrete as measured by Corley [C07-12] in

rotation capacity tests as a function of the transversal reinforcement ratio
Because the Mattock and Corley tests have a very high scatter, the calibration of the model parameters has
been performed considering the tests of Bigaj et al. [C07-9], Bosco et al. [C07-5] and Calvi et al [C07-10].
Table C7.3.2(5)- 2 shows the range of parameters covered by these tests which is fairly wide. Only the tests
by Bosco et al. have transversal reinforcement.

Table C7.3.2(5)‐2 Range of parameters covered in the tests by Bigaj, Bosco et al. and Calvi et al.
# tests Range Range Range Range Range
d [mm] xu/d k  w

Bigaj et 9 90‐450 0.06‐0.29 1.13‐1.18 0.28%‐1.12% 0

al.

Bosco et 16 176‐565 0.004‐0.42 1.08‐1.15 0.27%‐2.08% 0.13%‐0.38%

al.

Calvi et 5 140 0.067‐ 1.13‐1.24 0.23%‐0.34% 0

al. 0.088

The resulting calibration parameters are shown in Eq. (17). ky can the taken as nil because the values of
factor k1,pl are very small and the effect of the first term of the initial parenthesis on the rotation capacity is
negligible. This simplifies the model.

ky  0
kd  0.65
 d ref 
 cu , d ,    0.002  0.0015  3 w    cu ,max (17)
w
 d 
d ref  900 mm
 cu ,max  0.015

With these factors, the formulation to estimate the rotation capacity can be written as shown in Eq. (18).

1.3d    ud  cu , d , w   yd 
 pl , simpl   TS Mu min  ;   TS My 
  d  xu 
  xu d  x 
 (18)
1.35
 cu , d ,   0.002   3 w  0.015
w
d  mm 

Figure C7.3.2(5)-12 to -15 show that the calibrated model can reproduce the effects of the main variables
quite well.
In the tests by Bigaj et al. the size effect and the effect of reinforcement ratio (expressed here in terms of
xu/d) are clearly captured. For the 90 mm beam the effect of reinforcement ratio and the effect of size are
offset by the low ultimate strain of the steel used.

For citations, copying and referencing see Note on Front page! page 273
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Tests by Bigaj
80

70 Other effects are explained by concrete/bond

strength

60 d=180 mm

50

 [mrad] 40 d=450 mm
d=90 mm

30

20

10 su=3.6%
Others su9.2‐9.4%

0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
xu/d
Exp. Th.

Figure C7.3.2(5)‐12 Comparison of the test results and predictions obtained from the calibrated model
for the tests of Bigaj and Walraven

Tests by Bosco ‐ Steel Type B500H

70

Exp. h=600 mm
60

50

Exp. h=600 mm
 [mrad]

40 Exp. h=600 mm

30

20 Exp. h=400 mm

10 Exp. h=200 mm

Exp. h=600 mm Exp. h=400 mm

0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
xu/d
Exp. h=200 mm Exp. h=400 mm Exp. h=600 mm Theoretical

Figure C7.3.2(5)‐13 Comparison of the test results and predictions obtained from the calibrated model
for the tests of Bosco et al. for steel type B500H
Referring to Figure C7.3.2(5)-13 and to the elements with 600 mm of effective height, the effect of xu/d on
the rotation capacity is clearly captured. For very low reinforcement ratios, the rotation capacity increases
with the increase of , up to a maximum when the capacity starts decreasing. For xu/d equal to approximately
0.27 and 0.40, the size effect is clear with the 400 mm depth providing a larger rotation capacity than the
600 mm effective depth and the 200 mm effective depth providing a larger rotation capacity than the 400
mm depth. The same analysis can be done for Figure C7.3.2(5)-14.

For citations, copying and referencing see Note on Front page! page 274
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Tests by Bosco ‐ Steel Type B500N

70

Exp. h=200 mm
60

50

 [mrad]
40

Exp. h=400 mm
30 Exp. h=400 mm
Exp. h=200 mm
Exp. h=400 mm Exp. h=600 mm

20

10 h=600 mm
Exp.
Exp. h=600 mm Exp. h=400 mm

0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45
xu/d
Exp. h=200 mm Exp. h=400 mm Exp. h=600 mm Theoretical

Figure C7.3.2(5)‐14 Comparison of the test results and predictions obtained from the calibrated model
for the tests of Bosco et al. for steel type B500N

Tests by Calvi et al. (Only high bond tests and single load tests)
140
Higher reinforcement ratios lead to smaller
crack spacing and smaller TS effects. This is why
120 the curves show a growing tendency with xu/d su=7.99%
for small reinforcement ratios

100
 [mrad]

80

60
su=3.45%

40

20

0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
xu/d
Exp. Theoretical

Figure C7.3.2(5)‐15. Comparison of the test results and predictions obtained from the calibrated
model for the tests of Calvi et al
The tests by Calvi et al, shown in Figure C7.3.2(5)-15, are centred on the effect of the ultimate strain of steel
for elements with very low reinforcement ratio. The model correctly predicts a higher rotation capacity for
members built with the steel with the larger ultimate strain. Because the reinforcement ratios are very low,
the rotation capacity increases with increasing xu/d.

Figure C7.3.2(5)-16 shows a comparison between experimental and estimated rotation capacity values. The
slope of the correlation line is close to 1.00, if slightly on the conservative side.

Tests by Mattock and Corley have not been used for the calibration of the model because they show a much
larger scatter. Nonetheless, these results fall mainly on the conservative side, especially regarding the tests
by Mattock (see Figure C7.3.2(5)-17 and -18). This is most likely due to the limitation of the maximum
compressive strain.

For citations, copying and referencing see Note on Front page! page 275
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

Predicted vs Experimental
(Bigaj, Bosco, Calvi)
y = 1.1044x
R² = 0.9568
140

120

100

exp [mrad]
80

60

40

20

0
0 20 40 60 80 100 120 140
model [mrad]

Figure C7.3.2(5)‐16 Comparison between measured and estimated rotation capacity for the tests of
the calibration set.

Mattock ‐ Experiment/Prediction vs xu/d

7.00

6.00

5.00
exp /model

4.00

3.00

2.00

1.00

0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
xu/d

Figure C7.3.2(5)‐17 Ratio of experimental over estimated rotation capacity as a function of xu/d for
the tests of Mattock

For citations, copying and referencing see Note on Front page! page 276
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

All tests, divided by model uncertainty factor‐ Experiment/Prediction vs xu/d

20

18

16

14

exp /model
12

10

0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
xu/d

Figure C7.3.2(5)‐18 Ratio of experimental over estimated rotation capacity as a function of xu/d for
the tests of Corley

Model uncertainty factor, 

Because, despite the model showing adequate tendencies, there is still a very significant scatter, a model
calibration factor must be applied to guarantee an adequate level of reliability. There are many reasons for
the high scatter. Among them the following:
- Except for the tests by Bigaj et al, the ratio between the ultimate and yield strength of
reinforcement (factor k) are reported for the steel but not for each individual specimen. Thus, the
scatter in this parameter cannot be properly accounted for.
- The reported yield and ultimate loads are not always compatible with the reported values of the
material parameters. It occurs, for example that, in some cases, with the reported steel strength
the reported ultimate load cannot be reached.
- The measurement of the rotation can be done in several manners. There is no common
methodology, and this contributes to the observed scatter.
- As it is not reported in the test series, the crack spacing was determined using the formulation of
FprEN 1992-1-1:2023, Clause 9.2.3.
Table C7.3.2(5)- 3 shows the statistics of the performance of the model for the calibration tests. The mean
ratio between the experimental and estimated values of rotation capacity is close to 1.00 (1.06) and the CoV
is rather large at 38.1%.
Although, the formulation is not fully valid for such a high CoV, and pending a more detailed analysis
according to EN 1990, the formulation of FprEN 1992-1-1:2023, Annex F, formula (F.9), has been applied
to determine a first value of the model uncertainty factor, considering a sensitivity factor of 0.8:

e0.803.8CoV
   3.0 (19)

Table C7.3.2(5)‐3– Statistics of the ratio between experimental and estimated values
expth
 1.058
= 40.3%
CoV = 38.1%
 = 3.01

Using this value, the formulation has been compared with all the experimental results in Figure C7.3.2(5)-
19. Only three estimated values are unconservative (exp/model=0.58, 0.76 and 0.96). Of course, these
values would increase once characteristic values of material properties are considered, and material partial
factors are applied. This is an indication that the value of 3,0 for the model uncertainty factor is acceptable.

For citations, copying and referencing see Note on Front page! page 277
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

All tests, divided by model uncertainty factor‐ Experiment/Prediction vs xu/d

20

18

16

14

exp /model
12

10

0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
xu/d

Figure C7.3.2(5)‐19 Comparison of experimental results and the estimated value of the rotation capacity
divided by the model uncertainty factor
To illustrate the method two application examples are presented below, one in which failure occurs by
rupture of the steel, and one in which failure occurs by crushing of concrete.

Application Example 1 – Failure is by rupture of steel

Consider test T5B1 reported in (Bosco, 1993). The test consists in a simply supported beam subjected to
point loading with a span of 4.00 m, a rectangular cross section having a width of 200 mm, a total height of
400 mm and reinforced with one layer of 412 mm bars. The effective height is 365 mm. The mean crack
spacing is 88 mm. The reinforcement is Class B500N with the following experimental characteristics:
- yield strength of 595.6 MPa
- strain at yielding: 0.00298
- maximum tensile strength of 641 MPa
- strain at maximum strength: 0.0418
The transversal reinforcement is formed by 6 mm stirrups spaced at 200 mm. Also, the reported concrete
compressive strength is 30.9 MPa and the reported tensile strength, 2.97 MPa.
The estimated maximum compressive strain can be determined from Eq. (18):

62 1
w  2    1.414 103
4 200  200 (20)
1.35
 cu , d ,   0.002   3  1.414 103  0.00995
w
365

For ULS analysis a rectangular stress block may be assumed. Whether the sectional failure is governed by
steel or concrete can be determined by looking at the force in concrete and steel for the plane representing
simultaneous failure for both materials. If for this plane the force in concrete is larger than the force in steel,
then failure will be by steel, otherwise the failure will be by concrete.

 cu , d ,  9.95
xu ,lim  w
d 0.365  0.070 m
 cu , d ,    su
w
9.95  41.8
(21)
Fc,lim  0.8 f cm bxu ,lim  0.8  30900  0.2  0.07  346.08 kN
Fsu  As f su  4  1.13  64.1  289.73  Fc,lim
Failure occurs by steel, so the depth of the neutral axis can be determined from equilibrium of axial forces:

For citations, copying and referencing see Note on Front page! page 278
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

 ud ,ef   ud
f su ,ef  f su
(22)
Fsu 289.73
xu    0.058 m
0.8 f cm b 0.8  30900  0.2
The tension stiffening factor for ULS can then be determined by applying Eq. (14):


f s , ef  f yd  A

 641.0  595.60  103 1.13 104

5.13
 1.047  1
srm ,cal 2
0.088 2
4.90
 0.3 f 3   0.012 0.3   30.9   103
3
2
cm
2 (23)

1   yd  1  2.98 
TS M u  1  1    1  1    0.556
2   ud , ef  2  1.047  41.8 
and, likewise, the mean ultimate curvature can be determined as follows:

1  su 0.0418
   TS M u  0.556  0.0757 m -1 (24)
 r u , m d  xu 0.365  0.058

The depth of the neutral axis for the yielding moment is determined using the classical State II equations:

200000 200000
e  1
  6.71
29811
9500  30.9  3

4  1.13
  6.19  103 (25)
20  36.5
x  2   2 
  e   1  1  3
  6.71 6.19  10  1  1    0.250
d   e    6.71  6.19  10 3 

The tension stiffening factor for the yield moment can then be calculated as:

 f ctm bh 2 
 M cr   
TS M y  1  0.6    1  0.6  6 
M   1   x  
  yd
 2 Ecm d  x x b  d  3  
yd 2

  
 
 2970  0.20  0.402 
  (26)
 1  0.6  6 
 0.25  0.365 
0.365 
1 3 
  29811 10  0.00298   0.25  0.365  0.2
3 2

2 0.365  0.25  0.365 
15.84
 1  0.6  0.895
90.408
Finally, the rotation capacity follows from:
1.3d   1   yd 
 Rd      TS My 
 model   r u , m d  x 
 
1   0.895  0.00298   (27)
 1.3  0.365  0.0757     0.0104 rad
3.00    0.365  0.25  0.365   
   
 0.0066
 
0.0313

Application Example 2 – Failure is by crushing of concrete

Consider test T10A1 reported in (Bosco, 1993). The test consists in a simply supported beam subjected to
point loading with a span of 6.00 m, a rectangular cross section having a width of 300 mm, a total height of

For citations, copying and referencing see Note on Front page! page 279
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

600 mm and reinforced with 912 mm bars. The effective height is 565 mm. The calculated mean crack
spacing is 84 mm. The reinforcement is Class B500H with the following experimental characteristics:
- yield strength of 587.3 MPa
- strain at yielding: 0.00294
- maximum tensile strength of 672.2 MPa
- strain at maximum strength: 0.07
The transversal reinforcement is formed by 6 mm stirrups spaced at 150 mm, the reported concrete
compressive strength is 30.9 MPa and the reported tensile strength, 2.97 MPa.
The estimated maximum compressive strain is determined from Eq. (28):

62 1
w  2    1.257 103
4 150  300 (28)
1.35
 cu , d ,   0.002   3  1.257  103  0.00816
w
565

For ULS analysis a rectangular stress block may be assumed. Whether the sectional failure is governed by
steel or concrete can be determined by looking at the force in concrete and steel for the plane representing
simultaneous failure for both materials. If for this plane the force in concrete is larger than the force in steel,
then failure will be by steel, otherwise the failure will be by concrete.

 cu , d ,  8.16
xu ,lim  w
d 0.565  0.059 m
 cu , d ,    su
w
8.16  70
(29)
Fc,lim  0.8 f cm bxu ,lim  0.8  30900  0.3  0.059  437.54 kN
Fsu  As f su  9 1.13  67.22  683.63  Fc,lim
Failure occurs by concrete in this case. Due to steel hardening, the depth of the neutral axis has to be
determined iteratively by modifying its value until equilibrium of axial forces is fulfilled:

Fc  0.8 f cm bxu
 cu , d , 
 ud ,ef  w
 d  xu 
xu
(30)
f su  f y
f su , ef  f y 
 ud   y
 ud , ef y 

Fs  As f su ,ef  Fc

The result of this calculation provides a value of xu=87.8 mm:

Fc  0.8 f cm bxu  0.8  30900  0.3  0.0878  651.12 kN

 cu , d ,  0.00816
 ud ,ef  w
 d  xu    565  87.8  0.044
xu 87.8
(31)
f su  f y 672.2  587.2
f su ,ef  f y 
 ud   y
 ud , ef   y   587.3 
70  2.94
 44  2.94   639.34 MPa
Fs  As f su , ef  9  1.13  63.93  650.2 kN  Fc

The tension stiffening factor for ULS can then be determined from:


f s , ef  f yd  As

 639.34  587.3  103  1.13  104

5.880
 1.257  1
2 2
srm 0.084 4.677
 0.3 f cm
3
  0.012 0.3  30.9  10
3 3

2 2 (32)

1   yd  1  2.94 
TS M u  1   1    1 1    0.619
2   ud ,ef  2  1.257  70 

and, likewise, the mean ultimate curvature:

(33)

For citations, copying and referencing see Note on Front page! page 280
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

1  cu 0.00816
   TS M u  0.62  0.0576 m -1
 r u , m xu 0.0878
The depth of the neutral axis for the yielding moment is determined using the classical State II equations:

200000 200000
e  1
  6.71
29811
9500  30.9  3

9  1.13
  6.00  103 (34)
30  56.5
x  2   2 
  e   1  1  3
  6.71 6.00  10  1  1    0.246
3 
d  e    6.71 6.00  10 
The tension stiffening factor for the yield moment can then be calculated as:

 f ctm bh 2 
M   
TS M y  1  0.6  cr   1  0.6  6 
M   1   x  
  yd
 2 Ecm d  x x b  d  3  
yd 2

  
 
 2970  0.30  0.602  (35)
 
 1  0.6  6 
 0.246 
1
1 3 
  29811 10  0.00294   0.246  0.565   0.3
3 2

2 1  0.246 
53.76
 1  0.6  0.895
309.21
(36)
Finally, the rotation capacity follows from:

1.3d   1   yd 
 Rd      TS My 
 model   r u , m d  x 
 
1  
 (1)
0.895  0.00294
  1.3  0.565   0.0576     0.0126 rad
3.00     0.246  0.565   
 0.565
 0.735   
 0.0514

0.0377

For citations, copying and referencing see Note on Front page! page 281
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.3.2 (5) Background to FprEN 1992-1-1:2023

References [C07-5] Bosco, C., Debernardi, P.G. (1993). Influence of some basic parameters on the plastic rotation
of reinforced concrete elements. En R. E. (Ed.), Ductility - Reinforcement. Progress Report of
Task Group 2.2 'Ductility Requirements for Structural Concrete - Reinforcement'. Bulletin
d'Information nº 218 (págs. 25-44). Lausanne: CEB.
[C07-6] Eligehausen, R. L. (1987). Rotation capacity of plastic hinges and allowable degree of moment
redistribution. Struttgart: University of Stuttgarrt.
[C07-7] Marti, P. Á. (1998). Tension Chord Model for Structural Concrete. Structural Engineering
International,, 8(4), 287-298. doi:10.2749/10168669878048887
[C07-8] Bigaj, A., Walraven, J. (2002). Size Effects in Plastic Hinges of Reinforced Concrete Members.
Heron 47(1):53 – 75.
[C07-9] Bigaj, A. (1999).Structural Dependence of Rotational Capacity of Plastic Hinges in RC Beams
and Slabs. PhD Thesis. TNO.Delft University Press. 244 p.
[C07-10] Calvi, J.M., Cantù, E., Macchi, G., Magenes, G., (1993). Rotation capacity of reinforced
concrete slabs as a function of steel properties. Ductility - Reinforcement. Progress Report of
Task Group 2.2 'Ductility Requirements for Structural Concrete - Reinforcement'. Bulletin
d'Information nº 218 (págs. 25-44). Lausanne: CEB.
[C07-11] Mattock, A.H (1965). Rotational capacity of hinging regions in reinforced concrete beams.
Proceedings of the International Symposium, Flexural Mechanics of reinforced concrete.
American Society of Civil Engineers. P 143-181.
[C07-12] Corley, W.G. (1966). Rotational capacity of reinforced concrete beams. Journal of the
Structural Division. Proceedings of the American Society of Civil Engineers. Paper 4939, Vol.
92, ST5, p. 121-146.

List of versions of this document

27.10.2020, 1st draft
28.02.2021 2nd draft including first calibration of model parameters
31.12.2022 3rd draft – Adapted to final version of FprEN 1992-1-1:2023
27.03.2023 Peer review
28.03-2023 Final

For citations, copying and referencing see Note on Front page! page 282
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2 Project Team T1 FprEN 1992-1-1:2023

Background document to clause 7.4.3.2

Methods for 2nd order analysis

Simon Wijte, Eindhoven University of Technology

Alejandro Pérez Caldentey FHECOR, Spain

28.03.2023

For citations, copying and referencing see Note on Front page! page 283
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.4.3 Background to FprEN 1992-1-1:2023

Introduction
This document provides background information to paragraph 7.4.3.2 Simplified methods based on nominal
curvature and second order linear analysis and annex O Simplified approaches for second order effects.

7.4.3 Reasons for changes

Methods of analysis In section 5 of EN 1992-1-1:2004 three methods were described to determine the second order effects:
 a general method based on non-linear second order analysis
 a method based on nominal stiffness
 a method based on nominal curvature
where the last is mainly suitable for isolated members while the first two methods are suitable for both whole
structures and isolated members. Note that the two simplified methods applied to an isolated member could
result in significantly different reinforcement designs because they are based on different simplifications.
It should be recognized that the determination of second order effects will be performed more and more by
means of numerical calculations using analysis software. For this, both linear and non-linear elastic methods
are used. FprEN 1992-1-1:2023 should make this possible. On the other hand, there is the aspect of ease-
of-use and of the limitations of design rules that have to be considered. Also because the systematic review
showed that the method of the nominal curvature is highly appreciated, it was decided to delete the method
of the nominal stiffness. Note, however, that it is this nominal stiffness that should be used for the method
that uses linear analysis.
It is for this reason that in addition to the general method, clause 7 includes also a brief indication of how to
determine a stiffness that can be used for a linear elastic second order analysis. This stiffness is referred to
as the effective stiffness. Other methods to determine second order effects and slenderness criteria are
adressed in Annex O.
The general method is considered in clause 7.4.3.3 of FprEN 1992-1-1:2023 and is not significantly changed
with respect to paragraph 5.8.6 of EN 1992-1-1:2004.

For the second order linear analysis an effective stiffness should be chosen. This stiffness can conservatively
7.4.3.2
be derived from the point where yielding in the reinforcement will occur. For an isolated member this point
Simplified methods based can be derived from moment-curvature diagram in a governing section of the member. When more global
on nominal curvature effects should be considered in a case where plastic behaviour can occur in different sections in the structure,
it is established that the effective stiffness should be derived from the displacement that is found prior to the
occurrence of the last plastic hinge. These principles are shown in figure C7.4.3.2-1.

Figure C7.4.3.2‐1: Determination of effective stiffness from moment-curvature and force displacement
relations
It is allowed to use less conservative methods to determine the effective stiffness. However, in that case,
adequate justification of the stiffness used is required.

For citations, copying and referencing see Note on Front page! page 284
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
7.4.3 Background to FprEN 1992-1-1:2023

1. Walraven, J., Westerberg, B. (2008), e.o. Eurocode 2 – Commentary. European Concrete Platform
References
ASBL.
2. Hellesland, J. (2012). Evaluation of effective length formulas and applications in system instability
analysis. Engineering Structures, 45(12). 405-420.
3. Hellesland, J. Challamel, N. , Casandjian, C., Lanos, C. (2013) Reinforced Concrete beams, column
and frames – Section and slender member analysis, ISTE Ltd, London, and John Wiley & Sons,
Hoboken, N.J.
4. Westerberg, B. (2004) Second order effects in slender concrete structures – Background to the rules in
EC2, TRITA-BKN Report 77, Betongbyggnad, KTH, Stockholm

List of versions of this document

10-8-2021, 1st draft
2-9-2021 2nd draft
23-12-2022
27.03.2023 Peer review
28.03-2023 Final

For citations, copying and referencing see Note on Front page! page 285
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

FprEN 1992-1-1:2023

Background document to subsection 8.1.4

Confined concrete and confinement
reinforcement

Jesper H. Sørensen DTU, Technical University of Denmark

Linh C. Hoang
Aurelio Muttoni EPFL, Ecole Polytechnique Fédérale de Lausanne,
Switzerland
23.02.2023

For citations, copying and referencing see Note on Front page! page 286
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

Introduction
This document contains the explanation of the changes in the provisions related to strength increase of concrete
associated to transverse compressive stresses as well as the new provisions for designing confinement
reinforcement.

Strength increase of Reasons for change

concrete associated to
Current provision for confined concrete in EN 1992-1-1:2004 (clause 3.1.9(2)) accounts for the strength
transverse compressive
increase fck due to transverse minimum compressive stress 2 with a bilinear relationship, where:
stresses
8.1.4(2) ‐ the first branch, with a slope of 5, is valid for very low transverse stresses (2≤0.05fck) and
‐ the second branch, with a slope of 2.5 is valid for moderate and higher transverse stresses.
This relationship has been questioned by SRC GR14 as it provides too conservative resistances for moderate
transverse stresses (0.1fck ≤2≤0.6fck). For this reason, SRC GR14 suggested replacing current bilinear
relationship by the nonlinear equation of fib MC2010. This suggestion has been implemented in prEN 1992-1-
1:2017(D1) and following drafts, but in order to enhance ease of use, a simpler expression for usual cases
(c2d≤0.6fcd) has been proposed (fcd=4c2d).
Recent tests conducted at the Technical University of Denmark ([1], [2]) show a dependency of the composition
of the mixture on the strength of actively confined concrete and mortar. The test results show a decrease in
confined strength when the maximum aggregate size of the mixture is decreased. As shown in the following
paragraphs, a re-evaluation of other available tests shows the same results. Accounting for this effect is relevant
for practical applications, as mortar is often used in zones where confining pressures develop, e.g. in
connections between precast elements. For this reason, and addressing also D1C DK4, the relationships have
been corrected for mortar and concrete with Dmax< 16 mm (ddg< 32 mm).

Summary of the test results

The test results described below show that:
‐ For transverse pressures up to approximately 0.6∙fc, the failure envelope can be estimated with a
linear relationship.
‐ For transverse pressures above approximately 0.6∙fc, the gain in strength decreases for all mixture
compositions (both concrete and mortars).
‐ Test results from mixtures containing small maximum aggregate sizes show a smaller strength
increase compared to mixtures with larger aggregates in the test range considered.
‐ The test results obtained from a test setup where there is no pore pressure outlet (i.e. pore pressure
is allowed to build up in the specimen during deformation, such as a Hoek cell [3]) show a greater
strength increase compared to other setups containing a pore pressure outlet.

Conclusions
‐ The failure criterion should account for the composition of the mixtures, namely in terms of maximum
aggregate size.
‐ The failure criterion should be based on tests where the pore pressure was relieved.
‐ The current EN 1992-1-1:2004 failure criterion is not safe for mortar mixtures at higher transverse
pressures (test results (a)).
The proposed failure criterion for FprEN 1992-1-1:2023, which is derived from fib Model Code 2010, but with
some modifications to account for the effect of aggregate size, agrees reasonably well with test results and
captures the effect of maximum aggregate size.

Test results
In the following comparison of test results with failure criteria, test results from the literature have been selected
based on the following criteria:
‐ Test conducted on cylinders (disregarding hollow cylinders and cubes) subjected to active transverse
pressure (e.g. oil pressure).
‐ Cylinders with height to diameter ratio of approximately 2.
‐ Test results where the composition of the mixture is known.
The following aspects have not been directly addressed in the document (the parameters are however
evaluated in the comparison):
‐ Differences in the compressive strength of the mixtures.
‐ Dependency of specimen size (Li and Ansari [4] concluded that a slight size effect on failure strength
for HSC exists).

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 287
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

The test results are divided into three categories:

(a) concretes with Dmax ≥ 16 mm (see Figure C8.1.1),
(b) concretes with 4.75 mm < Dmax < 16 mm (see Figures C8.1.2 and C8.1.3),), and
(c) mortar with Dmax ≤ 4.75 mm (see Figure C8.1.4).
In the following diagrams, the line types refer to:
‐ Black line : Classical Coulomb criterion with k = 4 (ϕ = 37°):

‐ Blue dashed line : Formulae in FprEN 1992-1-1:2023

where the parameter ddg is :

(a) ddg  16  Dmax  40 mm for normal weight concrete fck ≤ 60 MPa

2
 60 
(b) ddg  16  Dmax    40 mm for normal weight concrete with fck > 60 MPa
 f 
 ck 

(c) ddg  16 mm for recycled aggregates

(d) Red dashed line : Formulae in EN 1992-1-1:2004

Figure C8.1.1: Comparison to test results – concrete with Dmax ≥ 16mm,

x - Specimens of size 150x300 mm
* - Specimens of size 100x200 mm
◊ - Specimens of size 76x152 mm

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 288
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

Figure C8.1.2: Comparison to test results - 4.75 mm < Dmax < 16 mm, All tests conducted with Hoek Cell
Setup
o – Specimens of size 55x110 mm (Hoek Cell)

Figure C8.1.3: Comparison to test results - 4.75 mm < Dmax < 16 mm, All tests conducted with Test setup
with pore pressure outlet
x - Specimens of size 150x300 mm
* - Specimens of size 100x200 mm
◊ - Specimens of size 76x152 mm

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 289
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

Figure C8.1.4: Comparison to test results from literature, mortar with Dmax ≤ 4.75 mm
x - Specimens of size 150x300 mm
* - Specimens of size 100x200 mm

Confinement Need for new provisions allowing to design confinement reinforcement

reinforcement
8.1.4(3) EN 1992-1-1:2004 contains only detailing rules on the confinement reinforcement, but no provisions on its
design and the calculation of the transverse stress to be used for calculating the strength enhancement related
to confinement.
Since in several European Countries, the confinement reinforcement designed according to EN 1998 is
governing, it is important that EN 1992-1-1 allows calculating the enhancement of resistance related to its
confinement effect under static actions.

Main assumptions and proposed model

The provisions in FprEN 1992-1-1:2023 are adapted from fib MC2010 and are consistent with the confinement
effectiveness factors in EN 1998-1:2004 (see background in [20]). These provisions are based on the following
assumptions:
(i) A stress in the confinement reinforcement corresponding to its yield strength fyd can be activated at
ULS.
(ii) Only the concrete area enclosed within the confinement reinforcement (“confinement core” defined by
the centreline of the reinforcement and approximated by replacing the parts of circles by corners) is
affected by the beneficial effect of confinement. The concrete cover outside the confinement core is
not affected by confinement, but it may be accounted for with its strength fcd .
(iii) The effect on the concrete core of circular confinement reinforcement (single circular hoops or
continuous spiral reinforcement) can be assumed as uniformly distributed, so that the full confinement
core may be considered (see Figure C8.1.5a which shows the cross section at the level of the
confinement reinforcement). In case of stirrups with square, rectangular or polygonal shapes (and
additional single-leg links), the transfer of forces between the confinement reinforcement and the
confined concrete occurs mainly at the bends (and at the anchorages of single-leg links). Along
straight segments between bends (and anchorages), the concrete within the confinement core
defined above is not fully confined. As a simplification, parabolic limits with 45° angles between fully
confined and non-confined concrete can be assumed as shown in Figures C8.1.5b and c, so that the
area to be deduced from the confinement core is:
1 2
 bi (C8.1.1)
6

where bi are the distances between bends (defined by their related corners) or anchorages of the
confinement reinforcement, see b1 and b2 in Figure C8.1.5c.
(iv) As shown in Figure C8.1.6, between confinement reinforcements, the outer regions of the
confinement core are less affected by the confinement effect than at the level of the confinement
reinforcements. This effect can be accounted for in a simplified manner by assuming again a
parabolic limit between fully confined core and the non-confined concrete with 45° slopes (see limits
between grey and white areas in Figure C8.1.6).

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 290
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

In prEN 1992-1-1:2023, the assumption (i) is considered in Formulae (8.11-14) whereas the effects (ii) and (iii)
are covered with coefficient kconf,b and effect (iv) with coefficient kconf,s in Formula (8.15):

fcd ,c  fcd  kconf ,b  kconf ,s  fcd (8.15)

where
fcd,c is the design value of the concrete compressive strength accounting for confinement which may be
assumed to be constant in the whole compression zone (coefficient kconf,b allows smearing the effect
of confinement)
fcd is the compressive strength increase due to triaxial action described in 8.1.4(2)
kconf,b and kconf,s are the coefficient described above

Figure C8.1.5: Influence of the shape of the confinement on the effectively confined concrete area for (a)
circular confinement reinforcements (b) squares and rectangular with simple confinement
reinforcement and (c) with additional confinement reinforcement fixing the straight segments
of the main confinement reinforcement (case with an intermediate one-leg link)

Figure C8.1.6: Influence of the spacing s of the confinement on the effectively confined concrete area for (a)
fully compressed columns and (b) compression zones

Remarks and consequences

With respect to the previous assumptions and the resulting model, the following comments can be made:
(e) The assumption (i) has been validated in [21]. As shown in Figure C8.1.7c, the comparison with tests
results shows fine agreement also for high yield strengths of the confinement reinforcement (the limit
of applicability of FprEN 1992-1-1:2023 is fyd = 610 MPa).
(f) Despite the fact that in some cases, spalling of the reinforcement cover can be observed before the
maximum resistance is attained (see for instance Figure 3h in [21]), the assumption (ii) is justified.
The comparison with the database described in [21] shows again fine agreement (see Figure C8.1.7b,
where good results are obtained also for small values of ratio Acs/Ag which refer to small columns with
relatively thick reinforcement covers). The simplification (ii) is also made possible with the introduction
of the strength reduction factor cc in 5.1.6 (see related background document and more detailed
explanation in [21]).

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 291
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

(g) The approximation with respect to the confinement core area in (ii) (rounded parts of the perimeter at
bends of the confinement reinforcement considered as corners) is not on the safe side, but it is
compensated by the approximation with respect to the definition of bi in (iii) (bi measured up to the
nominal corners as shown in Figure C8.1.5c) which is on the safe side. Here again, the comparison to
tests described in [21] (see Figure C8.1.7) shows that these simplifications are reasonable (similar
results for circular, square and rectangular columns in terms of experimental-to-calculated resistance
ratio).
(h) As shown in Figure C8.1.8a, the coefficient kconf,b , which accounts for the effects (ii) and (iii), strongly
depends on:
• The shape of the confinement reinforcement (circular reinforcements being significantly
more efficient than simple square or rectangular ones)
• The ratio bcs / b , which is associated to the thickness of the reinforcement cover and the
size of the column (in small columns, the reinforcement cover plays a major role)
(i) As depicted in Figure C8.1.8b, the coefficient kconf,s in Formula (8.8), which accounts for effect (iv),
shows a significant influence of the ratio s / bcs (spacing of confinement reinforcement / width of the
confinement core).

Figure C8.1.7: Comparison between tests and model for confinement reinforcement in FprEN 1992-1-
1:2023, influence of experimental-to-calculated ratio as a function of (a) the ratio of the
confinement reinforcement; (b) the ratio between the confinement core area Acs and the total
column area Ag, and (c) the yield strength of the confinement reinforcement (results of 264
centrically loaded columns, adapted from [21])

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 292
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

a)
1.20
large columns
1.10
1.00 small non‐exposed
0.90 columns
0.80

kconf,b
0.70 small exposed columns
0.60
0.50
0.40
0.30
0.20
0.10
0.00
0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.20 1.30 1.40
bcs / b

b)
1.20
1.10
1.00
0.90
0.80
kconf,s

0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10
s / bcs

Figure C8.1.8: Influences of the main geometrical parameters of circular and square columns on (a)
coefficient kconf,b and (b) coefficient kconf,s

The implication of the described model for practical cases are shown in Figure C8.1.9 for a square column
under axial force and bending. To allow for comparisons, the longitudinal reinforcement is kept constant.
Following cases are considered:
(a) fck = 40 MPa, simple confinement renforcement, parabola-rectangle
stress distribution 2
1  bcs 
kconf ,b   
(b) fck = 80 MPa, simple confinement renforcement, parabola-rectangle 3 b 
stress distribution
(c) fck = 40 MPa, multiple confinement renforcement , parabola-rectangle
stress distribution 2
7  bcs 
kconf ,b   
(d) fck = 80 MPa, multiple confinement renforcement , parabola-rectangle 9 b 
stress distribution
(e) – (h) same parameters as (a) – (d), but with stress-block distribution

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 293
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

(a) 0.600
0.40 x 0.40 m2,
Parabola-rectangle fck = 40 MPa,
0.500 4x3  20 B500

0.400

MRd [MNm]
0.300

FprEN 1992-1-1:2023
0.200 phi 8 @300
phi 8 @150
phi 10 @150
0.100 phi 12 @150
phi 14 @150
cc = 0.85
EN 1992-1-1:2004 a_cc0.85
cc = 1.00
EN 1992-1-1:2004 a1.00
0.000
0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10
NRd [MN]
(b) 0.600
0.40 x 0.40 m2,
Parabola-rectangle fck = 80 MPa,
0.500 4x3  20 B500

0.400
MRd [MNm]

0.300

FprEN 1992-1-1:2023
0.200 phi 8 @300
phi 8 @150
phi 10 @150
0.100 phi 12 @150
phi 14 @150
cc = 0.85
EN 1992-1-1:2004 a_cc0.85
cc = 1.00
EN 1992-1-1:2004 a1.00
0.000
0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10
NRd [MN]
(c) 0.600
0.40 x 0.40 m2,
Parabola-rectangle fck = 40 MPa,
0.500 4x3  20 B500

0.400
MRd [MNm]

0.300

FprEN 1992-1-1:2023
0.200 phi 8 @300
phi 8 @150
phi 10 @150
0.100 phi 12 @150
phi 14 @150
cc = 0.85
EN 1992-1-1:2004 a_cc0.85
cc = 1.00
EN 1992-1-1:2004 a1.00
0.000
0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10
NRd [MN]

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 294
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

(d) 0.600
0.40 x 0.40 m2,
Parabola-rectangle fck = 80 MPa,
0.500 4x3  20 B500

0.400

MRd [MNm]
0.300

FprEN 1992-1-1:2023
0.200 phi 8 @300
phi 8 @150
phi 10 @150
0.100 phi 12 @150
phi 14 @150
cc = 0.85
EN 1992-1-1:2004 a_cc0.85
cc = 1.00
EN 1992-1-1:2004 a1.00
0.000
0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10
NRd [MN]
(e) 0.600
0.40 x 0.40 m2,
Parabola-rectangle fck = 40 MPa,
0.500 4x3  20 B500

0.400
MRd [MNm]

0.300

FprEN 1992-1-1:2023
0.200 phi 8 @300
phi 8 @150
phi 10 @150
0.100 phi 12 @150
phi 14 @150
cc = 0.85
EN 1992-1-1:2004 a_cc0.85
cc = 1.00
EN 1992-1-1:2004 a1.00
0.000
0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10
NRd [MN]
(f) 0.600
0.40 x 0.40 m2,
Stress-block fck = 80 MPa,
0.500 4x3  20 B500

0.400
MRd [MNm]

0.300

FprEN 1992-1-1:2023
0.200 phi 8 @300
phi 8 @150
phi 10 @150
0.100 phi 12 @150
phi 14 @150
EN 1992-1-1:2004 cc = 0.85
a0.85
EN 1992-1-1:2004 cc = 1.00
a1.00
0.000
0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10
NRd [MN]

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 295
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

(g) 0.600
0.40 x 0.40 m2,
Stress-block fck = 40 MPa,
0.500 4x3  20 B500

0.400

MRd [MNm]
0.300

FprEN 1992-1-1:2023
0.200 phi 8 @300
phi 8 @150
phi 10 @150
0.100 phi 12 @150
phi 14 @150
EN 1992-1-1:2004 cc = 0.85
a0.85
EN 1992-1-1:2004 cc = 1.00
a1.00
0.000
0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10
NRd [MN]
(h) 0.600
0.40 x 0.40 m2,
Stress-block fck = 80 MPa,
0.500 4x3  20 B500

0.400
MRd [MNm]

0.300

FprEN 1992-1-1:2023
0.200 phi 8 @300
phi 8 @150
phi 10 @150
0.100 phi 12 @150
phi 14 @150
EN 1992-1-1:2004 cc = 0.85
a0.85
EN 1992-1-1:2004 cc = 1.00
a1.00
0.000
0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10
NRd [MN]

Figure C8.1.9: Influences of the main parameters on the NRd - MRd resistances of a typical square column,
cases (a)-(h) described above, DLower = 24 mm

One can observe that:

‐ The effect of the confinement reinforcement for simple stirrups in square (and rectangular) columns is
rather small. On the opposite, the enhancement due to multiple stirrups (or simple reinforcement in
circular columns) can be significant. This difference is related to coefficient kconf,b which is small for the
first case, but significantly higher for the second (see also Figure C8.1.8a).
‐ Compared to the resistance according to EN 1992-1-1:2004, FprEN 1992-1-1:2023 provides similar
resistances for fck ≤ 40 MPa without accounting for confinement in case cc = 1.00 has been adopted.
‐ For high concrete strengths (fck = 80 MPa), FprEN 1992-1-1:2023 provides similar resistances as EN
1992-1-1:2004 with cc = 0.85 for small eccentricities, but similar resistances as with cc = 1.00 for
larger eccentricities. This is due to the fact that EN 1992-1-1:2004 penalizes high concrete strength
for the case of bending only, whereas with FprEN 1992-1-1:2023, high concrete strengths are
penalized for both axial force and bending (due coefficient cc which applies for all cases).
‐ For high concrete columns with small eccentricities, the reduced resistance compared to EN 1992-1-
1:2004 with cc = 1.00 can be compensated by accounting for the confinement effect in case of
multiple stirrups or circular columns, but not for square columns with simple stirrups.
‐ The differences between the parabola-rectangle approach and the stress-block approach are rather
small according to FprEN 1992-1-1:2023, but can be significant according to EN 1992-1-1:2004. With
this respect, it has to be noted that a comparison with tests on eccentrically loaded columns with both
approaches according to EN 1992-1-1:2004 has shown that the stress-block approach provides better
and more reliable results [21].
‐ The abrupt slope change that can be observed in some EN 1992-1-1:2004 curves for very small
eccentricities refers to the pivot-point limitation to c2. It has been removed in FprEN 1992-1-1:2023
since it is not justified (compensated by long-term effects).

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 296
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.1.4 Background to FprEN 1992-1-1:2023

[1] B. Halldórsson, “Experimental study of strength and fracture of mortar under complex stress states,
References
Master Thesis, Technical University of Denmark,” Lyngby, 2016.
[2] F. Vind and F. Jensen, “Failure criteria for concrete and mortar subjected to triaxial stress states,
Bachelor Thesis, Technical University of Denmark,” Lyngby, 2017.
[3] E. Hoek and J. A. Franklin, “Simple triaxial cell for field or laboratory testing of rock,” Trans. Inst. Min.
Metall., vol. 77, pp. A22-26, 1968.
[4] Q. Li and F. Ansari, “High-strength concrete in triaxial compression by different sizes of specimens,”
ACI Struct. J., vol. 97, no. 6, pp. 684–689, 2000.
[5] D. Campbell-Allen, “Strength of concrete under combined stresses,” Constr. Rev., vol. 35, no. 4, pp.
29–37, 1962.
[6] D. J. Smee, “The Effect of Aggregate Size and Concrete Strength on the Failure of Concrete under
Triaxial Compression,” Civ. Eng. Trans. Inst. Eng. Aust., vol. 9, no. 2, pp. 339–344, 1967.
[7] K. K. B. Dahl, “A Failure Criterion for normal and High Strength Concrete, Series R No 286,”
Department of Structural Engineering, Technical University of Denmark, Lyngby, 1992.
[8] D. W. Hobbs, “Strength and deformation properties of plain concrete subject to combined stress -
Part 1: Strength results obtained on one concrete,” London, Cement and Concrete Association,
Report 42.451, 1970.
[9] K. Lahlou, P. C. Aïtcin, and O. Chaallal, “Behaviour of high-strength concrete under confined
stresses,” Cem. Concr. Compos., vol. 14, no. 3, pp. 185–193, 1992.
[10] J.-C. Chern, H. Yang, and H.-W. Chen, “Behavior of Steel Fiber Reinforced Concrete in Multiaxial
Loading,” ACI Mater. J., vol. 89, no. 1, pp. 32–40, 1992.
[11] J. Xie, A. E. Elwi, and J. G. MacGregor, “Mechanical Properties of Three High-Strength Concretes
Containing Silica Fume,” ACI Mater. J., vol. 92, no. 2, pp. 135–145, 1995.
[12] I. Imran and S. J. Pantazopoulou, “Experimental Study of Plain Concrete under Triaxial Stress,” ACI
Mater. J., vol. 93, no. 6, pp. 589–601, 1996.
[13] F. Ansari and Q. Li, “High-Strength Concrete Subjected to Triaxial Compression,” ACI Mater. J., vol.
95, no. 6, pp. 747–755, 1998.
[14] X. Lu and C. T. Hsu, “Stress-Strain Relations of High-Strength Concrete under Triaxial Compression,”
J. Mater. Civ. Eng., vol. 19, no. 3, pp. 261–268, 2007.
[15] F. Richart, A. Brandtzaeg, and R. L. Brown, “A Study of the Failure of Concrete under Combined
Compressive Stresses, Bulletin No. 26,” Univ. Illinois Bull., pp. 1–104, 1928.
[16] G. Balmer, “Shearing strength of concrete under high triaxial stress -- computation of Mohr’s
envelope as a curve, Structural Research Laboratory Report No. SP-23,” Denvar, Colorado, 1949.
[17] T. N. W. Akroyd, “Concrete under triaxial stress,” Mag. Concr. Res., vol. 13, no. 39, pp. 111–118,
1961.
[18] N. Gardner, “Triaxial Behavior of Concrete,” ACI J., vol. 66, no. 2, pp. 136–158, 1969.
[19] D. Sfer, I. Carol, R. Gettu, and G. Etse, “Study of the behavior of concrete under triaxial
compression,” J. Eng. Mech., vol. 128, no. 2, pp. 156–163, 2002.
[20] Fardis M., Seismic Design, Assessment and Retrofitting of Concrete Buildings: based on EN-
Eurocode 8, Springer, 2009, 742 p.
[21] Moccia F, Yu Q, Fernández Ruiz M, Muttoni A., “Concrete compressive strength: From material
characterization to a structural value”. Structural Concrete. 2020; 21 p..

List of versions of this document

2017.11.08, J.H.S., L.C.H., A.M. explanation of provisions on triaxial concrete strength
2021.05.31, A.M. explanation of provisions on confinement reinforcement implemented
2023.02.13, TJ review
2023.02.23, AM final review

J. H. Sørensen, L. C. Hoang, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 297
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to
clauses 8.2.1 and 8.2.2
Shear in members without shear
reinforcement

Aurelio Muttoni EPFL, Ecole Polytechnique Fédérale de Lausanne,

Miguel Fernández Ruiz Switzerland
Francesco Cavagnis
João Tiago Simões

Report EPFL-IBETON 16-06-R2 17.03.2023

For citations, copying and referencing see Note on Front page! page 298
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to clauses 8.2.1 and 8.2.2 and details the derivation of the closed-form
formulae for shear design of members without shear reinforcement based on the original formulation of the
Critical Shear Crack Theory (CSCT).

8.2.1 and 8.2.2 Current EN 1992-1-1:2004 clauses

Shear In the current EN 1992-1-1:2004, the formula for the shear resistance of members without shear reinforcement
is:

VRd ,c  CRd ,c  k  100  ρl  fck   k1  σ cp   bw  d

13
(C8.1)
 

where:
VRd,c is the shear resistance;
CRd,c is a constant equal to 0.18/γc (recommended value)
k is the size factor which is equal to 1+(200/d)1/2≤2;
ρl is the longitudinal reinforcement ratio (ρl≤2%);
fck is the characteristic compressive strength of concrete measured in cylinder;
k1 is a constant equal to 0.15 (recommended value);
σpd is the axial stress and it is equal to NEd/Ac < 0.2 fcd;
NEd is the axial force (NEd<0 in compression);
bw is the width of the cross section;
d is the effective depth of the member.

Reasons for change

Current clauses for shear design raised a large number of remarks in the systematic review comments to EN
1992-1-1:2004. On that basis, some significant aspects justify updating the current design procedure:
(i) the current design provisions underestimate the influence of the size effect, yielding to potentially
unsafe design solutions in the case of large values of the effective depth d (see Figure C8.1(a)). The
size-effect law has thus to be corrected in the next generation of EN 1992-1-1 (this was one of the
main scopes for the revision of EN 1992-1-1);
(ii) the current approach is too conservative for members subjected to axial tensile stresses (σp>0, see
Figure C8.1(b)) and may even yield negative values of the shear resistance for significant tensile
stresses;
(iii) the aggregate size dg, which governs the aggregate interlock capacity and whose influence on the
shear resistance has been experimentally verified [e.g. C8.1, C8.2] is neglected;
(iv) the effect of the shear slenderness is not considered: this can be potentially unsafe for slender beams
and yield over-designs for squat members;
(v) current Formula for the shear resistance is derived empirically on the basis of existing tests. Typically,
simply supported beams under point loads (not representative of the actual loading and support
conditions commonly found in practice) have been accounted for. This is potentially dangerous when
the original formulae are used outside of the ranges where the various parameters were empirically
calibrated.
(a) (b)
3
No. tests: 669 No. tests: 158
τR,test / τEN1992-1-1:2004

0
0 0.50 1.0 1.5 2.0 -20 -10 0 10 20
d [m] σp [MPa]
Figure C8.1: Ratio test/EN 1992-1-1: 2004 for the case of simply supported beams subjected to point load: (a) for
members without axial force as a function of the effective depth d (database by Reineck et al. [C8.3]); (b) for
members with axial force, as a function of the axial stress σp (database by Cavagnis [C8.6]); for comparison with
database, average material properties (fcm) have been used and the partial safety factor is assumed equal to 1.

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 299
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

Enhancement of ease of use

Ease of use has been reworked by proposing first a simple clause in clause 8.2.1 (Rdc,min) covering most usual
cases (whose strength can be checked based on a rather limited number of parameters, and not requiring the
knowledge of the longitudinal reinforcement ratio). When such check is not sufficient to ensure the shear
resistance without transverse reinforcement, a more detailed verification is provided (subsection 8.2.2. named
“detailed verification for members not requiring shear reinforcement”). This latter detailed verification is
formatted in a similar manner as current EN 1992-1-1:2004 clause and incorporates the influence of all relevant
mechanical and geometrical parameters. In addition, it should be noted that the detailed verification of the shear
resistance may be conducted with two different levels of approximation:
- first estimated without an explicit calculation of the mechanical shear slenderness and, if required,
- a more refined calculation of the shear resistance may be further performed based on the mechanical
shear slenderness,

Mechanical model for slender one-way members without shear reinforcement

The proposed design formulae of FprEN 1992-1-1:2023 are based on the Critical Shear Crack Theory (CSCT)
for shear, sharing the same background as the punching shear design provisions. A detailed description of the
theoretical model can be consulted elsewhere ([C8.1, C8.4, C8.5, C8.8, C8.10]). The original formulation used
the following hyperbolic failure criterion which describes the shear resistance as a function of the reinforcement
strain (without safety factors):
0.3
VR,c   fc  bw  d (C8.2)
1   v  d  kdg

where:
VR,c represents the shear resistance;
εv refers to the strain in the reinforcement;
kdg accounts for the concrete type and can be calculated as kdg = 48/(16+Dlower) = 48/ddg;
fc is the compressive strength of concrete measured in cylinder;
bw is the width of the cross section;
d is the effective depth of the member.
According to Muttoni and Fernández Ruiz [C8.1], a linear relationship between the acting bending moment and
the strain in the reinforcement (εv) can be assumed. The strain in the reinforcement (εv) can thus be calculated
as:
ME ME VE  acs
v    (C8.3)
z  As  Es z  l  bw  d  Es z  l  bw  d  Es

where:
ME is the acting bending moment at the control section;
VE is the acting shear force at the control section;
acs= │ME/VE│ is the effective shear span at the control section;
As is the area of the longitudinal reinforcement;
l is the longitudinal reinforcement ratio;
z is the effective level arm of the longitudinal internal forces;
Es is the modulus of elasticity of the longitudinal reinforcement.
Shear design by using the failure criterion (C8.2) and the load-deformation relationship (C8.3) is very simple as
shown in Figure C8.2, since it can be performed by checking the inequality VEd ≤ VR,c. Calculating the actual
resistance (VEd =VRd,c) requires yet solving the set of Formulae (C8.2) and (C8.3). This can be done iteratively or
by directly solving, in a closed-form manner, the resulting quadratic Formula (Fernández Ruiz et al. [C8.10])
which is not sufficiently simple for a code formula.

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 300
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

0.4

0.3

[√ MPa [
CSCT failure criterion

0.2

bd √ fc
VR,c
load-deformation
relationship
0.1

0
0 1 2 3 4 5
ε v·d·kdg [-]

Figure C8.2: Design using the CSCT failure criterion and the load-deformation relationship.

This problem can be solved on the basis of a recently published refinement of the mechanical of the CSCT
[C8.4, C8.5, C8.9], where an improved general failure criterion can be formulated as follows:
12
 fc  ddg 
VR,c  k     bw  d  VR,c 0 (C8.4)
 ε d
 v 

where:
k depends on the location and shape (inclination) of the critical shear crack and, based on [C8.4,C8.5,
C8.6], can be estimated to depend upon the ratio acs/d resulting into k = 0.015ꞏ(acs/d)1/4;
ddg = 16 mm+Dlower ≤ 40 mm refers to the average roughness dimension of the critical shear crack (see
definition in 8.2.1(4));
VR,c0 is the maximum shear resistance.
This failure criterion, in combination with the load-deformation relationship of Formula (C8.3), can be used to
derive closed-form expressions to calculate the shear resistance, as it will be shown in the following.

Figure C8.3: Comparison of hyperbolic (curve in black, Formula (C8.2) and power-law (curve in red, Formula
(C8.4) with a value of acs/d=2 (green), 4 (blue) and 6 (red)) failure criteria.

8.2.1(4) Minimum shear resistance

Formula (8.20) As the failure criterion (Formula (C8.4)) is strain dependent, a minimum shear resistance can be defined for the
maximum reinforcement strain in the design situation. Assuming that the reinforcement remains in the elastic
range (no plastic redistribution of internal forces), the maximum possible reinforcement strain is thus:
fy
v  (C8.5)
Es

where fy refers to the yield strength of the longitudinal reinforcement.

By introducing the value of the reinforcement strain (Formula (C8.5)) into the power-law failure criterion

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 301
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

(Formula (C8.4)), it results (assuming a value of acs/d = 4 for the slenderness ratio):
12
 Es  fc  ddg 
VR,c  0.021     bw  d (C8.6)
 fy  d 
 

Considering the value of the modulus of elasticity for the reinforcing steel as Es = 200’000 MPa and the fact that
according to Formulae (8.10a) and (8.10b), the design stress is calculated using the level arm z ≈ 0.9d, Formula
(C8.6) can be written in a design format as:
12
11  fck  ddg 
 Rdc,min    (C8.7)
 V  fyd  d 


where the partial safety factor V, it is defined in 4.3.3(1) and its justification can be found in the related
background document.
It has to be noted that the yield strength used in Formula (C8.7) is the strength that has been assumed for the
flexural design of the cross section. This means that if, for an old existing structure, the longitudinal
reinforcement has been designed using a lower yield strength than those used nowadays, the member will have
a higher mechanical longitudinal reinforcement ratio and, consequently, a higher shear resistance.

Physical meaning of the minimum shear resistance, cases with over-designed flexural
reinforcement or axial forces
It is important to note that, according to its definition, the minimum shear resistance ensures a member to reach
yielding of the flexural reinforcement and the shear resistance at the same load level. This presumes that
MEd = MRd at the control section for the shear verification. However, the available flexural reinforcement is
usually larger than the required one for bending conditions (i.e. MRd > MEd, verification for shear of sections
where bending is not governing for design, or cases where the reinforcement has been designed for other
criteria as SLS or fatigue for instance, or where the reinforcement layout has been simplified on the safe side).
For such cases, when the available flexural reinforcement (av) is larger than the required reinforcement for
flexural design (req), the minimum shear resistance could be increased as follows:
12
11  fck  ddg av 
 Rdc,min     (C8.8)
 V  fyd  d req 

or
12
11  fck  ddg 
 Rdc,min    (C8.9)
V   sd  d 

where sd refers to the reinforcement stress in the reinforcement for the design situation governing for the shear
verification.
This potential refinement is not included in clause 8.2.1 of FprEN 1992-1-1:2023. Formula (C8.7) is thus a
simple and safe estimate of the resistance, yet covering the shear demands for most usual members. Cases
with over-designed flexural reinforcement and where the shear resistance given by Formula (C8.7) might not be
sufficient, are addressed in a general manner in clause 8.2.2 (“detailed verification for members not requiring
shear reinforcement”) where the actual reinforcement ratio is accounted for explicitly.
The same comment applies also to members subjected to axial forces. For a given section, applying a
compressive axial force to a member whose reinforcement has been designed without accounting for it, would
reduce the strain in the longitudinal reinforcement, and would therefore increase the shear resistance. This can
be seen in Formula (C8.8), where applying a moderate compressive axial force reduces the amount of required
flexural reinforcement (req) for a given effective shear span, and allows thus to increase the value of Rdc,min if
the available flexural reinforcement (av) remains unchanged. In the case where the available reinforcement has
been designed accounting for the axial force, the available reinforcement is smaller, and the corresponding
shear resistance will consequently remain unchanged. This explains the reason why the effect of axial force
should not appear explicitly in Formula (C8.7).

Reduced shear resistance in case of plastic redistribution of internal forces

In case of plastic redistributions of internal forces, the reinforcement strain is potentially larger than assumed in
Formula (C8.5) and therefore, the shear resistance is theoretically lower than the value Rdc,min given by Formula
(C8.7). This has been confirmed by tests on statically indeterminate beams and by the theoretical
considerations of the CSCT (see Figure C8.4) [C8.12].
In a previous draft (prEN 1992-1-1:2017-04), this effect was considered with a reduced shear resistance in case
of plastic redistribution of internal forces. This provision has yet been criticized by some comments and for this
reason, CEN/TC250/SC2/WG1/TG4 has decided to remove it for simplicity reasons.

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 302
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

Figure C8.4: Shear resistance as a function of the plastic deformation (normalized plastic rotation on the
abscissa; all tests with constant cross section and longitudinal reinforcement ratio; figure from [C8.12]).

8.2.2(1) Location of control sections

In accordance with the last developments of the CSCT for shear failures of members without shear
reinforcement [C8.6], the control section for checking the shear resistance of one-way members is located at a
distance d (refer to figure C8.5) from a static discontinuity (intermediate support, end support, point of
contraflexure or concentrated load) or a geometric discontinuity (change of cross-section geometry or
reinforcement). It should be noted that this is applicable to the failure criterion of Formula (C8.4).
point of contraflexure

d1
d d d
d2
d d d

d2
d1

control sections

Figure C8.5: Locations of the control sections according to the CSCT (figure from [C8.6]).

8.2.1(6) Consideration for control sections in planar members

In order to account for both bending and shear redistributions, it is proposed to consider that the shear force can
be averaged over a distance 4d at the location of maximum shear (calculated by means of an uncracked linear
elastic analysis, preferably with =0 and G=Gel/8 [C8.15]). This approach accounts for the redistributions in the
shear field due to cracking in bending and shear [C8.15]. In the FprEN 1992-1-1:2023, the width where the
shear force may be averaged is defined as 2d on both sides from the peak of the shear force to account for the
case where the peak is located near to a slab edge.

8.2.2(2-3) Derivation of closed-form formulae for detailed verifications

Formulae (8.27) to (8.31) The power-law failure criterion presented in Formula (C8.4) in combination with the load-deformation
relationship allows deriving a closed-form expression for the calculation of the shear resistance. This can be
done by introducing the value of the reinforcement strain (Formula (C8.3)) into the power-law failure criterion
(Formula (C8.4)), and assuming that at failure, VE = VR,c [C8.4, C8.5, C8.9]:
12
 fc  ddg 
VR,c  0.015    ρl  bw  Es  z   bw  d (C8.10)
 VR,c  acs  d 
 

where acs is the shear span:

MEd
acs  d (8.19)
VEd

where the limit acs ≥ d replaces the upper limit VRc,0 in Formula (C8.4). Formula (C8.10) can be easily solved to

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 303
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

calculate the failure load in a closed-form manner [C8.4, C8.5, C8.9]:

13
 ddg z
VR,c  0.0152 3   ρl  Es  fc     bw  d (C8.11)
 acs  d d 


which can be written in an analogous format to EN 1992-1-1:2004 (considering the value of the modulus of
elasticity for the reinforcing steel as Es = 200’000 MPa) as [C8.6]:
13
 
 ddg 
VR,c  0.60   100  ρl  fc  bw  d (C8.12)
 acs 
 d 
 4 

For the code formulation, Formula (C8.12) needs to be adapted by using characteristic values and partial safety
factors. In addition, to enhance the ease of use, the calculation of the shear resistance can be performed in a
first step by adopting a value acs = 4ꞏd (generally safe for members subjected predominantly to distributed
loading when shear is governing considering the exponent associated with the parameter acs). In that case,
Formula (C8.12) can thus be written in a design format as:
13
VRd ,c VRd ,c 0.66  ddg 
 Rd ,c      100  l  fck  (8.16)
bw  z bw  0.9  d V  d 

where the value 0.66 replaces 0.60 in Formula (C8.12) since Rd,c is calculated on the basis of the lever arm
z  0.9  d according 8.2.1(3).

For a more refined calculation, parameter d in Formula (8.16) may be replaced by av  acs  d / 4  d (see
8.2.2(3).

Comparison with datasets

Figure C8.6 presents a systematic comparison of the shear resistances calculated using Formula (C8.12)
against the experimental results of 669 simply supported beams or cantilevers subjected to point load with an
explicit calculation of av. Table C8.1 shows the results that could be obtained using both levels of approximation
(8.2.2(3) and (4)), this is, calculating the shear resistance without and with explicit calculation of av. The
database used is that of Reineck et al. [C8.3], completed with the tests by Cavagnis et al. [C8.8], where only
beams with rectangular cross section and with known aggregate size has been considered. The results
presented in Figure C8.6 show that the calculated shear resistances are in good agreement when compared
with test results, with no visible trends for the parameters investigated. In addition, the results presented in
Table C8.1 show that a reasonable estimate of the shear resistance can be obtained without an explicit
calculation of the mechanical shear slenderness (Approach (1) in Table C8.1 according to 8.2.2(2)). The results
further show that, if the shear resistance is calculated accounting for the influence of the slenderness (Approach
(2) in Table C8.1 according to 8.2.2(3)), better estimates are obtained in terms of both average and coefficient
of variation of the experimental-to-calculated shear resistance ratios.
It has also to be noted that some tests with small slenderness and very high reinforcement ratios have
significantly higher experimental resistances than predicted. If the tests with Vtest / Vcalc > 1.40 are removed from
the database, the statistical values can be improved (mean value = 1.02 instead of 1.03 and COV = 0.140
instead of 0.162).

Table C8.1: Summary of experimental-to-calculated shear resistance ratio of 669 simply supported beams or
cantilevers subjected to point load (database of Reineck et al. [C8.3], completed with the tests by Cavagnis et
al. [C8.8]).
Number of tests Approach Mean *) COV [%] *)

Simply supported beams or cantilevers (1) 1.13 19.0

669
subjected to point load (2) 1.02 14.0

(1) Without calculation of av, 8.2.2(2), Formula (8.16),

(2) With calculation of av, Formula (8.16) with d replaced by av  acs  d / 4 (8.2.2(3))
*) tests with Vtest / Vcalc > 1.4 not considered (but shown and considered in Figure C8.6)

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 304
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

Figure C8.6: Ratio R,test/R,FprEN2023 as a function of the main mechanical and geometrical parameters for the
case of simply supported beams subjected to point load: (a) shear span-to-effective depth ratio a/d; (b)
longitudinal reinforcement ratio ρ; (c) effective depth d; (d) width-to-effective depth ratio b/d; (e) cylinder
compressive strength of concrete fc; (f) aggregate size Dlmax

8.2.2 (4) Influence of axial forces

The influence of axial forces is accounted for by suitably evaluating the strain in the reinforcement [C8.9]. The
strain v is related to the force in the longitudinal reinforcement, which is influenced by the presence of the axial
force accounting for equilibrium conditions. When the axial force is applied in the centroid of the section (located
at h/2 for a rectangular section), the reinforcement strain results [C8.9]:

ME  NE   z  d  h / 2  ME  NE  d / 3
v   (C8.13)
z  Al  Es z  Al  Es

This formula can be expressed by defining an effective shear span which accounts for the presence of the axial
force:

ME N d
acs,N   E  (C8.14)
VE VE 3

It can be noted that the effective shear span accounting for the axial force has a physical meaning as shown in
Figure C8.7(a).
To allow for the simplified use of Formula (8.16) (using d instead of av), the corrected shear span acs,N of
previous Formula can be replaced by coefficient kvp which may be used to reduce the value of d in Formula
(8.16) or av  acs  d / 4 in case of a refined calculation:

NEd d
kvp  1    0.1 (8.20)
VEd 3  acs

In case of tensile axial force, the same coefficient should be used to increase d or av.

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 305
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

NEd d
acs = a' +
VEd 3 d
(a)

h d
NEd /VEd z-d+ ~
2 3
- NEd z

a
VEd VEd a' + NEd z - d + h/2 =VEd cs
z z z
a'
Pd
acs = a' - z
VEd d
(b)
MEd = VEd ∙a' - Pd∙(dp - h/2)
h
2 Pd Pd
dp acs = a' - (dp - h/2) - (z - d + h/2)
VEd VEd
ep Pd Pd
= a' - z ~ a' - (d/3 + ep)
VEd VEd
Pd
VEd
a'

acs d
(c)
Hd

acs unchanged

VEd

Figure C8.7: equilibrium of internal forces: members subjected to (a) centred axial force; (b) prestressing force
on the tension side; (c) prestressing force on the compression side (from [C8.9]).

Consideration of prestressing
Prestressing has three potential influences on the shear resistance:
(i) the vertical component of the prestressing force related to the longitudinal axis may be reduced from
the design value of the shear force VEd according to 8.2.1(8) in case prestressing is considered as an
imposed deformation between prestressing steel and concrete (see approach (a) in 7.6.1(1)). This
effect is accounted for implicitly by considering prestressing as a set of self-equilibrated system of
actions (see approach (b) in 7.6.1(1)). In this case, the shear force VEd is reduced by the effect of
deviation forces and the vertical component of anchorage forces;
(ii) the horizontal component influencing the axial force NEd, and thus the shear resistance according to
8.2.2(4). This effect is accounted for implicitly if prestressing is considered according approach (b) in
7.6.1(1) (as external action) or by adding the term -Pd∙cosp to NEd if prestressing is considered
according approach (b) in 7.6.1(1).
(iii) the eccentricity of the tendon, influencing the reinforcement force and thus the shear resistance. This
effect is accounted for implicitly if prestressing is considered according approach (b) in 7.6.1(1) (as
external action) or by adding the term -Pd∙ep∙cosp to MEd if prestressing is considered according
approach (b) in 7.6.1(1).
Figure C8.7(b) shows that the effect of tendon eccentricity is relevant for prestressed members, as it reduces
the shear span with respect to members subjected to a centred axial force (Figure C8.7(a)). On the contrary,
prestressing on the compression side of the member shows negligible influence on the shear response (Figure
C8.7(c)).

In conclusion, the three effects (i-iii) described above are accounted for implicitly if prestressing is considered as
an external action (approach (b) in 7.6.1(1)) or, if prestressing is considered as an imposed deformation
between prestressing steel and concrete (approach (a) in 7.6.1(1)), by adapting the internal forces according to
8.2.1(8).

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 306
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

Comparison with datasets

Figure C8.8 shows the ratio of experimental-to-predicted shear resistances obtained by with Formula (8.27), for
the case of simply supported beams under point load and with an axial force applied in the centroid of the
section (Figure C8.8). The database by Cavagnis [C8.6] has been used. The comparison shows that the
proposed approach yields good estimates of the shear resistance.

Figure C8.8: Ratio R,test/R,FprEN 1992-1-1:2023 for simply supported beams subjected to point load and to centred
axial force (database by Cavagnis [C8.6]).

8.2.2 (8) Loads acting close to control section

The shear enhancement in the case of distributed loads acting near supports is also covered by the FprEN
1992-1-1:2023. In the presence of distributed loads pushing against the member on the tension side (see Figure
C8.9(b) near to the intermediate support), a fraction of the loads ΔVEd = qd∙Δx may be subtracted from the shear
force acting at the control section VEd. The amount of load which may be subtracted from the shear force at the
control section can differ between a cantilever or an intermediate support due to the different location and shape
of the critical shear crack at failure, but, for the sake of simplicity, an equal fraction of the load which can be
deduced is considered. It is thus considered that in the case of loads pushing against the member on the
tension side, refer to an intermediate support (Figure C8.9(b)),x can be assumed equal to d. Moreover, ΔVEd
shall not be larger than VEd/4, yielding ΔVd = min(VEd/4; q∙d).
Finally, no amount of load can be subtracted from the shear force in the case of distributed loads near simple
supports (as the loads act on the compression side, refer to figure C8.9(a)), thus corresponding to Δx=0,
yielding ΔVEd = 0.
(a) (b)

d d Δx d
q q
d

control section control section control section

end support end support intermediate support

MEd MEd

VEd VEd

lc lc lt

Figure C8.9: Definition of Δx for the case of distributed loads acting (a) near simple supports of end spans
(Δx=0) and (b) near intermediate supports or on cantilevers (Δx=d).

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 307
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

Comparison with datasets

A fine agreement can be observed when comparing Formula (8.16) (accounting for the value of the mechanical
shear slenderness av according to 8.2.2(3)) to the available test results of members subjected to distributed
loads near simple supports of end spans (see Figure C8.10(a)-(c), refer to [C8.6] for database). It can be
observed that the results are consistent for the main mechanical and geometrical parameters: control section,
shear span-to-effective depth ratio acs/d (Figure C8.10(a)), length-to-effective depth ratio lc/d (Figure C8.10(b),
where lc is the distance from the end support to the section of zero shear force, refer to Figure C8.9(a)), and
effective depth d (Figure C8.10(c)). It shall nevertheless be noted that the approach underestimates the shear
resistance of members with short spans (lc/d ≤ 5) in the case of distributed loads near simple supports of end
spans. However, it is important to emphasize that for larger values (lc/d > 5), corresponding to the most common
situations in daily practice, the approach leads to very good results (blue dots in Figure C8.10).

Figure C8.10: Ratio R,test/R,FprEN 1992-1-1:2023 for members with distributed loads near simply supports of end
spans (load on the compression side) as a function of (a) acs/d; (b) the span-to-effective depth ratio lc/d (refer to
Figure C8.9); (c) the effective depth d (database by Cavagnis [C8.6]).
Figure C8.11 illustrates the comparisons of the shear resistances calculated by using Formula (8.16) (again,
accounting for the value of the mechanical shear slenderness av according to 8.2.2(3)) against the test results of
members with distributed loads near intermediate supports or on cantilevers as a function of control section
shear span-to-effective depth ratio acs/d (Figure C8.11(a)) and length-to-effective depth ratio lt/d (Figure
C8.11(b)), where lt is the distance from the intermediate or fixed support to the section of zero shear force (see
Figure C8.9(b)). In this comparison, also the load VEd according to 8.2.2(8) is accounted for. The results show
that this approach agrees consistently with the experimental results.
Table C8.2 presents the summary of the experimental-to-calculated shear resistance ratios of members with
distributed loads (both near simply supports of end spans and intermediate supports or cantilevers), where the
shear resistance is calculated with Formula (8.16) without (approach (1), acs=4d) and with an explicit calculation
of the mechanical shear slenderness (approach (2) accounting for the actual value of acs ≤ 4). The results show
again that a satisfactory estimation of the shear resistance can be first obtained considering acs=4d, while an
explicit calculation of this parameter will normally lead to a better prediction (both in terms of average and
coefficient of variation).

Figure C8.11: Ratio R,test/R,FprEN 1992-1-1:2023 for members with distributed loads near intermediate supports or on
cantilevers (load on the tension side) as a function of (a) acs/d; (b) the span-to-effective depth ratio lt/d defined in
Figure C8.9 (database by Cavagnis [C8.6]).

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 308
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

Table C8.2: Summary of experimental-to-calculated shear resistance ratio of members with distributed loads
near simply supports of end spans and members with distributed loads near intermediate supports or
cantilevers (databases by Cavagnis [C8.6]).
Number of tests Approach Average COV [%]
1.18 14.5
(1)
Members with distributed loads near simply lc / d ≥ 5 : 68 (1.36) (23.3)
supports of end spans (all tests: 117) 1.00 13.5
(2)
(1.16) (24.3)
Members with distributed loads near (1) 1.30 16.3
43
intermediate support of cantilevers (2) 1.12 13.1

(1) Without calculation of av, Formula (8.27) on the basis of d according to 8.2.2(2)
(2) With calculation of av, Formula (8.27) with d replaced by av  acs  d / 4 according to 8.2.2(3)

[C8.1] Muttoni A., Fernández Ruiz M., Shear strength of members without transverse reinforcement as
References
function of critical shear crack width. ACI Structural Journal, Vol. 105, N° 2, 2008, pp. 163-172.
[C8.2] Muttoni A., Shear and Punching Strength of Slabs Without Shear Reinforcement, (in German:
Schubfestigkeit und Durchstanzen von Platten ohne Querkraftbewehrung, Beton- und Stahlbetonbau)
Beton-und Stahlbetonbau, Vol. 98, 2003, pp. 74-84.
[C8.3] Reineck K., Bentz E.C., Fitik B,. Kuchma D.A., Bayrak O., ACI-DAfStb Database of Shear Tests on
Slender Reinforced Concrete Beams without Stirrups. ACI Structural Journal, Vol. 110, N° 5, 2013,
pp. 867–876.
[C8.4] Cavagnis F., Simões J., Fernández Ruiz M., Muttoni A., “Shear Strength of Members without
Transverse Reinforcement Based on Development of Critical Shear Crack.”, ACI Structural Journal
(2020), 103-118
[C8.5] Cavagnis F., Fernández Ruiz M., Muttoni A., “A mechanical model for failures in shear of members
without transverse reinforcement based on development of a critical shear crack.”, Engineering
Structures (2018), 200-315
[C8.6] Cavagnis F., Shear in reinforced concrete without transverse reinforcement: from refined
experimental measurements to mechanical models. PhD Thesis. Thesis No. 8216. Lausanne,
Switzerland: EPFL, 2017.
[C8.7] Cavagnis F., Fernández Ruiz M., Muttoni A., Shear failures in reinforced concrete members without
transverse reinforcement: An analysis of the critical shear crack development on the basis of test
results, Engineering Structures, Vol. 103, 2015, pp.157–173.
[C8.8] Cavagnis F., Fernández Ruiz M., Muttoni A., An analysis of the shear transfer actions in reinforced
concrete members without transverse reinforcement based on refined experimental measurements,
Structural Concrete, 2017, doi: 10.1002/suco.201700145
[C8.9] Muttoni A., Fernández Ruiz M., Cavagnis F., Shear in members without transverse reinforcement:
from detailed test observations to a mechanical model and simple expressions for codes of practice,
fib International Workshop on Beam Shear, Zurich, Switzerland, September 5-6, 2016, fib Bulletin 85.
[C8.10] Fernández Ruiz M., Muttoni A., Sagaseta J., Shear strength of concrete members without transverse
reinforcement: a mechanical approach to consistently account for size and strain effects. Engineering
Structures, Vol. 99, 2015, pp. 360-372.
[C8.11] Leonhardt F., Walther R., Shear Tests on Beams With and Without Shear Reinforcement, Deutscher
Ausschuss für Stahlbeton, No. 151, 1962, 83 pp.
[C8.12] Vaz Rodrigues R., Muttoni A., Fernández Ruiz M., “Influence of Shear on Rotation Capacity of
Reinforced Concrete Members Without Shear Reinforcement”, ACI Structural Journal, V. 1072010.
[C8.13] Muttoni A., The Applicability of the Theory of Plasticity in the Design of Reinforced Concrete, (in
German: Die Anwendbarkeit der Plastizitätstheorie in der Bemessung von Stahlbeton), Report No.
176, Zürich, Switzerland, 1989, 159 pp.
[C8.14] Muttoni A., Thürlimann B., Shear Tests on Beams and Slabs Without Shear Reinforcement, Institut
für Baustatik und Konstruktion, Zürich, Switzerland, 1986, 12 pp.
[C8.15] Natário F., Fernández Ruiz M., Muttoni A., Shear strength of RC slabs under concentrated loads near
clamped linear supports. Engineering Structures, Vol. 76, 2014, pp. 10-23.
[C8.16] Pérez Caldentey A., Padilla P., Muttoni A., Fernández Ruiz M., Effect of load distribution and variable
depth on shear resistance of slender beams without stirrups, ACI Structural Journal, Vol. 109, N°. 5,
2012, pp. 595-603.
[C8.17] Reineck, K.-H.; Dunkelberg, D., ACI-DAfStb databases 2015 on shear tests for evaluating
relationships for the shear design of structural concrete members without and with stirrups. Report for

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 309
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.1 and 8.2.2 Background to FprEN 1992-1-1:2023

Research Project DAfStb V479. (Eds.). DAfStb H. 617, Beuth Verl. Berlin, 2015.
[C8.18] Reineck K., Todisco L., Database of Shear Tests for Non-slender Reinforced Concrete Beams
without Stirrups. ACI Structural Journal, Vol. 111, N° 6, 2014, pp. 1363.1372.
[C8.19] Sigrist V., Bentz E. C., Fernández Ruiz M., Foster S. J., Muttoni A., Background to the Model Code
2010 Shear Provisions - Part I: Beams and Slabs. Structural Concrete, Vol. 14, No. 3, 2013, pp. 204-
214.

List of versions of this document

2016.06.03, A.M., M.F.R., F.C. 1st draft (submitted to TC250/SC2/WG1/TG4)

2016.08.18, A.M., M.F.R., F.C. 2nd draft (submitted to TC250/SC2/WG1/TG4)
2017.11.30, A.M., M.F.R., F.C. changes in prEN 1992-1-1:2017-10(D2) implemented
2018.06.29, A.M., M.F.R., J.S. changes in prEN 1992-1-1:2018(D3) implemented
2021.05.27, A.M., J.S. changes in prEN 1992-1-1:2020(D7) implemented
2023.03.02, T.J. review, references to FprEN 1992-1-1:2023 implemented
2023.03.17, A.M., J.S final review

A. Muttoni, M. Fernández Ruiz, F. Cavagnis, J.T. Simões 17.03.2023

For citations, copying and referencing see Note on Front page! page 310
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG104 FprEN 1992-1-1:2023

Background document to subsection 8.2.2(5)

Shear with normal forces - consideration of
normal compressive forces with k1∙σcp as
alternative to standard formulation

Maximilian Schmidt, M.Sc. Institute of structural concrete (IMB)

Univ.-Prof. Dr.-Ing. Josef Hegger RWTH Aachen University

03.01.2023

For citations, copying and referencing see Note on Front page! page 311
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.2 (5) Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared in order to give a detailed background to the proposed EC2-provisions for
considering the effects of normal compressive forces on the shear capacity of members without shear
reinforcement. The shear capacity is increased by the stress level required for decompression in the control
section.

8.2.2(5) Reasons for change

Shear with normal The current formulation for shear capacity of RC members without shear reinforcement in EN 1992-1-1:2004 [1]
compressive forces is based on the approach in Model Code 1990, which again is based on the semi-empirical multiplicative
formulation derived by Zsutty [2]. EN 1992-1-1:2004 does consider the effects of normal compressive forces
through an additive term k1∙σcp. The reasons for changing the design concept to the proposal in
FprEN 1992-1-1:2023 are explained in detail in the general Background Document to 8.2.2.
The factor k1 in EN 1992-1-1:2004 is based on the approach by Hedman & Losberg [3], who assume that a
prestressed beam can be regarded as a reinforced beam after reaching the decompression moment. For
simplification, a constant value k1 representing a safe lower bound for shear failure has been introduced in
EN 1992-1-1:2004, as explained in [4]. This assumption was verified by a database evaluation. Normal tension
is considered equivalently using the same k1, while only attributing an unfavourable effect to it.
In Eqs. (8.27) – (8.31) of FprEN 1992-1-1:2023, the effective shear span acs is modified depending on the acting
normal force (tension or compression), as explained in the general Background Document to 8.2.2. Due to the
modification of acs to account for normal forces, the shear capacity is increased (compressive forces) or
decreased (tension forces).
Eqs. (8.32) to (8.34) are an alternative to allowing for a simplification for members with normal compressive
forces due to prestress or external axial forces. The positive effect of compression forces on the shear capacity
is considered as an additive term k1, comparable to EN 1992-1-1:2004. Here, the shear capacity is increased by
the stress level required to reach decompression, also comparable to the formulation of EN 1992-1-1:2004 but
based on a linearization of the general failure law from Critical Shear Crack Theory, see the general
Background Document to 8.2.2.
Detailed information on the background of the formulation in 8.2.2(5) with the belonging derivation can be found
in [5] and thus will be not repeated here.

[1] EN 1992-1-1:2004: Eurocode 2: Design of concrete structures - Part 1-1: General rules and rules for
References
buildings. 2004.
[2] Zsutty, T.C.: Beam Shear Strength Prediction by Analysis of Existing Data. In: ACI Journal, Vol. 65
(1968), Iss. 11, pp. 943-951.
[3] Hedman, O.; Losberg, A.: Design of Concrete Structures with Reagrd to Shear Forces. In: Regan,
P.E.; Taylor, H.P.J. (Hrsg.): Shear and Torsion – Explanatory and Viewpoint Papers on Model Code
chapters 11 and 12 prepared by members of CEB Commission V, CEB Bulletin d'Information 126,
1978, S. 184-209.
[4] Walraven, J.C.: Background document for EC-2, Chapter 6.2 Shear. Delft University of Technology,
Delft 2002.
[5] Miguel, P.; Fernández, M.; Hegger, J.; Schmidt, M: Shear resistance of members without shear
reinforcement in in presence of compressive axial forces in the next Eurocode 2. In: Hormigón y
Acero, 2023

List of versions of this document

16.03.2021. 1st draft
09.08.2021 2nd draft
03.01.2023 Final Draft
01.03.2023 Changes based on peer review approved

J.Hegger, M.Schmidt 03.01.2023

For citations, copying and referencing see Note on Front page! page 312
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to subsection 8.2.3

Members requiring design shear
reinforcement

Aurelio Muttoni Ecole Polytechnique Fédérale de Lausanne,

Miguel Fernández Ruiz Switzerland
Marko Pejatovic

Report EPFL-IBETON 17-01-R1 23.02.2023

For citations, copying and referencing see Note on Front page! page 313
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 Background to FprEN 1992-1-1:2023

Introduction This document has been prepared to clarify the background of the changes introduced in the design provisions
of FprEN 1992-1-1:2023 with respect to Section 8.2.3 (“Members requiring design shear reinforcement”). These
provisions inherit the approach of former Section 6.2.3 of EN 1992-1-1:2004, maintaining the variable-angle
truss model approach grounded on limit analysis. It however updates the values for some parameters, notably
of efficiency factors, as well as the provisions for direct support.

Reasons for change Current provision for shear design of members with more than minimum amount of shear reinforcement in EN
1992-1-1:2004 (clause 6.2.3) is based on a variable-angle truss model (see a detailed description of the method
in [C8.5.1-3]). This model is a lower-bound solution based on limit analysis, allowing to select the angle of
inclination of the compression field. Such approach is sound, provides the engineer with a clear understanding
of the load-carrying actions and is consistent with section 8.5 (design based on strut-and-tie and stress field
models). In addition, it allows considering also members with variable depth, geometric discontinuities (e.g.
dapped-end beams and beams with openings), concentrated force introductions (loads and anchor forces),
indirect supports and other cases which are not uncommon in practice. The content of this subsection is also
fully consistent with Annex G (Design of membrane-, shell- and slab elements).
For FprEN 1992-1-1:2023, the core of the EN 1992-1-1:2004 approach is maintained, but some changes are
nevertheless introduced to update them to current state-of-the-art with respect to a number of topics:
i. Minimum allowable angle: as lower angles than those provided in EN 1992-1-1:2004 have been
consistently observed in experimental programmes of prestressed beams [C8.5.4,5]
ii. Efficiency factors for concrete compressive strength: allowing for a detailed consideration of the strain
state, consistently with Section 8.5 and Annex G of prEN 1992-1-1:2020-11 (D7)
iii. Consideration for loads applied near supports: providing a design model based on the same theory as
the variable-angle truss model [C8.5.6], replacing current empirical formulation of Section 6.2.3(8) in
EN 1992-1-1:2004, which provides in addition poor fitting to available experimental data.

Minimum angle of the The minimum angle of the compression field has been reduced for members subjected to significant axial
compression field compression forces. This change is ground on several experimental evidences [C8.5.4,5] and allows for a more
economic design.
8.2.3(4)
Conversely, the angle has been limited for members in tension, following the considerations of MC2010
[C8.5.7,8], which makes the provision also consistent with strain-based analyses.
In addition, the minimum angle of the compression field has been increased (cotmin reduced) for the case of
brittle shear reinforcement (ductility class A). The reasons for this change are described in detail in the
background document to clause 12.2(4) (minimum shear reinforcement).

Design formulae for shear These provisions have been reworked with respect to the formulation of current EN 1992-1-1:2004. Their
resistance verification of theoretical ground remains however unchanged (variable-angle truss model). The proposed formulation allows
concrete stress in the for a direct calculation of the optimum angle of the compression field (simultaneous crushing of compression
compression field field and yielding of stirrups, minimising the amount of shear reinforcement to be arranged, refer to Formula
8.2.3(5) (8.31)).

Efficiency factor  for the The value of the efficiency factor is modified with respect to EN 1992-1-1:2004:
effective concrete
 The influence of concrete brittleness (reduction of  for high concrete compressive strength) is
strength in the
considered in FprEN 1992-1-1:2023 in the definition of fcd and thus has been removed from the
compression field
efficiency factor  of the compression field;
8.2.3(6-7)
 The proposed constant value of the efficiency factor ( = 0.5, 8.2.3(6)) has been adapted to match the
strain-based definition of that factor in 8.2.3(7) (similarly as performed in MC2010 [C8.5.7,8], following
the researches of Vecchio and Collins [C8.5.9]) by considering that the flexural reinforcement is at
yielding (x ≈ 0.002/2 = 0.001, safe estimate), refer to Figure C8.2.3.1.
0.8

0.7 Formula (8.45) (with x = 0.001)

0.6

0.5
Constant value §8.2.3(6)
0.4


0.3

0.2

0.1

0
0 10 20 30 40 50 60
[°]

Figure C8.2.3.1: Comparison of constant and variable value of concrete efficiency factor 

A. Muttoni, M. Fernández Ruiz, M. Pejatovic 23.02.2023

For citations, copying and referencing see Note on Front page! page 314
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 Background to FprEN 1992-1-1:2023

Disturbance by ducts in A new case has been introduced to consider the disturbance of plastic ducts in the resistance of the
the web compression field as considered in MC 2010 [C8.5.7] (for details, refer to Muttoni et al. [C8.5.9]).
8.2.3(10)

Concentrated loads near Remark: The material presented hereafter is a concise version of [C8.5.6] which is referred to for further details.
to supports
8.2.3(12) This clause replaces the empirical approach of EN 1992-1-1:2004 (§6.2.3(8)) with an approach based on a
stress field consistent with the variable-angle truss model and the stress field approach presented in 8.5.
The approach of EN 1992-1-1:2004 (§6.2.3(8)) is based on a reduction of the action for concentrated loads
applied in between 0.5d and 2d from a support to account for a direct strutting (so called “arching action”,
development of a direct strut between load and the support). In addition, only 75% of the shear reinforcement in
the region between support and concentrated load can be accounted for. Such approach has an empirical
basis, and its performance is in addition poor when compared to test results, see Figure C8.2.3.2. This is related
to the fact that there is no transition between the case with direct strutting (av<2d) and the case without direct
strutting (av≥2d). As a consequence, for concentrated loads in between 1.2 and 2d, the model is overly
conservative (due to the fact that only 75% of the reinforcement is accounted for).

Figure C8.2.3.2: Comparison of current EN 1992-1-1:2004 to tests (from [C8.5.6])

In an effort to improve this situation, it has been proposed to calculate the resistance of loads near supports on
the basis of a consistent stress field. Such stress field consists of the consideration of a direct strut combined
with a fan region activating the shear reinforcement, see Fig. C8.2.3.3a.

Figure C8.2.3.3: Considered stress field in prEN 1992-1-1:2020-11 (D7): (a) detail of stress field; and (b)
consideration of concrete and reinforcement contributions (from [C8.5.6])
On that basis, by analysing the free-body shown in Fig. C8.2.3.3b, the total shear resistance can be calculated
as the sum of a contribution related to the direct strut (drawn in blue in Fig. C8.2.3.3b) plus a contribution related
to the stirrups (drawn in red in Fig. C8.2.3.3b):

VR  VR , strut  VR , stirrups (C8.2.3.1)

As it can be noted, the inclination of this free-body corresponds to the angle  defining the clear shear span and
thus is not equal to  (angle of the compression field) as usually performed for design of stirrups in slender
members. By introducing the value of the two components, it results (see Fig. C8.2.3.3b):

cot   cot  Asw

V Rd  bw  z   f cd  z   swd cot  (C8.2.3.2)
1  cot 2  s
where bw refers to the effective width of the web, z to its lever arm, to its efficiency factor (calculated

A. Muttoni, M. Fernández Ruiz, M. Pejatovic 23.02.2023

For citations, copying and referencing see Note on Front page! page 315
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 Background to FprEN 1992-1-1:2023

consistently with the design of slender members), Asw to the area of stirrups at a given horizontal position (sum
of multiple legs if applicable), s to the stirrup spacing and swd to the stress in the stirrups. This Formula can be
expressed in terms of the shear stress resistance (Rd = V Rd /(bwz)) as:

cot   cot 
 Rd    f cd   w   swd cot  (C8.2.3.3)
1  cot 2 
The term swd can be calculated in a general manner accounting for the compatibility of deformations in the clear
shear span region (by making use of the Mohr’s circle for instance, see Figure C8.2.3.4):

 swd  Es   y  Es cot 2   ( x  0.001)  0.001  f ywd (C8.2.3.4)

Figure C8.2.3.4: Calculation of vertical strain in the clear shear span region (from [C8.5.6])
For practical purposes, when cot> 1, this term can be assumed to be equal to the yield stress of the
reinforcement (swd ≈ fywd), which is in agreement to experimental evidence and also elastic-plastic analyses of
this region. Thus, the resulting expression results:

cot   cot 
 Rd    f cd   w  f ywd cot  (C8.2.3.5)
1  cot 2 
which corresponds to Formula (8.55). In addition, by performing a free-body limited by a vertical line (parallel to
the stirrups) at mid –distance of the clear shear span, it will also result the crushing condition of the fans and
direct strut (all having the same inclination ):

cot   cot  cot 

 Rd    f cd   w  f ywd cot     f cd (C8.2.3.6)
1  cot 
2
1  cot 2 
The optimum solution (leading to higher resistance for a given amount of shear reinforcement which is,
according to limit analysis, the exact solution where the stress field correspond to the resistance of a
mechanism) can be obtained by maximizing the previous equation in terms of the parameter . This can be
performed for variable or constant values of . In the latter case, an analytical solution can be obtained,
resulting into:

cot   cot   1  cot 2   cot  min (C8.2.3.7)

where the latter inequality condition will refer to the strength outside of this region as governing. For the case of
variable values of  (8.2.3(7), Formula (8.45), where  can be calculated on the basis of the compression field
inclination  and the longitudinal strain x), an analytical solution is not possible and the optimum compression
field inclination could be obtained by numerical iteration. A first estimate of this value can be obtained in a
simple manner by approximating as follows:

cot   1.3  a / z (C8.2.3.8)

As it can be observed, for values cot<1, design can be performed on the basis of Formula (C8.2.3.2),
combined with Formula (C8.2.3.4) and by optimizing the resistance (for a variable or constant value of ),
depending on the case and performed approach.
The comparison of the FprEN 1992-1-1:2023 to pertinent tests [C8.5.6] shows a neat improvement on the
accuracy of the model, as well as a uniform level of safety for squat and slender members, see Figure C8.2.3.6.
It has to be noted that for the approach of prEN 1992-1-1:2020-11(D7), the limit between with and without direct
strutting depends on the inclination cot with respect to cot. This means that for members with low shear
reinforcement ratio, this limit can be significantly higher than a/d = 2.

A. Muttoni, M. Fernández Ruiz, M. Pejatovic 23.02.2023

For citations, copying and referencing see Note on Front page! page 316
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 Background to FprEN 1992-1-1:2023

Figure C8.2.3.6: Comparison of FprEN 1992-1-1:2023 to test data; (a) case with constant efficiency factor 
(0.50) ad (b) case with variable effciency factor  according to formula (8.32), (red marks
refer to tests with direct strutting (cot < cot); blue marks indicate no direct strutting analysed
according to 8.2.3(3-6)) (from [C8.5.6])

Other considerations, as the increase of tension force in the flexural reinforcement (introduced as an increase of
the design bending moment MEd) or the consideration of inclined shear reinforcement, can be obtained on the
basis of the same stress field.

Inclined shear For members with inclined shear reinforcement (45° ≤ w < 90°), design formulae are derived on the basis of the
reinforcement stress field approach.

8.2.3(13) It has to be noted, that these provisions shall not be used in case of shear reinforcement arranged with an
inclination w > 90°. This is due to the fact that in this case, the principal tensile strain can be much larger than
for the case of vertical shear reinforcement (vertical with respect to the member axis), leading to wide cracks at
SLS and reduced efficiency factor  at ULS. For these reasons, this situation, which could arise in slightly
inclined beams where the stirrups would be arranged vertically or for members with inclined shear reinforcement
with reverse shear force, should be avoided.

References [C8.5.1] Nielsen M. P., Braestrup M. W. and Bach F. “Rational Analysis of Shear in Reinforced Concrete
Beams.” IABSE Colloquium Proceedings P-15, Vol. 2, 1978, 16 p.
[C8.5.2] Thürlimann B. “Plastic Analysis of Reinforced Concrete Beams.” IABSE Colloquium - Plasticity in
Reinforced Concrete, Reports of the working commissions, Vol. 28, 1979, p. 71-90.
[C8.5.3] Nielsen, M. P., and Hoang, L. C. (2011), Limit Analysis and Concrete Plasticity, third edition, CRC
Press, Boca Raton, FL, 2011, 796 pp.
[C8.5.4] Fernández Ruiz, M., Muttoni, A., Shear strength of thin-webbed post-tensioned beams, American
Concrete Institute, Structural Journal, Vol. 105, No. 3, 2008, pp. 308—317
[C8.5.5] Rupf, M., Fernández Ruiz, M., Muttoni, A., Post-tensioned girders with low amounts of shear
reinforcement: shear strength and influence of flanges, Engineering Structures, Elsevier, Vol. 56,
2013, pp. 357—371
[C8.5.6] Pejatovic, M., Fernández Ruiz, M., Muttoni, A., Design of slender and squat reinforced concrete
members with shear reinforcement, Structural Concrete, November 2022.
https://doi.org/10.1002/suco.202200887
[C8.5.7] International Federation for Structural Concrete, “fib Model Code for Concrete Structures 2010,”
Ernst & Sohn, Berlin, Germany, 434 p., 2013.
[C8.5.8] Sigrist, V., Bentz, E., Fernández Ruiz, M., Foster, S.J., Muttoni, A., Background to the Model
Code 2010 Shear Provisions - Part I: Beams and Slabs, Structural Concrete, Vol. 14, No. 3,
2013, pp. 204--214
[C8.5.9] Vecchio, F. J., and Collins, M. P., “The Modified Compression-Field theory for Reinforced
Concrete Elements Subjected to Shear,” ACI Journal Proceedings, V. 83, No. 2, Mar.-Apr. 1986,
pp. 219-231.
[C8.5.10] Muttoni, A., Burdet, O. L., and Hars, E., “Effect of duct type on shear strength of thin webs”, ACI
Structural Journal, V. 103, No. 5, 2006, pp.729-735.

A. Muttoni, M. Fernández Ruiz, M. Pejatovic 23.02.2023

For citations, copying and referencing see Note on Front page! page 317
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 Background to FprEN 1992-1-1:2023

List of versions of this document

01.2.2021, AM 1st draft
26.2.2021, AM 2nd draft
03.02.2023, TJ peer review and update of references to FprEN 1992-1-1:2023
23.02.2023, AM final review

A. Muttoni, M. Fernández Ruiz, M. Pejatovic 23.02.2023

For citations, copying and referencing see Note on Front page! page 318
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1: 2023

Background Document to
Clause 8.2.3(9): Members with circular
cross section

Linh C. Hoang Technical University of Denmark

03.02.2023

For citations, copying and referencing see Note on Front page! page 319
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 (9) Background to FprEN 1992-1-1: 2023

Introduction
This document has been prepared to clarify the background for the provisions introduced in clause 8.2.3(9) of
prEN-1992-1-1:2021 for shear design of members with circular cross section.

Reason for chosen approach

In the literature, there exists several methods to determine the shear capacity of members with circular cross
section, see e.g. [1 - 7]. Task Group 4 of CEN/TC250/SC2/WG1, however, has chosen the model developed in
[7] as the basis for the provisions in FprEN-1992-1-1:2023. This choice was made primarily because the model
is fully consistent with the Variable Angle Truss Model, which forms the basis for shear design in Eurocode 2.
The model is thereby also fully consistent with the general limit analysis method of Annex G in FprEN-1992-1-
1:2023. The consistency makes it easier to adopt the same provisions to design shear reinforcement in
members with large compressive normal forces as well as adopting the same guidelines to determine the
reduction factor  and the limitations on the truss angle - regardless of the shape of the cross section.

Theoretical background
The design rules given in clause 8.2.3(9) of FprEN-1992-1-1:2023 are based on a rigid-plastic lower bound
model presented in [7], see also [8 & 9] for details. To illustrate the model, a circular beam subjected to constant
shear is considered (Figure 1). It is assumed that the shear force V is transferred to the shear span, a, by
means of a sufficiently strong diaphragm (to simplify the boundary conditions). The shear reinforcement consists
of circular stirrups with diameter Dh, closely placed at spacing s. In the longitudinal direction, reinforcement bars
are evenly distributed along the inner perimeter of the stirrups. The first step to establish an admissible stress
field (lower bound solution) to carry V is to assume that the stirrups are at yield and confine the concrete within
Dh, i.e. the shaded area shown in Figure 1(b). The concrete cover, which is not confined, may spall of prior to
the ultimate state. The cover is therefore disregarded.

Figure 1. Cantilever beam with circular cross section and reinforced with circular stirrups.

(a) (b)
Figure 2. Average confinement pressure p = fc on the concrete core within Dh.
As illustrated in Figure 2(a), the average confinement pressure p due to yielding of the stirrups can be
determined by:

2 Ash 2 Ash f y (1)

p f y   fc & 
Dh s Dh s f c
where Ash is the cross section area of one stirrup leg and fy is the reinforcement yield stress. In Equation (1) the
mechanical degree of shear reinforcement, , has also been introduced. Within the confined circular cross
section, a rectangular core is chosen to carry the shear force V, see Figure 2(b). As illustrated in the figure, the
four sides of the core experience a plan hydrostatic stress, fy, introduced through the four circular segments

L. C. Hoang 24.05.2021
For citations, copying and referencing see Note on Front page! page 320
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 (9) Background to FprEN 1992-1-1: 2023

that encapsulate the core. Within the top- and bottom segment, concentrated chord forces, C and T, are
assumed to develop to carry the bending moment. The two remaining segments have no function, other than to
provide confinement to the inner core. This also means that the longitudinal bars in these two segments are not
interacting with the inner core. They can therefore, in this lower bound model, not be used to carry the bending
moment.
The inner core, with dimension h and bw, is the cross section of the shear zone (Figure 3(b)). In this zone, a
homogenous tri-axial concrete stress field is assumed to develop, see Figure 3(a), where the chords carrying T
and C are also shown. It is assumed that the internal lever arm for the chords is approximately the same as the
height of the shear zone, i.e. z  h. In the x,y-plane, the shear zone carries a compressive normal stress, fcc,
inclined at the angle  with the beam axis, and a compressive normal stress, con, acting perpendicular to fcc.
These stresses are superimposed by the above-mentioned confinement pressure,fy, acting in direction of the
z-axis.

Figure 3. Triaxial stress field in circular beam subjected to constant shear (adapted from [7-9]).
Free body diagrams of the shear zone, the tension and compression chords, the load introduction diaphragm
and the four circular segments are shown in Figure 3(c – d). As illustrated in the free body diagram for the shear
zone, the boundaries are acted upon by the constant shear stress, , which from equilibrium conditions is given
by  = V/bwh  V/bwz. Furthermore, the horizontal boundaries also have to transfer the above-mentioned normal
pressure fy while the vertical boundaries parallel to the y,z-plane carry the compressive normal stress |x|. For
any chosen value of , the available equilibrium conditions can be used to establish relations that involve the
quantities , fcc, con, fy, C and T, see [7-9] for details. Furthermore, by applying the classical Modified Coulomb
failure criteria, see e.g. [9], and by requiring fcc ≥ fy ≥ con, the following relationship between fcc and con can be
established:

f cc   f c  4 con (2)

where fcc can be identified as the tri-axial crushing strength of the concrete and fc is the effective uniaxial
compression strength of concrete (as known from the shear provisions for rectangular beams).
The assumed stress field will lead to a lower bound for the shear capacity of the beam, where failure is
governed by crushing of the concrete core and yielding of the stirrups. This lower bound solution is only valid if
the chords are able to carry C and T. For the sake of determining the maximum shear capacity, it will in the

L. C. Hoang 24.05.2021
For citations, copying and referencing see Note on Front page! page 321
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 (9) Background to FprEN 1992-1-1: 2023

following be assumed that the beam is not critical to bending failure, i.e. C and T can be carried.
To determine the maximum shear capacity, optimisation of the stress field with respect to the angle  as well as
the stress con must be performed. The detailed calculations may be found in [7-9]. The optimisation leads to the
following results:

    1
 1   ;  (with  con  0) (a)
      5 (3)
u 

 fc 
1 3  1
  ;  (with  con  (5 f c  f c ) / 8) (b)
 4 4   5

  / 
 ;  1/ 5
 1  /   (4)

tan   
 1 
 ;  1/ 5
 2 

where u = Vu/bwz represents the maximum average shear stress that the shear cross section can carry. As can
be seen, for shear reinforcement content up to /0.20, it is optimal for the beam to carry the shear action
without the need of the in-plane confinement stress, i.e. con = 0. Solution (3a) is therefore identical to the so-
called web crushing criterion for shear in rectangular members, which is obtained by Expressions (8.44) and
(8.45) in prEN-1992-1-1: 2021. The web crushing criterion has been illustrated in Figure 4 as the dotted line,
which coincides with solution (3a). As indicated in the figure, for /0.20 the circular beam carries the shear
force by means of a biaxial compression field (fcc in the x,y-plane and fy in the direction of the z-axis) while the
tri-axial strength is only beneficial for /0.20. Comparison of solution (3) with test results with circular
members can be found in [7]. It should be noted that solution (3) is also valid when the entire circular area
within Dh is utilised to carry . This, however, requires that the bending moment must be carried solely by the
longitudinal reinforcement (both in compression as well as tension) and that all the longtidinal bars are engaged.
In this case, the lower bound solution will also be an exact solution as it coincides with the upper bound solution.
Detailed discussions of this exact solution can be found in [8].

Figure 4. Optimised rigid-plastic lower bound solution (Ac = bwz)

Formulation of code provisions

The fact that the optimal solution (Equation (3a)) requires con = 0 when /< 0.20 has been utilised to formulate
a simple design provision, which is fully consistent with the design rules for rectangular cross section. In most
practical cases, the shear reinforcement content will correspond to a value of / less than 0.20. Furthermore,
even for a shear reinforcement degree as high as / = 0.40, the benefit of working with a tri-axial stress field in
the shear zone is not significant compared to the plane stress model for rectangular cross sections (see Figure
4). Therefore, for most practical purposes, con = 0 can be used. This means that fcc in Figure 3 shall be
replaced with a compression stress c (which cannot be larger than fc). The free body diagram of the shear
zone shown in Figure 3(c) can therefore be modified as shown in Figure 5. For this stress field, the following
relationships can be established from the equilibrium conditions in the x,y-plane:

   f c cot  (5)

L. C. Hoang 24.05.2021
For citations, copying and referencing see Note on Front page! page 322
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 (9) Background to FprEN 1992-1-1: 2023

 c    cot   tan   (6)

 x   cot  (7)

By inserting  as defined in Equation (1), Equation (5) takes the following form:
2 Ash (8)
 f y cot 
Dh s
Furthermore, by adopting the definition of the shear reinforcement area, Asw, according to FprEN-1992-1-
1:2023, i.e. Asw = 2Ash, Equation (8) can be rewritten as:

bw  Asw  (9)
  f y cot  
Dh  bw s 
It can be seen that the term within the bracket is identical to Expression (8.42) in FprEN-1992-1-1:2023, i.e. the
shear capacity of beams with rectangular web in case where the transverse stirrups are rectangular and at yield.
Hence, according to Equation (9), the shear capacity of circular beams reinforced with circular stirrups can be
determined by means of Expression (8.42) in FprEN-1992-1-1:2023, provided that the area Asw is multiplied by
the ratio bw/Dh and provided that the nominal shear stress is taken as  = V/zbw, where bw can be chosen freely,
however not greater than Dh (see Figure 6). Equation (9) is thus the theoretical background for the parts of
clause 8.2.3(9) in FprEN-1992-1-1:2023, which deal with members with circular cross section. As noted in this
clause, the internal lever arm, z, is based on a cross section fitted into the circular cross section, where the
tension bars within bw define the tension chord. Fulfilment of this clause would correspond to the fulfilment of the
assumptions made, when developing the lower bound model as explained in the above.

Figure 5. Bi-axial compression of the shear zone (con = 0).

Figure 6. Effective cross section, where the tension bars within the chosen width bw define the tension chord.
It appears that Equation (6) is identical to Expression (8.44) in FprEN-1992-1-1:2023. Verification of the stress
in the compression field for members with circular cross section can therefore be performed as for rectangular
members.
Finally, according to Equation (7), the compression field in the shear zone results in a compressive axial force,
which due to equilibrium must be outbalanced by tensile axial forces in the chords. Hence, by multiplying both
sides of Equation (7) with the cross section area of the shear zone, zbw, the following expression for the
additional tensile force due to shear is obtained:

N v  V cot  (10)

which is identical to Expression (8.50) in FprEN-1992-1-1:2023. This force can be added to the chords in the
same manner as described in clause 8.2.3(8) of FprEN-1992-1-1:2023.

L. C. Hoang 24.05.2021
For citations, copying and referencing see Note on Front page! page 323
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.3 (9) Background to FprEN 1992-1-1: 2023

References

[1] Ang YB, Priestley MJN and Paulay T. Seismic shear design of circular reinforced concrete columns.
ACI Structural Journal, Vol. 86, No. 1, 54-59, 1989.
[2] Schwartz J. Stress field design of reinforced concrete members with circular cross section. Structural
Engineering International, Vol. 12, No. 1, 36-39, 2002.
[3] Kim JH, Mander JB. Theoretical shear strength of concrete due to transverse steel. Journal of
Structural Engineering, Vol. 131, No. 1, 197-199, 2005.
[4] Kim JH, Mander JB. Closure of theoretical shear strength of concrete due to transverse steel. Journal
of Structural Engineering, Vol. 132, No. 7, 1178-1179, 2006.
[5] Turmo J, Ramos G, Aparicio AC. Shear truss analogy for concrete members of solid and hollow
circular cross section. Engineering Structures, Vol. 31, No. 2, 455-465, 2009.
[6] Orr JJ. Darby AP, Ibell TJ, Denton SR, Shave JD. Shear design of circular cross sections using the
EC2 truss model. The Structural Engineer, Vol. 88, No. 23/24, 26-33, 2010.
[7] Hoang LC, Jensen UG. Rigid plastic solutions for the maximum shear capacity of confined RC
members. Magazine of Concrete Research, Vol. 62, No. 9, 625-636, 2010.
[8] Jensen, UG: Limit analysis of reinforced concrete bridge substructures – Shear strength of piers and
piles with circular cross section and strength of pile caps. PhD thesis, Inst. of Technology and
Innovation, University of Southern Denmark, 242 pp., 2010.
[9] Nielsen, MP, and Hoang, LC: Limit Analysis and Concrete Plasticity. Third edition, CRC Press, Boca
Raton, FL, 796 pp., 2011.

List of versions of this document

24.05.2021 1st draft
03.02.2023 Peer review and update of references to FprEN 1992-1-1:2023

L. C. Hoang 24.05.2021
For citations, copying and referencing see Note on Front page! page 324
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.6 Background to FprEN 1992-1-1:2023

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to subsection 8.2.6

Shear at interfaces

Josef Hegger, Dominik Kueres, RWTH Aachen University, Germany

Maximilian Schmidt
Norbert Randl Carinthia University of Applied Sciences, Austria

01.03.2023

J. Hegger, D. Kueres, M.Schmidt, N. Randl 01.03.2023

For citations, copying and referencing see Note on Front page! page 325
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.6 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to subsection 8.2.6 (Shear at interfaces).

8.2.6(1) Extension of scope

Compared to current EN 1992-1-1, the scope of the subsection does not only account for composite action, but
also for shear at interfaces in a broader sense. In this context, the subsection shall be applied where the static
equilibrium depends on shear transfer across a given concrete plane, such as interfaces between two concretes
cast at different times and interfaces between concrete and similar materials (e.g. concrete cast on rock
surfaces).

8.2.6(2) Omission of verification

The extended scope of the subsection allows for the verification of shear at interfaces in a more general
manner. As a result, further verifications might govern the design. To avoid additional work for designers in
cases where it is not needed, further guidelines are introduced in clause 8.2.6(2). In this context, the verification
may be omitted for very rough interfaces with sufficiently anchored minimum reinforcement crossing the
interface (e.g. construction joints between webs and flanges of bridge deck slabs).

8.2.6(3) Definition of resisting shear stress

The design approach implemented in current EN 1992-1-1 deals with shear at interfaces without interface
reinforcement and with sufficiently anchored interface reinforcement (ensuring yielding of the reinforcement). As
the existing approach shows consistent agreement with test results, it is also implemented in the revised
subsection for the aforementioned cases (Eq. (8.76)). Nevertheless, if yielding of the interface reinforcement
due to insufficient anchorage cannot be provided (e.g. concrete toppings), the current approach is not
applicable. In these cases, a new approach based on fib MC2010 is introduced considering shear forces to be
transmitted via interfaces by a combination of clamping effect and dowel action of the reinforcement (Eq. (8.77),
Figure 8.14). Clause 8.2.6(3) provides an orientation on the differentiation between both design expressions.

8.2.6(4) Definition of shear stress

Current EN 1992-1-1 only deals with composite action in flexural members. To allow for a more general
application of the subsection, Eq. (8.74) is introduced to calculate the decisive shear stress based on the area of
the interface Ai (Figure 8.14). Also, additional provisions for shear keys are given. The decisive shear stress for
composite members (Eq. (8.75)) is defined as a special case.

Figure 8.14: Examples of interfaces

J. Hegger, D. Kueres, M.Schmidt, N. Randl 01.03.2023

For citations, copying and referencing see Note on Front page! page 326
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.6 Background to FprEN 1992-1-1:2023

Figure C8.1: Detailed examples for shear in longitudinal direction

Figure C8.2: Detailed examples for shear in perpendicular direction

J. Hegger, D. Kueres, M.Schmidt, N. Randl 01.03.2023

For citations, copying and referencing see Note on Front page! page 327
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.6 Background to FprEN 1992-1-1:2023

8.2.6(5) Design shear resistance without reinforcement or with sufficiently anchored reinforcement
The design shear resistance is defined complying with current EN 1992-1-1. An introductory clause is added to
clarify the use of the design expression. Compared to current EN 1992-1-1, the adhesion term of Eq. (8.76) is
expressed as a function of √fck/C instead of fctd to simplify the design procedure:

Such approach requires full anchorage of the reinforcement crossing the interface. The comparison between the
revised (FprEN 1992-1-1) and current (EN 1992-1-1) design shear resistance without interface shear
reinforcement is shown in Figure C8.3. As a result of the definition of fctd with respect to fck, the ratio of Rdi
according to FprEN 1992-1-1 and EN 1992-1-1 varies with varying concrete strength fck. While both expressions
yield similar results for fck ≈ 30 MPa, larger differences can be observed for lower and higher concrete strengths.
Nevertheless, as the differences are moderate (between +10% and -8%) for common concrete strength classes
(C20/25 – C60/70), this simplification seems justified.

4,0
very smooth EN 1992‐1‐1
smooth prEN 1992‐1‐1
3,0
τRdi (adhesion) [MPa]

rough
very rough

2,0 keyed

1,0

0,0
0 20 40 60 80 100 120
fck [MPa]

2,0

very smooth smooth rough

τRdi,prEN1992/τRdi,EN1992 [‐]

1,5 very rough keyed

+10%
1,0
‐8%

0,5

0,0
0 20 40 60 80 100 120
fck [MPa]

Figure C8.3: Comparison between revised (FprEN 1992-1-1) and current (EN 1992-1-1) design shear
resistance without shear reinforcement (Eq. 8.76)

J. Hegger, D. Kueres, M.Schmidt, N. Randl 01.03.2023

For citations, copying and referencing see Note on Front page! page 328
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.6 Background to FprEN 1992-1-1:2023

The shear reinforcement may be staggered as shown in Figure 8.15d. The lower limitation for the inclination of
interface reinforcement was extended to 35°≤  ≤ 135° (current EN 1992-1-1: 45°≤  ≤ 90°) for all surfaces
except for very smooth surfaces where 35°≤  ≤ 90°, since the crack opening is not relevant for reinforcement
tension. According to product standards and due to experience, lattice girders may be considered also for
smooth surfaces. The angle limitations are consistent with EN 13747:2010. However, for shear reinforcement,
the current limits (45°≤  ≤ 90°) still apply.

Figure 8.15d: Shear-diagram with required shear reinforcement

The upper limitation of Eq. 8.76 is required according to RANDL [C8.1].
Definition of “very rough” surface conditions was added instead of “indented” surface conditions to avoid
confusions with “keyed” interfaces. Also, “keyed” interfaces are introduced to explicitly account for shear keys.
In this context, the factors cv1 and v given in Table 8.2 shall be applied for the area of each key (not a smeared
approach).
The category “very rough” covers in fact by means of a “smeared approach” also indented surface conditions
with small keys not exceeding the order of magnitude of the definition provided in clause (6) (6mm roughness
from peak to valley at short distances (see Figure 8.15 a)), but not larger keys in the range of or even exceeding
the size of the coarse aggregate. Such a surface typology is then covered by the new category of “keyed”
interfaces.

8.2.6(6) Definition of roughness categories

Taking into account the aforementioned changes, clause 8.2.6(6) provides definitions and explanations of the
roughness classes. Consequently, Figure 8.15a) illustrating the classification of interfaces was extended. Also,
the definition of anchorage length and inclination of interface reinforcement was separated from the geometrical
limitations of keyed interfaces for the sake of clarity (Figure 8.15 (b,c)).

Figure 8.15: Classification of interfaces and definition of interface reinforcement

8.2.6(7) Design shear resistance with insufficiently anchored reinforcement

The provisions for interface shear design implemented in current EN 1992-1-1 consider yielding of the
reinforcement crossing the interface. Nevertheless, full tensile yielding of the interface reinforcement cannot be

J. Hegger, D. Kueres, M.Schmidt, N. Randl 01.03.2023

For citations, copying and referencing see Note on Front page! page 329
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.6 Background to FprEN 1992-1-1:2023

achieved in case of limited and insufficient anchorage length in the old and / or new concrete layer (e.g.
concrete toppings). For these cases, reduced anchorage requirements apply and the resisting shear stress at
the interface can be described by a combination of clamping effect and dowel action of the reinforcement
perpendicular to the interface. In the revised subsection, these effects are considered by the following
expression:

where cv2, v, kv, and kdowel are factors depending on the roughness of the interface, fck is the smallest strength of
the concretes at the interface, n is the stress caused by the minimum external normal force across the interface
that acts simultaneously (permanent stresses caused by confinement of surrounding structural parts may be
taken into account),  is the degree of interface reinforcement, fyd is the design value of the yield strength of the
interface reinforcement, and fcd is the design value of the concrete compressive strength.
Again, the upper limitation of Eq. 8.77 is required according to RANDL [C8.1].
The new approach follows up on the work by RANDL AND WICKE and the background of the design expression is
described in detail in [C8.1], [C8.2], and [C8.3]. Interface shear design following this approach may result first of
all in somewhat higher degrees of interface reinforcement. This is however in many situations compensated by
the following two accompanying clauses as per FprEN 1992-1-1:2023 :
- In the case of horizontal shear transfer in slab members with cast-in-place structural toppings and
rough or very rough interfaces, the coefficient cv2 may be increased by a factor of 1,2 for determining
the design value τRdi of the interface shear resistance.
- The interface reinforcement should be anchored for a stress of at least 0,5fyd with a minimum length of
embedment of 8φ if no other methods of anchorage than by straight bars are applied.
This clause leads therefore to a reduced anchorage length in comparison to Eq. 8.76 and thus allows for shorter
reinforcing bars.

8.2.6(9) Detailing
Compared to current EN 1992-1-1, additional spacing rules are defined in clause 8.2.6(9) if interface
reinforcement is required in composite slabs to ensure sufficient detailing. In this context, the spacing of the
bars crossing the interface shall not exceed 2.5h ≤ 300 mm in the direction of shear transfer and 5h ≤ 750 mm
(≤ 375 mm) perpendicular to the direction of shear transfer (with h being the slab depth). The provisions were
derived by FURCHE AND BAUERMEISTER [C8.4] and were also incorporated in current German Annex of EN 1992-
1-1.
To avoid delamination of concrete toppings along free edges, an additional minimum interface reinforcement is
defined. The minimum reinforcement as,min shall be calculated as:

where tmin is the smaller value of the thickness of both concrete layers, fctm is the mean value of the concrete
tensile strength of the respective concrete layer, and fyk is the characteristic value of the yield strength of the
interface reinforcement. Such reinforcement along edges of concrete layers covers peeling forces (tensile forces
perpendicular to the interface) resulting from differential shrinkage or temperature effects, moreover adhesive
bond tends to be somewhat weaker towards edge regions than inside the member. The approach is based on
an 45°-inclined compressive strut equilibrating the horizontal forces along the edge; with its vertical component
to be anchored by as,min.
There exist however various solutions to cope with the peeling forces along such edges, even structural
solutions are possible (Fig. C8.26).

Figure C8.2.6: Methods to prevent delamination along edges of concrete overlays

References [C8.1] Randl, N., Design recommendations for interface shear transfer in fib Model Code 2010, Structural
Concrete, Vol. 14, No. 3, 2013, pp. 230-241.
[C8.2] Randl, N., Tragverhalten einbetonierter Scherbolzen, Beton- und Stahlbetonbau, Vol. 100, No. 6,

J. Hegger, D. Kueres, M.Schmidt, N. Randl 01.03.2023

For citations, copying and referencing see Note on Front page! page 330
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.2.6 Background to FprEN 1992-1-1:2023

2005, pp. 467-474. (in German)

[C8.3] Randl, N., Wicke, M., Schubübertragung zwischen Alt- und Neubeton – Experimentelle
Untersuchungen, theoretischer Hintergrund und Bemessungsansatz, Beton- und Stahlbetonbau, Vol.
95, No. 8, 2000, pp. 461-473. (in German)
[C8.4] Furche, J., Bauermeister, U., Elementbauweise mit Gitterträgern, Betonkalender 2009/1, Berlin: Ernst
& Sohn, pp. 337-498. (in German)

List of versions of this document

18.03.2021. 1st draft
17.08.2021 2nd draft
03.01.2023 Final Draft
01.03.2023 Changes based on peer review approved

J. Hegger, D. Kueres, M.Schmidt, N. Randl 01.03.2023

For citations, copying and referencing see Note on Front page! page 331
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1: 2023

Background Document to
Subsection 8.3: Torsion and Combined
Actions

Linh C. Hoang Technical University of Denmark

08.02.2023

For citations, copying and referencing see Note on Front page! page 332
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.3 Background to FprEN 1992-1-1: 2023

Introduction
This document has been prepared to clarify the background for the provisions and changes introduced in
subsection 8.3 of FprEN-1992-1-1:2023. The subsection provides methods for ULS design of members
subjected to torsion as well as combinations of torsion, bending, shear and normal forces.

Reasons for change

Subsection 8.3:Torsion and Combined Actions of FprEN-1992-1-1:2023 replaces Subsection 6.3: Torsion in the
1st generation of EC2 (EN 1992-1-1:2004). The provisions and changes are the result of work conducted by
TG4 addressing a large number of comments received in the systematic review of EN-1992-1-1:2004. The
reasons for changes include:
1) The equation to determine the transverse reinforcement for torsion is missing in EN-1992-1-1:2004
2) The rules to design members for (V,T)-interaction are incomplete in EN-1992-1-1:2004
3) Design rules for general (T, V, M, N)-interaction are missing in EN-1992-1-1:2004
4) A clear link between the provisions for torsion design and the general method for ULS design of
membrane elements is missing in EN-1992-1-1:2004.
When addressing the mentioned topics as well as other topics in Section 8, TG4 has worked with the roadmap
illustrated in Figure 1 showing the framework for ULS design of members with adequate deformation capacity.
The general method in Annex G can be related to specific cases treated in Section 8 as shown by the roadmap.
Topics written in red font are new and not treated in EN 1992-1-1:2004. The provisions in Subsection 8.3 have
largely been formulated to be consistent with Annex G. Furthermore, cross references from 8.3 to Annex G are
added to make it much more clear (compared to EN 1992-1-1:2004), that the various provisions in subsection
8.3 are based on the same lower bound approach of plasticity theory.
Explanations to specific changes and provisions are given in the following.

Figure 1: Roadmap to link the various provisions in Section 8 with Annex G. Topics written in red font are new
and not included in EN 1992-1-1:2004.

Subsection 8.3.1
The navigation clauses in EN 1992-1-1:2004 have been reworded and slightly changed. Compared to EN 1992-
1-1:2004, clause 8.3.1(3) allows the distribution of the acting torsional moments over the sub-sections in
proportion to their uncracked torsional stiffnesses or in proportion to their maximum possible torsional capacity.
Both methods give similar result but the distribution in proportion to the maximum strength of each sub-wall is
much easier for design.

Subsection 8.3.4
Subsection 8.3.4 presents the complete set of torsional capacity equations, where the equation to determine
t,Rd,sw is missing in EN 1992-1-1:2004. Note that the equations are formulated as capacity equations (and not
design equations) to make it easier when dealing verification of members subjected to combined actions. The
capacity has been formulated in terms of a shear stress to be consistent with subsection 8.3.2 and Annex G.

L. C. Hoang 18.05.2021
For citations, copying and referencing see Note on Front page! page 333
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.3 Background to FprEN 1992-1-1: 2023

To avoid the risk of overestimation of the torsional capacity due to partial spalling of the concrete cover prior to
the ultimate state (spalling of corners), clause 8.3.4(5) introduces a maximum nominal cover, c, to be used
when determining the thin-walled cross section that carries torsion. The well-known negative effect of corner
spalling on the torsional capacity is most pronounced for square cross sections. Explanation of this
phenomenon can be found in e.g. [1] while an extensive comparison of test results with calculations
demonstrating this effect can be found in [2].

Subsection 8.3.5

The provisions in EN-1992-1-1:2004 for combined actions are incomplete and have too narrow scope. These
provisions can for instance only be used for case (a) shown in Figure 2, but not for case (b). Even in case (a),
the provisions of EN-1992-1-1:2004 do not provide a method to verify the compression field in all of the four thin
walls, but only a check of the two walls carrying V and T. This can be seen in Figure 3, which shows the
procedure to verify the compression field in the left sub-wall for a cross section subjected to shear and torsion.
The calculation of cd follows the provisions of subsection 8.3.2 and Annex G. As stated in the figure, the check
is identical to Eq. (6.29) in EN-1992-1-1:2004. However, a similar verification of the bottom wall should also be
performed, since the -factor at this position may be significantly reduced due to large tension strains generated
by M and T. Such a verification is not included in Eq. (6.29) of EN-1992-1-1:2004.

Figure 2: Square cross sections subjected to combinations of shear, bending and torsion (V,M,T).

Figure 3: Procedure to verify the compression field in a wall of a cross section subjected to combinations of
bending, shear and torsion.

To enable design and capacity verification in the general case with combinations of biaxial bending, biaxial
shear, torsion as well as normal forces (My, Mz, Vy, Vz, T, N), clause 8.3.5(2) allows the designer to use
equilibrium methods to determine stress distributions, which are statically equivalent to the sectional forces,
without the need to fulfill compatibility conditions. The stresses can thereafter be used to design each thin-
walled part by applying Annex G. As an example of this concept, which is grounded on the lower bound
theorem of limit analysis, Figure 4 shows simple distributions of normal- and shear stresses, which are statically
equivalent to the actions (My, Mz, Vy, Vz,, T). Note that the shown stress distributions are just one of many
admissible sets of stress distribution. A linear elastic calculation with fulfilment of compatibility conditions would
lead to the well-known Navier- and Grashof distribution for normal stresses and shear stresses, respectively.

L. C. Hoang 18.05.2021
For citations, copying and referencing see Note on Front page! page 334
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.3 Background to FprEN 1992-1-1: 2023

Figure 4. Example of a set of normal- and shear stress distributions that is statically equivalent with the
sectional forces.

Subsection 8.3.6
The general method given in clause 8.3.5(2) is especially useful for the design of members in civil
engineering structures such as box bridge girders. For members, where torsion is not dominant, it will in
most cases be sufficient to adopt a simpler and conservative approach. For this reason, FprEN-1992-1-
1:2023 also includes a simple linear interaction formula for verification. The following formula is given in
clause 8.3.6(1),

 S Ed 
 S  1
(1)
 Rd i
where SEd and SRd represent the individual design action and, respectively, the corresponding individual
design resistance of the cross section. The linearized interaction formula has been illustrated in Figure
5 for combinations of shear, bending and torsion. The formula is conservative since yield conditions in
limit analysis are in general convex. As noted in clause 8.3.6(2), the interaction between shear action
and bending action do not need to be verified, provided that the bending capacity is based on an area

L. C. Hoang 18.05.2021
For citations, copying and referencing see Note on Front page! page 335
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.3 Background to FprEN 1992-1-1: 2023

of reinforcement, that does not include the portion of the reinforcement reserved to carry the additional
tension, Nvd, generated by the shear action. In this case, the interaction formula illustrated in Figure 5 is
split into the following two criteria:

M Ed TEd (2)
 1
M Rd TRd

VEd TEd (3)

 1
VRd TRd

It appears that Equation (3) is similar to Eq. (6.29) in EN-1992-1-1:2004. The linear interaction formula
in clause 8.3.6(2) therefore contains the interaction rule in EN-1992-1-1:2004 as a special case.

Finally, it is noted that the fully linearized interaction formula can be replaced by other approximate
interaction diagrams developed on the basis of the provisions of clause 8.3.5(2). This is stated in clause
8.3.6(3). This clause basically allows the designer to use more accurate interaction models in
combination with linearization. The idea is illustrated in Figure 6 showing a (M,T,N)-interaction diagram,
where the knowledge of the accurate (M,N)-diagram can be used when adding a linear interaction with
torsion.

Figure 5. Illustration of linear interaction diagram for combinations of bending, shear and torsion.

Figure 6. Example of an alternative interaction diagram for combinations of bending, axial compression
and torsion, allowed according to clause 8.3.6(3).

L. C. Hoang 18.05.2021
For citations, copying and referencing see Note on Front page! page 336
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.3 Background to FprEN 1992-1-1: 2023

References

[1] Nielsen, MP, and Hoang, LC: Limit Analysis and Concrete Plasticity. Third edition, CRC Press, Boca
Raton, FL, 796 pp.
[2] Feddersen, B: Jernbetonbjælkers bæreevne -rene og kombinerede påvirkninger [Load-bearing
capacity of reinforced concrete beams subjected to pure and combined actions], PhD thesis,
Department of Structural Engineering, Technical University of Denmark, Report No. R-251 and R-
252, 1990.

List of versions of this document

18.05.2021 1st draft
08.02.2023 Peer review and update of references to FprEN 1992-1-1:2023

L. C. Hoang 18.05.2021
For citations, copying and referencing see Note on Front page! page 337
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to Clause 8.4

Punching

Aurelio Muttoni EPFL, Ecole Polytechnique Fédérale de

Miguel Fernández Ruiz Lausanne, Switzerland
João Tiago Simões
Diego Hernández Fraile
Josef Hegger RWTH Aachen University, Germany
Carsten Siburg
Dominik Kueres
Report EPFL - 17-01-R5 27.3.2023

For citations, copying and referencing see Note on Front page! page 338
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

Introduction
This document, which is a concise version of reference [C8.4.37], provides the background of the provisions
presented in Clause 8.4 of FprEN 1992-1-1:2023 for punching shear design. The main formulae for the punching
shear resistance are derived from the fib MC2010 [C8.4.1] and are based on the Critical Shear Crack Theory
(CSCT) [C8.4.2,C8.4.3,C8.4.4,C8.4.5,C8.4.6, C8.4.7,C8.4.8,C8.4.9,C8.4.11,C8.4.12].
This document introduces first the reasons for changing the current design method (EN 1992-1-1:2004) and briefly
presents the theoretical background of the proposed approach. Thereafter, the newly introduced design formulae
in Clause 8.4 of FprEN 1992-1-1:2023 are presented, justified and compared with experimental results.

8.4 Reasons for change

Punching Punching design in current EN 1992-1-1:2004 was one of the chapters that collected more systematic review
comments before starting the 2nd generation process. Many reasons supported an in-depth review of this clause,
mostly to address scientific and design concerns (update of state-of-the-art) as well as to enhance the ease-of-
use. Numerous works criticizing the consistency of the method for punching shear design according to EN 1992-
1-1:2004 have been published in the scientific literature [e.g. C8.4.4,C8.4.13,C8.4.14,C8.4.15,C8.4.16,C8.4.17,
C8.4.18,C8.4.19,C8.4.20,C8.4.21]. Some of the most important criticisms [C8.4.19] are concisely summarized
below:
 The verification of punching shear resistance according to EN 1992-1-1:2004 is different for slabs and
foundation slabs. While the control section is defined at 2.0d for slabs, an iteration is required for
foundation slabs to search the control section that minimizes the resistance. If, on the one hand, the
location of the control perimeter at 2.0d for flat slabs is not physically consistent with experimental
observations, on the other hand, the iteration required for the case of foundation slabs is not suitable
for practice [C8.4.19]. In addition, it has been shown that the different methods for foundation slabs and
slabs prescribed by the current approach also lead to significant differences with respect to their safety
level [C8.4.19]. In this sense, a consistent treatment of both foundation slabs and slabs is recommended
for the next generation of EN 1992-1-1. Additionally, a re-definition of the control section consistent with
the experimental observations, equal for both cases and without any iterative procedure is also
recommended [C8.4.19].
 Recent works have shown that the size effect law included in the current approach does not suitably
describe the corresponding phenomena [C8.4.20,C8.4.21]. The approach of EN 1992-1-1:2004 can
underestimate the effect of the size of the member and, consequently, overestimate the punching
resistance for large size members. Recent comparisons of the current approach against datasets of
experimental tests confirm the mentioned trend [C8.4.19]. The size-effect law has to be corrected in the
next generation of EN 1992-1-1 (this was the 5th sub task of PT SC2.T1).
 The current approach does not consider any slenderness effect [C8.4.4]. Recent works assessing the
performance of the current approach against datasets of experimental tests have shown that slab
slenderness plays a role in the punching resistance [C8.4.19]. In addition, this effect can also be
predicted by mechanical models (e.g. [C8.4.3,C8.4.4]). Including this parameter in the design method
should be considered for the next generation of EN 1992-1-1.
 A recently performed comparison between the design method of EN 1992-1-1:2004 and a dataset of
experimental tests suggests that the punching resistance of slabs with shear reinforcement can be
overestimated by the current approach [C8.4.19]. Some works on the topic questioned the validity of
the calculation of the effective stress in the shear reinforcement only as a function of the effective depth
[e.g. C8.4.14], while others discussed on the general validity of the verification of the concrete struts at
the face of the supported area [e.g. C8.4.19, C8.4.17]. This latter verification can underestimate the
punching resistance in some cases (as foundation slabs [C8.4.19] and small columns [C8.4.17]) and
overestimate in others (as large columns [C8.4.17]). For these reasons, a complete revision of the
design of shear reinforcement is recommended for the next generation of EN 1992-1-1. In addition, this
revision should also include provisions for other widely used shear reinforcement systems (e.g., studs)
[C8.4.19,C8.4.18]. Finally, this revision incorporates a new formulation for the verification of the
maximum punching capacity of shear-reinforced slabs, thus assessing the maximum compression that
can be resisted by the strut at the face of the column.

Enhancement of ease of use

The punching resistance of foundations (column bases) and slabs is calculated consistently with the same method
in FprEN 1992-1-1:2023 avoiding a member-based design. The control section is located at the same distance
(0.5dv, where dv refers to the newly introduced definition of shear-resisting effective depth) from the supported
area in both cases without any minimization being required to perform the punching shear verification of
foundation slabs (column bases). In addition, the punching resistance may be calculated in a first approach without
an explicit definition of the slenderness of the member (as for instance performed in the current approach). An
increase of the punching resistance is nevertheless allowed for short-span members as a function of their
slenderness.

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 339
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

Theoretical background
The new provisions for punching shear design of FprEN 1992-1-1:2023 are based on fib Model Code 2010
[C8.4.1], which has a pre-normative character. A detailed description of their theoretical bases can be consulted
in references [C8.4.4,C8.4.5,C8.4.6,C8.4.8,C8.4.9,C8.4.11,C8.4.12]. For further details on the provisions for
punching shear of fib Model Code 2010, please refer to reference [C8.4.7]. In the following, only the main
equations grounding the analytical derivation of the expressions included in the FprEN 1992-1-1:2023 are briefly
presented.

Overview of the punching shear verification according to the CSCT

The punching shear resistance and the deformation capacity can be calculated by intersecting the load-rotation
relationship of the slab-column connection and a failure criterion [C8.4.2,C8.4.3,C8.4.4]. The former relationship
describes the flexural behaviour of the slab-column connection (defining the state of deformation in the shear-
critical region of the slab associated to the applied shear force), while the latter relationship establishes the
maximum shear force that can be carried by the slab-column connection associated to a given state of
deformations, refer to Figure C8.4.1(a) [C8.4.8]. Such methodology is followed for both slabs without and with
shear reinforcement (while the load-rotation relationship can be calculated similarly for slabs without and with
shear reinforcement, a different failure criterion has to be considered for each failure mode, see Figure C8.4.2 for
failure criteria and failure modes) [C8.4.5].

Figure C8.4.1: (a) Calculation of punching shear resistance and rotation at failure, with different possible punching
shear failure regimes (figures from [C8.4.8]); (b) experimental validation of the hyperbolic failure criterion of CSCT
(Eq. (C8.4.2)) proposed by Muttoni in 2008 [C8.4.4] .

Figure C8.4.2: Failure criteria associated to the different failure modes (Eqs. (C8.4.3) to (C8.4.5)) and intersection
with load-rotation relationship (figure adapted from [C8.4.11]; failure modes adapted from [C8.4.5])

Slab-column connection response

The load-rotation relationship describing the flexural behaviour of the slab-column connection can be calculated
following a number of different approaches (analytical formulae [C8.4.22,C8.4.2,C8.4.4]; simplified formulas
[C8.4.4,C8.4.3]; simplified numerical tools [C8.4.23,C8.4.24] or nonlinear finite element methods [C8.4.25,
C8.4.26]). For the development of the design provisions included in the FprEN1992-1-1:2023, the simplified

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 340
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

formula specifically derived from the general law (based on a quadri-linear moment curvature relationship) for
practical purposes (detailed analytical derivation in [C8.4.7]) was retained:

𝜓 𝑘 (C8.4.1)

where ap refers to the distance between the axis of the supported area and the line of zero radial moment, d is
the effective depth, fy and Es are respectively the yield strength and the modulus of elasticity of the flexural
reinforcement, km is a factor depending on the level of the refinement of the approach used to estimate the acting
bending moment in the support width, ms is the acting bending moment in the support strip width (bs) and mR is
the average moment capacity in the support width. According to the fib Model Code 2010 [C8.4.1], the acting
moment in the support strip width mS can be calculated using a linear Finite Element Analysis or, in a first step,
can be estimated as a function of the acting shear force.

Failure criterion for slabs without shear reinforcement

In 2003 Muttoni [C8.4.3, C8.4.4, C8.4.6, C8.4.7] proposed a hyperbolic failure criterion relating the punching
resistance and the crack width (represented by the product ∙d) and roughness (accounting for dg) according to
the following expression (average value without safety factors) [C8.4.4]:
, .
⋅ (C8.4.2)
⋅ , ⋅ ⋅

where VR,c is the punching shear resistance (concrete contribution), ke is the reduction factor (≤1) accounting for
the concentration of shear forces along the control perimeter [C8.4.6], b0,5 is the control perimeter at a distance
of dv/2 from the column face (round corners in case of square or rectangular columns), fc refer to the cylinder’s
concrete compressive strength and dg0 is the reference aggregate size (dg0 = 16 mm for normal weight concrete).
Intensive research carried out in the last 20 years has shown that Eq. (C8.4.2) has a very general scope and it
can be used to calculate the punching shear resistance in a general manner [C8.4.6] (see also Figure C8.4.1(b)).
One interesting advantage for the practical use of the CSCT is that the slab’s load-rotation relationship (which
results mostly from the slab flexural curvatures and thus depends on the load level) can be calculated in different
manners as described above.

Failure criteria for slabs with shear reinforcement

The arrangement of shear reinforcement is one of the most suitable solutions to enhance the strength and
deformation capacity of slabs [e.g., C8.4.5, C8.4.18, C8.4.27, C8.4.28]. The CSCT was extended by Fernández
Ruiz and Muttoni [C8.4.5] to address these cases.
Consistently with the detailing rules of FprEN 1992-1-1:2023, three potential failure modes can govern (see Figure
C8.4.2) [C8.4.5]: (i) crushing of concrete struts in the column vicinity (maximum punching resistance); (ii) failure
within the shear-reinforced area and (iii) failure outside the shear reinforced area. The failure criterion proposed
for each one of these failure modes by Fernández Ruiz and Muttoni [C8.4.5] in the framework of the CSCT are
briefly presented in the following (further details can be consulted in [C8.4.5]). The punching resistance of slabs
with shear reinforcement is governed by the failure mode leading to the lowest resistance.

 Failure by crushing of concrete struts (maximum punching resistance)

Fernández Ruiz and Muttoni [C8.4.5] proposed to calculate the maximum punching resistance (crushing of
concrete struts) considering that it results from an enhancement of the punching resistance of the corresponding
element without shear reinforcement (failure governed by the capacity of cracked concrete to transfer shear
forces). The increase of punching resistance is a function of the shear reinforcement system used, notably of the
anchorage conditions, geometry and detailing rules [e.g. C8.4.5, C8.4.18, C8.4.27, C8.4.28, C8.4.29], as well as
on the geometry of the connection (column size and the effective depth) [C8.4.11, [C8.4.30]. To calculate the
punching resistance associated to this failure mode, Fernández and Muttoni [C8.4.5] proposed a failure criterion
for the strain-based approach of MC2010 consisting on the multiplication by a factor sys,sb (describing the
enhancement associated to the shear reinforcement system) of the failure criterion of slabs without shear
reinforcement:
𝑉 , 𝜂 , ⋅𝑉 , (C8.4.3)
where VR,crush refers to the punching resistance associated to crushing of the concrete struts, VR,c is the failure
criterion of the corresponding slab-column connection without shear reinforcement and sys,sb is considered to be
a function of the type of shear reinforcement (mainly related to its anchorage performance).

 Failure outside the shear-reinforced area

To calculate the punching resistance of failures outside of the shear-reinforced area, Fernández Ruiz and Muttoni
[C8.4.5] considered that the same amount of rotation concentrates in the critical shear crack developing for the
case of slabs without shear reinforcement (crack in the vicinity of the supported area) and in the critical shear
crack developing outside the shear-reinforced area. According to Fernández Ruiz and Muttoni [C8.4.5], the failure

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 341
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

criterion to be used for failures outside the shear-reinforced area is therefore similar to the one used to calculate
the punching resistance of slabs without shear reinforcement:
, .
⋅ (C8.4.4)
, ⋅ , , ⋅ , ⋅

where b0,5,out refers to the outer control perimeter (defined at 0.5dv,out from the last perimeter of shear reinforcement
and accounting for some limitations in the effective perimeter between the shear reinforcement units), dv,out is the
outer shear-resisting effective depth (thus accounting for the development of a horizontal crack along the
compression reinforcement that reduces the shear-resisting effective depth (see [C8.4.5] for details) and ke,out is
the factor accounting for the concentration of shear forces along the outer control perimeter. Conservatively, all
rotations are assumed to develop outside of the shear-reinforced area according to [C8.4.5].

 Failure within the shear-reinforced area

In this case, in the framework of the CSCT, the punching resistance is given by the sum of concrete VR,c,cs and
shear reinforcement VR,s,cs contributions [C8.4.5]:
𝑉 , 𝑉 , , 𝑉 , , (C8.4.5)
where the concrete contribution VR,c,cs is given by the failure criterion of the corresponding element without shear
reinforcement (therefore a function of the rotation) and the shear reinforcement contribution (VR,s,cs) is given by
[C8.4.5]:
𝑉 , , 𝑘 𝜎 𝐴 𝑘 𝑓 𝐴 (C8.4.6)
where σsw refers to the average stress in the shear reinforcement developing inside the punching cone (considered
in a simplified manner to develop with an inclination of 45º and its tip at the column face) and Asw is the total area
of activated shear reinforcement in the punching cone (assumed within 0.35dv and dv according to fib Model Code
2010 [C8.4.1]). The average stress in the shear reinforcement system can be calculated, following the hypothesis
of the CSCT, as a function of the rotation as follows [C8.4.5] (rotation κꞏψ concentrated at the critical shear crack
– inclined at 45˚ - with a centre of rotation at the tip of the crack):

𝜎 𝜓 𝑓 𝑓 (C8.4.7)

where fb is the average bond stress, w and fyw are respectively the diameter and the yielding strength of the shear
reinforcement. For members governed by shear deformations (as foundations or prestressed slabs)
[C8.4.31,C8.4.35], the maximum punching resistance can be found for large shear deformations with the concrete
contribution cancelling (Vc→0) and the stress in the shear reinforcement tending to the yielding strength (σs→fy;
shear strength governed by yielding of shear reinforcement).

Prestressed slabs and slabs with in-plane forces

The influence of the prestressing effects on the punching shear resistance and associated deformation capacity
has been investigated in the framework of the CSCT by Muttoni et al. [C8.4.32], Clément [C8.4.33] and Clément
et al. [C8.4.34,C8.4.35].

Different effects due to the introduction of prestressing influence the punching resistance [e.g., C8.4.34]: (i)
introduction of in-plane forces; (ii) moments due to eccentric prestressing cables; (iii) deviation of forces in the
shear-critical region. As discussed by Clément et al. [C8.4.34,C8.4.35], the fib Model Code 2010 [C8.4.1] and the
CSCT account for all these effects.

As previously shown, the deformation and load carrying capacities can be calculated in the framework of the
CSCT by equalling the shear demand to the shear strength [C8.4.4]. This procedure consists on intersecting the
load-rotation relationship (representing the shear demand) with the failure criterion (corresponding to the shear
capacity). A simplified load-rotation relationship was proposed in the framework of the CSCT to account for the
effects of prestressing (eccentricity of the tendons and in-plane stresses) as follows [C8.4.34] (this formula is an
extension of a previous work of Muttoni [C8.4.4] for slabs without prestressing (Eq. (C8.4.1))):

𝜓 𝑘 (C8.4.8)

where mD is the decompression moment, calculated according to [C8.4.34] (mD positive if it leads to compressive
stresses on the top face of the slab):

𝑚 𝑛 𝑒 (C8.4.9)

where n represents the axial force (σdꞏh, with h referring to the height of the slab) and e is the eccentricity of the
normal forces. The punching shear design of prestressed slabs according to the fib Model Code 2010 [C8.4.1]
and SIA 262:2013 [C8.4.36] (which are based on the CSCT) consist on the application of Eq. (C8.4.8) (with
km=1.5) and an hyperbolic failure criterion similar to Eq. (C8.4.2) (with characteristic values).

A refined calculation of the punching shear resistance of prestressed slabs in the framework of the CSCT was
also proposed by Clément et al. [C8.4.35]. By considering the effects of the prestressing as external forces, the
mentioned authors proposed the calculation of the load-rotation relationship of the slab based on compatibility

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 342
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

and equilibrium conditions of a slab sector and by adopting a multi-linear moment-curvature relationship for the
sectional response (further details on the calculation of the load-rotation relationship may be found in [C8.4.35]).

8.4.3 (1) through (3) Derivation of closed-form formulae for punching shear design of flat slabs and foundation
Formulae (8.92) to (8.98) slabs without shear reinforcement

General considerations
To enhance the ease-of-use while applying the fib Model Code 2010 [C8.4.1] punching shear provisions (based
on the strain-based approach of the CSCT) in the FprEN 1992-1-1:2023, it was decided to keep the consistency
with the format of the current EN 1992-1-1:2004 and to formulate the punching resistance in terms of closed-form
solutions instead of intersecting the failure criteria and load-rotation relationships (strain-based approach). In
addition, it was further decided:
 to use systematically the shear-resisting effective depth dv, instead of distinguishing between the flexural
effective depth (d, theoretically to be used in the load-rotation relationship) and the shear-resisting
effective depth (dv, governing for the location of the control section, length of the control perimeter and
calculation of the acting shear stress E)
 to consider the effect associated with the concentration of shear forces along the control perimeter on the
action side (increasing the acting shear stress by a factor βe>1; similarly to the approach of the current EN
1992-1-1:2004), rather than including it on the resistance side (by reducing the control perimeter by a
factor ke<1) as included in the fib Model Code 2010 [C8.4.1].
The closed-form solution is first presented in a general format (the analytical derivation can be found in references
[C8.4.11,C8.4.37]) and simplified thereafter based on the two simplifications stated above (d=dv and ke=1). A
discussion on the impact of such simplifications on the derived equations and results is presented elsewhere
[C8.4.11,C8.4.37].

Load-rotation relationship adopted for the derivation of provisions of Section 8.4

With respect to the load-rotation relationship, Muttoni [C8.4.4] proposed a simplified power-law equation to
approximate the highly non-linear behaviour (accounting for cracking, tension-stiffening effects and reinforcement
yielding) of two-way slabs (Eq. (C8.4.1)). Recent investigations suggest that the influence of ap/d on the previously
mentioned formulae can be slightly modified to better approximate not only the results of the load-rotation
relationship obtained with more refined tools (e.g., integration of quadri-linear moment-curvature relationship
[C8.4.4]), but also the punching shear resistance calculated with the refined model [C8.4.9]). Accounting for such
developments, a simplified load-rotation relationship as follows is eventually adopted [C8.4.11,C8.4.37]:

𝜓 𝑘 ⋅ 8 (C8.4.10)

where ms/mR ≈ VE/Vflex is considered to simplify its application.

Failure criterion associated to the concrete contribution adopted for the derivation of
provisions of Section 8.4
Closed-form solutions for the punching resistance cannot be derived using a hyperbolic failure criterion. To that
aim, Muttoni et al. [C8.4.8, C8.4.38] showed that the hyperbolic failure criterion of Eq. (C8.4.2) can be replaced
by a power-law equation with very similar results (see comparison in Figure C8.4.3):

𝑉 , 0.55 𝑘 𝑏 , 𝑑 𝑓 0.50 𝑘 𝑏 , 𝑑 𝑓 (C8.4.11)

⋅

where ddg represents the reference value of roughness of the critical shear crack and is computed as follows
[C8.4.8, C8.4.39]:

𝑑 𝑑 𝑑 ⋅ 𝑚𝑖𝑛 ,1 40mm (C8.4.12)

where dg0 = 16 mm. It can be noted that a value of 0.50 is used in Eq. (C8.4.11), instead of the value of reference
[C8.4.8] (where a discussion on that value suggests that it can vary within the range from 0.45 to 0.65), for
simplicity and yielding almost identical results (limit governing for small values of the effective depth, large values
of the flexural reinforcement ratios and/or small supporting areas).

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 343
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

Figure C8.4.3: Comparison of original hyperbolic failure criterion (Eq. (C8.4.2)) with failure criterion based on a
power law (Eq. (C8.4.11)) (figure adapted from [C8.4.8]).

Derivation of closed-form solution for members without shear reinforcement

Replacing Eq. (C8.4.10) in Eq. (C8.4.11) and solving the equation in order to obtain VR,c by considering VE = VR,c
in Eq. (C8.4.10) to calculate the intersection point between the two curves, and with all the aforementioned
assumptions (neglecting the effect of the concentrations of shear forces on the reduction of the control perimeter;
flexural effective depth equal to shear-resisting effective depth; km ≈ 1.2 [C8.4.7]), the following closed-form
expression can be obtained to calculate the punching resistance (analogous derivation to that presented by
Muttoni et al. [C8.4.38, C8.4.32, C8.4.11, C8.4.12]; ):

,
4 𝑎 100𝜌 𝑓 0.50 𝑓 (C8.4.13)
, ⋅ , ⋅

Where:
 a = Vflex/mR;
. .
 mR is the flexural resistance which is approximated as 𝑚 0.75𝑑 𝜌 . 𝑓 𝑓 [C8.4.38, C8.4.8];

 Es=200’000 MPa was adopted;

 the coefficient 4 accounts for the constant values and for the influence of the parameters whose
exponents are rounded.
Eq. (C8.4.13) can be written in a design format by:
 considering d=dv as previously referred in the General Considerations
 considering that the influence of the shear force concentrations along the control perimeter is
accounted for by increasing the acting shear stress (by a factor βe≥1) rather than by reducing the
control perimeter by a factor (ke≤1) as mentioned above in the General Considerations
 using characteristic values and partial safety factors to account for all uncertainties
 adopting a large (and, thus, safe) value for ap equal to 8ꞏd (accounting for the exponent associated
with the parameter ap, its calculation can even be avoided in a first step)
leading to the following expression:

. .
𝜏 , 𝑘 100𝜌 𝑓 𝑓 (C8.4.14)

where ddg is calculated according to Eq. (C8.4.12), l is the longitudinal flexural reinforcement ratio (with 𝜌
𝜌 , ⋅ 𝜌 , , where subscripts x,y refer to two orthogonal directions) and Rd,c refers to the design punching shear
stress (to be compared with the acting shear stress Ed=e·VEd/(b0,5·dv) considering the effect associated to the
concentration of shear forces on the action side). With respect to coefficient kpb in Eq. (C8.4.14), it accounts for
the strength enhancement due to the shear field gradient in the control section, and in the general format (i.e.,
distinguishing between d and dv, and considering ke) it is written as:

𝑘 𝑎 (C8.4.15)
. ,

which is simplified to (previously adopted assumption of d=dv):

𝑘 5 𝑎 5 𝜇 (C8.4.16)
, ,

where p accounts for the shear force-to-bending moment ratio in the shear-critical zone as well as for shear force
concentrations (considering that a=VEd/msd , where the value of msd is the acting moment computed in the width
of the support strip [e.g. C8.4.1, C8.4.4, C8.4.7], and accounting for the fact the factor ke considers the
concentration of shear forces along the control perimeter). The definition of factor p thus yields:

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 344
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

𝜇 𝑎 (C8.4.17)

For assessment of existing structures, the calculation of μp can be done in a tailored manner by using Eq.
(C8.4.17). Nevertheless, for design purposes, this formulation can be further simplified adopting the following
values of μp for the different cases (see [C8.4.1] for details):
p=8 for internal columns;
p=4 for edge columns and ends of walls;
p=2 for corner columns and corners of walls;

This factor p can also be expressed as a function of the two main control perimeters defined in FprEN 1992-1-
1:2023, namely b0 and b0,5, where b0 corresponds to the perimeter of the support region (perimeter at the column
edge minimised for re-entrant corners and columns near slab edges, see Figure 8.18 of FprEN 1992-1-1:2023)
and b0,5 is the length of the control perimeter for the shear verification located at a distance 0,5dv from b0.
𝜇 ⋅ ⋅𝑑 𝑏 , 𝑏 (C8.4.18)

Introducing Eq. (C8.4.18) in (C8.4.16), a simplified formula for kpb based on a geometrical rule can be derived as
follows (8.4.3(1) of FprEN1992-1-1:2023):

1 𝑘 3.6 1 2.5 (C8.4.19)

,

It can be noted that the shear-gradient enhancement factor kpb represents a smooth transition between one- and
two-way shear. The upper bound of kpb is required to limit the increase of shear resistance for very small supported
areas.
According to 8.4.3(2) of FprEN1992-1-1:2023, for distances between the axis of the supporting area and the
location of the zero radial bending moment smaller that ap≤8d, the punching resistance calculated according to

Eq. (C8.4.14) may be further increased by replacing the parameter dv by 𝑑 ⋅ , where:

𝑎 𝑎 , ⋅𝑎 , 𝑑 (C8.4.20)

with ap referring to the location where the radial bending moment is equal to zero with respect to the support axis.
Finally, with respect to the parameter βe accounting for the concentration of shear forces due to moment transfer
between the slab and the column, it is now defined in FprEN 1992-1-1:2023 for wall supports and slab-column
connections. In the case of wall corners and ends of walls, parameter βe has been derived from the provision of
fib Model Code 2010 [C8.4.1] of eccentricity parameter ke, and is proposed as a constant value which is different
for ends of walls and wall corners. For columns, it can be estimated using an approximate approach, in which the
values for internal, edge and corner columns consist of a constant value, equal to the values in the current version
of EN 1992-1-1:2004, which are as well consistent with the provisions of fib Model Code 2010 [C8.4.1].
Additionally, in the case of slab-column connections, parameter βe can also be calculated using a more refined
procedure based on the work by Vollum and co-workers [C8.4.40]. In this work, an assessment of typical building
arrangements (in terms of span lengths, slab thickness and column size and type) was performed systematically
by means of Finite Element analyses. For each case, parameter βe was computed as the ratio between the shear
stress calculated using the Finite Element analysis (making use of shear fields) and the average shear stress
around the reduced control perimeter as per EN 1992-1-1:2004. The values obtained allowed the authors to
propose an expression for the calculation of parameter βe for corner columns, that was made extensible for inner
and edge connections and incorporated to FprEN 1992-1-1:2023, as shown in Eq. (C8.4.21). The proposal is as
well compared with current provisions of EN 1992-1-1:2004, showing that it is slightly more conservative at the
same time as it covers a wider range of scenarios than the current constant values.
𝛽 1 1.1 ⋅ 1.05 (C8.4.21)

where eb is the eccentricity of the resultant of shear forces with respect to the centroid of the control perimeter,
calculated for internal, edge and corner columns with Eqs. (C8.4.22a) through (C8.4.24c) respectively; and bb
represents the dimensions of the control perimeter and is calculated as the geometric mean of the minimum and
maximum overall widths of the control perimeter. With respect to the eccentricities, according to [C8.4.40], they
can be defined as follows:
𝑒 𝑒 𝑒 for internal columns (C8.4.22a)

𝑒 0.5 ⋅ |𝑒 | 𝑒 for edge columns (C8.4.22b)

𝑒 0.27 ⋅ |𝑒 | 𝑒 for corner columns (C8.4.22c)

It is important to highlight that, in addition to its simplicity, the design formulation expressed in Eqs. (C8.4.14) to
(C8.4.21) has a mechanical background which allows dealing with more complex cases (e.g., assessment of
existing structures). Finally, it should be noted that coefficient 0.60 in Eq. (C8.4.14) is based on a reliability analysis
that also covered the partial safety factor γv (see Background Document to 4.3.3(1) and Annex A [C8.4.41]).

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 345
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

Comparison with datasets

In this clause, the design provisions according to Eqs. (C8.4.14), (C8.4.19), (C8.4.21) and (C8.4.22a-c) are
compared to the following test datasets:
 121 tests of slender slabs without shear reinforcement (database of reference [C8.4.8] – updated with
[C8.4.4] and [C8.4.24]);
 34 tests of foundation slabs without shear reinforcement (database of reference [C8.4.8]);
 47 tests of slender slabs supported on inner columns with unbalanced moment (database of reference
[C8.4.42]).
Specimens not respecting the detailing rules or reaching the flexural capacity were excluded from the databases.
The comparisons have been performed with average values of concrete properties (fck=fcm) and considering all
partial safety factors equal to one. Table C8.4.1 presents a summary of the results.

Table C8.4.1: Summary of experimental-to-calculated punching shear resistance for the different investigated
cases of specimens without shear reinforcement.
Number of tests Approach Average COV [%]

(1) 1.11 11.4

Slabs with concentric loading 121
(2) 1.09 10.6
(1) 1.48 12.7
Foundation slabs with concentric loading 34
(2) 1.17 11.8
(1) 1.10 14.3
Slabs with unbalanced moment 47
(2) 1.09 14.5
(1) Without calculation of ap, Eq. (C8.4.14)
(2) With calculation of ap, Eq. (C8.4.14) with dv replaced by 𝑑 ⋅ 𝑑 where ap=B/2 with B representing
the side length or diameter of the isolated specimens

The experimental-to-calculated resistance ratios are shown as a function of the main parameters governing the
punching shear resistance in Figures C8.4.4 to C8.4.6 for slabs and foundation slabs with concentric loading and
slabs with unbalanced moment, respectively. In the mentioned figures, the punching resistance was calculated
using Eq. (C8.4.14), still accounting for the favourable effect of the slenderness. It is important to note that no
noticeable trend can be clearly identified.
When the database of slabs without shear reinforcement is compared to EN 1992-1-1:2004, it is observed that
the accuracy of the punching resistance predictions is increased (in terms of coefficient of variation, 13.9% using
EN 1992-1-1:2004 compared to 11.2% for the same database) and no noticeable trends can be observed
(contrary to the case of EN 1992-1-1:2004, where, according to Kueres et al. [C8.4.19] using a different database,
trends are observed for the column size-to-effective depth ratio and depth of the member).

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 346
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

Figure C8.4.4: Experimental-to-calculated punching resistance ratio for a total of 121 specimens (database of
[C8.4.8]) representing slender slabs without shear reinforcement subjected to a concentric loading (punching
resistance calculated with Eq. (C8.4.14) accounting for the effect of slenderness replacing dv by (dvꞏap/8)1/2).

Figure C8.4.5: Experimental-to-calculated punching resistance ratio for a total of 34 specimens (database of
[C8.4.8]) representing foundation slabs without shear reinforcement subject to an uniform loading (punching
resistance calculated with Eq. (C8.4.14) accounting for the effect of slenderness replacing dv by (dvꞏap/8)1/2).

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 347
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

Figure C8.4.6: Experimental-to-calculated punching resistance ratio for a total of 47 specimens (database of
[C8.4.42]) representing slender slabs without shear reinforcement subject to a combined effect of shear forces
and unbalanced moment (factor e according to Eq. (C8.4.21) with eb=MR,test/VR,test; punching resistance calculated
with Eq. (C8.4.14) accounting for the effect of slenderness replacing dv by (dvꞏap/8)1/2).

8.4.3 (4) Punching shear resistance of prestressed slabs or slabs with axial forces
Formulae (8.99 to 8.103) Prestressing in slabs can influence the punching shear resistance due to three effects (see e.g. [C8.4.35]): (i)
deviation forces associated with prestressing tendons; (ii) moments due to eccentric prestressing forces and (iii)
compressive axial forces. The first effect is accounted for by calculating the acting shear forces at the control
perimeter (thus allowing to remove deviation forces inside of it, or considering the shear forces at the control
perimeter). The second effect is normally neglected in code formulations, while the third is often accounted for as
an increase of the unitary shear resistance as a function of the axial compressive stresses (see e.g., [C8.4.34] for
a discussion on the topic). In the fib Model Code 2010 [C8.4.1] and in the framework of the CSCT, the second
and third effects are accounted for by means of a decompression moment, which modifies the simplified load-
rotation relationship [C8.4.34, C8.4.35]. However, the analytical derivation of closed-form solutions for the
punching shear design of prestressed slabs combining the load-rotation relationship of Eq. (C8.4.8) (considering
the second and third previously mentioned effects) and the failure criterion of Eq. (C8.4.11) leads to complex
expressions that are not suitable for daily practice.
For that reason, the approach proposed in FprEN 1992-1-1:2023 results from an approximation of the results of
the mechanical model of the CSCT for slabs with in-plane forces [C8.4.35]. The effects of prestressing or axial
forces (compressive or tensile) are considered in the calculation of the punching shear resistance by multiplying
the value of kpb by the factor kpp calculated as follows:
𝑘 𝑘 if σd < 0 (compressive axial forces) (C8.4.23a)

𝑘 if σd > 0 (tensile axial forces) (C8.4.23b)

where the factor kN is calculated as follows:

. | |
𝑘 1 1 6 (C8.4.23)
,

where ep is the eccentricity of the tendons at the axis of the supported area with the respect to the centre of gravity
of the cross section (it is to be considered as positive for tendons on the tensile side and equal to zero for
prestressing forces centred with the cross section), and d is the effective depth in bending.

8.4.4 (6) and I.8.5.1 (1) Influence of compressive membrane action and slab continuity
Formula (I.15) Einpaul et al. [C8.4.43, C8.4.44] have shown that the slab continuity and compressive membrane action might
have a significant beneficial influence on the punching behaviour and strength of slab-column connections. These

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 348
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

effects are neglected in current EN 1992-1-1:2004 but have been accounted for in FprEN 1992-1-1:2023 for the
assessment of existing structures (Annex I). This is done consistently with the references [C8.4.43, C8.4.44,
C8.4.38, C8.4.8] where it is shown that the influence of slab continuity and compressive membrane action can be
analytically introduced in the closed form solution presented in Eq. (C8.4.13) and, therefore, in the design form
Eq. (C8.4.14).
The analytical derivation briefly presented in the following is grounded on the works presented by Muttoni et al.
[C8.4.38, C8.4.8], which can be consulted for further details. Einpaul et al. [C8.4.44] suggested a modification of
the load-rotation curve presented in Eq. (C8.4.1) to account for the stiffer behaviour (resulting from the
consideration of slab continuity) by multiplying the expression by a factor kcs. Analogously, Eq. (C8.4.10) can be
adapted to account for slab continuity as follows:

𝜓 𝑘 𝑘 8 (C8.4.24)

where kcs is a factor accounting for the influence of slab continuity and compressive membrane action. Based on
these works, Muttoni and Fernández Ruiz [C8.4.38] suggested a formulation for this factor which is a function of
the ratio mcr/mR (supported on the evidence that the confinement at the column region is provided by the
surrounding concrete during the crack development stage):

𝑘 0.08 1 (C8.4.25)

where mcr is the cracking moment per unit length (without considering an axial force). Eq. (C8.4.13) can therefore
be rewritten accounting for slab continuity and compressive membrane action as (following the same assumptions
used in the derivation of Eq. (C8.4.13) and considering ke=1 and d=dv as previously introduced):

,
4𝑎 100𝜌 𝑓 0.50 𝑓 (C8.4.26)
, ⋅ , ⋅ ⋅

It can therefore be seen that Eq. (C8.4.26) is analogous to Eq. (C8.4.13), considering that the value of ap is
corrected (reduced) to account for the slab continuity and compressive membrane action by the factor kpm=kcs2.
Considering that the flexural capacity per unit length mR can be approximated as 𝑚 0.75𝑑 𝜌 . 𝑓 . 𝑓 .
[C8.4.38, C8.4.8] and that the cracking moment per unit length can be computed as 𝑚 ℎ (considering
h=d/0.9), the factor kpm can be simplified as follows:
. . .
.
𝑘 𝑘 ≅ 0.08 ⋅ 6 ⋅ 1.0 (C8.4.27)

Simplifying the exponents (with minor impact) and considering fc≈30 MPa, fy≈500 MPa, ρ≈0.01, and fct=0.53ꞏfck1/2,
Eq. (C8.4.27) can written as:

𝑘 100𝜌 1.0 (C8.4.28)

This enhancement due to compressive membrane action can be expressed in terms of a coefficient that multiplies
factor kpb, thus improving the ease of use of the formulation. This coefficient, defined as ηpm, can be directly
obtained as the ratio between Equations (C8.4.26) and (C8.4.13):

, ,
𝜂 0.95 (C8.4.29)
, ⋅

The punching shear resistance of slabs in existing structures can therefore be calculated accounting for slab
continuity and compressive membrane action using Eq. (C8.4.14) with the parameter kpb being replaced by kpbꞏηpm
where the factor ηpm is calculated with Eq. (C8.4.29).

8.4.4 (1) Punching shear resistance of slabs with shear reinforcement

Formulae (8.104) to The analytical derivation of the design expressions for the calculation of the punching shear resistance of slabs
(8.107) with shear reinforcement (failures within the shear-reinforced zone) included in the FprEN 1992-1-1:2023 is
presented and discussed in reference [C8.4.11,C8.4.37] following the principles of the CSCT discussed above.
The main steps of such derivation are the following:
(1) Based on the load-rotation relationship of Eq. (C8.4.10), it is possible to establish the following
relationship: ψE/ψR,c = (VE/VR,c)3/2, where ψR,c and VR,c correspond respectively to the rotation at failure
and punching resistance of the corresponding member without punching reinforcement
(2) Other than the punching resistance of members without shear reinforcement, it is also possible to derive
the closed-form solution for the calculation of the rotation at failure for members without shear
reinforcement ψR,c (introducing Eq. (C8.4.13) in Eq. (C8.4.10)), see reference [C8.4.11,C8.4.37].
(3) Considering that VR,c/VE=ηc (with ηc ≤ 1), the relationship in (1) can be written as ψE = ψR,c (1/ ηc)3/2.
Introducing this relationship into Eq. (C8.4.7) (replacing in addition ψR,c by its closed-form expression)

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 349
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

allows the calculation of the stress in the shear reinforcement as a function of a number of known
parameters.
(4) Using the descending branch of the power-law failure criterion of Eq. (C8.4.11), and still considering
the relationship VR,c/VE=ηc established in (3), it can be shown that the concrete contribution VRc,E
(associated to the rotation ψE corresponding to the shear force VE) is in fact given by VRc,E= VR,c ∙ ηc
Based on the steps (1) to (4) above presented (please refer to [C8.4.11,C8.4.37] for details), it is possible to show
that the punching shear resistance associated to failures within the shear-reinforced region can be calculated as
follows (already expressed in a design format) [C8.4.11,C8.4.37]:
𝜏 , 𝜂 ⋅𝜏 , 𝜂 ⋅𝜌 ⋅𝑓 𝜌 ⋅𝑓 (C8.4.30)
where

𝜂 15 0.8 is a factor accounting for the activated shear reinforcement stress (upper
⋅ ⋅

limit referring to both yielding and amount of shear reinforcement that can
be effectively activated, see Figure C8.4.7)
,
𝜂 is the concrete contribution reduction factor accounting for the increased
deformation of slabs with shear reinforcement at ultimate limit state with
respect to slabs without shear reinforcement, see Figure C8.4.7

It should be noted that the additive factor in Eq. (C8.4.30) can only be accounted for provided that the
⋅
shear reinforcement consists of ribbed or indented bars.

Figure C8.4.7: Definition of contributions of concrete and shear reinforcement for punching failure within the
shear-reinforced region (from [C8.4.11]).

8.4.4 (5) and I.8.5.2 (2) Maximum punching shear resistance

Formulae (8.109) to Consistently with the principles of the CSCT, it can be shown that the maximum punching shear resistance of a
(8.111) and (I.19) slab with shear reinforcement is proportional to the punching shear resistance of the corresponding member
without shear reinforcement:
𝜏 , 𝜂 ⋅𝜏 , (C8.4.31)
where the factor ηsys accounts for a number of parameters [C8.4.30]: (i) the type of punching reinforcement (i.e.
the anchorage performance of the punching reinforcement system), (ii) the size of the column, (iii) the position of
the first perimeter of punching reinforcement and (iv) the detailing of the anchorages (mostly related to the
adequate enclosure of the third or the fourth layer of flexural reinforcement with the punching reinforcement and
the spacing of the subsequent perimeters). This factor has been analysed by means of a mechanical model
[C8.4.30] based on the fundamentals of the Critical Shear Crack Theory [C8.4.4, C8.4.5], consistently with the
work performed in [C8.4.9]. By means of a parametric analysis, the dependency of ηsys on different variables was
assessed, finding the most influential parameters in the characterisation of the maximum punching resistance of
a reinforced concrete element. Further information on the parametric analysis can be found in [C8.4.30]. As a
result of this parametric analysis, a general design formulation has been derived (Eq. (C8.4.32)) and included in
the Annex I.8.5.2(2), fitting the predictions of the mechanical model and showing the most influential parameters
in the definition of ηsys (the coefficients of each of the three factors involved in the following equation were
calibrated on the basis of test results):

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 350
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

𝜂 1.15 0.63 0.85 1.0 (C8.4.32)

where dsys is defined as in Figure I.3 of FprEN 1992-1-1:2023 and represents the anchorage performance of the
punching reinforcement system and its detailing, b0 is the perimeter of the supported region and represents the
size of the column, and s0 is the distance from the column face to the axis of the first perimeter of shear
reinforcement.
To increase the ease-of-use, Eq. (C8.4.32) has been simplified in Clause 8.4 for studs, links and stirrups fulfilling
the detailing rules of Clause 12.5.1:

𝜂 0.70 0.63 1.0 for studs (C8.4.33)

𝜂 0.50 0.63 1.0 for links and stirrups (C8.4.34)

It has to be noted that for this coefficient, two different symbols are used:
- sys is the coefficient which applies to the resistance calculated with the closed-form formulation
(clauses 8.4 and I.8.5.2(2)) whereas
- sys,sb is the coefficient which applies to the failure criterion in the strain-based approach (clause 8
I.8.5.2(3)). Since the resistance results from the intersection between the failure criterion and the load-
rotation relationship, the ratio between the maximum punching resistance and the resistance of a similar
slab without shear reinforcement is always smaller than sys,sb .

8.4.4 (7) and (8) Development of a closed-form formula for outer control perimeter
Formula (8.112) The punching shear resistance outside the shear reinforced region shall be calculated in accordance to Eq.
(C8.4.14), considering the reduced shear-resisting effective depth dv,out (function of the shear reinforcement
system) and the outer control perimeter b0,5,out, located at dv,out/2 from the outer perimeter of shear reinforcement
with a length of the straight segments not exceeding 3ꞏdv,out. This approach is consistent with the model that
grounds the expressions, thus considering all the rotation to concentrate outside of the shear-reinforced region.
The expression therefore is (following similar considerations as previously adopted for the derivation of Eq.
(C8.4.14)):

. .
𝜏 , , 𝑘 , 100𝜌 𝑓 𝑓 (C8.4.35)
⋅

where the value of kpb,out is given by (similarly to Eq. (C8.4.16)):

𝑘 , 5 𝜇 , (C8.4.36)
. ,

where dv is the shear-resisting effective depth and dv,out the outer shear-resisting effective depth (accounting for
the possibility of delamination in compression surface). It should be noted that the dv in the parts of the Eq.
(C8.4.35) related to kpb and to the size- and strain-effect is not replaced by dv,out as the dv in those cases is
associated to the flexural effective depth (simplified to the shear-resisting effective depth as discussed above).
The resulting Eq. (C8.4.35) thus comes from the analytical derivation of Eq. (C8.4.13) introducing d=dv but not by
dv,out in kpb and strain- and size-effect parts. Such consideration is on the safe side as shown in reference [C8.4.37]
(see also d/dv ratio in Eq. (C8.4.15)), where a discussion on this issue is presented.
Eq. (C8.4.36) which can be further simplified in the same manner as kpb in Eq. (C8.4.19), i.e., assuming that:
𝜇 , ⋅ ⋅𝑑 , 𝑏 , , 𝑏 , (C8.4.37)

where b0,out is the control perimeter at the outermost perimeter of shear reinforcement, with a length of the straight
segments not exceeding 3ꞏdv,out. Introducing Eq. (C8.4.37) in (C8.4.36), a simplified formula for kpb,out based on a
geometrical rule can be derived as follows (clause 8.4.4(4) of FprEN1992-1-1:2023) in an analogous manner to
the previous derivation of kpb in Eq. (C8.4.19):

,
1 𝑘 , 3.6 1 2.5 (C8.4.38)
, , ,

Comparing Eq. (C8.4.38) for the calculation of kpb,out to Eq. (C8.4.19) for kpb, it results that an increase of kpb by
multiplying it by factor (dv/dv,out)1/2 is possible for the verification outside the shear-reinforced region (as allowed in
Clause 8.4.4(4)). The consideration of such increase on the factor kpb can be shown to be justified and on the safe
side (see Eq. (C8.4.15) and reference [C8.4.37] for further discussion on this topic).
One can derive a closed-form solution to calculate the minimum length of the outer control perimeter to avoid a
failure outside the shear-reinforced region by equalling VRd,c,out (calculated with Eq. (C8.4.35)) to VEd as follows
(all rotation considered to develop outside the shear-reinforced area and strength criterion considered not to be
governing):

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 351
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

𝑑 𝑏 , 𝑘 , 𝑑 𝑏 , 𝑏 ,
𝜏 𝜏 , ⇒𝜏 𝜏 ,
𝑑 , 𝑏 , , 𝑘 𝑑 , 𝑏 , , 𝑏 , ,

⇒𝑏 , , 𝑏 , (C8.4.39)
,

considering again that c=Rd,c/Ed.

8.4.4 Punching shear of structural members with shear reinforcement

Experimental validation In this clause, the design provisions according to Eqs. (C8.4.30) through (C8.4.38) are compared to the following
databases of tests:
 135 tests of slender slabs with shear reinforcement (dataset of the database of [C8.4.45] completed
with the tests of [C8.4.46]);
 12 tests of foundation slabs with shear reinforcement and subject to uniform loading.
Specimens not respecting the detailing rules or reaching the flexural capacity were excluded from the databases.
The comparisons have been performed with average values of concrete properties (fck=fcm) and considering all
safety factor equal to the one. Table C8.4.2 presents a summary of the results.
Table C8.4.2: Summary of experimental-to-calculated punching shear resistance for the different investigated
cases of specimens with shear reinforcement.
Number of tests Approach Average COV [%]

(1) 1.21 11.9

Slabs with SR with concentric loading 135 (2) 1.19 11.4
(3) 1.17 11.8
(1) 1.68 6.8
Foundation slabs with SR with concentric
12 (2) 1.32 7.6
loading
(3) 1.06 11.7
(1) Without calculation of ap and factor ηsys calculated according to Eqs. (C8.4.33) and (C8.4.34).
(2) With calculation of ap, dv in Eq. (C8.4.14) replaced by 𝑑 ⋅ 𝑑 where ap=B/2 with B representing
the side length or diameter of the isolated specimens (same approach for ηsys as in (1))
(3) Approach (2) with coefficient ηsys calculated according to Eq. (C8.4.32)

The experimental-to-calculated resistance ratios for members with shear reinforcement are shown as a function
of the main parameters governing the punching shear resistance in Figures C8.4.8 and C8.4.9 for slabs and
foundation slabs with concentric loading, respectively. Again, no noticeable trend can be clearly identified.

With respect to approach (3) in Table C8.4.2 (refined coefficient ηsys according to I.8.5.2(2) for the assessment of
existing structures), in the case of flat slabs with details not fulfilling the rules of Clause 12, no notable increase
of the accuracy compared to the simplified formula in Clause 8.4 can be observed (approach (2) in Table C8.4.2).
Nevertheless, the detailed formula according to I.8.5.2(2) is particularly useful for the assessment of slabs not
fulfilling anymore the detailing rules of Clause 12.

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 352
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

Figure C8.4.8: Experimental-to-calculated punching resistance ratio for a total of 135 specimens (dataset of
database of [C8.4.45] completed with the tests of [C8.4.46]) representing slender slabs with shear reinforcement
subjected to a concentric loading, using approach (2) as defined in Table C8.4.2: 4 tests of [C8.4.47]; 1 test of
[C8.4.48]; 1 test of [C8.4.49]; 6 tests of [C8.4.50]; 1 test of [C8.4.51]; 7 tests of [C8.4.52]; 7 tests of [C8.4.53]; 6
tests of [C8.4.54]; 2 tests of [C8.4.55]; 2 tests of [C8.4.56]; 5 tests of [C8.4.57]; 10 tests of [C8.4.58]; 3 tests of
[C8.4.59]; 8 tests of [C8.4.60]; 17 tests of [C8.4.61]; 9 tests of [C8.4.62]; 6 tests of [C8.4.63]; 2 tests of [C8.4.64];
1 test of [C8.4.65]; 8 tests of [C8.4.66]; 2 tests of [C8.4.67]; 5 tests of [C8.4.14]; 11 tests of [C8.4.28]; 11 tests of
[C8.4.46].

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 353
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

Figure C8.4.9: Experimental-to-calculated punching resistance ratio for a total of 12 specimens representing
foundation slabs with shear reinforcement using approach (2) as defined in Table C8.4.2: 3 tests of [C8.4.13]; 5
tests of [C8.4.69]; 4 tests of [C8.4.31].

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 354
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

References [C8.4.1] Fédération internationale du béton, fib Model Code for Concrete Structures 2010, Ernst & Sohn,
Germany, 2013, 434 p.
[C8.4.2] Muttoni, A., and Schwartz, J., Behaviour of Beams and Punching in Slabs without Shear
Reinforcement, IABSE Colloquium, V.62, Zürich, Switzerland, 1991, pp. 703-708
[C8.4.3] Muttoni A., Shear and Punching resistance of Slabs Without Shear Reinforcement (In German:
Schubfestigkeit und Durchstanzen von Platten ohne Querkraftbewehrung, Beton- und Stahlbetonbau),
Beton-und Stahlbetonbau, V.98, 2003, pp. 74-84
[C8.4.4] Muttoni A., Punching shear strength of reinforced concrete slabs without transverse reinforcement, ACI
Struct J, Vol.105, No.2, 2008, pp. 163-172
[C8.4.5] Fernández Ruiz M., Muttoni A., Applications of the critical shear crack theory to punching of R/C slabs
with transverse reinforcement, ACI Struct J., V.106, No.4, 2009, pp. 485-494
[C8.4.6] Muttoni A., Fernández Ruiz M., Shear in slabs and beams: should they be treated in the same way?,
fib bulletin 57, Lausanne, Switzerland, 2010, pp. 105-128
[C8.4.7] Muttoni A., Fernández Ruiz M., Bentz E., Foster S., Sigrist V., Background to fib Model Code 2010
shear provisions–part II: punching shear, Struct Conc., V.14, No.3, 2013, pp. 204-214
[C8.4.8] Muttoni A., Fernández Ruiz M., Simões J.T., The theoretical principles of the Critical Shear Crack
Theory for punching shear failures and derivation of consistent closed-form design expression, Struct
Concr., V. 19, 2018, pp. 174-190, https://doi.org/10.1002/suco.201700088
[C8.4.9] Simões J.T., Fernández Ruiz M., Muttoni A., Validation of the Critical Shear Crack Theory for punching
of slabs without transverse reinforcement by means of a refined mechanical model, Struct Conc, V. 19,
2018, pp. 191–216. https://doi.org/10.1002/suco.201700280
[C8.4.10] Muttoni A, Fernández Ruiz M., Simões J. T., Recent improvements of the Critical Shear Crack Theory
for punching shear design and its simplification for code provisions, Proceedings of 5th fib Symposium,
Melbourne, 2018, 14 p.
[C8.4.11] Muttoni A., Simões J.T., Faria D.M.V., Fernández Ruiz M., A mechanical approach for the punching
shear provisions in the second generation of Eurocode 2, Hormigón y Acero, 2023, doi:
https://doi.org/10.33586/hya.2022.3091
[C8.4.12] Muttoni A., Simões J. T., “Shear and punching shear according to the Critical Shear Crack Theory:
background, recent developments and integration in codes,” Rev. IBRACON Estrut. Mater. (submitted
for publication)
[C8.4.13] Hegger J., Ricker M., Sherif A. G., Punching resistance of Reinforced Concrete Footings, ACI Struct
J., V.106, No.5, 2009, pp. 706-716
[C8.4.14] Vollum R.L., Abdel-Fattah T., Eder M., Elghazouli A.Y., Design of ACI-type punching shear
reinforcement to Eurocode 2, Mag conc Research, V.62, No.1, 2010, pp. 3-16
[C8.4.15] Siburg C., Ricker M., Hegger J., Punching shear design of footings: critical review of different code
provisions, Struct Conc., V.15, 2014, pp. 497-508
[C8.4.16] Ricker M., Siburg C., Punching shear strength of flat slabs – critical review of Eurocode 2 and fib Model
Code 2010 design provisions, Struct Conc., V.17, 2016, pp. 457-468
[C8.4.17] Einpaul, J., Bujnak, J., Fernández Ruiz, M., Muttoni, A., Study on influence of column size and slab
slenderness on punching resistance, ACI Struct J., V.113, No.1, 2016, pp. 135-145
[C8.4.18] Einpaul J., Brantschen F., Fernández Ruiz M., Muttoni A., Performance of Punching Shear
Reinforcement under Gravity Loading: Influence of Type and Detailing, ACI Struct J., V.113, No.4,
2016, pp. 827-838
[C8.4.19] Kueres D., Siburg C., Herbrand M., Classen M., Hegger J., Uniform Design Method for punching shear
in flat slabs and column bases, Eng Struct., V.136, 2017, pp. 149-164
[C8.4.20] Fernández Ruiz M., Muttoni A., Size effect in shear and punching shear failures: differences between
statically determinate members and redundant structures, Struct Concr., 2017, pp. 1-11, DOI:
10.1002/suco.201700059
[C8.4.21] Dönmez A., Bazant Z.P., Size effect on punching resistance of reinforced concrete slabs with and
without shear reinforcement, ACI Struct J., V.114, No.4, 2017, pp. 875-886
[C8.4.22] Kinnunen S., Nylander H., Punching of Concrete Slabs Without Shear Reinforcement, Transactions of
the Royal Institute of Technology, No.158, 1960, 112 p.
[C8.4.23] Guandalini S., Symmetric punching shear of reinforced concrete slabs (In French), Doctoral Thesis,
EPFL, Lausanne, Switzerland, No.3380, 2005, 289 p.
[C8.4.24] Einpaul J., Punching resistance of Continuous Flat Slabs, Doctoral Thesis, EPFL, Lausanne,
Switzerland, No.6928, 2016, 209 p.
[C8.4.25] Belletti B., Walraven J.C., Trapani F., Evaluation of compressive membrane action effects on punching
shear resistance of reinforced concrete slabs, Eng Struct., V.95, 2015a, pp. 25-39

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 355
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

[C8.4.26] Belletti B., Pimentel M., Scolari M., Walraven J.C., Safety assessment of punching shear failure
according to the level of approximation approach, Struct Conc., V.16, No.3, 2015b, pp. 366-380
[C8.4.27] Beutel R., Hegger J., The effect of anchorage on the effectiveness of the shear reinforcement in the
punching zone, Cement and Concrete Composites, V.24, 2002, pp. 539-549
[C8.4.28] Lips S., Fernández Ruiz M., Muttoni A., Experimental Investigation on Punching resistance and
Deformation Capacity of Shear-Reinforced Slabs, ACI Struct J., V.109, 2012, pp. 889-900
[C8.4.29] Brantschen F., Faria D. M. V., Fernández Ruiz M., Muttoni A., Bond Behaviour of Straight, Hooked, U-
Shaped and Headed Bars in Cracked Concrete, Struct Conc., V.17, No. 5, 2016, pp. 799-810
[C8.4.30] Hernández Fraile D., Simões J. T., Fernández Ruiz M., Muttoni A., A mechanical approach for the
maximum punching resistance of shear-reinforced slab-column connections, Proceedings of the fib
Symposium, Lisbon, 2021, pp. 1628-1639.
[C8.4.31] Simões J.T., Bujnak J., Fernández Ruiz M., Muttoni A., Punching shear on compact footings with
uniform soil pressure, Structural Concrete, Struct C., No.4 , 2016
[C8.4.32] Muttoni A., Fernández Ruiz M., Guandalini S., Poinçonnement des ponts-dalles, ASTRA, 2007.
[C8.4.33] Clément T., Influence de la précontrainte sur la résistance au poinçonnement de dalles en béton armé,
Doctoral Thesis, EPFL, Lausanne, Switzerland, 2012, No. 5516, 222 p.
[C8.4.34] Clément T., Pinho Ramos A., Fernández Ruiz M., Muttoni A, Design for punching of prestressed
concrete slabs, Struct Conc., V.14, No. 2, 2013, pp. 157-167
[C8.4.35] Clément T., Pinho Ramos A., Fernández Ruiz M., Muttoni A., Influence of prestressing on the punching
resistance of post-tensioned slabs, Eng Struct., V.72, 2014, pp. 59-69
[C8.4.36] SIA. Code 262 for Concrete Structures. Zürich, Switzerland: Swiss Society of Engineers and Architects;
2013, 102 p.
[C8.4.37] Muttoni A., Simões J.T., Hernández Fraile D., The punching shear provisions of the 2º generation of
Eurocode 2: a discussion on the derivation of the formulae, adopted simplifications and their
consequences, under preparation, 2023
[C8.4.38] Muttoni A., Fernández Ruiz M., The Critical Shear Crack Theory for punching design: from a
mechanical model to closed-form design expressions, ACI/fib International Punching Shear
Symposium, Bulletin fib 81, 2007, pp. 237-252
[C8.4.39] Cavagnis F. Fernández Ruiz M., Muttoni A., A mechanical model for failures in shear of members
without transverse reinforcement based on development of a critical shear crack, Eng Struct, V. 157,
2018, pp. 300-315
[C8.4.40] Abu-Salma D., Vollum R. L., Maccorini L., Design of biaxially loaded external slab column connections,
Engineering Structures, V. 249, 2021, 16 p.
[C8.4.41] Muttoni A., FprEN 1992-1-1:2023 – Background document to clause 4.3.3 and Annex A – Partial safety
factors for materials, CEN/TC250/SC2/WG1/TG6, 2023, 21 p.
[C8.4.42] Tassinari L., Non symmetric punching of reinforced concrete slabs (In French : Poinçonnement
symétrique des dalles en béton armé avec armature de poinçonnement), Doctoral Thesis, EPFL, No.
5030, Lausanne, Switzerland, 2011, 197 p.
[C8.4.43] Einpaul J., Ospina C. E., Fernández Ruiz M., Muttoni A., Punching Shear Capacity of Continuous
Slabs, ACI Struct J., V. 113, No. 4, 2016, pp. 861-872
[C8.4.44] Einpaul J., Fernández Ruiz M., Muttoni A., Influence of moment redistribution and compressive
membrane action on punching resistance of flat slabs, Eng Struct., V.86, 2015, pp. 43-57
[C8.4.45] Walkner R., Critical review of EC 2 regarding punching and improving the design approach (In German:
Kritische Analyse des Durchstanznachweises nach EC2 und Verbesserung des
Bemessungsansatzes), Doctoral Thesis, Innsbruck University, 2014.
[C8.4.46] Ferreira M. P., Melo G. S., Regan P. E., Vollum R. L., Punching of Reinforced Concrete Flat Slabs with
Double-Headed Shear Reinforcement, ACI Struct J., 2014, pp. 363-374
[C8.4.47] Andersson J.L., Punching of Concrete Slabs with Shear Reinforcement, Transactions of the Royal
Institute of Technology, Stockholm, Sweden, 1963
[C8.4.48] Marti P., Pralong J., Thürlimann B., Schubversuche an Stahlbeton-Platten, Institut für Baustatik und
Konstruktion, No.7305-2, Zürich, Switzerland, 1977, 123 p.
[C8.4.49] Seible F., Ghali A., Dilger W. H, Preassembled shear reinforcing units for flat plates, ACI J Proc., V.77,
No.1, 1980, pp. 28-35
[C8.4.50] Van der Voet A.F., Dilger W.H., Ghali A., Concrete Flat Plates with Well-Anchored Shear
Reinforcement Elements, Canadian Journal of Civil Engineering, V.9, No.1, 1982, pp. 107-114
[C8.4.51] Müller F.X., Muttoni A., Thürlimann B., Punching shear tests on slabs with openings (In German:
Durchstanzversuche an Flachdecken mit Aussparungen), Institut für Baustatik und Konstruktion,
No.7305-5, Zürich, Switzerland, 1984, 117 p.

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 356
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.4 Background to FprEN 1992-1-1:2023

[C8.4.52] Mokhtar A.S., Ghali A., Dilger W.H., Stud Shear Reinforcement for Flat Concrete Plates, ACI J., No.82-
60, 1985, pp. 676-683
[C8.4.53] Chana P.S., Desai S.B., Design of shear reinforcement against punching, The Structural Engineering,
V.70, 1992, pp. 159-164
[C8.4.54] Yamada T., Nanni A., Endo K., Punching Shear Resistance of Flat Slabs: Influence of Reinforcement
Type and Ratio, ACI Struct J., V.88, 1992, pp. 555-563
[C8.4.55] Chana P.S., A prefabricated shear reinforcement system for flat slabs, Proceedings of the Institution of
Civil Engineers: Structures & Buildings, V.99, 1993, pp. 345-358
[C8.4.56] Ladner M., Schaeidt W., Gut W., Durchstanzversuche an Flachdeckenausschnitten, EMPA Report No.
419, 1996, 38 p.
[C8.4.57] Lee S.C., Lee S.B., Teng S., Morley C.T., Punching shear tests on high strength concrete slabs,
Proceedings of the 5th international symposium on utilisation of high strength-high performance
concrete, 1999, pp. 401-410
[C8.4.58] Gomes R.B., Regan P.E., Punching Resistance of RC Flat Slabs with Shear Reinforcement, ASCE
Journal of Struct Eng., V.125, No.6, 1999, pp. 684-692
[C8.4.59] Oliveira D.R., Melo G.S., Regan P.E., Punching resistances of Flat Plates with Vertical or Inclined
Stirrups, ACI Struct J., V. 97, 2000, pp. 485-491
[C8.4.60] Regan P.E., Samadian F., Shear Reinforcement against punching in reinforced concrete flat slabs, The
Structural Engineer, 2001, pp. 24-31
[C8.4.61] Beutel R., Durchstanzen schubbewehrter Flachdecken im Bereich von Innenstützen, RWTH, Aachen,
Germany, 2002, 267 p.
[C8.4.62] Birkle G., Punching of Flat Slabs: The Influence of Slab Thickness and Stud Layout, Doctoral Thesis,
University of Calgary, Calgary, Canada, 217 p., 2004
[C8.4.63] Rojek R., Keller T., Punching Tests using Reinforcement with Enhanced Bond (In German), Structural
Analysis and Design Approaches, Beton-und Stahlbetonbau, V.102, No.8, 2007, pp. 584-556, 2007
[C8.4.64] Broms C. E., Ductility of Flat Plates: Comparison of Shear Reinforcement Systems, ACI Struct J.,
V.104, No.6, 2007, pp. 703-711
[C8.4.65] Hegger J., Häusler F., Ricker M., Zur maximalen Durchstanztragfähigkeit von Flachdecken, Beton -
und Stahlbetonbau, V.102, 2007, pp.770-777
[C8.4.66] Feix J., Schustereder C., Durchstanzen nach EN 1992-1-1, Durchstanzversuche zur Festlegung der
Nachweisgrenzen, Bauingenieur, V.82, 2007, pp. 135-142
[C8.4.67] Etter S., Heinzmann D., Jäger T., Marti P., Versuche zum Durchstanzverhalten von Stahlbetonplatten,
IBK Bericht, 64, Zürich, Switzerland, 2009
[C8.4.68] Lips S., Muttoni A., Fernández Ruiz M., Punching of flat slabs: Design example, EPFL, Lausanne,
Switzerland, 2011, 13 p.
[C8.4.69] Siburg C., Hegger J., Experimental Investigations on Punching Behaviour of Reinforced Concrete
Footings with structural dimensions, Struct C., V.15, 2014, pp. 331-339

List of versions of this document

03.06.2016, A.M., M.F.R., J.S. D1 1st draft (submitted to SC2/WG1/TG4)
18.08.2016, A.M., M.F.R., J.S. D2 2nd draft: substantial revision of the 1st draft, inclusion of slab continuity and compressive membrane
action; prestressed slabs; edge and corner columns (submitted to SC2/WG1/TG4)
30.11.2017, A.M., M.F.R., J.S. D3 new template, changes in prEN 1992-1-1:2017-10 implemented (submitted to SC2 & WG1)
29.06.2018, A.M., M.F.R., J.S., D.K. D4 changes in prEN 1992-1-1:2018 implemented (Background document to prEN 1992-1-1:2018 (D3),
distributed by PT1)
20.05.2021, A.M., M.F.R., D.H.F. D7 changes in prEN 1992-1-1:2021 implemented (Background document to prEN 1002-1-1:2020-11 (D7)
17.03.2022, A.M., J.S., D.H.F. Substantial revision of previous draft, changes in FprEN 1992-1-1:2023 implemented
20.03.2023, TJ Review of the document
27.03.2023, AM Final review

A. Muttoni, M. Fernández Ruiz, J.T. Simões, D.H. Fraile – EPFL 27.3.2023

J. Hegger, C. Siburg, D. Kueres - RTWH
For citations, copying and referencing see Note on Front page! page 357
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to subsection 8.5

Design with strut-and-tie models and stress
fields

Miguel Fernández Ruiz Ecole Polytechnique Fédérale de Lausanne,

Aurelio Muttoni Switzerland

Report EPFL-IBETON 17-01-R2 23.02.2023

For citations, copying and referencing see Note on Front page! page 358
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.5 Background to FprEN 1992-1-1:2023

Introduction This document has been prepared to clarify the background of the design provisions of FprEN 1992-1-1:2023
with respect to Section 8.5 (“Design with strut-and-tie models and stress fields”). These provisions inherit the
approach of former Section 6.5 of EN 1992-1-1:2004, but updates its content and design clauses based on the
evolutions of the last 20 years. Also, the section has been reworked to have an enhanced consistency with
Section 8.2.3 (“Members requiring design shear reinforcement”) and Annex G (“Design of membrane-, shell-
and slab elements”) of FprEN 1992-1-1:2023.

Reasons for change Current provision for strut-and-tie models in EN 1992-1-1:2004 (clause 6.5) gave rules for the design of D-
regions (discontinuity regions “where a non-linear strain distribution exists” as stated in the code) based on the
strut-and-tie method. These provisions were to a large extent consistent with the vision by Schlaich and co-
workers [C8.5.1], that represented the state-of-the-art in the 1980’s. Since then, many advances in the
knowledge have been consolidated in this field, leading to a more comprehensive understanding of several
topics as:
i. analogies of strut-and-tie models with the stress field method [C8.5.2]
ii. explicit consideration of the strain state in the calculation of efficiency factors [C8.5.3]
iii. integration of strut-and-tie models and stress fields within automated design procedures and
consistency with design of B-regions [C8.5.4]
Such advances have shown limitations and deficiencies of the EN1992-1-1:2004 approach, advising to revise it.
A detailed review of current state-of-the-art in this domain has recently been completed by a dedicated group of
fib (WP2.2.4), whose results have been condensed within a specific Bulletin [C8.5.5]. The revision of the
FprEN1992-1-1:2023 provisions are to a large extent based on this knowledge and such document shall be
consulted for extended clarifications. The fib Bulletin [C8.5.5] can thus be considered as a more
comprehensive background document for clause 8.5 of FprEN 1992-1-1:2023.

8.5.1 General This clause has been revised and extended to clarify the definitions used and the analogy between strut-and-tie
models and stress fields. In general, the strut-and-tie models can be considered as the equilibrium models
where the location of the struts and ties represent the resultant of compression fields and reinforcement
respectively, see Fig. C8.5.1.

Figure C8.5.1: Analogy between: (a) stress fields; and (b) strut-and-tie models

8.5.2 Struts and Other than an explicit definition of the stress in the strut, this clause incorporates, with respect to EN 1992-1-
compression fields 1:2004, a variable value of the efficiency factor  for the concrete strength depending on the cracking state of
the compression field:
 Constant values for a compression field developing at a given angle with a tie (stressed
reinforcement, §8.5.2(4)). This approach is particularly suitable for a preliminary design, where the
exact value of the angle between struts and ties is not known a priori, but its range can be easily
estimated
 Variable value for a compression field developing at a given angle with a tie or tension field (Formula.
(8.119)). This approach allows for a tailored calculation of the resistance, particularly for a more
refined calculation in a final check or automated design procedure
 Local value of the efficiency factor based on its principal tensile strain, Formula (8.121). This
approach is mostly addressed for an automated calculation (as a FE stress field analysis or for the
use of Annex G)
The different approaches are compared in Figure C8.5.2. As it can be noted, the constant values are aimed at a
simple and safe estimate of the efficiency factor  while Formula (8.119) provides in general higher estimates of
the strength.

M. Fernández Ruiz, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 359
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.5 Background to FprEN 1992-1-1:2023

Figure C8.5.2: Comparison of different approaches to calcualate the concrete efficiency factor

It is interesting to note that Formula (8.119) can be consistently derived from the local formulation of  (Formula
(8.121)), by estimating the principal strain (1) assuming that the compressive strain in concrete is 0,001 and
that the strain tensor is parallel to the stress field (applies when one assumes that cracks, which have the same
inclination  as the compression field, open without sliding):
1= x +( x + 0.001)∙cot2 (C.11.3.1)
where the principal strain value is calculated for a given level of horizontal strain ( x) and an inclination of the
compression field equal to  (refer to Fig. C8.5.3). As it can be noted, this assumption is consistent with the
procedure for shear design presented in § 8.2.3(7) and Formula (8.45). In this case, for calculation of Formula
(8.119), it is considered that  x ≈ 0.001 (horizontal strain at mid-height of a compression field where one side
reaches the yield strength and the deformations on the other are neglected, refer to Formulae. (8.47) - (8.49)).

Figure C8.5.3: Mohr’s circle of strains

It can be noted that the formulation proposed in Formula (8.119) covers a region with a constant state of strains.
Its application is thus direct for parallel stress fields. However, for fan-shaped stress fields, this requires
identification of potential governing control sections. A typical example (end support region of a beam) is shown
in Figure C8.5.4. A peculiar state of strains is found near to the nodal region (control section 1, no tensile strains
due to stirrups in tension) and the  factor has thus to be selected according to the provisions for such type of
node (§8.5.4.3). However, when the fan is intersected by the stirrups (refer to control section 2 in Fig. C8.5.4),
higher tensile strains result. In this case, design according to Formula (8.119) applies, considering the angle 
between the resultant of forces and the tension tie (refer to Fig. C8.5.4 and to §8.5.2(4)) and a width bc in
Formula (8.113) corresponding to the length of the control section.

M. Fernández Ruiz, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 360
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.5 Background to FprEN 1992-1-1:2023

Figure C8.5.4: Example of identification of control sections

In addition to these aspects, several considerations have been incorporated covering the contribution of
compression reinforcement (§8.5.2(7)) and also local enhancement of the compressive strength due to
confinement (§8.5.2(6))

8.5.3 Ties The clause has been reworked to enhance clarity and to clarify how to account for prestressing units.

8.5.4 Nodes Design for nodes has been integrated as part of the general verification of the compression fields reaching the
nodes (see example of Fig. C8.5.2).
The provisions for the nodal regions are consistent with current draft of Model Code 2020 (based on the work by
fib’s WP 2.2.4 [C8.5.5]) with respect to the conditions for the strength reduction factors of cracked concrete. For
CCT nodes, a simplification has been adopted for the value of  in cases with partial anchorage within the node.
This simplification consists of assuming a linear interpolation between the value for a fully smeared nodal region
(parallel compression field) and a concentrated node (full anchorage outside of the nodal region). This choice
(not explicit in current draft of MC2020) avoids requiring a detailed check of the nodal region.
With respect to the stress states of the nodes, they refer normally to pseudo-hydrostatic stress conditions (equal
stresses in two directions, with no out-of-plane stress). This ensures that no explicit check of the stress state in
the nodes has to be performed (refer to Kupfer’s response of biaxially-loaded concrete [C8.5.2]). For some
cases, as for the partial anchorage case shown in Figure 8.28c, a more complex stress state may be provided.
A detailed verification of such state is also not required provided that the values of the  factors are respected
for the compression fields converging in the node.
For other cases, similar treatment can be followed, as for instance when multiple struts and ties converge in a
node (which can be subdivided into a number of smaller nodes, see Fig. C8.5.5)

Figure C8.5.5: Example of a node where multiple struts converge and detail of nodal region

8.5.5 Spreading of struts This clause is based on the design approach of previous EN1992-1-1:2004 for §8.5.5(1-2) but making the
design assumptions consistent between different cases. It is also allowed to apply alternative (tailored)
approaches based on specific knowledge.
A new clause 8.5.5(3) has been added. This clause refers to a reinforcement for local spreading of a load
applied near to an edge. This can be a typical situation of columns supported on short walls. The reinforcement
detailed in §8.5.5(3) prevents local cracking to develop in an uncontrolled manner near the loading region.
Respecting this condition does not however exclude to develop a complete load-carrying model for the element.
Alternatively, a detailed analysis of the D-region can be performed ensuring proper detailing.

M. Fernández Ruiz, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 361
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.5 Background to FprEN 1992-1-1:2023

References [C8.5.1] Schlaich, J.; Schäfer, K.; and Jennewein, M., “Toward a Consistent Design of Structural Concrete,”
PCI Journal, V. 32, No. 3, 1987, pp. 74-150.
[C8.5.2] Muttoni, A.; Schwartz, J.; and Thürlimann, B., Design of Concrete Structures with Stress Fields,
Birkhäuser, Basel-Boston-Berlin, 1997, 143 pp.
[C8.5.3] Vecchio, F. J., and Collins, M. P., “The Modified Compression-Field theory for Reinforced Concrete
Elements Subjected to Shear,” ACI Journal Proceedings, V. 83, No. 2, Mar.-Apr. 1986, pp. 219-231.
[C8.5.4] Fernández Ruiz, M., and Muttoni, A., “On Development of Suitable Stress Fields for Structural
Concrete,” ACI Structural Journal, V. 104, No. 4, July-Aug. 2007, pp. 495-502.
[C8.5.5] fib, WP2.2.4, The International Federation for Structural Concrete, “Design and assessment with
strut-and-tie models and stress fields: from simple calculations to detailed numerical analysis”,
Bulletin 100, fib, 2021.

List of versions of this document

3.2.2021. 1st draft
11.02.2021 Corrections
24.2.2021 Minor changes
23.02.2023 Update to FprEN 1992-1-1:2023

M. Fernández Ruiz, A. Muttoni 23.02.2023

For citations, copying and referencing see Note on Front page! page 362
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.6 and I.8.6 Background to FprEN 1992-1-1:2023

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to
clauses 8.6 and I.8.6
Partially loaded areas

Alejandro Pérez Caldentey FHECOR, Spain

Francesco Marchetto UPM, Spain
08.03.2023

A. Pérez Caldentey, FHECOR, F. Marchetto, UPM 08.03.2023

For citations, copying and referencing see Note on Front page! page 363
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.6 and I.8.6 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to subsection 8.6 (Partially loaded areas).

8.6 and I.8.6 Reasons for change

Partially loaded areas The model of current EN 1992-1-1.2004 section 6.7 is very much on the safe side with respect to compression
failure for two reasons:
- The maximum triaxial compression capacity is 3,0fcd. The origin of this limit can be traced back to
Leonhardt [C08-1], who, while working the method of admissible stresses, proposed a reduction
factor of 3,0 for the admissible tri-axial concrete strength. This factor has not been accounted for
when translating the safety format from admissible stresses to the partial factor method. In spite of
this fact, Section 8.6 has maintained the factor of 3,0, “unless a larger resistance can be justified
based on refined analysis” due to the opposition of some countries to introduce the less
conservative value. This excess of safety may be addressed in the National Annexes.
- The concrete resisting area is determined as having a similarity relationship to the plate
dimensions. The application of this criterion to the experimental data results in predictions which
are too conservative.
Furthermore, the code forces the designer to provide tension reinforcement. This cannot always be done,
especially when concentrated loads arise after construction or due to construction (for instance the placing
of a large crane on an already built structure). Because of this, an expression allowing the user to determine
the expected design failure in tension, in absence of reinforcement, is given in Annex I.

Enhancement of ease of use

For existing structures, the proposal widens the range of application of the formulation to cases where no
tension reinforcement is present, thereby providing a solution to a fairly common engineering problem.
The less conservative approach avoids non-compliance problems which can be encountered with the current
formulation due to its clearly over-conservative character.

8.6 and I.8.6 Description and justification of the proposed expressions

Expressions (I.22), (I.23) The basis for the proposed model for the resistance to partially loaded areas without reinforcement was
and (I.24) and published in Engineering Structures in 2013 [C08-2]. It was validated by ad hoc tests described in reference
considerations on the [C08-3] and has been the object of a PhD thesis [C08-4] read in 2015. Following is a description of the basis
compressive capacity for the model.
Stresses generated by partially loaded areas may lead to failure due to tension stresses generated by the
diffusion of compression forces from the plate to the concrete around it or by uniaxial compression, if the plate
dimensions are close to the actual concrete block dimensions, or by triaxial compression if the plate
dimensions are very small.
The forces acting on concrete due to the plate can be analysed by using the classical formulation for
concentrated loads acting on a concrete block. These expressions date back to research conducted by
Leonhardt [C08-1] (see Fig. 1).

Fig. 1– The problem of concentrated loads (after reference [C08-1]).

A. Pérez Caldentey, FHECOR, F. Marchetto, UPM 08.03.2023

For citations, copying and referencing see Note on Front page! page 364
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.6 and I.8.6 Background to FprEN 1992-1-1:2023

This problem is solved in most modern design codes (EN-1992-1-1, MC-2010, ACI-318) by strut-and-tie
modelling using the model of Fig. 2, where a1 is the block dimension, and a0 is the dimension of the plate
parallel to a1.

a1
a0
a1  a0
N Sd
a1

a1/2
a1
a1/2
N Sd a1  a0 N Sd a1  a0
2 a1 2 a1

Fig. 2 – Strut-and-tie modelling for concentrated loading.

The force, which is concentrated within the area of the bearing, will spread out until uniform compression (d)
is achieved in the concrete block at approximately one block depth, a1. Not all the force has to spread out
since a part of it – that which is already distributed within the bearing width (a0) can follow a straight path. The
total force NSd, can be expressed as in Eq. (1).

NSd   Sd a1

The part of this force that does not need to be deflected N*Sd can be expressed as in Eq. (2).
*
NSd   Sd a0

Therefore, the force that does need to be deflected can be determined as in Eq. (11).

NSd  NSd
*
  Sd  a1  a0   NSd
 a1  a0 
a1

To determine the tensile force needed to achieve the diffusion of the force at the bearing, NSd can be divided
into two halves each located at a distance from the centre of the block of (a1-a0)/4. This force must be
transferred to a1/4 (measured horizontally) from the centre of the block. It is assumed that this is accomplished
in a distance equal to the width of the block, the centroid of the tension force being located at a1/2 (measured
vertically). More generally a certain diffusion angle mod can be assumed. The value of the tangent of this
angle, if the diffusion distance is taken as a1, as in the case represented in Fig. 2, would be tanmod =0,5.
More generally, the tensile force generated by the spreading of the compression stresses can be expressed
as in Eq. (4).

Nsd  a  a0  Nsd  a 
TSd  tg  mod   1  tg  mod   1  0 
2  a1  2  a1 

For a three-dimensional problem, with a bearing plate of dimensions a0b0 and a concrete block of dimensions
a1b1 (b1 is defined later), this equation can be generalized as in Eq. (5) where the geometric mean of the
ratio between plate and block dimensions is used.

NSd  a0 b0 
TSd  tg  mod   1  
2  a1b1
 
In existing structures, the presence of sufficient transverse reinforcement cannot always be guaranteed, so
that this tensile force may have to be taken by the tensile resistance of concrete. For this, it may be assumed1

1
Other more complex assumptions are possible, such as modelling the failure surface with a pyramidal shape as proposed by De Vries [C08-5]
in his “physical model”. However, due to the uncertainties regarding stress distribution in concrete within the failure surface, such an assumption
only leads to more complexity which is not rewarded by a better fit.

A. Pérez Caldentey, FHECOR, F. Marchetto, UPM 08.03.2023

For citations, copying and referencing see Note on Front page! page 365
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.6 and I.8.6 Background to FprEN 1992-1-1:2023

that the area of concrete resisting this force has a depth of one block dimension (a1) and an effective width,
b1, which will be dependent on the width of the bearing plate and, again, on the height of the resisting section,
a1, and can be expressed as in equation (6).

b1  bh   a1  b

Where b is the maximum block dimension and b0 is the dimension of the plate parallel to b, and  is another
model parameter (like mod). Finally, it must be considered that the distribution of tensile stresses within this
area will not be uniform as shown on the right-hand side of Fig. 1. Therefore, the mean admissible tensile
stress can be expressed as fctd, where  is a stress distribution factor which must be smaller than 1.0 and
fctd is the design tensile strength of concrete. With these definitions, the resistance condition can be expressed
in terms of the ultimate load which produces side blowout failure, NRd,2, as in Eq. (7).

Nsd  ab 
Tsd  TRd  tg  mod   1  0 0    fctd a1b1
2  a1b1 

fctd a1b1
NRd ,2  Nsd  NRd ,2  2  
tg  mod   ab 
 1  0 0 
 a1b1 

Eq. (7) provides an expression for the resistance of the concrete block to side blowout failure. The model
parameters values which are adopted are the following:
- tan mod will be taken as 0,5 as is normally done in code standards.
-  will be taken a 2/3 assuming a parabolic distribution of tensile stresses.
- Finally,  will be taken as 1.00, assuming a lateral diffusion angle within the resisting section also
equal to mod at failure. This value is consistent with the assumed diffusion angle in the
perpendicular direction.
With these values Eq. (7) simplifies to Eq. (8);
8 a1b1
NRd ,2  fctd 
3 a0 b0
1
a1b1
NRd ,2 8 fctd a1b1
 sd
'
 
a0 b0 3 a0 b0 ab
1 0 0
a1b1

Because the tensile strength is subjected in FprEN 1992-1-1:2023 to a sustained load coefficient (kct) which
is not to be applied here, as it would make the model overconservative, the tensile strength has, instead,
been expressed in terms of the square root of the compressive strength throughout the code (as, for instance
for shear resistance). Thus the expressions in FprEN 1992-1-1:2023 are expressed as in Eq. (9)

fck a1b1
NRd ,2  
c ab
1 0 0
a1b1
NRd ,2 f ab 1
 sd
'
  ck 1 1
a0 b0  c a0 b0 ab
1 0 0
a1b1

Besides the tension blow-out failure, two types of compression failures need to be checked, as contemplated
in all code standards regarding concentrated loads. If the loading plate is very small, failure can occur in
triaxial compression at the plate surface. If the plate is large with respect to the concrete block and the force
too high, a uniaxial compression failure can occur. These failure modes are expressed in Eq. (10).

 ab 
NRd ,1  min  1 1 ;7,0  a0 b0fcd 
 a0 b0 
N  ab a b
 sd
'
 Rd ,1  min  1 1 ;7,0  0 0 fcd
As a b
 0 0  As

The limit of 3,0fcd for triaxial compression currently in EN 1992-1-1:2004 was increased to 7,0 fcd in the initial
proposal, as shown in the previous equation. The reason for this is that the current value is extremely
conservative as shown by tests (see tests by Spieth [C08-6], or the Niyogi localized load on concrete cubes
[C08-7] data). Both show an increment in concrete strength up to 7 times the uniaxial concrete strength fcd
when specimens were subjected to high confinement. This can also be seen by applying concrete models for
triaxial stresses such as the model by Menetrey-Williams [C08-8]. The origin of this conservative limit can be
traced back to Leonhardt [C08-1], who, while working with the method of admissible stresses, proposed a
reduction factor of 3,0 for the admissible tri-axial concrete strength. This factor has not been accounted for
when translating the safety format from admissible stresses to the partial factor method. It should be now

A. Pérez Caldentey, FHECOR, F. Marchetto, UPM 08.03.2023

For citations, copying and referencing see Note on Front page! page 366
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.6 and I.8.6 Background to FprEN 1992-1-1:2023

time to account for it, the value of 7,0 probably still being conservative. Unfortunately, this value was again
reduced to the traditional value of 3,0 to achieve consensus withing SC2, with the allowance given in section
8.6 (“unless larger resistance can be justified based on refined analysis including tensile stresses due to load
or restraint, where relevant”). These considerations should be accounted for in the National Annexes.
The three failure criteria (tension, uniaxial compression and triaxial compression) are summarized in Fig. 3.
For very small plates, the dominant failure mode is triaxial compression, while for very large plates, with
respect to the concrete dimensions, the failure mode is uniaxial compression. For intermediate plate
dimensions, tension failure dominates.

a0/a1

Fig. 3 – Ultimate load represented against the relative size of the plate and dominating zones for each failure
criterion (adapted from ref [C08-4])
For uniaxial compression failure the actual block dimensions are taken since any reduction thereof
significantly reduces correlation to test data. Experimental observation by Niyogi [C08-7] confirms that
defining the effective unloaded area as a projection of the load plate tends to under-predict the measured
capacities of eccentric specimens.
Even though design codes recommend using a concrete resisting area having a similarity relationship to the
plate dimensions, the application of such criteria to the experimental data available would predict compression
failure in almost all cases whereas tension side blowout was the type of failure observed in all the tests carried
out by De Vries.
Another important point is that the model assumes uniform distribution of stresses on the plate and therefore
it should be ensured that the plate is rigid. This condition can be met by providing a plate depth greater than
half the distance between the bar edge and the plate edge.
The comparison of the model to experimental data is extensively reported in reference [C08-4] and will not
be repeated here.
Some amendments to account for cases where the 3-dimensional behaviour does not occur (cases where
the width of the plate is similar to the width of the concrete block or cases where the dimensions of the plate
are very different (i.e. long plates)), which were pointed out by the Swiss Mirror Group ([C08-9]) were
introduced during discussions.
To account for plate lengths that are similar in width to the dimension of the concrete block, the definition of
the effective dimensions of the concrete block (a1, and b1) are defined as shown in Eq. (11). Dimension a1 of
the contributing concrete area is taken as a which is the length of the load introduction block parallel to a0.
The perpendicular dimension, b1, is taken as the plate dimension in that direction, b0, plus the difference
between a1 and the plate dimension parallel to a1, a0, but not larger than the actual block dimension, b. In this
way if a0=a1 the behaviour is conservatively considered as unidirectional (see Fig. 4b) and when a0 is
significantly smaller than a1 3-dimensional effects are accounted for.

a1  a taken as the length of the load introduction block parallel to a0

b1  min  b0   a1  a0  ; b 

A. Pérez Caldentey, FHECOR, F. Marchetto, UPM 08.03.2023

For citations, copying and referencing see Note on Front page! page 367
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.6 and I.8.6 Background to FprEN 1992-1-1:2023

Fig. 4 Definition of the contributing concrete area

To account for the case of a long plate where one dimension is much longer than the other, which influences
the tensile capacity of the block, the formulation is split up into 2 cases (see Eq. (12)):
- Those in which the maximum plate dimension is less than 3 times the minimum plate dimension.
In this case 3-dimensional behaviour is assumed to occur.
- Those in which the maximum plate dimension is larger than three times the minimum plate
dimension. This this case the behaviour is assumed to be 2-dimensional, which is a good
approximation for the central length of the plate and is conservative for the behaviour occurring at
the ends of the plate.

 Ac1 1
A if max(a0 , b 0 )  3d 0
 c 0 1  Ac 0
f  Ac1
 Rd ,t  ck 
 c  d1 1
d if max(a0 , b 0 )>3d 0
d
0 1 0

 d1
d 0  min(a0 , b 0 )
d1  block dimension parallel to d 0
With these amendments, tests carried out at the ETH for which the formulation of draft D5 provided unsafe
results, turn into conservative estimations of the ultimate capacity as shown in Fig. 5.

Tests with long plates carried out at ETH

250

200

150
U,model [MPa]

100

50

0
0 50 100 150 200 250
u,exp [MPa]
Formulation in D5 Corrected Equations 1:1 slope

Fig. 5 Comparison of tests carried out at ETH with the formulation for partially loaded area of the D5 draft and
with the corrected model.

A. Pérez Caldentey, FHECOR, F. Marchetto, UPM 08.03.2023

For citations, copying and referencing see Note on Front page! page 368
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.6 and I.8.6 Background to FprEN 1992-1-1:2023

eb
b1

ea

a1

Fig. 6 Effective plate area for eccentrically loaded plates

For cases in which the applied load is not concentric with the plate, an effective plate area is defined as shown
in Fig. 6. The classical Meyerhoff approach [C08-10] is used for this.
FprEN 1992-1-1:2023 also defines in figure 8.33 (reproduced below in Fig. 7) how to determine the
contributing concrete area for cases where two plates of different dimensions are placed adjacent to each
other. The proposed criterion is derived from splitting the central spacing between the two plates in half and
defining the external limit of the contributing concrete area by symmetry with respect to the centre of the plate:

 a0, F1 a0, F2 
 a0, F s   
c2, F1  2 1
 2 2   a  s  a0, F1  a0, F2  s  a0, F1  a0 , F2
0 , F1
 2 2  2 2 2
 
 
 a0, F1 a0, F2 
 a0, F s    a0 , F1 a0, F2 a0, F2  a0, F1
 2  2 2 a
c2, F2 0, F 2  s    s
2

 2 2  2 2 2
 
 

A. Pérez Caldentey, FHECOR, F. Marchetto, UPM 08.03.2023

For citations, copying and referencing see Note on Front page! page 369
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
8.6 and I.8.6 Background to FprEN 1992-1-1:2023

Fig. 7 Definition of contributing concrete area for cases where two adjacent plates have different dimensions

References
[C08-1] F. Leonhartd. Vorlesunger über Massivbau. 1973. Springer Verlag, Berlin/Heidelberg.
[C08-2] Alejandro Pérez Caldentey, Francesco Marchetto, Hugo Corres Peiretti, Jesús Iglesias Villareal,
Plate-anchored reinforcement bars: A new simple and physical model for practical applications, In
Engineering Structures, Volume 52, 2013, Pages 168-178, ISSN 0141-0296,
https://doi.org/10.1016/j.engstruct.2013.02.019.
[C08-3] Marchetto, F., Pérez Caldentey, A. and Corres Peiretti, H. (2016), Structural performance of corner
joints subjected to a closing moment using mechanical anchorages: an experimental study. Structural
Concrete, 17: 987–1002. doi:10.1002/suco.201500222.
[C08-4] Marchetto, F. Use of headed reinforcement bars in construction. A theoretical approach to determine
the dimensions of anchorage plates and experimental tests on knee joints subjected to closing moment. PhD
Thesis. Technical University of Madrid, 2015.
[C08-5]DeVries, R.A., Anchorage of Headed Reinforcement in Concrete. 1996, The University of Texas at
Austin: Austin. p. 314.
[C08-6] Hans-Peter Spieth. Das Verhalten von Beton unter hoher örtlicher Pressung. Beton und
Stahlbetonbau, 56(11):257–263, November 1961.
[C08-7] S.K. Niyogi, Bearing Strength of Concrete - Geometric variations. Journal of structural division. July
1973.
[C08-8] P. Menetrey, K. J. William. Triaxial failure criterion for concrete and its generalization. ACI Structural
Journal, 1994, 92 (3), pp 311-318. 1995.
[C08-9] W. Kaufmann, T. M., Markić, T. and Morger, F. (2020). Comment on Section 8.6 Partially loaded
areas, addressing inconsistencies that would partly result in unsafe designs. CEN/TC 250/SC 2/WG 1 N 845,
16.
[C08-10] Meyerhof, G. G. (1963). Some Recent Research on the Bearing Capacity of Foundations. Canadian
Geotechnical Journal, Vol. 1, No. 1, 1963, pp. 16-26.

List of versions of this document

28.11.2017, APC – 1st draft (submitted to SC2/WG1, 30.11.2017)
19.02.2021, APC – 2nd draft (submitted to SC2/WG1, 28.02.2021)
08.01.2023, APC – 3rd draft – Adapter to final version of FprEN 1992-1-1:2022
08.02.2023, TJ Peer review and update of references to FprEN 1992-1-1:2023
08.03.2023, APC – Minor editorial revisions.

A. Pérez Caldentey, FHECOR, F. Marchetto, UPM 08.03.2023

For citations, copying and referencing see Note on Front page! page 370
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to clause 9.2.3

Refined control of cracking

Alejandro Pérez Caldentey, FHECOR, Technical

Roberto García, Viktor University of Madrid and
Gribniak, Arvydas Rimkus Vilnius Gediminas Technical
University

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 371
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to clause 9.2.3 (Refined control of cracking).

9.2.3 Reasons for change

Refined control of As it will be shown below, the formulation in EN 1992-1-1:2004 is not well calibrated for tension members
cracking (being over-conservative) and does not account for important variables such as the effect of casting position
or the increase in crack width in bending members due to curvature, whose effect is evident in the available
experimental data and important when assessing the crack width of members with large bar diameters.
Additionally, the current definition of the effective area needs to be amended for the following reasons:
1. The effective area around a bar must be limited. Its current definition leads to the absurdity that a
member with only one bar is fully effective in a huge area of concrete, while it is known that a bar can
only control cracking in its vicinity.
2. The (h-x)/3 limit to the effective height in the current definition can lead, when covers and ratios are
large, to the absurd result that the steel bars are outside of the effective area. Also, this limit can lead to
huge differences in the effective area of otherwise identical members subjected to tension and flexure,
which does not make any sense.
3. The dependence of hc,ef on just (h-d) does not perform well when there are multi-layer sections, hence
the need to introduce the spacing between bars for these cases (sy)
4. The current definition has no other basis than it has been used for a long time. But in fact, the value
“2.5” is just a calibration factor. The above considerations require a change of this value to improve
agreement with experimental results.
The proposed model solves the above issues, introduces the effect of important variables, which are currently
neglected and achieves a significantly smaller scatter, when compared with experimental results, which
means that the reliability of the model will be higher.

Enhancement of ease of use

From the point of view of ease of use, the effort needed to apply the proposed formulation is equivalent to
that of EN 1992-1-1:2004. One simplification is the elimination of the (h-x)/3 limit for the height of the effective
section. This simplifies matters. The consideration of the variable stresses within the effective area prior to
cracking, the consideration of casting position, and the consideration of the increase in crack width due to
curvature, while they do not simplify the model, lead to more accurate predictions which are better adapted
to specific design situations. This allows not to penalize favourable conditions in order to have a sufficiently
conservative model.

Proposed corrections to EN 1992-1-1:2004

Consideration of the variation of stresses in members subjected to flexure

Fig. 1 shows a comparison of the experimentally measured crack spacing against the prediction made using
the formulation of EN 1992-1-1:2004. The approximation for bending is reasonable as it is close to the 45º
line but the crack spacing for tension is significantly overpredicted, resulting in a slope of the correlation line
of 1.41.
The reason for this can be traced to the definition of factor k2. Factor k2 accounts for the variation of stresses
within the effective tension area prior to cracking, which affects the transfer length. In EN 1992-1-1:2004 [C09-
6], coefficient k2 is defined as follows (Eq. (1)):

k2 
 1   2 
(1)
21
where 1 and 2 are the greater and lesser tensile strains (2=0 if part of the section is compressed) in the
section. With this definition k2=1.0 for pure tension and 0.5 for pure flexure.
The rationale behind the above factor is that, in bending, the transfer length will be shorter because the
tension force (coloured area in Fig. 2 times the width of the section) that is needed to produce a new crack
will be half than that in tension because the strain of the least tensioned fibre is zero, therefore the transfer
length would be half. The current Eurocode 2 (EN 1992-1-1:2004) formulation, however, instead of providing
small crack spacings in flexure, is calibrated in such a way that the model results in notoriously exaggerated
crack spacings for tension members, as shown already in Fig. 1.
Note that the concept of an effective area is meant to account for the fact that the transfer length is a zone of
load introduction and therefore Navier’s hypothesis is not applicable. The consideration of an effective area
is meant as a way to account for this fact while still assuming the validity of Navier’s hypothesis. But there is
still a difference in the way the stresses are distributed in tension and bending, and this should be accounted
for.

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 372
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

EN 1992‐1‐1:2004
400
Flexure
y = 1.06x
350 R² = 0.94
Tension
y = 1.41x
300 R² = 0.95

sm, EN 1992‐1‐1:2004 [mm]

250

200

150

100

50

0
0 50 100 150 200 250 300 350 400
sm,exp [mm]
Flexure Tension 45º line Linear (Flexure) Linear (Tension)

Fig. 1 Comparison of crack spacing predicted by EN 1992-1-1:2004 and the actual measured values

Fig. 2 Effect of distribution of stresses prior to cracking on transfer length according to EN 1992-1-1:2004
The above reasoning, as shown in Fig. 2, is not sound, because what matters is not the stress gradient within
the full tensile zone itself but rather the tensile gradient within the effective area around the bar. In a bending
member of significant size, the effective area can represent only a small part of the cross-section and the
approximation of a tie, as done in MC 2010 can be reasonable. However, in small members, the model of EN
1992-1-1:2004 would be better. Fig. 3 demonstrates that, for tension, the mean stress, mean, in the effective
area will be fctm when the next crack occurs. However, in the case of flexure, the mean stress can be
determined as shown in Eq. (2),

1 1 f ct ,ef 
 mean 
2
 f ct ,ef   c,min,ef    f ct ,ef 
2  h  xg
 h  xg  hc,ef   

1   h  xg  hc ,ef   (2)
 f ct ,ef 1  
2 h  xg 


k fl

In the above equation xg is the depth of the neutral axis in the uncracked section. This expression results in
a value to 1.0 for pure tension (xg=) and a value of 0.5 in flexure if hc,ef is equal to (h-xg).
For a rectangular cross section and pure bending, the expression for kfl of Eq. (2) simplifies to Eq. (3):

 h 
1   h  xg  hc ,ef
k fl  1 
   1 1   2  h c , ef  1
   2  2hc ,ef  
 
2 h  xg  2 h  2 h 
    (3)
 2 
h  hc ,ef

h

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 373
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

Fig. 3 Consideration of the effective tensile area around a bar in flexural elements

Effect of casting position

It is a well-established fact (e.g., see [C09-6], [C09-24], and [C09-25]) that casting position affects the required
anchorage length of reinforcing bars. In ULS, the anchorage length has traditionally been increased by a
factor of 1.4 for bars in horizontal members which are close to the top surface. This modification has to do
with the appearance of voids under the top bars due to plastic settlement and bleeding, which reduces the
bond perimeter of the bar. Even though cracking has to do with bond (and a bond factor has been present in
codes – for instance factor k1 in EN 1992-1-1:2004 [C09-6] which accounts for the different bond properties
of ribbed and smooth bars) casting position has – to the knowledge of the authors – never been considered
in models dealing with crack spacing [C09-26].
However, tests carried out at the Universidad Politécnica de Madrid (UPM), show clear evidence that casting
position has a substantial effect on crack spacing. The flexural specimens 12-70-00 and 12-70-F tested in
2009 [C09-18] and 2017 [C09-8], respectively, with identical geometry (though slightly different concrete mix
proportions), showed substantial differences in crack spacing (see Fig. 4). Besides the concrete mix (which
is known not to have an important effect on cracking), the only difference was the casting position of the
tension reinforcement. In the beam tested in 2009, the tensile reinforcement was cast in ‘poor’ casting position
(top), while in the test carried out in 2017, it was cast in ‘good’ casting position (bottom). The beam tested in
2009 showed a mean crack spacing of 236 mm, while the element tested in 2017 showed a mean crack
spacing of only 162 mm.

Fig. 4 Cracking patterns of beam cast in ‘good’ casting position (top) and ‘poor’ casting position (bottom)

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 374
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

The effect of casting position has been confirmed by analysis of the cracking pattern of ties, where the face
cast in good casting position has a definite tendency to develop a cracking pattern with more closely spaced
cracks. Fig. 5 provides further illustration of this fact using results of tie specimen 16-20-T tested in 2017.

Fig. 5 Cracking pattern of a tie showing different crack spacing on different faces
Experimental evidence, reported in [C09-8], seems to show by comparison of the cracking patterns of two
side faces of the tie, one concreted in ‘poor’ casting position and one concreted in ‘good’ casting position,
that both cover and bar diameter influence the crack pattern. However as the experimental evidence is still
very limited, a very simple model to account for this effect was adopted to the revision of EC2, meant
essentially at recognizing casting position as an important factor in determining crack spacing, to be refined
in the future when sufficient experimental evidence is available.
the introduction of a coefficient to account for casting position is proposed, as follows:

1.2 for poor bond conditions

kb   (4)
0.9 for good bond conditions
A value of 0.9 is adopted for good bond conditions (instead of 1.00) foe aesthetical reasons because then the
calibration coefficient for the bond term coincides with the value that is derived from bond theory.

New definition of the effective concrete area

As discussed above, corrections are needed to the definition of the effective area concept. The introduction
of coefficient kfl makes unnecessary the limit of (h-x)/3 for the effective depth for elements subjected to flexure
which is introduced in the current definition in EN 1992-1-1:2004. It seems that this value was there to account
for the same effect, but the new factor is much more transparent and has physical meaning.
For an isolated bar, the proposed definition of the effective area is given in Fig. 6 and the following equation:
Ac ,eff ,bar  bc ,eff  hc ,eff
hc ,eff  min  ry  5 ;10 ;3.5ry   h  x
(5)

bc ,eff  min  rx  5 ;10 ;3.5rx 

Fig. 6 Effective tension area of concrete around an isolated bar, Ac,ef,bar

When individual tension areas of different bars overlap, the effective reinforcement ratio should be considered
for the group of bars, as shown in Fig. 7 and the following equation:
Ac ,eff , group  bc ,eff  hc ,eff
(6)
hc ,eff  min  ry  5 ;10 ;3.5ry   s y  h  x
bc ,eff  b

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 375
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

Note that a more precise definition would lead to a different height of the effective area where two rows of
reinforcement are present than where there is only one layer. This difference is waived in FprEN1992-1-
1:2023 to improve ease-of-use with a small sacrifice in precision.

Fig. 7 Effective tension area of concrete around a group of bars, Ac,ef,group (to be applied when effective tension
areas of adjoining bars overlap)
A limit on the effective area around a bar is given as a linear function of the bar diameter. This condition
accounts for the fact that a bar can only control cracks within its proximity. With the definition of the effective
area given in MC 2010 and EN 1992-1-1, a single bar placed in the middle of a large rectangle of concrete
would have an effective area equal to the area of concrete, and the value of the effective area would increase
indefinitely with the dimensions of the cross section. This does not make sense and a limit is therefore
necessary for consistency.

Effect of curvature on crack width

Regarding crack width, it has been well established that, in bending, the value of the crack opening increases
from the level of the reinforcement towards the most tensioned face (see for instance the tests reported in
reference [C09-18]). The increase in the crack opening is proportional to coefficient k1/r, defined, as follows:

hx
k1  (7)
r dx

where:
h is the section height
d is the effective depth, and
x is the depth of the neutral axis of the cracked section
This coefficient was included in the left-hand side comments to MC 2010, but not in the right-hand formulation
to avoid explicitly increasing crack width predictions. As an illustration, Fig. 8 shows two tests of the
experimental campaign described in reference [C09-8] with 32 mm cover (represented by circles) and 82 mm
cover (represented by squares). The black symbols show the crack width measured at the level of the bar,
while the white symbols show the crack measured at the most tensioned concrete fibre. The grey symbols
result from multiplying the values of crack width at the bar level by k1/r. As can be seen, the grey points land
almost on top of the white symbols (in some cases the white symbols are not visible anymore), thereby
showing that factor k1/r adequately addresses the issue of increase of crack opening due to curvature.

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 376
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

Increase of crack width with curvature ‐ Experimental results

Performance of correction factor (h‐x)/(d‐x)
0.70

0.60

Measured crack width [mm]

0.50

0.40

0.30

0.20

0.10

0.00
0 100 200 300 400 500 600
Theoretical steel stress ‐ State II [MPa]
16‐20‐F At bar level c=32 mm 16‐20‐F At most tensioned fibre of concrete c=32 mm
16‐20‐F Bar level * (h‐x)/(d‐x) 16‐70‐F At bar level c=82 mm
16‐70‐F AT most tensiones fibre of concrete c=82 mm 16‐70‐F Bar level * (h‐x)/(d‐x)

Fig. 8 Illustration of the performance of factor k1/r on two beam tests with different covers

0.70

0.60
wm, est=TS sm,exp s/Es [mm]

0.50
Flexure
sm,est = 0.58wm,exp
0.40 R² = 0.65

Tension
0.30 sm,est = 0.76wm,exp
R² = 0.73

0.20

0.10

0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
wm,exp [mm]

Flexure Tension Linear (Flexure) Linear (Tension)

Fig. 9 Comparison between measured crack width and estimated crack width without accounting for factor
k1/r.
By comparing the measured crack width against the estimated crack width (calculated by multiplying the
experimentally observed crack spacing by the strain in the reinforcement at the crack and by a tension
stiffening factor (TS), according to EN 1992-1-1:2004), see Fig. 9, two effects can be observed:
– The effect of shrinkage happening before the test. Shrinkage compresses the reinforcement before
the tests. This strain will then be released when cracking occurs and, as a result, the crack width
due to the load is overestimated if no correction is introduced to account for the pre-existing
compressive strain in steel. This explains why the estimated cracks widths fall below the measured
values
– The effect of curvature, since the underestimation of the crack width is much larger for bending
tests than for tension tests.

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 377
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

0.70

0.60

wm, est=k1/rTS sm,exp s/Es [mm]

0.50 Flexure
sm,est = 0.73wm,exp
R² = 0.75
0.40
Tension
sm,est = 0.76wm,exp
0.30 R² = 0.73

0.20

0.10

0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
wm,exp [mm]

Flexure Tension Linear (Flexure) Linear (Tension)

Fig. 10 Comparison between measured crack width and estimated crack width accounting for factor k1/r.
If coefficient k1/r is used to correct the estimated crack width, the skew between tension tests and flexure tests
shown in Fig. 9 is corrected, as is shown in Fig. 10. Even so, the slope of the trendline is still not satisfactory.
As mentioned above, the fact that the slope is still smaller than 1, is attributed to the shrinkage of concrete
occurring before the tests. These effects can, unfortunately, not be adequately corrected since the shrinkage
strains developing before the tests are not generally reported in the literature.

Calibration and comparison to experimental data

The formulation is calibrated to determine the coefficients to be applied to the cover term (kc) and the bond
term (k/). The expressions for the mean crack spacing for flexure and tension can be written as follows
f ctm  
srm,cal  kc  c  k /   k fl  kb    1.5c  k fl  kb (8)
 bms  p ,eff 7.2  p ,eff
where:
kc is an empirical parameter to take account of the influence of the concrete cover; as a simplification, kc
= 1.5 can be assumed;
c is the maximum concrete cover. The maximum value has been adopted because recent research [C09-
27] confirms that, when vertical and horizontal covers are different, crack spacing is much better
correlated to the maximum cover than to minimum cover. When the effective area concept applies to
a single bar located in the perimeter of the section the maximum cover of this bar applies. When the
effective area applies to a group of bars, the most unfavourable value of cover of the bars located in
the perimeter of the section should be considered;
k/ is an empirical parameter to account for the influence of bond; as a simplification, k/ =0.25 can be
assumed.
The ratio fctm/bms can be assumed, as a simplification, to be equal to a constant factor of 1.8 (as proposed by
MC 2010 [C09-6])
The design crack width is obtained from the following equation:

wk , cal   w k1/ r srm , cal ( sm   cm ) (9)

where:
w is a factor to obtain a design value of the crack width from the mean value, which can be taken equal
to 1.7. This value is proposed to be a nationally determined parameter (NDP) since it has to do with
the risk the owner is willing to take of finding surface cracks which are larger than the established limit;
εsm is the average steel strain over the length srm,cal;
εcm is the average concrete strain over the length srm,cal;

1  f 
 sm   cm    s  kt ctm 1   e  p , eff   (10)
Es   p , eff 
where:

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 378
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

Es is the modulus of elasticity of steel;

s is the stress of steel at the crack;
kt is an empirical coefficient to assess the mean strain over the transfer length, equal to 0.6 for short-term
analysis and equal to 0.4 for long-term analysis or repeated loading;
fctm is the mean tensile strength of concrete;
e is the modular ratio=Es/Ec;

FprEN 1992‐1‐1:2023 ‐ Calibration of proposal

400

350

Flexure
300 sm,FprEN 1992 = 0.98sm,exp
sm, Fpr EN 1992‐1‐1:2023 [mm]

R² = 0.95
250 Tension
sm,FprEN1992 = 0.99sm,exp
R² = 0.97
200

150

100

50

0
0 50 100 150 200 250 300 350 400
sm, exp [mm]
Flexure Tension 45º line Linear (Flexure) Linear (Tension)

Fig. 11 Comparison between predicted and measured crack spacing

Fig. 11 shows the comparison between predicted values, srm,cal FprEN1992-1-1:2023, and experimental values of
crack spacing, sm,exp. The correlation lines show a high coefficient of determination of 0.95 in flexure and 0.97
in tension and a slope close to 1.00. By comparison of Fig. 1 and Fig. 11, it can be observed that scatter is
reduced as it will now be demonstrated.
Table 2 shows the statistical parameters referred to the 73 tests used for the calibration for crack spacing
(see Annex A of reference [C09-10]). The table includes the mean squared error, the minimum value of the
ratio of model prediction and experimental values (min), its maximum value (max), its mean value (), its
standard deviation () and its coefficient of variation (COV).
It can be seen that the proposed corrections improve the prediction quality in statistical terms, for all the
considered statistical parameters, both in flexure and in tension.

Table 2 Statistical analysis of crack spacing

EN 1992-1-1:2004 FprEN 1992-1-1:2023

Flexure Tension Flexure Tension

nº of tests 37 36 37 36

2/N) 35.25 83.01 29.01 30.97

min(sm,model/sm,exp)= 0.60 0.68 0.53 0.67

Model/Measured

max(sm,model/sm,exp)= 1.74 2.04 1.47 1.34

 1.08 1.43 1.00 1.01

 0.27 0.30 0.24 0.158

COV = 24.69% 21.25% 23.57% 15.62%

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 379
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

Verification of the calibration using an alternative test data set

The model was calibrated with the same data set (calibration set) as was used for the original calibration of
MC 2010, with the addition of several experimental series (data of Annex A of reference [C09-10]). The
robustness of this calibration was verified using a different set of data (test set – see Annex B of reference
[C09-10]). The independent database includes a total of 144 specimens tested in flexure. This database
consists of the following tests:
– Clark (54 specimens) [C09-12]
– Rehm&Rüsch (30 specimens) [C09-17]
– Gribniak (6 specimens) [C09-28]
– Gilbert & Nejadi (12 specimens) [C09-30]
– Calderón (14 specimens) cast in poor casting position [C09-31]
– Wu (4 specimens) – 2 Tests with excessive side cover were discarded [C09-32]
– Frosch (2 specimens) – Other tests with excessive side cover were discarded – Poor casting
position [C09-33]
– Beeby (16 specimens) – Tests with mild reinforcement were discarded [C09-34]
– Klakauskas (6 specimens) [C09-35]
The details of the specimens are available in Annex B of reference [C09-10].
Fig. 12 demonstrates the performance of the proposed model on the independent testing set. Even though
there is scatter, the model proves to be well-calibrated. Table 3 shows the corresponding statistical
parameters. It can be observed that the application of the proposed modification reduces the coefficient of
variation from 24.6% to 18.7% with respect to EN 1992-1-1:2004. This result corroborates the need for the
proposed corrections.

FprEN 1992‐1‐1: 2023 ‐ Verification of Calibration with Test Data Set

250
Flexure
srm,FprEN 1992 = 1.00sm,exp
R² = 0.96

200
srm, FprEN 1992:2023 [mm]

150

100

50

0
0 50 100 150 200 250
sm, exp [mm]
Test Data Set ‐ Flexure

Fig. 12 Comparison between predicted and measured crack spacing for the independent test series (Annex
B)

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 380
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

Table 3 Statistical analysis of crack spacing for the independent data set

Test data set (flexure)

EN 1992-1-1:2004 FprEN 1992-1-1:2023

nº of tests 144 144

 /N)
2 26.30 23.55

min(sm,model/sm,exp)= 0.51 0.53

Model/Measured
max(sm,model/sm,exp)= 1.63 1.76

 0.99 1.04

 0.24 0.20

COV = 24.56% 18.73%

Summary
In this background document, reasons have been given for why the cracking model of EN 1992-1-1:2004 had
to be revised, the most important being the uneven behaviour regarding crack width in tension and flexure.
The main proposals to improve the cracking model are the following:
– A new coefficient has been proposed to account for the effects of non-uniform stress distribution
(kfl). The introduction of kfl by replacing k2 with a better value, does away with the gross
overestimation of crack spacing in tension and allows to delete the (h-x)/3 limit to the effective
height currently implemented in the EN 1992-1-1:2004 model.
– Recent experimental evidence demonstrates that casting position has a strong influence on crack
spacing, increasing it when tension bars are concreted in ‘poor’ casting position. This effect has
been accounted for by introducing factor kb.
– The definition of the effective area has been reformulated to eliminate some of the inconsistencies
of the current model.
– The calibration, made using the database used in the original calibration of MC 2010, has been
tested against an independent database and has proven to be robust.
– It has been observed that the model for the crack width, which derives the mean crack width from
the product of the mean crack spacing and the mean strain difference of steel and concrete
underestimates the measured crack width because of the generally unreported effect of shrinkage
of concrete occurring before testing. It is therefore not possible to make a meaningful comparison
between models and experimental databases in terms of crack width.

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 381
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

[C09-5] Corres Peiretti, H., Pérez Caldentey, A., Petschke, T. PrEN Chapter 7 – Serviceability Limit State.
References
Cracking. Supporting Document. May 2003.
[C09-6] fib Model Code for Concrete Structures 2010 (2013). Fédération internationale du béton (fib),
Lausanne, Switzerland. ISBN 978-3-433-03061-5
[C09-7] CEN-TC250 (2004), EN 1992-1-1. Eurocode 2. Design of concrete structures. General rules for
buildings.
[C09-8] García. R. Pérez Caldentey, A. (2018). Cracking of Reinforced concrete: Tension vs. Flexure: Report
on cracking tests of large ties. Technical University of Madrid (UPM).
[C09-9] fib Bulletin 92 (2019), Serviceability Limit State of Concrete Structures. Lausanne, Switzerland, ISSN:
1562-3610, pp 292. doi.org/10.35789/fib.BULL.0092.
[C09-10] Pérez Caldentey, A, García, R, Gribniak, V, Rimkus, A. Tension versus flexure: Reasons to modify
the formulation of MC 2010 for cracking. Structural Concrete. 2020; 21: 2101– 2123.
https://doi.org/10.1002/suco.202000279
[C09-11] Broms, Bengt B., (1964) Crack Width and Crack Spacing in Reinforced Concrete Members, JL62-
67 Part II.
[C09-12] Clark, A.P. (1956), Cracking in Reinforced Concrete Flexural Members. ACI Journal, Proc. vol. 27,
No. 8, Apr. 1956, pp.851–862.
[C09-13] CUR (1994) Dutch Centre for Civil Engineering, Research and Codes Report no. 37 (Out
of print).
[C09-14] Farra, B., Jaccoud, J.P., (1993), Influence du Béton et de l’armature sur la fissuration des structures
en Beton. Rapport des Essais de Tirants sous déformation imposée de courte durée. EPFL, Lausanne.
Département de Génie Civil, École Polytechnique Fédérale de Lausanne, pub. No. 140.
[C09-15] Farra, B., Jaccoud, J.P. (1996) Rapport des Essais de Tirants sous déformation imposée de courte
durée.
[C09-16] Hognestad (1962). Journal of PCI Research and Development Laboratories. 1962.
[C09-17] Rüsch, E.H., Rhem G. (1963-64), Versuche mit BetonformStählen. (1963), pt. II (1963), pt. III (1964).
Deutscher Ausschuss für Stahlbeton, No. 140.
[C09-18] Pérez Caldentey, A., Corres Peiretti, H., Peset Irribaren, J. and Giraldo Soto, A. (2013), Cracking of
RC members revisited: influence of cover, φ/ρs,ef and stirrup spacing – an experimental and theoretical study.
Structural Concrete 01/2013. Nr. 16, pp. 69-78.
[C09-19] Rimkus A. Effects of Bar Reinforcement Arrangement on Deformations and Cracking of Concrete
Elements. Doctoral Dissertation. Vilnius: Technika, 2017. 143 p
[C09-20] Jakubovskis, R.; Kaklauskas, G.; Gribniak, V.; Weber, A.; Juknys, M. 2014. Serviceability analysis
of concrete beams with different arrangement of GFRP bars in the tensile zone. ASCE Journal of Composites
for Construction, 18(5): 04014005-1–10. DOI: 10.1061/(ASCE)CC.1943-5614.0000465.
[C09-21] Krips, M. (1984), Rissbreitenbeschränkung im Stahlbeton und SpannBeton. 1984. Doctoral thesis
[C09-22] Viktor Gribniak, Alejandro Pérez Caldentey, Gintaris Kaklauskas, Arvydas Rimkus, Aleksandr
Sokolov (2016) Effect of arrangement of tensile reinforcement on flexural stiffness and cracking, Engineering
Structures, Volume 124, 2016, Pages 418-428, ISSN 0141-0296.
[C09-23] Rimkus, A.; Gribniak, V. 2017. Experimental Investigation of cracking and deformations of concrete
ties reinforced with multiple bars. Construction and Building Materials, 148: 49–61. DOI:
10.1016/j.conbuildmat.2017.05.029.
[C09-24] Steyl L, (2016) Plastic cracking of concrete and the effect of depth. PhD thesis. Stellenbosch
University.
[C09-25] Combrinck, R., Steyl, L., Boshoff, P. (2018), Interaction between settlement and shrinkage cracking
in plastic concrete. Construction and Building Materials 185:1-1
[C09-26] Borosnyói, A., Balázs, G. (2005), Models for flexural cracking in concrete: the state of the art.
Structural Concrete Vol. 6, Issue 2. pp. 1464-4177. Thomas Telford, fib.
[C09-27] Annette Beldholm Rasmussen. (2019). Modelling of reinforced concrete in the serviceability limit
state : A study of cracking, stiffness, and deflection in flexural members. PhD Thesis. Aarhus University.
[C09-28] Pérez Caldentey, A., Jones. T., Goodchild, C. Consideration of shrinkage strain in cracking
calculations. Background Document to Clause 9.2.4 of prEN 1992-1-1:2020. CEN TC 250/SC2/WG
[C09-29] Gribniak, V.; Jakubovskis, R.; Rimkus, A.; Ng, P.-L.; Hui, D. 2018. Experimental and numerical
analysis of strain gradient in tensile concrete prisms reinforced with multiple bars. Construction and Building
Materials, 187: 572–583. DOI: 10.1016/j.conbuildmat.2018.07.152.
[C09-30] R. I. Gilbert, S. Nejadi. An experimental study of flexural cracking in reinforced concrete members
under short term loads. Report. University of South Wales. 2015.

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 382
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 Background to FprEN 1992-1-1:2023

[C09-31] Calderón Bello, E. (2008). Estudio experimental de la fisuración en piezas de hormigón armado
sometidas a flexión pura. PhD Thesis. Technical University of Madrid.
[C09-32] Wu, M.H.Q (2010). Tension stiffening in reinforced concrete – instantaneous and time-dependent
behaviour. PhD Thesis. The University of Southern Wales.
[C09-33] Frosch, R. J., Blackmann D.T, Radabaugh, R.D. (2003). Investigation of bridge deck cracking in
various bridge superstructure systems. Report No. FHWA/IN/JTRP-2002/25. Purdue University West
Lafayette, Indiana 47907. February 2003.
[C09-34] Base, C.D., Beeby, A.W., Taylor, P.J. (1966). An investigation of the crack control characteristics of
various types of bar in reinforced concrete beams. Research report 18. Cement and Concrete Association.
[C09-35] Kaklauskas, G. (2019) Personal communications (unpublished matter).

List of versions of this document

28.02.2021, APC 1st draft
24.03.2022, APC 2nd draft – Correction of erratum in title of Table 3 (EN 1992-1-1:2004 for MC 2010)
31.12.2022 APC 3rd draft – Adapted to final version of FprEN 1992-2-2:2022

Pérez, García, Gribniak and Rimkus 31.12.2022

For citations, copying and referencing see Note on Front page! page 383
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to clause 9.2.3(4)

Consideration of shrinkage strain in

cracking calculations

Alejandro Pérez Caldentey FHECOR, Spain

Tony Jones The Concrete Centre, UK
Charles Goodchild The Concrete Centre, UK

03.01.2023

For citations, copying and referencing see Note on Front page! page 384
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

Introduction
This document provides background to clause 9.2.3 Refined control of cracking and addresses under which
circumstances shrinkage strain increases crack width. The document centres on a tie and evaluates the
problem considering increasing degrees of complexity, as follows:
– Tie subjected to shrinkage but free to move
o linear elastic problem: the tie does not crack due to redistribution of stresses
o cracked analysis
– Tie subjected to shrinkage but with movement at the ends fully restrained. An external axial force
develops due to shrinkage
o linear elastic analysis
o cracked analysis
As a result of these analyses, it will be concluded that shrinkage would tend to increase the mean strain of
steel with respect to concrete by the shrinkage strain reduced by a tension stiffening factor. However, further
analysis will reveal that, for the crack formation stage, the increase in crack width will be very small due to
the reduction in the cracking force produced by the tensile stresses generated by shrinkage in the uncracked
concrete. Furthermore, it will be concluded that, for stabilised cracking, the development of shrinkage will
result in a reduction of the transfer length, and therefore, in the formation of a new crack between the most
widely spaced ones. For these reasons, FprEN 1992-1-1:2023 does not require the consideration of
shrinkage strains in crack width calculations.

9.2.3 Reasons for change

Refined control of The proposal for ignoring the shrinkage strain for crack width calculation does not constitute a change as
cracking compared to EN 1992-1-1:2004 (CEN, 2004). However, whether or not shrinkage should be considered has
been a matter of debate, especially in the light of its inclusion in MC 1990 (CEB-FIP, 1993) and MC 2010
(fib, 2013). This background paper is aimed at clearly developing the technical arguments that have resulted
in the absence of the shrinkage effect in the formulation for crack width calculation in FprEN 1992-1-1:2023.

Enhancement of ease of use

The absence of shrinkage is not only technically and physically sound, but also improves ease-of-use by
simplifying the model.

Tie subjected to shrinkage but free to move – linear analysis

Consider the structure of Fig. 1. The concrete is subjected to an imposed shrinkage strain cs. The strain that
develops in the element, , however, will be smaller than the shrinkage strain. The reason for this is that
the steel will partially constrain the concrete which will not be able to move freely due to this internal constraint.
As a result, tensile strains will develop in concrete. The problem can easily be analysed writing equilibrium
and compatibility equations, assuming a linear elastic behaviour of materials and accounting for creep.

L


Fig. 1 – Simply supported tie free to move longitudinally

Eq. (1) shows the linear elastic constitutive laws assumed for steel and concrete. Effects of creep relaxation
are accounted for by using the age-adjusted effective modulus method (Trost, 1991). Here  is the ageing
coefficient (which is taken as 0.8 in FprEN 1992-1-1:2023).

 s  Es  s
 c  c Ec
 c      cs   c       (1)
Ec Ec 1    c cs
Compatibility conditions dictate that the strain variation in reinforcement, s, should be equal to the strain
variation in concrete, c, and equal to the strain variation of the reinforced concrete section,  (See Eq.(2))

 s   c   (2)

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 385
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

Finally, equilibrium implies that the variation in axial force, N, due to the development of shrinkage should
be zero, since the element is free to move and, therefore, the reaction will be zero. The tension force that will
develop in concrete due to the fact that concrete will shorten less than the shrinkage strain, must be balanced
by an equal compressive force in steel (see Eq. (3))

N  0  Ac  c  As  s (3)

By introducing the conditions of compatibility into the constitutive laws and entering these into the equilibrium
equation, the strain of the element can be determined. Then, by introducing this strain into the constritutive
laws, the stresses in the materials can be calculated (see Eq.(4)). This equation will result in tension in
concrete since ||<|cs |.

 s  Es 
Ec
 c      cs 
1   
N  Ac  c  As  s  0 
 
Ec   Es 1     
(4)
  Ac  As    cs Ac   0
1      Ec


 
 Ah , 
A  cs  cs
    cs c  
Ah , E 1    1   e , 
1 s 
Ec

Therefore, if the shrinkage strain is high enough so that the stress in concrete exceeds fct,eff, the section
would crack. The next paragraph develops the analysis for this case.

Tie subjected to shrinkage but free to move – cracked analysis

Consider the structure of Fig. 2, in which, now, due to the internal restraint, concrete has cracked. For normal
values of shrinkage strain the element will be in the crack formation stage. There will be few cracks, spaced
relatively far apart. Far away from the cracks there will be strain compatibility and chunks of the tie will behave
linearly as outlined in the previous paragraph. At the crack however, the shrinkage strain will be liberated,
and steel will expand, since it will not be restrained by concrete and concrete will shorten, since the tensile
axial force to which it was subjected by internal restraint will disappear.

Fig. 2 – Simply supported tie free to move longitudinally with crack in concrete
Therefore, at the crack, the expressions of Eq. (5) will govern behaviour. Immediately adjacent to the crack
there will be a difference between the strain in the steel and the strain in concrete equal to –cs (tension)

s  0 
   s   c   cs
 c   cs 
(5)
s  0
c  0
One transfer length away from the crack, compatibility will be reestablished. Tension stiffening effects can be
accounted for as follows:

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 386
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

  cs 
 sm   s  kt   s   sE   0  kt  0  
 1   e,  s 
  cs  (6)
 cm   cs  kt   cs  
 1   e,  s 
 sm   cm   cs 1  kt 
where:
sE is the strain in the reinforcement at the section where compatibility between concrete and steel is re-
established.
If kt is taken as 0,4, as in FprEN 1992-1-1:2023 and EN 1992-1-1:2004, only 60% of the shrinkage strain will
contribute to increase the relative mean strain of steel with respect to concrete.

Tie subjected to shrinkage, fully restrained – linear analysis

Consider now a fully restrained tie subjected to shrinkage with linear elastic behaviour (see Fig. 3). In this
case, there will be no movement, so the strain of the tie, which is equal to the strain in concrete and also
equal to the strain in the reinforcement due to compatibility conditions, must be zero.

Fig. 3 Embedded tie: fully restrained. Linear behaviour

Therefore, since there is shrinkage, an external axial tension force must appear to compensate for the
shrinkage strain, as shown in Eq. (7). In a linear elastic analysis, this force will be proportional to the shrinkage
strain and to the age-adjusted axial stiffness of the concrete section.
Since the strain of the reinforcement is nil, the stress in the reinforcement will also be nil. Concrete, however,
will be tensioned because the shrinkage strain will not be free to develop, so cracking is likely to occur.

 cs N
 s   c     0
E 1    Ec  Es 1    
1 s  Ac  As 
Ec
 1     Ec

Ac Ah ,
 cs
Ah ,

Ec Ac (7)
 N   cs
1   
 s  E s  s  0
Ec Ec
 c   0   cs       
1    1    cs
The next section deals with analysis after concrete cracks.

Tie subjected to shrinkage, fully restrained – cracked analysis

When the restrained tie cracks (see Fig. 4), the axial force will be reduced because the stiffness will be
reduced, and because forces due to imposed strains are proportional to the stiffness. If the tie is in the crack
formation stage, which is the case for normal imposed deformations (temperature and shrinkage), the axial
force will be, at most, equal to the force that cracks the concrete (see Fig. 5 and references (Scott, 2020) and
(Farra, 1996)).

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 387
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

Fig. 4 Embedded tie: fully restrained. Cracked behaviour.

At the crack, the reinforcement will take the full cracking force, so the strain will be equal (at most) to the
cracking force divided by the axial stiffness of the reinforcement. The strain of concrete will be equal to the
free shrinkage strain (See Eq. (8)).

N cr 
 s   N
Es As    s   c  cr   cs
Es As
 c   cs 
(8)
N
 s  cr
Es As
 c  0

Fig. 5 Behaviour of a tie subjected to an imposed strain

The effect of tension stiffening for this case can be evaluated as done before, by adapting Eq. (6) as follows
(Eq. (9)). This equation is in fact general and also valid for cases when restraint is not rigid. For a fully
restrained member, the condition that the mean steel strain be nil, need not apply between every two cracks
but only at the structural level along the full length of the element. This condition would allow to determine the
value of the shrinkage strain that would produce a new crack. This value would increase with the reduction
of the stiffness due to the appearance of new cracks.

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 388
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

  
  
N N  cs N cr  
 sm   s  kt   s   sE   cr  kt  cr   
Es As  Es As  1   e ,  s Ec Ah ,  



 1     
  
   (9)
 cs N cr  
 cm   cs  kt   cs   
  1   e ,  s Ec Ah ,  



 1     
 N cr 
 sm   cm     cs  1  kt 
 Es As 
Again, tension stiffening applies to shrinkage in the same way as it applies to loads. It is stressed that for
axially constrained members, shrinkage will produce both an axial force equal to the cracking force, thereby
contributing to the value of tensile strain in reinforcement at the crack (s) and a compressive strain in
concrete, thereby increasing by these two effects the relative steel strain with respect to concrete at the crack.
The analysis, however, is not yet concluded, since the cracking force in the tie will in fact not be constant as
shown in Fig. 5, but will show a tendency to decrease. This is because shrinkage induces tensile stresses in
concrete between cracks and this makes it easier to produce new cracks in the crack formation stage. The
external force will, therefore become smaller with time, as shown in Fig. 6.

N Crack formation Stabilized cracking stage

stage

Ncr


Fig. 6 Behaviour of a tie subjected to an imposed strain accounting for redistribution of forces due to shrinkage
In order to try to quantify this effect four fully embedded ties representing elements with high and low
reinforcement ratios and large and small heights have been analysed using step-by-step time-dependent
analysis, using the software described in (Pérez Caldentey, 1996). Table 1 shows the main characteristics
of the sections that were considered. All the sections were assumed to be cast with a C30/37 concrete. To
determine the shrinkage strain, 50% relative humidity was assumed.

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 389
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

Table 1. Characteristics of the sections analysed

Section 1 2 3 4
width [mm] 1000 1000 1000 1000
height [mm] 250 250 1000 1000
Reinf. 512 1025 10 1032
 0.28% 2.46% 0.21% 0.85%
hc,ef [mm] 110 167 110 175
 cs (t=100y)[mm/m]= 0.622 0.622 0.575 0.575
Fig. 7 shows the evolution of the force needed to produce a new crack in the tie as function of time. The
cracking force increases in the short term due to the increase in the concrete strength. However, this tendency
is reversed after a while by the tensile stresses generated by the development of shrinkage. The reduction
on the cracking force is much more pronounced for elements with higher reinforcement ratios than for
elements with small reinforcement rartios. This is logical since stresses in concrete develop because
reinforcement does not allow the free development of shrinkage. The larger the reinforcement ratio the more
restrained the shrinkage strain and therefore the larger the tensile stresses in concrete needed to fulfil
compatibility conditions.

Crack Formation Stage ‐ Force needed to produce a new crack

1 000

900

800

700

600
Fcr [kN]

500

400

300
28 days
200

100

0
1 10 100 1000 10000
Time [days]
h=25 cm ‐ 10phi12 h=25 cm ‐ 10phi25 h=100 cm ‐ 10phi16 h=100 cm ‐ 10phi32

Fig. 7 Evolution of the cracking force with time

Fig. 8 shows as continuous or discontinuous lines the crack width at different moments accounting both for
the reduction in the cracking force as shown in Fig. 7 as well as for the effect of shrinkage as shown in Eq.
(9). The crack width increases with time because the effect of shrinkage represented in Eq. (9) offsets the
reduction in the cracking force. In the same figure, the calculation ignoring the effect of shrinkage is given as
black dots placed at the end of the calculation time. It is evident that ignoring shrinkage leads to non-
conservative results for the crack formation stage.

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 390
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

Crack Formation Stage ‐ Evolution of crack width with time

0.50

0.45

0.40

0.35

0.30

wk,cal [MPa] 0.25

Results if shrikage is ignored
0.20

0.15

0.10

0.05

0.00
1 10 100 1000 10000
Time [days]
h=25 cm ‐ 10phi12 h=25 cm ‐ 10phi25 h=100 cm ‐ 10phi16 h=100 cm ‐ 10phi32 No shrinkage

Fig. 8 Crack width as a function of time, accounting for and ignoring shrinkage strain
Nonetheless, as shown in Table 2 and Fig. 9 the error, is rather small, no larger than 18%, with the error
being larger for the lower reinforcement ratios as was to be expected. It was therefore agreed within SC2/WG1
to neglect this effect given the large uncertainties surrounding crack width calculations.

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 391
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

Table 2. Comparison of crack width calculations performed considering and neglecting the effect of

shrinkage in the crack formation stage.

wk,ca l ,NoSH wk,ca l ,SH

Section h [m] Reinf. ratio
[mm] [mm]
1 0.250 1012 0.376 0.443 1.18
2 0.250 1025 0.105 0.115 1.09
3 1.000 10 0.260 0.304 1.17
4 1.000 10 0.083 0.093 1.12

Crack formation stage ‐ effect of shrinkage

0.500

0.450

0.400

0.350

0.300
wk,cal [mm]

0.250

0.200

0.150

0.100

0.050

0.000
25cm‐10phi12 25cm‐10phi25 100cm‐10phi16 100cm‐10phi32
Section
Shrinkage ignored Shrinkage considered

Fig. 9 Comparison of calculated crack width accounting for and neglecting shrinkage
If stabilised cracking is reached, then the effect of shrinkage can be more easily neglected. As it will be
demonstrated, in such a case, shrinkage would end up by producing an intermediate crack, whose effect
would more than compensate for the effect of shrinkage on the largest crack width by reducing the crack
spacing to half.

Fig. 10 Equilibrium of the bar between the section at the crack and the section at the end of the transfer length
Fig. 10 shows the equilibrium of a reinforcing bar belonging to a tie between the section of a crack and the
section at the end of the transfer length (ls) where a new crack would form. The equilibrium condition of the
bar leads to the expression of the transfer length given in Eq. (10), where bm is the mean bond strength. bm is
used for simplicity as a “mean value” of time dependent bond strengths over the length between the crack
and the end of the transfer length.

 s   SE (t )
ls (t )   (10)
4 bm (t )
Fig. 11 shows the equilibrium between the section at the crack where all the force is resisted by the
reinforcement only and the section at the end of the transfer length where part of the force is resisted by

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 392
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

concrete in tension. The tensile capacity of concrete (fctm×Ac,eff) is reduced by the tension induced between
cracks by shrinkage (c×Ac,eff). From this equilibrium condition it is possible to obtain the variation of stress
in the reinforcement bar (s2 -sE), as shown in Eq. (11), and by introducing this expression in Eq. (10), the
transfer length can be obtained.

Fig. 11 Sectional equilibrium between the section at the crack and the section at the end of the transfer length,
accounting for the effect of shrinkage

Ac ,eff
 s   SE (t )   f ctm (t )   c 
As
 Ec  1   Ac ,eff
 s   SE (t )   f ctm (t )   cs 1  
 1    1      A
  e ,  ef   s
(11)
 Ec 
   SE (t ) 
1 f (t )  (t )  1  
 ls (t )  s    ctm  bm  cs 1   
4 bm (t ) 4   bm (t ) 1    1  
   eff
   e ,  eff

 1.8 
Eq. (11) shows that the transfer length is reduced by the development of shrinkage, in spite of the increase
in tensile strength, because this increase is compensated by an increase in the bond strength. This remark
leads to the conclusion, that as shrinkage develops, new cracks will form in the specimen, thereby reducing
the maximum crack width.

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 393
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.2.3 (4) Background to FprEN 1992-1-1:2023

Conclusions
As a result of the above considerations, the following conclusions can be drawn:
– If the effects of shrinkage on restraining loads and its effect on the transfer length are neglected, it
is found that shrinkage increases the mean relative strain of steel with respect to concrete. This
increase can be mitigated by accounting for tension stiffening effects, applied to the shrinkage
strain.
– In the crack formation stage, the above effect is partially compensated by the reduction with time
of the force needed to crack the section due to the development of tensile stresses in the uncracked
concrete due to shrinkage. This means that subsequent cracks will form for a lower value of the
restraining axial force and will result in smaller stresses in the reinforcement. Although this effect
does not fully compensate for the increase in crack width due to shrinkage, the errors in the crack
width estimate are below 20%. Given the large uncertainties around crack width calculations it is
considered an acceptable simplification to neglect altogether the effect of shrinkage for this case.
– For the stabilised cracking stage, shrinkage leads to reduction of the transfer length. This means
that new cracks will form between the cracks that are most widely spaced resulting in the reduction
of the crack width at that location by a factor of 2. This justifies neglecting the effect of shrinkage in
the stabilised cracking stage.
– The above considerations allow to fully neglect the effect of cracking for crack width calculations.

References

Bibliography
[1] CEB-FIP. (1993). Model Code 1990. Design Code. Lausanne: Thomas Telford.
[2] CEN. (2004). EN 1992-1-1. Eurocode 2 Design of concrete structures - Part 1-1: General rules and rules
for buildings . Brussels: Comité Européen de Normalisation.
[3] Farra, J. J. (1996). Rapport des Essais de Tirants sous déformation imposée de courte durée. Lausanne:
EPFL.
[4] fib. (2013). Model Code for Concrete Structures 2010. Lausanne: Fédération Internationale du Béton.
[5] Pérez Caldentey, A. (1996). Comportamiento en servicio del hormigón estructural (PhD Thesis - Technical
University of Madrid). Madrid: ache.
[6] Scott, R. J. (2020). Influence of shrinkage strains on the cracking of reinforced concrete beams. The
Structural Engineer: journal of the Institution of Structural Engineer, págs. 25-29.
[7] Trost, H. (1991). Creep, Relaxation and Shrinkage of Structural Concrete. . lABSE Colloquium Structural
Concrete. Stuttgart: 1991.

List of versions of this document

09.09.2019, APC. 1st draft for internal discussion.
03.11.2019, APC 2nd draft
13.05.2021, APC 3rd draft – accounting for mitigating positive effects of shrinkage strain on crack width
03.01.2023, APC 4th draft – Adapted to final version of FprEN 1992-1-1:2023
22.02.2023 Reviewed and Final draft
20.02.2023 update of title

Pérez Caldentey, Jones, Goodchild 03.01.2023

For citations, copying and referencing see Note on Front page! page 394
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to clause 9.3.3 and

clause 9.3.2
Simplified calculation of deflections for
reinforced concrete building structures
Simplified deflection control by span/depth-
ratio

Alejandro Pérez Caldentey FHECOR, Spain

04.01.2023

For citations, copying and referencing see Note on Front page! page 395
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to clauses 9.3.3 (Simplified calculation of deflections for reinforced concrete
building structures) and 9.3.2 (Simplified deflection control by span / depth-ratio for buildings).

9.3.3 Reasons for change

Simplified The general formulation of EN 1992-1-1:2004 (-method) is kept due to the request of several NSBs. However, its
calculation of long- rigorous application is cumbersome and complex when dealing with real cases. It is, in fact, much more complex
term deflections for than other code proposals, such as Branson’s method included in ACI-318.
reinforced concrete
building structures For this reason, an alternative procedure which provides virtually the same results as the -method has been
included. There is therefore no risk of inconsistencies arising from the chosen method. This procedure is based on
determining closed-form factors which allow to correct a linear elastic calculation. Linear elastic analysis is the type
of analysis that is used in normal engineering practice, so this approach is convenient for engineers. In this way, the
method allows to consider all the complex geometry of a real case, obtain linear elastic values of the deflection and
correct theses values by coefficients to obtain the values accounting for cracking, tension-stiffening, differences
between the gross and the homogenised properties of concrete, reduced increase of curvature due to creep in
cracked sections and load history. All the main variables are duly considered and accounted for.

Enhancement of ease of use

The practical application of the proposed alternative method is much simpler than the application of the general
formulation (-method) with little sacrifice in numerical precision.

9.3.3 Introduction
Expression (9.23) Deflection calculations are complex due to the non-linear behaviour of concrete which cracks, has tension stiffening
effects and has a complex time-dependent behaviour. To consider this behaviour, concrete codes, such as EN 1992-
1-1, propose to determine the deflection of RC elements by nonlinear interpolation between cracked and uncracked
states (-method). Creep is taken into consideration by an effective modulus of elasticity while a separate calculation
is needed to consider the effect of shrinkage. In its more rigorous form, this method allows the calculation of
curvatures in different sections while the deflections must be determined by double integration of curvatures. The
application of this method becomes very complex in practical applications due to complex load histories and the fact
that the calculation is nonlinear. Furthermore, the uncertainties involved in construction times contribute to question
whether the significant computational effort required for a rigorous application of the method leads to better deflection
predictions. Yet another shortcoming of this method is that it is very complex to apply to continuous structures, since
the complexity is enhanced by the complexity of the structure’s geometry (possibly variable spans, variable widths,
presence of voids, etc.).
As is well known by building designers, deflections determine the depth of concrete floors, which must be optimized
to maximize the available free space. It is very important to have good tools to determine for each case the required
minimum depth at an early stage of the project and this topic has been addressed in several papers ([C09-6][C09-
7][C09-8][C09-9]), and incorporated, with more or less fortune into major design codes ([C09-10][C09-11][C09-12]).
Calculation of deflections in reinforced concrete beams is, however, not an easy task due to the complex material
behaviour (cracking, tension stiffening, time-dependent effects, etc.) and the dependence of the results on the
construction sequence and time of application of loads. Many complex models have been developed to this end, as
for instance those of Marí et al. [C09-13], Ghali et al. [C09-14] or more recently Ulm et al. [C09-15]. The effect of
shear deformations on deflections has been studied by Debernardi et al. [C09-16], leading to the conclusion that this
effect is only important for very stocky members and therefore negligible for common flexural members and will not
be considered here.
Due to the complexity of these calculations, there has been a search for simplified methods. Most of them are based
on simplifications, such as those incorporated in MC 1990[C09-17] or EHE-08[C09-18]. The original basis of this last
method can be found in reference [C09-19] with corrections introduced later on and documented in reference [C09-
20].

9.3.3 Derivation of the formulation for deflection calculations

Expression (9.23) The proposed formulation is meant to improve ease-of-use for designers by allowing them to use linear elastic
models considering the complex geometry of real cases and to correct the deflection results to consider cracking of
concrete and tension stiffening effects. The total long-term deflection  is obtained by correcting the linear elastic
deflection due to loads LOADS by a factor kI and the linear elastic deflection due to a constant curvature produced by
shrinkage, cs by a factor kIks, as shown in Expression (1).

  kI  LOADS  kS   cs
(1)

cs is determined by applying on the linear elastic model a curvature given by Expression (2):

1 E S
   s  cs s (2)
r
  cs E c, ef Ig

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 396
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

where:
Ss is the first order moment of the provided tension and compression reinforcements with respect to the
centroid of the gross concrete cross section,
Ig is the inertia of the gross concrete cross section,
Es is the modulus of elasticity of steel and,

1, 05Ecm
Ec,ef is the effective modulus of elasticity of concrete: Ec ,eff  ,where mean is the mean creep
1  mean
coefficient. If it is assumed that the dead load (qDL) is applied at age t0, the superimposed dead load (qSDL) is applied
at age t1, and the quasi-permanent live load (2qLL) is applied at age t2, mean can be estimated from Expression (3).

 , t 0 q DL   , t1 qSDL   , t 2  2 q LL

 mean  (3)
q DL  qSDL   2 q LL
The linear elastic model should have gross cross section properties and a modulus of elasticity equal to Ec,eff.
kI represents the ratio between the gross moment of inertia Ig and the effective moment of inertia Ie, obtained
according to the -method. It can be derived as a function of the ratio between the fully cracked inertia Icr and the
gross inertia. The -method is based on the uncracked, State I inertia, II, which, depending on the degree of
approximation desired can be taken as the gross inertia or as the homogenised inertia, Ih, and on the State II, or fully
cracked inertia (III=Icr). The deflection is determined as follows. From these expressions, the ratio between the State I
inertia and an equivalent inertia Ie, accounting for cracking and tension-stiffening effects can be derived:

k k k
    II  1     I    1    
Ie I II II
(4)
II I Ig Ig
kI    I  1       1   
Ie I II Ie I cr

qL4
Where k is the expression for the deflection divided by the equivalent inertia (for example k  ).
Ec ,eff
The ratio between the gross inertia and the cracked inertia can be analytically determined as follows, as a function of
the homogenised reinforcement ratio (e,ef), which accounts for creep effects:

1 3
   e ,ef  1   
2
3
I cr 3 0.6  d 

Ig
 3
 2.7   e , ef    
h (5)
1 h
 
12  d 
where:
=x/d is the ratio of the cracked neutral axis to the effective depth of the section
=As/bd is the reinforcement ratio referred to the tension reinforcement
d/h is the ratio between the effective and the total height of the section
The approximation to the ratio between the fully cracked and gross inertia given in Eq. (5) is plotted in Fig. 1. The
approximation is quite close to the theoretical values.

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 397
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

2.50

2.00

1.50
Icr/Ig

1.00

0.50

0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
e,ef
Icr/Ig, aprox Icr/Ig

Fig. 1 Ratio between the fully cracked and gross inertia and proposed approximation
The -factor can be determined assuming that the member is subjected to bending but no axial force, and that the
calculation is for long-term analysis as:
2
M 
  1  0.5  cr  (6)
 Mk 
where:
Mcr is the moment that cracks the critical section
Mk is the characteristic moment at the critical section

Then factor kI is defined as:

1.00 if M k  M cr

kI   Ig
 I  1    if M k  M cr
(7)

 cr
kS Se/Sg is aimed to estimate the ratio between the first order moment in the cracked section, accounting for tension-
stiffening effects, and the first order moment of reinforcement in the gross section. It has been fitted to a mean value
for concrete strengths classes C30, C60 and C100 (see Fig. 2) using an estimation of the equivalent first order
moment of the area of reinforcement with respect to the centroid of the section, Se, determined by interpolating
between cracked and uncracked states, using parameter  defined in paragraph 7.4.3 of EN 1992-1-1:2004. The
results are not very sensitive to strength classes and this approach is deemed sufficient. The resulting formulation is
given in Expression (8). As can be seen, a distinction is made between reinforcement ratios, for which the section will
crack under the characteristic combination of actions, and situations for which it will not.

1,00 if M k  M cr
kS   (8)
455  35  1.6 if M k  M cr
2

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 398
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

1.80

1.60

1.40

1.20

1.00
Se/Sg

0.80

0.60

0.40

0.20

0.00
0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00% 3.50% 4.00%

C30 C60 C100 Proposed adjustment kS

Fig. 2 Adjustment for factor kS.

Verification
A verification of the formulation described above has been carried out by computing the deflection of simply
supported beams covering the full range of practical variables. For this, the following values were used:
 Concrete classes C30, C60 and C100
 Geometrical reinforcement ratios of 0,3%, 0,5%, 1%, 1,5%, 2% and 2,5%
 Effective heights of 0,2; 0,3; 0,4; 0,5; 0,6; 0,7; 0,8; 0,9 and 1,00. Total heights were assumed to be equal to
effective heights plus 0,05 m
 qLL/qTot ratios were taken as 0,5, 0,4 and 0,25
All possible combinations of the above values were considered so that calculations were made for 3693=486
cases. The rheological parameters were determined for each one. The creep coefficients were determined using
Expression (3) assuming that:
 The self-weight is applied at 7 days
 The superimposed dead load is 15% of the self-weight and is applied at 60 days
 The quasi-permanent live load is applied at 365 days
 The relative humidity is 50%
 The deflection is determined for a design life of 100 years
Table 1 shows a selection of the rheological parameters used for the calibration

Table 1. Sample of rheological parameters used in the study

fck [MPa] d [m] h/d qLL/qTot mean cs [mm/m] (36500,7) (36500,60) (36500,365)
30 0.2 1.25 0.5 2.84 0.617 3.2 2.29 1.74
60 0.2 1.25 0.5 1.62 0.528 1.84 1.31 0.96
100 0.2 1.25 0.5 1.07 0.47 1.21 0.86 0.62
30 0.2 1.25 0.4 2.91 0.617 3.2 2.29 1.74
60 0.2 1.25 0.4 1.66 0.528 1.84 1.31 0.96
100 0.2 1.25 0.4 1.1 0.47 1.21 0.86 0.62
30 0.2 1.25 0.25 2.99 0.617 3.2 2.29 1.74
100 0.2 1.25 0.25 1.13 0.47 1.21 0.86 0.62
30 0.5 1.1 0.5 2.56 0.565 2.9 2.07 1.54
60 0.5 1.1 0.5 1.5 0.492 1.7 1.21 0.88
100 0.5 1.1 0.5 1 0.448 1.14 0.8 0.58

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 399
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

30 0.5 1.1 0.4 2.63 0.565 2.9 2.07 1.54

60 0.5 1.1 0.4 1.54 0.492 1.7 1.21 0.88
100 0.5 1.1 0.4 1.03 0.448 1.14 0.8 0.58
60 0.5 1.1 0.25 1.59 0.492 1.7 1.21 0.88
100 0.5 1.1 0.25 1.06 0.448 1.14 0.8 0.58
30 1 1.05 0.5 2.38 0.461 2.7 1.92 1.41
60 1 1.05 0.5 1.42 0.42 1.62 1.14 0.82
100 1 1.05 0.5 0.96 0.403 1.1 0.77 0.55
30 1 1.05 0.4 2.45 0.461 2.7 1.92 1.41
100 1 1.05 0.4 0.99 0.403 1.1 0.77 0.55
60 1 1.05 0.25 1.51 0.42 1.62 1.14 0.82
100 1 1.05 0.25 1.02 0.403 1.1 0.77 0.55

With the above formulation it was possible to obtain a quite good approximation to the -method currently proposed
by EN 1992-1-1:2004 and which is kept as the general method in FprEN 1992-1-1:2023. Table 2 shows the general
statistics of the ratio of the deflection predicted by the simplified model and that predicted by the -method, including
coefficient of variation (Cv), standard deviation (), the mean value (), as well as the minimum and maximum
values, and 5% and 95% quantiles assuming a normal distribution. These values are given for the three concrete
classes considered together and separately. The simplified formula provided consistent values.
Table 2 Statistics for the simplified method
Mcr determined with Ig
All tests C30 C60 C100
 9.3.3 method/‐method

Max. Value 1.820 1.196 1.820 1.186

Min. Value 0.945 1.024 0.991 0.945
Mean () 1.054 1.090 1.066 1.007
Standard dev . () 0.070 0.039 0.093 0.034
CoV 0.067 0.035 0.087 0.033
+1.64 1.170 1.153 1.218 1.063
1.64 0.939 1.027 0.914 0.952

Mean root sq Error [mm] 4.087 3.406 8.905 1.728

Fig. 3 shows the comparison between both methods. The regression line is close to the 45º line, and the coefficient
of determination (R2) is very high. Some of the results seem to be over-conservative. These correspond to low
reinforcement ratios as can be seen from Fig. 4. This is not surprising as for these reinforcement ratios the section is
on the limit of cracking or not, and, of course, this makes a very significant difference in the value of the estimated
deflection when the assumption is not correct.
Fig. 4 through Fig. 8 show the variation of the ratio of the deflection predicted by the simplified method and by the -
method with respect to the reinforcement ratio, the concrete strength, the beam height, the ratio of total height to
effective height and the live load to total load ratio. No significant bias is observed with respect to any of these
variables.
200

180

160
Deflection Simpl. Formula [mm]

140

120
y = 1.0587x
R² = 0.9937
100

80

60

40

20

0
0 20 40 60 80 100 120 140 160 180 200
Deflection ‐method [mm]

Series1 45º Line Regression Line

Fig. 3 – Comparison between proposed formulation and -method

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 400
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

1.8

1.6

1.4

Simpl. Formula/‐method
1.2

0.8

0.6 y = 0.5892x + 1.0468

R² = 0.0043
0.4

0.2

0
0 0.005 0.01 0.015 0.02 0.025 0.03
Geometrical Reinforcement ratio, 

Fig. 4 – Results with respect to reinforcement ratio – no significant bias

2

1.8

1.6

1.4
Simpl. Formula/‐method

1.2

0.8

0.6
y = ‐0.0012x + 1.1304
0.4 R² = 0.2396

0.2

0
5 15 25 35 45 55 65 75 85 95 105
fck [MPa]

Fig. 5– Results with respect to concrete strength – no significant bias

2

1.8

1.6

1.4
Simpl. Formula/‐method

1.2

1
y = 0.0067x + 1.0504
R² = 0.0006
0.8

0.6

0.4

0.2

0
0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10
Effective height [m]

Fig. 6 – Results with respect to effective beam height – no significant bias

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 401
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

1.8

1.6

1.4

Simpl. Formula/‐method
1.2

0.8
y = ‐0.0615x + 1.1225
R² = 0.0029
0.6

0.4

0.2

0
1.00 1.05 1.10 1.15 1.20 1.25 1.30
h/d

Fig. 7 – Results with respect to total height/effective height ratio – no significant bias

1.8

1.6

1.4
Simpl. Formula/‐method

1.2

y = ‐0.0053x + 1.0565
0.8
R² = 6E‐05

0.6

0.4

0.2

0
0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55
qLL/qTot

Fig. 8 – Results with respect to LL/TL ratio – no significant bias

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 402
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

9.3.2 Converting the formula for deflection calculations into a formula for the slenderness limit
Expression (1) can easily be converted into a general formulation to determine the slenderness limit. In this
Simplified
paragraph it will be shown how to do this. Expression (9) shows the condition for the slenderness limit, i.e., that the
deflection control
deflection be limited to a fraction (1/a) of the span. It also shows the expressions of the deflections due to a uniformly
by span/depth ratio
distributed load and a constant curvature due to shrinkage.

L
  kI  LOADS  kS     cs
a
5 qqp L4
 LOADS  K
384 E 1 3
bh
c ,eff
12 (9)
 h h 
A d    As    h  d   2
Es s  2  2 L
  Kcs cs
cs
Ec ,eff 1 3 8
bh
12

where:
K is a factor that considers the support conditions for the deflection due to uniformly distributed loads and can
be determined from Expression (10) (for a detailed derivation see [C09-9]):

f simply,sup
K3 (10)
f real , sup,cond
where:
fsimply,sup is the linear elastic deflection of the simply supported member of arbitrary span subjected to a
uniformly distributed load, and
freal,sup,cond is the linear elastic deflection of the member with the actual support conditions with the same
arbitrary span and subjected to the same uniformly distributed load.
Kcs is a factor that considers the support conditions for the deflection due to shrinkage and can be determined
from Expression (11) (for a detailed derivation see [C09-9]):

fcs , simply,sup
Kcs  (11)
fcs ,real ,sup,cond
where:
fcs,simply,sup is the linear elastic deflection of the simply supported member of arbitrary span subjected to a
constant curvature, and
fscs,real,sup,cond is the linear elastic deflection of the member with the actual support conditions with the same
arbitrary span and subjected to a constant curvature.
Assuming that the provided reinforcement is that strictly needed in ULS, the value of the quasi-permanent load can
be determined from the ultimate bending resistance as shown in the last equation of Expression (12).

L2  Af 
M Rd  qRd  As f yd  d  0,5 s yd 
8  bfcd 
 
 As f yd  8
qRd  As f yd d 1  0,5 
 bdfcd  L2

  
 LL 
1  
 TL   (12)
LL 2  LL  LL
qqp G   2Q 1    2
  TL   TL  TL
 kDL
qRd  GG   QQ  LL   LL  LL
 1    G 1  
 Q
G 
TL   TL  TL
Q
LL
TL
8
qqp  kDL As f yd d 1  0,5  2
L

In expression (12) it is assumed that there is no need for compression reinforcement in ULS.
Introducing Expression (12) into Expression (9), and developing, the slenderness limit can be obtained as shown in

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 403
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

Expression (13). The final expression proposed in Eq. (14) assumes that there is no compression reinforcement
since the effect of compression reinforcement on deflections is limited, and the increase in precision for this case is
not worth the complication.

L2 
qqp  kDL qRd  kDL As f yd d 1  0,5  
8 
As'  0 

 8 
 5 kDL As f yd d L2 1  0,5  L
4

K 1 3 
 384 Ec ,eff bh 
12  L
kI 
 h ' h   a
 As  d    As    h  d   2  (13)
k K  Es  2 2 L 
 S cs cs Ec ,eff 1 3 8
 bh 
12
 
3 
L Ec ,eff  h   1 
  
d 12akI  d   5 E   h   h  h 
 KkDL f cd  1  0,5   kS Kcs cs s   1     '     1   
 48 8   2d   2d  d    

The final expression for the slenderness limit, assuming (’=0) is shown in (14).

 
3 
L Ec ,ef  h  1
    5 E  h 
 (14)
d kI 12a  d  
KkDL f cd  1  0,5   kS Kcs cs s  1   
 48 8  2d  
The slenderness limits shown in Table 9.3 of FprEN 1992-1-1:2022 can be derived from the above expression.

[C09-6] Branson, D.E.: Deformations of concrete structures. McGraw-Hill, New York (1977)
References
[C09-7] Gilbert, R.I. Deflection Calculation for Reinforced Concrete Structures—Why We Sometimes Get It Wrong.
ACI Structural Journal. Vol 96, No. 6, pp 1027-1032, Nov. 1999.
[C09-8] Bischoff, P.H., Scanlon, A. Effective Moment of Inertia for Calculating Deflections of Concrete Members
Containing Steel Reinforcement and Fiber-Reinforced Polymer Reinforcement. ACI Structural Journal.
Vol. 104, No. 1, pp. 68-75, Jan-Feb 2007.
[C09-9] Pérez Caldentey, A., Mendoza Cembranos, J., Corres Peiretti, H.: Slenderness Limits for deflection
control: a new formulation for flexural RC elements. Structural Concrete (2016). DOI:
10.1002/suco.201600062
[C09-10] ACI Committee 318: Building Code Requirements for Structural Concrete (ACI 318-08) and commentary
(318R-08), American Concrete Institute, Farmington Hills, Mich., (2008), 430 p.
[C09-11] CEN: EN-1992-1-1. Eurocode 2. Design of concrete structures – Part 1-1. General rules and rules for
buildings, (2004).
[C09-12] fib: fib Model Code for Concrete Structures 2010. Ernst & Sohn. ISBN 978-2-88394-105-2 and ISBN 978-
2-88394-106-9.
[C09-13] Marí A.R.: Nonlinear geometric, material and time dependent analysis of three dimensional reinforced and
prestressed concrete frames. Report UCB-SESM-84/12. Berkeley, USA: Univ. of California; 1984
[C09-14] Ghali A. and Elbadry M.: User's manual and computer program CPF: Cracked plane frames in prestressed
concrete. Res. report CE85-2. Alberta, Canada: Dept. of Civil Eng., Univ. of Calgary; 1985
[C09-15] Ulm F.J., Clement J.L. and Guggenberger J.: Recent advances in 3-D nonlinear FE-analysis of R/C and
P/C beam structures. In: Proc. ASCE Struc. Cong. XII. 1994. p. 427-1433
[C09-16] Debernardi, G., Guglia, M. and Taliano, M.: Parametric analysis of the influence of shear on the deflection
of RC beams in service. Magazine of Concrete Research Volume 64, Issue 4, April 2012, pp 335-350.
[C09-17] CEB-FIP: Model Code 1990. Thomas Telford, 1993.
[C09-18] Ministerio de Fomento: EHE-08. Instrucción de Hormigón Estructural. EHE-08. (5ª edición) (2011).
[C09-19] Murcia Vela, J.: Cálculo práctico de flechas diferidas en estructuras de hormigón armado. Hormigón y
acero nº2015 (2000).
[C09-20] Mari, A., Bairán, J., Duarte, N.: Long-term deflections in cracked reinforced concrete flexural members.
Engineering structures.32 – 3, pp. 829 – 842.03/2010. ISSN 0141-0296

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 404
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
9.3.3 and 9.3.2 Background to FprEN 1992-1-1:2023

List of versions of this document

28.11.2017, APC 1st draft (submitted to SC2/WG1, 30.11.2017)
02.04.2018, APC 2nd draft (submitted to SC2/WG1, 30.04.2018)
03.03.2021 APC 3rd draft, editorial changes (submitted to SC2/WG1 03.03.2021)
04.01.2023 APC 4th draft – Adapted to final version of FprEN 1992-1-1:2022

A. Pérez Caldentey, FHECOR 04.01.2023

For citations, copying and referencing see Note on Front page! page 405
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to clause 10 and

Annex E
Fatigue

Josef Hegger, Matthias RWTH Aachen University, Germany

Hillebrand, Maximilian Schmidt

24.01.2023

For citations, copying and referencing see Note on Front page! page 406
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
10 and Annex E.4 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to clause 10 (Fatigue) and Annex E (Additional rules for fatigue
verification).

10 and E Enhancement of ease of use

General change Clause 10 (former clause 6.8) and Annex E (former Annex NN) were reorganized to enhance the ease-of-use.
The corresponding sections of EN 1992-1-1 and EN 1992-2 were merged together following a decision of WG1
and SC2. After defining some general conditions, the main clause 10 for fatigue design now only contains the
simplified verifications. The more complex verifications using damage equivalent stresses or PALMGREN-MINER
rule were moved to Annex E. In addition, damage equivalent verifications for road and railway bridges were moved
to Annex K.

Present order of EN 1992-1-1:2004 New order

6.8.1 Verification conditions 10.1 General
6.8.2 Internal forces and stresses for

Chapter 10 of EC2-1-1
fatigue verification 10.2 Combinations of actions

6.8.3 Combinations of actions 10.3 Internal forces and stresses for fatigue
6.8.4 Verification procedure for reinforcing verification
and prestressing steel
- Steel, level 3 Simplified verifications (level 1)
10.4 Steel
6.8.5 Verification using damage equivalent 10.5 Concrete Compression
stress range 10.6 Concrete Shear
- Steel, level 2 10.7 Shear Interfaces
6.8.6 Other verifications
- Steel, level 1 E.1 General
6.8.7 Verification of concrete under damage equivalent stress range (level 2)
compression or shear E.2.1 General
- Concrete, compression, level 2

Annex E
E.2.2 Steel
- Concrete, compression, level 1 E.2.3 Concrete Compression
- Concrete, shear, level 1
explicit verification of damage strength (level 3)
EC 2-2 6.8
E.3.1 Verification conditions
- Concrete, level 3
E.3.2 Steel
6.2.5 (5) Shear at interface E.3.3 Concrete Compression

Figure C10.1: New order of section 10 and Annex E

10.3 Internal forces and stresses for fatigue verification

The former equation (Eq. C10.1) takes into account the effect of different bond conditions of prestressing and
reinforcing steel by increasing the stress range of the reinforcing steel [C10.1]. For the derivation, it is assumed
that prestressing and reinforcing steel have the same inner lever arm.

A s  Ap
η
A s  Ap ξ( s / p )
(C10.1)
The new equation (Eq. C10.2) is derived in agreement with clause 9.2.2(3) by decreasing the activated
prestressing steel area and calculating the stress range considering different inner lever arms of prestressing and
reinforcing steel.

s
Ae  Ap ξ (C10.2)
p
Stress gradient
Clause 10.3(3) was added to take into account a redistribution of concrete stresses under compression for a given
stress gradient. Figure C10.2 shows the definition of the stress gradient and the design concrete stress c for
fatigue. The limit of c  2/3 c2 given in Figure C10.2 applies for fck = 55 MPa and increases to c  0,9 c2 for fck
= 100 MPa.

J. Hegger, M.Hillebrand, M. Schmidt

RWTH Aachen University 24.01.2023
For citations, copying and referencing see Note on Front page! page 407
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
10 and Annex E.4 Background to FprEN 1992-1-1:2023

Figure C10.2: Definition of stress gradient and design concrete stress c

10.4 Simplified verification of reinforcing and prestressing steel

The values of adequate fatigue resistance were supplemented and updated. In the former code the simplified
verification was restricted to unwelded reinforcing bars. The new clause 10.4 provides adequate values of fatigue
resistances also for welded reinforcing bars as well as prestressing steel.
A maximum of 108 load cycles was defined for the simplified verification. The values for s at N=108 were derived
from the S-N curves according in Annex E (E.4.2(1)). For the simplified verification a partial safety factor of s=1,15
was applied. The values for prestressing steel were derived in a similar manner.

10.5 Simplified verification of concrete under compression

The design fatigue strength fcd,fat of former Eurocode 2 (Eq. C10.3 [C10.1]) was improved with respect to high
strength concretes. Also, the effect of sustained loading was considered. While in Eq. C10.3 the design concrete
strength fcd=fck/1,5 was reduced by the factors 0,85 and (1-fck/250), the new equation for fcd,fat (Eq. C10.4) captures
the influences of sustained loading and the more brittle behaviour of higher strength concrete by the factor cc,fat
= min (0,85cc; 0,8). Where cc is the factor according to Eq. C10.6 taking into account the differences between
the undisturbed compressive strength of a cylinder and the effective compressive strength that can be developed
in a structural element. The factors 0,85 and 0,8 consider the reduction under cyclic loading compared to
monotonic short-term loading.
ck
𝑓cd,fat 0,85 ⋅ 𝛽cc 𝑡 ⋅ 𝑓cd ⋅ 1 (C10.3)
250

ck
𝑓cd,fat 𝛽cc 𝑡 ⋅ ⋅ 𝜂cc,fat (C10.4)
c

where
𝜂 , 𝑚𝑖𝑛 0,85 ⋅ 𝜂 ; 0,8 (C10.5)
⁄
𝜂 1 (C10.6)

Figure C10.3 depicts the development of fcd and fcd,fat according to the former and new versions of EN 1992-1-1.

J. Hegger, M.Hillebrand, M. Schmidt

RWTH Aachen University 24.01.2023
For citations, copying and referencing see Note on Front page! page 408
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
10 and Annex E.4 Background to FprEN 1992-1-1:2023

90
fcd,fat fcd ‐ EN 1992‐1‐1:2004
fcd ‐ FprEN 1992‐1‐1
80
fcd,fat ‐ EN 1992‐1‐1:2004
fcd ‐ FprEN 1992‐1‐1
70 fcd,fat ‐ FprEN 1992‐1‐1

fcd ‐ EN 1992‐1‐1:2004
60

50

40

30
fcd,fat ‐ FprEN 1992‐1‐1
20
fcd,fat ‐ EN 1992‐1‐1:2004
10

fck
0
0 20 40 60 80 100 120

Figure C10.3: Development of fcd and fcd,fat according to FprEN 1992-1-1:2023 and EN 1992-1-1:2004

10.7 Simplified verification of shear at interfaces

The former clause for fatigue design for the verification of shear at interfaces was originally located in clause
6.2.5(5) [C10.1]. The former design rules were extended and moved to clause 10 fatigue.
As adhesion is omitted for the fatigue verification of interfaces with crossing reinforcement, it is assumed that the
inclination of the compression struts results in about 25°. This favourable effect is covered by the factor 0,45 in
the term for the stress of the interface reinforcement.
Since no tests under cyclic loads are available, the new verification for insufficiently anchored reinforcement
crossing the interfaces is not allowed for fatigue verification. Nevertheless, in these cases, the verification can be
performed by neglecting the insufficiently anchored interface reinforcement.

E.4 Parameters for S-N curves

The parameters for S-N curves for reinforcing steel are based on [C10.1] in combination with new research results
from the Netherlands [C10.4]
The parameters for S-N curves for prestressing steel have been adjusted according to [C10.2] and [C10.3].

[C10.1] European committee for standardization (CEN): EN 1992-1-1, Eurocode 2: Design of concrete
References
structures Part 1-1, 2004.
[C10.2] Deutsches Institut für Bautechnik (DIBt): Richtlinie für Zulassungs- und Überwachungsprüfungen für
Spannstähle (guideline for approval and monitoring tests for prestressing steel), 2004. (in German)
[C10.3] Deutsches Institut für Bautechnik (DIBt): Ergänzung zur Richtlinie (supplement to guideline), 2014. (in
German)
[C10.4] Breedijk, Theo; Dieteren, Gerrie: Background document to the Dutch comments NL
118/119/126/127/128 on prEN 1992-1-1 draft D4-6 / D5-11 section 10.4, table C.1 and table E.1:
Fatigue strength reinforcing steel, 24.07.2020.

J. Hegger, M.Hillebrand, M. Schmidt

RWTH Aachen University 24.01.2023
For citations, copying and referencing see Note on Front page! page 409
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/Dutch FprEN 1992-1-1:2023

Mirror Committee

Background document to clause 10.4,

Annex C.4 and Annex E.4
Fatigue strength of reinforcing steel

Theo Breedijk Independent expert reinforcing steels, Netherlands

Gerrie Dieteren
TNO, Netherlands
06.02.2023

For citations, copying and referencing see Note on Front page! page 410
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
10.4 and C.4 and E.4 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to FprEN 1992-1-1:2023, clause 10.4, Annex C.4 (table C.1) and
Annex E.4 (table E.1).

Reasons for change

For some years the importance of fatigue performance of reinforcing steel has shifted towards the smaller
diameters 8 – 16 mm by the increasing application of pre-tensioned steel as main reinforcement. Also windmills
and floors in storage buildings carrying transport lorries need the smaller sizes. Therefore, the known higher
fatigue strength for these smaller diameters should be used in practice.
Furthermore the smaller bars 6 – 20 mm are today efficiently produced as mechanically straightened bars from
coils (de-coiled bar). This processing may influence the fatigue properties by hardly avoidable surface damages
and/or from the subsurface microstructure. Therefore, the standard must reflect fatigue properties still valid for
de-coiled bars including process control on the mechanical straightening process.
Also efficiently prefabricated reinforcement (constructions) uses today tack welding instead of binding the bars
with wires. This enables robust reinforcement constructions and respect the required labour conditions.
However, the tack welding lowers the fatigue strength and therefore this fatigue strength should be carefully
checked for still possible application. Some “easy” prohibition may have great and costly consequences.

Background
For above mentioned reasons over 500 fatigue tests have been performed on mainly B500B diameter 12 and
16 mm de-coiled bars from different mechanical straightening processes and cross-welded bars by resistance
welding and CO2 tack welding. The results show that the differences in straightening processes and cross
welding methods are limited and are being overruled by the scatter in the fatigue test results.
The fatigue tests are performed with constant stress amplitude with upper stress limit 0,6*fyk, load cycle
frequency 50 - 100 Hz and mostly stopped at N = 10 million load cycles (run-outs). The results are statistically
evaluated by 5 % linear regression from the failures (the k1 region) as well as by 5 % probabilistic approach with
Bayesian estimation from the failures + run-outs (the k1 region up to the horizontal CAFL) which provides better
insight on the knee-point N* (CAFL = constant amplitude fatigue limit). Further background is given in [1] and
[2].
As in EN1992-1-1:2004 the S-N curves are based on a fixed relation between k1 and k2 (Haibach – 1970): k2 =
2*k1 -1. For not welded bars there was a tendency for k1 = 4 but this would result in k2 = 7 which is too critical.
Therefore, for not welded bars k1 = 5 and k2 = 9 is maintained as in EN1992-1-1:2004. For welded bars, like in
EN1992-1-1:2004, k1 = 3 and k2 = 5 is maintained since the results for k1 did not show a need for adjustment.
The test results (see below) on the straightened bars show better fit with the results for N* = 2 x 106 than for N*
= 106 cycles, which therefore should be respected. On the welded bars the better fit for N* is somewhat higher
than 2 x 106. For the good sake of most testing in practice up to N = 2 x 106 and to prevent misunderstanding
between N* from the S – N curve and the N from the declared value by testing, it was proposed to apply N* = 2
x 106 for not welded and welded reinforcing steel. Also, already several countries are (were) used to specify
∆σRsk at N = 2 x 106 (UK at N = 5 x 106).
As expected, the test results on the mechanically straightened bars as well on the welded bars show for
diameter 12 mm relevant higher fatigue strength than for 16 mm which should be used in practice. For 16 mm
the test results on the mechanically straightened bars are (from the effect of de-coiling) somewhat lower and for
the welded bars about the same as the fatigue strength in EN 1992-1-1:2004 taken at N ≥ 2 x 106. Therefore, for
bars including de-coiled bars (are now ruling) 3 groups of bar sizes are proposed in order to use the best
possible properties. However the group of bigger bar sizes including de-coiled bar may interfere with the straight
produced bars which are not suffering from the mechanical straightening process. Therefore, arbitrary d = 20
mm is taken as the max. diameter for the mechanical straightening. The few possible straightening of bigger
sizes (up to 25 mm) should be addressed with special certification.

Conclusions

Based on the performed tests and evaluation of the results adjustment of the design parameters for S-N curves
is proposed as presented in Table C10.4.1.

Table C10.4.1 : Design parameters for S-N curves for carbon reinforcing steel (Table E.1 in FprEN1992-1-
1:2023)

T. Breedijk, G. Dieteren 18.05.2021

For citations, copying and referencing see Note on Front page! page 411
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
10.4 and C.4 and E.4 Background to FprEN 1992-1-1:2023

Type of Diameter ∆σRsk 5 % fractile (Test σmax 0,6fyk)

reinforcement (mm) ∆σRsk (MPa) N* (cycles) kf1 – kf2
Bars including de‐coiled bars1 ≤ 12 160
12 < ϕ ≤ 16 140 2 x 106 5–9
16 < ϕ ≤ 20 130
Straight bars1)2) > 20 130
Tack welded bars3)4) and ≤ 12 100 2 x 106 3–5
welded wire fabrics > 12 80
Couplers ‐‐ 35 107 3–5
Note 1) : Values for bent parts of bars should be obtained using a reduction factor ζ = 0,35 + 0,026.ϕmand/ϕ.
The reduction factor ζ may be omitted for shear reinforcement with 90° stirrups ϕ ≤ 16 mm and depth h ≥ 600 mm.
Note 2) Values for ∆σRsk only to be applied for straight bars and not for de-coiled bars.
Note 3) Values for ∆σRsk of tack welded bars apply for a distance of 5 d at each side of the weld.
Note 4) Values for ∆σRsk of tack welded bars apply only for CO2 welded bars.

By making the table a NDP different fatigue properties can be set in the National Annex if higher values of the
fatigue stress range and/or the number of cycles are confirmed by testing in accordance with EN 10080. It
should be noted that effect of de-coiling should be investigated if de-coiled bars are being applied.

Note 3 is added to table C.10.4.1 since the influence of welding is not present in the whole bar and in practice
this has been discussed many times.

To ensure the design parameters as presented in table C.10.4.1 the fatigue requirements in table C.1 of
FprEN1992-1-1:2023 are as follows:

Fatigue stress range 2σa [MPa] at N≥ 2 x 106 160 for bars and de‐coiled bars ϕ ≤ 12 mm
cycles based on testing with an upper stress limit 140 for bars and de‐coiled bars ϕ ≤ 16 mm
130 for bars and de‐coiled bars ϕ > 16 ≤ 20 mm
of 0,6*fyk (10% quantile)
130 for straight bars ϕ > 20 mm
100 for wire fabrics ϕ ≤ 12 mm
80 for wire fabrics ϕ > 12 mm

The adjustment of the design parameters for S-N curves also affects the simplified verification method
presented in 10.4. Required adjustments are as follows:
“a) reinforcing steel bars (for bent bars footnote a) of table E.1(NDP) should be applied):
- Δσsd ≤ 90 MPa unwelded reinforcing bars ø ≤ 12 mm
- Δσsd ≤ 73 MPa unwelded reinforcing bars ø > 12 mm
- Δσsd ≤ 40 MPa butt and tack welded reinforcing bars ø ≤ 12 mm
- Δσsd ≤ 30 MPa butt and tack welded reinforcing bars ø > 12 mm
- Δσsd ≤ 19 MPa couplers

Note: For practical reasons the split between diameter groups is limited at 12 mm.

T. Breedijk, G. Dieteren 18.05.2021

For citations, copying and referencing see Note on Front page! page 412
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
10.4 and C.4 and E.4 Background to FprEN 1992-1-1:2023

Test results

Figure C10.4.1 Test results for 12 mm B500B mechanically straigthened bars tested with σmax= 0,6*fyk with
proposal FprEN1992-1-1:2023 (Proposal 160/2 – 5/9) and comparison with EN1992-1-1 :2004.

Figure C10.4.2 Test results for 16 mm B500B mechanical straigthened bars tested with σmax= 0,6*fyk with
proposal FprEN1992-1-1:2023 (Proposal 140/2 – 5/9) and comparison with EN1992-1-1 :2004.

T. Breedijk, G. Dieteren 18.05.2021

For citations, copying and referencing see Note on Front page! page 413
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
10.4 and C.4 and E.4 Background to FprEN 1992-1-1:2023

Figure C10.4.3 Test results for 12 mm B500B welded bars tested with σmax.= 0,6*fyk with proposal
FprEN1992-1-1:2023 (Proposal 100/2 – 3/5) and comparison with EN1992-1-1 :2004.

Figure C10.4.1 Test results for 16 mm B500B welded bars tested with σmax= 0,6*fyk with proposal FprEN1992-1-
1:2023 (Proposal 80/2 – 3/5) and comparisson with EN1992-1-1 :2004.

[1] D.L. Allaix, F.B.J. Gijsbers Bayesian estimation of characteristics S-N curves for reinforcement bars
References
and proposal for the national annex of NEN-EN 1992-1-1, TNO-report TNO 2016 R10819, 21 June
2016
[2] T. Breedijk, New standard requirements on fatigue strength of reinforcing steel, Cement 5, 2020,
page 44 – 49 + annex page 1 – 5 (in Dutch).

T. Breedijk, G. Dieteren 18.05.2021

For citations, copying and referencing see Note on Front page! page 414
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
10.4 and C.4 and E.4 Background to FprEN 1992-1-1:2023

List of versions of this document

18-05-2021 1st draft
31-05-2021 Final draft
06.02.2023 Peer review and update of references to FprEN1992-1-1:2023

T. Breedijk, G. Dieteren 18.05.2021

For citations, copying and referencing see Note on Front page! page 415
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to Clause 11.3

Permissible mandrel diameters for bent
bars

Aurelio Muttoni EPFL, Ecole Polytechnique Fédérale de Lausanne,

Miguel Fernández Ruiz Switzerland
Frédéric Monney
Report EPFL-IBETON 17-01-R4 28.03.2023

For citations, copying and referencing see Note on Front page! page 416
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.3 Background to FprEN 1992-1-1:2023

Introduction This document has been prepared to clarify the background of the design provisions of FprEN 1992-1-1:2023
with respect to clause 11.3 concerning the permissible mandrel diameters for bent bars. These provisions update
current Section 8.3 of EN 1992-1-1:2004.

Reasons for change The current provisions for permissible mandrel diameters for bent bars of EN 1992-1-1:2004 (Section 8.3)
include a number of rules to suitably bend the reinforcement without damaging it or crushing the concrete inside
the bend. These rules have been identified to have two shortcomings:
1. Formula (8.1) of EN 1992-1-1:2004 relating the mandrel diameter and maximum stresses developed
in the reinforcement has been observed to have a poor performance to estimate the actual response
of bent details [C.11.3.1] (see also next paragraph).
2. Adapting the mandrel diameter according to the types of bends requires the use of different mandrels
for the same bar diameter. In addition, bending for large radii typically requires the use of machines
(based usually on the three-roller principle) different than those for normal mandrel diameters (where
the bars are bent around mandrels with Ømand = 4Ø or Ømand = 7Ø according to 11.3(2)). To avoid
these shortcomings and to increase the efficiency in bending the reinforcement, bends with larger
radii can be replaced by multiple bends with normal mandrels spaced by sufficiently long straight
segments. This requires a rational model to derive consistent rules easy to use.

Permissible mandrel Following these observations, a revision of the subsection has been performed. A detailed description of the
diameter, general case, theoretical principles and its implementation can be consulted in Monney et al. [C.11.3.1]. In this background
11.3(4) document, only the main assumptions and results will be highlighted.
The provisions of EN 1992-1-1:2004 provide a formula to evaluate the maximum force that can be developed in
a bar without damaging the concrete accounting for the bar diameter (Ø), the concrete strength (fcd), the concrete
cover (cd) and the mandrel diameter (Ømand). The origin of that formula is not completely clear, probably adapted
from a former British Standard [C.11.3.2] (see [C.11.3.1]). Although the basic parameters considered are
relevant, the performance of the expression is relatively poor, particularly for low Ømand /Ø ratios. Also, the
generality of the formula is penalized by the fact that it does not consider the bending angle of the detail. A
comparison of the performance of EN 1992-1-1:2004 with relevant test results is shown in Figure C11.3.1.

Figure C11.3.1: Performance of Formula (8.1) of EN 1992-1-1:2004 compared to tests (the first graph refers
to tests conducted by the authors, the second one shows tests found in the literature; the ratio
on the ordinates refers to the steel stress related to concrete crushing or spalling measured
in tests and the value calculated according to the code provision with average material
strengths and without partial safety factor), figure from Monney et al. [C.11.3.1].
In order to propose a more consistent formula for design, FprEN 1992-1-1:2023 has been based on a
mechanical model accounting for potential spalling failures. These failures, Fig. C11.3.2, occur due to the thrust
of a wedge of concrete forming at the bend region (blue region in Fig. C11.3.2). Such wedge is confined, thus
increasing significantly its resistance with respect to uniaxial conditions. The confinement pressures
(perpendicular to the bending plane) are in equilibrium with concrete tensile stresses developing in a confining
region (red region in Fig. C11.3.2a). As shown by detailed measurements (see [C.11.3.1]), a splitting crack
develops between the confined and the confining zone (white area between red and blue areas in Fig. C11.3.2).
Nevertheless, the load can be further increased and finally, failure occurs when the splitting crack develops in
an unstable manner leading to spalling of the concrete cover (Fig. C11.3.2).
The extent of the different areas has been observed by detailed measurement and analysis of test results
[C.11.3.1] to depend on a number of parameters:
 Ratio between bar diameter and concrete cover

A. Muttoni, M. Fernández Ruiz, F. Monney, EPFL 28.03.2023

For citations, copying and referencing see Note on Front page! page 417
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.3 Background to FprEN 1992-1-1:2023

 Deviation angle of the bend (bend in Fig. C11.3.2a)

 Size of the detail (size effect)
On that basis, by suitably analysing the contribution of the different parameters to the confining and confined
areas, the following relationship can be mechanically derived through equilibrium considerations [C.11.3.1]:
1/3
2 Ø d   c 1  45 Ø 
σ s   mand  ηcc  f c  f c   dg       32   0.7  mand   fy (C.11.3.1)
π Ø  Ø   Ø 2   α  Ø 

 

Figure C11.3.2: Mechanical model for FprEN 1992-1-1:2023: (a) plan view with confined (blue), confining
(red) and spalled areas (white); (b) detail of stresses and forces; and (c) detail of the concrete
wedge with the confinement stresses. (figure from [C.11.3.1])
This expression is the basis of Formula (11.1) of FprEN 1992-1-1:2023, where 2/ has been rounded to 0.65:
1/3
Ø f d  c 1  Ø 
σ sd  0, 65  f cd mand  ck   dg       kbend  0.7  mand 

(11.1)
Ø γC  Ø 
  Ø 2  Ø 

where k 45
bend  32 
α
The performance of this Formula is compared in Fig. C11.3.3 to the same tests as those of Fig. C11.3.1. The
results show consistent agreement, suitably reproducing all design situations and a low scatter.

Figure C11.3.3: Performance of Eq. (C.11.3.1) (basis of FprEN 1992-1-1:2023) compared to test results,
figure from Monney et al. [C.11.3.1] (the first graph refers to tests conducted by the authors, the second one
shows tests found in the literature).

The linear Formula (11.1) can also be solved to determine directly the required mandrel diameter:
Ømand σ sd  kbend  σc,conf ,d (C.11.3.2)

Ø 0, 65  fcd  0, 7  σ c,conf , d

A. Muttoni, M. Fernández Ruiz, F. Monney, EPFL 28.03.2023

For citations, copying and referencing see Note on Front page! page 418
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.3 Background to FprEN 1992-1-1:2023

1/3
f d   cd 1
where σ
c ,conf , d  ck   dg    
γC  Ø  Ø 2
 

Simplified detailing rules, To enhance ease of use, simplified detailing rules are given in 11.3(3) for the cases where fyd ≤ 25ꞏfcd, covering
11.3(3) most practical cases. The verification of the concrete inside the bend may be omitted for all bends with an angle
bend ≤ 45° at a clear distance cx ≥ 2,5Ø from a free edge parallel to the bending plane and at a clear distance
between parallel bars cs ≥ 5Ø (equal to twice cx). As shown in Fig. C11.3.4 on the basis of Formula (11.1) of
FprEN 1992-1-1:2023, the limit cx ≥ 2,5Ø covers all practical cases. For diameter not larger than 32mm, this limit
could even be reduced to cx ≥ 2Ø.

Figure C11.3.4: Results of Formula (11.1) of FprEN 1992-1-1:2023 compared to the simplified detailing rule
cx ≥ 2,5Ø in 11.3(4).

Rules for multiple bends, Following the implications of Formula (11.1), it can be observed that the deviation angle of the bend (parameter
11.3(4) bend) plays a significant role. Small deviations allow in fact to develop large stresses, potentially ensuring
yielding of the reinforcement without cover spalling.
Based on this idea, bending of a bar can be performed as a succession of multiple smaller bends with the same
radius, see Figure C11.3.5. This concept allows to use the same mandrel as for normal bends (Ømand = 4Ø or
Ømand = 7Ø) and the same machine, allowing to bend the bars for all situations in a more efficient manner.

Figure C11.3.5: Performance of a detail with multiple bends, figure from Monney et al. [C.11.3.1].

In order to ensure suitable performance of the detail, two criteria shall nevertheless be respected:
 Ensuring no damage of the reinforcement while bending of the bare bar (fulfilling the requirement of
11.3(2), namely Ømand = 4Ø for Ø ≤ 16 mm and Ømand = 7Ø for Ø >16 mm).
 A minimum distance between the bends is provided (lmand in Fig. C11.3.5)
A general expression to calculate minimum values of lmand based on the mechanical model of Formula (C.11.3.1)
is derived in [C.11.3.1] and plotted in Fig. C11.3.6 for variable bending angle bend, bar diameter and concrete
cover for the cases where fyd ≤ 25ꞏfcd, covering most practical situations. Such expression has been simplified
into FprEN 1992-1-1:2023 where the distance between the bends should not be shorter than 4Ø.

A. Muttoni, M. Fernández Ruiz, F. Monney, EPFL 28.03.2023

For citations, copying and referencing see Note on Front page! page 419
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.3 Background to FprEN 1992-1-1:2023

Figure C11.3.6: Minimum distance between bends lmand as a function of: (a) the bending angle; and (b) the
concrete cover with the detailing rule of FprEN 1992-1-1:2023.

Simplified rules for According to EN 1992-1-1:2004, the verification of the concrete strength inside the bend of hook and bend
standard hook and bend anchorages may be omitted if (i) the anchorage of the bar does not require a length more than 5Ø after the end
anchorages, of the bend and (ii) the bar is not positioned at the edge (plane of bend close to concrete face). Based on
theoretical considerations using Formula (C.11.3.1) confirmed by tests (see Figure C11.3.7, [C.11.3.1]), these
11.3(3) requirements can be mitigated by defining a minimum clear distance cx ≥ 1,5Ø between the bar and a free edge
parallel to the bend. The reduction of cx compared to normal bends can be explained theoretically by the fact
that the tensile stress is lower at the start of the bent because a part of the tensile force is transferred by bond.

Figure C11.3.7: Maximum tensile stress at failure for different hook and bend anchorages (tensile stress
measured at the start of the bent, see definition of lb).

Enhancement in case of If transverse bars are placed within the bend, a portion of the deviation forces can be transferred to the
transverse bars placed transverse bars as shown in Figure C11.3.8 below. In this case, the limit to the steel stress sd in Formula (11.1)
within the bend may be increased by the factor defined in Formula (11.2) (see [C.11.3.3] for its development).

11.3(5)

Figure C11.3.8: Effect of transverse bars placed inside the bend (figure from Monney [C.11.3.3]

References [C.11.3.1] Monney, F.; Fernández Ruiz, M.; Muttoni, A., “Design against splitting failures in reinforced concrete
due to concentrated forces and minimum bend diameter of reinforcement”, Engineering Structures,
245 (2021) 112902, p. 21
[C.11.3.2] BS 8110, Structural use of Concrete - Code of Practice for Design and Construction, British
Standards Institution, Part. 1, London, UK, 1985.
[C.11.3.3] Monney F., “Revisiting Detailing Rules for Bending Bars, Anchorage and Minimum Shear
Reinforcement in Concrete Structures”, PhD Dissertation, École Polytechnique Fédérale de
Lausanne (EPFL), Switzerland, 2022, p. 184.

A. Muttoni, M. Fernández Ruiz, F. Monney, EPFL 28.03.2023

For citations, copying and referencing see Note on Front page! page 420
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.3 Background to FprEN 1992-1-1:2023

List of versions of this document

13.03.2020, AM, MFR, FM 1st draft

03.03.2021, AM 2nd draft
24.08.2021, AM Minor revision
28.03.2023, AM updated with reference to FprEN 1992-1-1:2023, reference to 11.3(5) added

A. Muttoni, M. Fernández Ruiz, F. Monney, EPFL 28.03.2023

For citations, copying and referencing see Note on Front page! page 421
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to clauses

11.4 and 11.5
Anchorage and laps of bars in tension and
compression

Aurelio Muttoni EPFL, Switzerland

John Cairns Heriot-Watt, Great Britain
Charles Goodchild Great Britain
Hans Rudolf Ganz Switzerland
27.03.2023

For citations, copying and referencing see Note on Front page! page 422
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to clauses 11.4 (“Anchorage of reinforcing steel in tension and
compression”) and 11.5. (“Laps of reinforcing steel in tension and compression and mechanical couplers”).
However, mechanical couplers (Clause 11.5.6) are not covered in this paper.
Please note that background to the following clauses is provided in separate background documents:
- 11.4.7 “Anchorage of headed bars”;
- 11.5.4 “Laps using U-bar loops”:
- 11.5.5 Laps using headed bars;
- 11.6.3 “Minimum radius of curvature and straight length of tendons adjacent to anchorages”;
- 11.7 “Deviation forces due to curved tensile and compressive chords”..

11.4.2 Reasons for change

Anchorage of straight The entire clause on anchorages and laps in particular have been reworked addressing the significant number
bars of Systematic Review Comments (SRC) it received.
In the EN 1992-1-1:2004 approach, the anchorage length is calculated in three steps: (i) calculation of the
ultimate bond strength; (ii) calculation of the basic anchorage length and (iii) calculation of the design anchorage
length. In addition, equations for steps (i) and (iii) contain a high number of parameters, so that also for the
simplest cases, a relatively long calculation process is required for determining the anchorage length. In order
to enhance ease of use, the provisions have been organized in a different manner.

Enhancement of ease of use

The clause on anchorage length of straight bars starts with: Clause (1) a provision with some definitions followed
by Clause (2) a simple table allowing a fast determination of the anchorage length for common cases which are
clearly defined ( ≤ 32 mm for ribbed bars and  ≤ 14 mm for indented bars, sd = 435 MPa, cd ≥ 1.5∙ and a
correction factor for poor bond conditionsFor cases not complying with the requirements of Clause (2), or for a
more detailed calculation, Clause (3) allows the required anchorage length of straight reinforcing steel bars to
be determined directly by Formula (11.3) as a function of the geometrical and mechanical parameters including
coefficients for bond conditions in slipform construction and under bentonite or similar slurries. Clause (4) gives
the definition of so-called “good and poor bond conditions”.
Additional clauses may be used to reduce the anchorage length accounting for some favourable effects:
- Clause (5) presence of confinement reinforcement, transverse reinforcement or external pressure and
- Clause (6) for anchorages in compression.
Since the steel stress is directly considered in Formula (11.3) of FprEN 1992-1-1:2022, the beneficial effect of
lower stresses than fyd on the anchorage length is considered directly.
Clause (7) specifies transverse and confinement reinforcement, additional to that for shear, for anchorage of
bars 32 mm where transverse compression is not present. These rules were taken from EN 1992-1-1:2004.

11.4.2(3) Derivation of the direct design expression

Expression (11.3) Equation (C11.3) below is derived by simplification from fib MC2010 provisions and its background document
[C11.1]. Although the original equation is semi-empirical, this choice is justified as the reference has a pre-
normative character and it is currently accepted as state-of-the-art.
The original equation (refer to Eq. (3.2) in [C11.1]) gives the tensile stress in the reinforcement s which can be
developed by bond along the anchorage length as a function of the concrete compressive strength fc, anchorage
length ℓb, bar diameter ø, minimum and maximum covers cmin and cmax as well as the influence of transverse
reinforcement or transverse compressive stresses on the surface. Neglecting the latter effect and the influence
of cmax, the original equation becomes (based on mean values):

1 11 1 1
 fc  4  lb  20  25 mm  5  cd  4
 sd  54         (C11.1)
 25 MPa          
Where cd is defined in Figure C11.1a.

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 423
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

Figure C11.1: (a) definition of nominal cover cd; (b) non-uniform distribution of bond stresses and (c) nonlinear
distribution of reinforcement force in the development length compared to the usual assumption of a constant
bond stress along the development length (dotted blue lines)
It can be noted that the influence of the anchorage length ℓb is not linear as it accounts for the non-uniform
distribution of bond stresses along the development length as well as a strain effect (decreased bond strength
as a function of reinforcement strain [C11.2, C11.3], see figures C11.1b and c).
Equation (C11.1) can be rewritten to express directly the anchorage length as a function of the stress in the
reinforcement s and other relevant parameters:

11 1 1 1
 lb  20  25 MPa  4 s   5    4
   f      (C11.2)
   c  54 MPa  25 mm   cd 

or
5 20 4 5
 25 MPa 11  s  11   11   11
lb            (C11.3)
 fc   54 MPa   25 mm   cd 
As the original equation has a semi-empirical origin, it is reasonable to round the exponents of the design
equation to simpler fractions. In addition, reference values for sd,  and cd are adapted to correspond to the limit
values for the simple table (clause (2)). Finally, when switching from mean test values to design level, design
stresses and characteristic concrete strength are used and a calibration parameter klb is added to arrive at the
format used in Formula (11.3) in FprEN 1992-1-1:2022 (Note: coefficient for casting position kcp is not shown in
(C11.4), see text to Clause 11.4.2(4) below):
1 3 1 1
 25 MPa  2   sd  2    3  1.5  2
lbd  klb            (C11.4)
 fck   435 MPa   20 mm   cd 
Some of the terms in this equation have to be limited due to insufficient validation by testing (lack of experimental
data) or because of a mechanically justified threshold. According to [C11.1], the following limits should be
defined for calculation purposes:
- 15 MPa ≤ fcm ≤ 110 MPa: An upper limit has been introduced for fcm in FprEN 1992-1-1:2022 with the
ratio (25/fck)  0,3 since Formula (11.3) tends to underestimate anchorage length for concrete
strengths fcm  90 MPa. The lower limit has been neglected as it applies mostly to existing structures
(potentially to be accounted for in old existing structures);
- 0.5 ≤ cd/ ≤ 3.5: the lower limit is neglected because it is already covered by clause 6.5.2.3; the upper
limit is rounded to 3.75 (1.5  / cd ≥ 0.4);
-  ≥ 12.5 mm: according to [C11.1], this limit reflects experimental evidence (probably reflecting a low
relative rib area for small bars). Although, based on Figure 3-12 in [C11.1], this limit could probably
be removed, it has been fixed to /20 mm ≥ 0.6 in FprEN 1992-1-1:2022.
As shown in the figure C11.2 below, the rounding of the exponents of Eq. (C11.4) has only a low to negligible
influence on the results. The same diagrams show also the trends when compared to EN 1992-1-1:2004
(EC2:2004). It has to be observed that only the trends are compared, not the absolute values of the anchorage
lengths, as parameter klb in Expression C11.4 is subject to calibration against test data for the relevant design
situation, see below.

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 424
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

a) b)

c) d)
Figure C11.2: Influence of a) concrete strength, b) bar diameter, c) developed tensile steel stress and d)
concrete cover on the anchorage length (note: only the influence is compared, not the absolute values as
parameter klb in Expression C11.4 is subject to calibration)
Figure C11.2b) shows that according to fib MC2010 (and FprEN 1992-1-1:2022), the influence of bar diameter
on the anchorage length is more than linearly proportional (influence of the size effect on the bond strength).
With respect to the strain effect and the influence of non-uniform bond stress (neglected in EC2:2004), Figure
C11.2c) shows that according to fib MC2010 and FprEN 1992-1-1:2022, relative to the requirement for a bar
stress of 435MPa, the anchorage length can be substantially reduced for moderate and low stress levels. This
could explain why only few problems related to insufficient anchorage length have been reported so far
(anchorages and laps are in practice only occasionally subjected to high stress levels).

Comparisons
Figure C11.3 shows a comparison of the required anchorage length according to EC2:2004 and FprEN 1992-1-
1:2022 with klb = 50 (assumed parameters are defined in the figure caption). The required anchorage length at
reinforcing steel stresses sd  350 MPa is comparable. For higher reinforcing steel stresses, FprEN 1992-1-
1:2022 requires longer anchorage lengths. For sd = 435 MPa, the same required anchorage length as given in
EC2:2004 would be obtained for these parameters with klb = 39.
Characteristic strength of ribbed bars was around 400MPa when MC78, which forms the basis of current EC2
design rules, was drafted. The main steel Grade in current practice is now 500MPa, an increase of 25%. To
maintain a consistent level of safety with that when these rules were originally compiled, anchorages for Grade
500 bars should now be (500/400)1.5 times longer than for Grade 400 bars, according to Expression [C11.3],
corresponding to a 40% increase in anchorage and lap length. The increase in bond length required by current
EC2 provisions has been only (500/400)1.0, i.e. 1.25 times longer. With the introduction of higher Grade 500 bars
a further 12% increase in anchorage and lap lengths over current values should have been introduced to
maintain the level of safety intended when current provisions for bond lengths were formulated. Based on
[C11.1], the required further increase would be 20%. This should be kept in mind when interpreting Figure C11.3.
Also shown in Figure C11.3 is the required anchorage length according to prEN 1992-1-1:2021. The required
anchorage length is identical to the values of FprEN 1992-1-1:2022 for reinforcing steel stresses sd  435 MPa.
However, below that stress the required anchorage length was longer because an exponent on the reinforcing
steel stress of n = 1,0 was used in prEN 1992-1-1:2021.
For the actual calibration of klb versus test results as a function of the reliability factor , see specific paragraph
“Test results and calibration of klb against test data” below.

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 425
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

Figure C11.3: Comparison of the required anchorage length according to EN 1992-1-1:2004 and FprEN 1992-
1-1:2022 (fck = 30 MPa,  = 16 mm, cd/ =1.5, αct=1.0,C=1.5, good bond conditions and no confinement)

Design example
The anchorage length of previous example (fck = 30 MPa,  = 16 mm, cd/ =1.5, αct=1.0, C=1.5, good bond
conditions and no confinement) is designed for sd = 435 MPa:
According to EN 1992-1-1:2004 According to FprEN 1992-1-1:2022

This example shows that FprEN 1992-1-1:2022 is significantly easier to use. Already the first calculation (clause
(2)) allows for a fast and accurate result. If it were important, refinements using Clause (3) or Clauses (3) and
(5) allow for a reduction of the required anchorage length under certain conditions.

Strength equation and average bond strength

Eq. (C11.4), which is a design expression allowing a direct calculation of the required anchorage length, can be
reformulated as a strength equation, where the steel stress sR that can be activated at anchorage failure is
expressed as a function of the geometrical and mechanical parameters including lb :

2 1 2 1
 l  3  fck  3  20 mm  9  cd  3
 sR  435 MPa  b        (C11.5)
 klb     25 MPa      1.5 
The concept of bond strength fbd has been used in the past in order to dimension the anchorage length on the
basis of equilibrium and assuming a constant value of the bond stress throughout the development length:
2

fbd      lb   s (C11.6)
4
Conversely, Eq. (C11.6) allows also determining an average bond strength from Eq. (C11.5):

MPa mm
𝑓 ⋅ (C11.7)
⋅ MPa .
Or as a function of the steel stress sd:

MPa mm
𝑓 ⋅ (C11.8)
35 MPa MPa .
The physical meaning of this average strength is represented in Figures C11.1b and c (dotted blue lines). As
shown in Figure C11.4, the average bond strength is very sensitive to the anchorage length or to the maximum
steel stress. This confirms that this value cannot be used for dimensioning the anchorage length, or only
assuming a sufficiently conservative value which can be developed for long anchorages.

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 426
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

a) Bond strength as a function of anchorage length b) Bond strength as a function of maximum

shown for klb = 40 stress shown for klb = 40

Figure C11.4: Average bond strength as a function of (a) the anchorage length and (b) the maximum steel
stress (klb = 40)
Figure C11.4a also confirms that high average bond stresses may be developed in short anchorages. Assuming
for instance lb=5ꞏ (typical bond tests on short anchorages), the average bond strength becomes for klb = 40:

1 2 1
 f ck  3  20 mm  9  cd  3
f bd  5.4 MPa        (C11.9)
 25 MPa      1.5 
     
which gives higher values than generally assumed (dot in Figure C11.4a).

During the first hours after casting, fresh concrete behaves like a saturated soil [C11.4] and consolidates. Water
11.4.2(4)
in excess migrates from the bottom fresh concrete towards the top (so-called “concrete bleeding”) and the
Good and poor bond concrete settles (between 0 and 3 mm for a fresh concrete depth of 1 m). As a consequence of concrete bleeding
conditions – effect of and concrete settlement, porous concrete develops under the reinforcement bar and a void can even form in
casting position case the reinforcement bar is restrained to settle with the concrete [C11.5]. The water /binder ratio also increases
near the top of the fresh concrete and reduces tensile strength of the concrete cover. These phenomena, which
are more pronounced in the upper part of the concrete members, are associated to a loss of the bond properties
(effect known as “top bar effect” or “poor bond condition”).
EC2:2004 accounts for this effect with parameter 1 which applies to the design bond strength (1 =1.0 for good
bond conditions, 1 =0.7 for poor bond conditions). ACI 318 applies a similar factor (1.3), but to the development
length. A recent literature review [C11.6] has confirmed this effect, but some findings are still to be verified.
Recent tests have shown a significant reduction of the anchorage strength already for moderate depths [C11.7]
which is potentially related to high consistency class. The influence of the concrete consistency has been verified
by further tests, see clause “Test results” below. In addition, it has to be noted that the governing failure mode
observed in the [C11.7] tests is pull-out. Work is ongoing on the verification whether the effect of poor bond
conditions in case of splitting or spalling failures is the same.
For these reasons, in FprEN 1992-1-1:2022, the concept of EN 1992-1-1:2004 is kept, but its parameters have
been adjusted and recalibrated based on recent testing [C11.9]. A coefficient kcp has been introduced in Formula
(11.3) of FprEN 1992-1-1:2022 to account for good and poor bond conditions directly in the formula. The
coefficient takes the following values:
kcp = 1,0 for good bond conditions
kcp = 1,2 for poor bond conditions and for all bars used in slipform construction (unless it is shown that the vertical
bars cannot move during casting)
kcp = 1,4 for bars executed under bentonite or similar slurries (unless data is available for the specific slurry
used).
The factor kcp = 1.20 for ‘poor’ bond (accounting for the effects of casting on bond) is based on work by Suryanto,
Aitken and Cairns [C11.9], where it was concluded that for a member with overall depth greater than 250 mm a
top cast factor of 1/0.85 could be considered for concrete with a consistency S2 and 1/0.75 for S5 concrete. A
factor of 1.20 (=1/0.83) aligns with usual design assumption that S2 consistence (50 – 90 mm slump) concrete
will be used.
Slipform construction is aligned with ‘poor’ bond conditions. There are perceived issues of long unsupported
vertical bars and dragging of concrete over horizontal bars during construction. No test data is known.
The provision for bentonite slurry walls is adapted from MC2010 and Jones and Holt [C11.19]. This paper
advocates a factor of 1.4 - the contemporary factor pertaining to ‘poor’ bond.
Further recent testing [C11.10] has confirmed that indented bars show similar difference to ribbed bars between
good and poor bond conditions, see clause “Test results” below.

In FprEN 1992-1-1:2022, a clear distinction is made between confinement reinforcement (reinforcement

11.4.2(5)
perpendicular to the free surface and enclosing the bar to be anchored, see Figure C11.5a) and transverse
Reduction of the reinforcement (reinforcement perpendicular to the bar to be anchored, but parallel to the free surface, see Figure
anchorage length in C11.5b). The latter can provide confinement to the anchorage, only if it is placed between the bar to be anchored
presence of confinement and the free surface as shown in Figure C11.5b. In addition, the transverse reinforcement can be activated only
in case the reinforcement to be anchored is spaced not less than cs = 8cy (Fig. C11.5b), otherwise, a splitting

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 427
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

crack parallel to the free surface as shown in Fig. C11.5a becomes governing and the transverse reinforcement
cannot be activated. To account for this effect, the potential splitting cracks are shown in the code figure. The
parameter nc (number of legs of reinforcement providing reinforcement and crossing the potential splitting crack)
as well as parameters for transverse reinforcement in the design equation are defined in a different manner from
fib MC2010 and EN 1992-1-1:2004. An active pressure acting on the free surface (Figure C11.5c) can also
provide confinement and can be accounted for in a similar manner.

Figure C11.5: Definition of cases where the anchorage length may be reduced due to confinement provided
by (a) confinement reinforcement (nc=1, nb=2); (b) transverse reinforcement and (c) external pressure

fib MC2010 accounts for confinement by increasing the design bond strength (Eq. 6.1-21 of fib MC2010 with
slightly different symbols and definitions, influence of cmax neglected and influence of minimum transverse
reinforcement neglected as it is not to be considered in the calibration of klb):
⁄ .
𝑓 𝑘 𝜌 𝑓 , 2𝜎 2𝑓 , 0,4𝜎 𝑓 (C11.10)

Where ρconf is the ratio of the reinforcement providing confinement referred to the diameter of the bar to be
anchored or spliced:
∑ , 
𝜌 (C11.11)
⋅ ⋅ℓ ⋅ ⋅

ϕ is the diameter of the bar to be anchored or lapped;

ϕc is the diameter of the reinforcement providing confinement;
nc is the number of legs of reinforcement providing confinement and crossing the potential
splitting failure surface (see Figures C11.5a);
nb is the number of anchored bars or pairs of lapped bars in the potential splitting failure
surface;
sc is the spacing of reinforcement providing confinement along the bar to be anchored;
ccd is the design value of the mean compression stress perpendicular to the potential splitting
failure (see Figure C.11.5c))
kd is an effectiveness factor depending on the reinforcement detail:
- kd= 20 for confinement reinforcement crossing the potential splitting surface
and fulfilling the requirement of Figure C11.5a (net distance ≤5ϕ), value form fib
fib MC2010 §6.1.3.3
- kd= 5 for transverse reinforcement within the cover cy and fulfilling the
requirement of Figure C11.5b (cs≥8cy). This value is adapted from fib MC2010, Figure
6.1-14 b) in which the value of kd = 10 is associated with the transverse bar crossing two
splitting surfaces. So, for use in Exp (C11.10) above, kd = 10/2 = 5. (See Figure C11.5b)
- kd= 0 in other cases
In the approach of FprEN 1992-1-1:2022, the effects of confinement and transverse reinforcement on the
required anchorage length have been separated. Transverse reinforcement is considered in cd,conf as an increase
in the cover cy limited to 3,75 while confinement reinforcement together with compression stress are
considered in the term cd which is presented as follows:

2𝑡 𝑐𝑠
𝑐 , min 𝑐𝑥 ; 𝑐𝑦 25 ;
𝑠𝑡 2
; 3.75 𝑐𝑑  6 (C11.12)

where 𝑐 70𝜌 12 𝜎 ⁄ 𝑓 ⋅

Parameters 70 and 12 in cd have been determined based on calibration against test results. The limit of cd,conf
 6ϕ for cases with confinement reinforcement or with compression stress perpendicular to the potential splitting
failure was confirmed based on the same test results. Formula (C11.12) is considered much simpler to use with
little loss of accuracy. Figure C11.6 shows the influence of confinement reinforcement on the required anchorage
length for an example.

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 428
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

Figure C11.6: Required anchorage length as a function of the amount of confinement and transverse
reinforcement (fck = 25 MPa, bar to be anchored  = 20 mm, cd = 30 mm, klb = 50, good bond conditions and
designed for sd = 435 MPa, nb = 1, nc = 1 for confinement reinforcement)
Note: nb=1 and nt=1 was used to determine conf = (ntt2/4nbst) for transverse reinforcement; and nb=1 and
nc=1 was used to determine conf = (ncc2/4nbsc) for confinement reinforcement

For bars in compression, a reduction of the anchorage length by 15 may be taken into account if the end face
11.4.2(6)
of the bar is at a distance of at least 5 from the free concrete surface in the direction perpendicular to the end
Anchorage of bars in face of bar to avoid punching of the cover in front of the bar end face. However, an absolute minimum anchorage
compression length of 10 applies in FprEN 1992-1-1:2022.

These provisions for transverse and confinement reinforcement for the anchorage of straight bars   32 mm
11.4.2(7)
correspond to the provisions in EN 1992-1-1:2004, Clause 8.8(6). However, the reinforcing steel area As has
Transverse and been replaced by (2/4).
confinement
reinforcement for
anchorage of bars   32
mm
These provisions for anchorage of bundles are consistent with EN 1992-1-1:2004, Clause 8.9.1 and 8.9.2.
11.4.3
However, while EC2:2004, 11.4.3, implicitly assumed all bars within a bundle were of the same size. FprEN
Anchorage of bundles 1992-1-1:2022 makes provision for bundles comprising bars of mixed size.

Deductions for standard bends and hooks in tension are in accordance with fib MC2010 Cl 6.1.3.4 and 5.
11.4.4(1), (3)
Where cover is less than 3.5ϕ, the effect of bends and hooks on the anchorage capacity of bars in compression
Anchorage of bars with
should generally be neglected because of the risk of concrete spalling due to transverse deflections of the bar
bends and hooks in
tension and compression close to a bend or hook in compression. FprEN 1992-1-1:2022 permits an exception when all free concrete
surfaces, perpendicular and parallel to the bar, have a distance of at least 3,5 from the bar surface and end
face. If those distances are satisfied spalling is prevented and the same reduction of the anchorage length by
15 applies as for straight bars in compression, however, to the straight projected length to the bend or hook,
with the same absolute minimum anchorage length of 10.

Anchorage of bars with welded transverse reinforcement in current EN 1992-1-1:2004 has the same anchorage
11.4.5
length reduction factor  as given for standard hooks. Hence, in FprEN 1992-1-1:2022 the same reduction of
Anchorage of bars with the anchorage length by 15 is permitted as for bars with hooks under the specified conditions for the size and
welded transverse number of transverse bars, however, with an absolute minimum anchorage length of 5.
reinforcement
The minimum requirements for the strength of the weld are given in Annex C.4.1(3).

For U-bar loops subject to pure tension, i.e. both legs having the same tension force, anchorage may be
11.4.6
assumed to be provided at the end of the bends if the loop details (bends) comply with Clause 11.3 of FprEN
Anchorage of U-bar loops 1992-1-1:2022 for permissible mandrel diameters.
Alternatively, or for the case where the legs of the loop have different forces, the anchorage length may be taken
as the one of straight bars according to (11.3) but reduced by 20 but with an absolute minimum anchorage
length of 10. The bends have to comply with minimum permissible mandrel diameter.

See separate background document.

11.4.7
Anchorage of headed
bars in tension
The anchorage length of bonded post-installed bars is increased when compared with cast-in bars by dividing
11.4.8
the anchorage length according to FprEN 1992-1-1:2022, Clause 11.4.2, by a bond efficiency factor, kb,pi which
Anchorage of bonded may range between 0,7 and 1,0. kb,pi = 1,0 means an equivalent anchorage performance as for a cast-in bars.
post-installed reinforcing The product performance criteria for the bond efficiency factor kb,pi are given in Annex C.8.
steel
Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023
For citations, copying and referencing see Note on Front page! page 429
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

If, according to clause (4), laps are located away from sections where inelastic deformations in the reinforcing
11.5.2(4), (5)
steel (plastic hinges) are expected to occur, laps may be detailed with up to 100% of bars lapped at any section
Laps and they may be designed for the maximum value of the reinforcement design stress sd. The same applies for
laps with bars in compression.
In cases where the previous requirement is difficult to satisfy, clause (5) allows design of laps which are located
near sections where inelastic deformations in the reinforcing steel are expected to occur. In these cases, the
designer has basically three options:
i. a confinement reinforcement is arranged according to 11.5.2(8); or
ii. the laps are staggered; or
iii. the laps are designed for a design stress 1.2 times the design value of the reinforcement stress.
Exceptions i. and ii. are justified by the fact that confinement reinforcement and staggering the laps reduce the
brittleness of lap failure and better maintain residual strength after peak load [C11.8], [C11.15]. The percentage
of lapped bars allowed depends on the member type (35% for linear members and 50% for planar members).
This is justified by the increased redundancy of planar members.
The factor 1.2 applied in exception iii. is justified by the fact that plastic deformations of the reinforcement lead
to a reduction of the lap resistance. Alternatively, an increase by a factor of 1.2 results in reinforcement stresses
near the tensile strength.

The required lap length in FprEN 1992-1-1:2022 is now equal to the required anchorage length multiplied by a
11.5.2(2) and Table 11.3
factor kls=1,2 (NDP), i.e. a calibration coefficient of (klb x kls). This is different to fib MC2010 and is based on
Required lap length for recent individual, separate calibration of tests on anchorages and tests on laps. This recent calibration work also
bars in tension and uses a single exponent n = 1,5 for the entire range of reinforcement stresses sd in Formula (11.3). Since the
compression value of coefficient klb depends on the choice of n in Formula (11.3) both values klb and n are now NDPs.

Provisions of EN 1992-1-1:2004 have been simplified addressing several Systematic Review Comments (SRC).
11.5.2(8) and (9)
For compression laps, at least one transverse bar or confinement link should be placed at each end of the lap,
Reinforcement providing within a maximum distance of 50mm or 2, from the ends of the lap. This is required to control potential cracking
confinement in laps of due to the introduction of compression forces by end bearing from the bar into the concrete.
straight bars
Figure 11.12 has been modified showing the difference between transverse reinforcement (Figures 11.12a and
11.12b) and confinement reinforcement (Figures 11.12c and 11.12d). In addition, two options are given
(reinforcement providing confinement distributed over the lap length (a) and c)) or concentrated at the ends of
the spliced bars (b) and d)).

Design example
The tension lap length for the beam represented below with fck = 30 MPa,  = 25 mm, cnom = 25 mm, αct=1.0,
C=1.5, good bond conditions is designed for sd = 435 MPa assuming that the lap is located in an area with
potentially inelastic deformations in the reinforcing steel. It is assumed, that the laps are not staggered.

In a first step, the lap length is calculated neglecting the confinement provided by the stirrups:
According to EN 1992-1-1:2004 According to FprEN 1992-1-1:2022

Note: Table 11.1 may not be used here since cd <1,5

In a second step, the confinement effect provided by the stirrups is accounted for. The calculation according to
FprEN 1992-1-1:2022 is performed designing two options: first considering the given stirrups ϕ12@150 and as
a second option, by arranging sufficient confinement reinforcement according to 11.5.2(8) so that the lap may
be designed without increasing the design stress (1st option according to 11.5.2(5)).

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 430
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

According to EN 1992-1-1:2004 According to FprEN 1992-1-1:2022

For cases where bars   20 mm are lapped or the percentage of lapped bars is  25%, a minimum amount of
11.5.2(10)
transverse or confinement reinforcement should be provided which corresponds to the value specified in the
current EN 1992-1-1:2004.

The provisions for lapping bundles of bars of EN 1992-1-1:2004 are maintained with two modifications:
11.5.3
- there is provision for bundles of mixed diameter;
Laps of bundles
- there is no distinction anymore between laps with bars smaller and laps with bars equal or larger than
32mm diameter consistent with the remainder of FprEN 1992-1-1:2022;
- lap length of bundles with 2 or 3 bars may be calculated based on individual bar diameter either when
laps are staggered or when an additional “lap-bar” is added, [C11.16].

Calibrations were performed for FprEN 1992-1-1:2022 Formulae (11.3) and (11.4) at RWTH Aachen, [C11.22]
Test results and
and [C11.23], and at EPFL, [C11.24], each considering anchorages and laps separately. These calibrations
calibration of klb against
were based on test data compiled by ACI 408 supplemented by additional results added by TG2.5 [C11.17],
test data
from RWTH Aachen and EPFL for comparison with the strength predicted by FprEN 1992-1-1:2022 Formulae
(11.3) and (11.4). For these calibrations, the reliability-based methodology of Mancini et al [C11.13], was used
to analyse the data.
Calibrations at EPFL and at RWTH Aachen performed for anchorages found a mean value of coefficient klb,m =
22 and recommended for design klb = 50 and klb = 55, respectively, for a target reliability factor of  = 3,8, see
Figure C11.7. A coefficient of klb = 50 was retained as recommended values in FprEN 1992-1-1:2022.

a) Calibration at EPFL b) Calibration at RWTH Aachen

Figure C11.7: Test results for anchorages compared against theory mean values and design values

Figure C11.8 which is based on data from [C11.11] illustrates that the modifications in FprEN 1992-1-1:2022
Formula (11.4) for the effect of confinement reinforcement and compression stress in cd may be used up to
cd,conf = 6.

a) Range of cd,conf b) Effect of compression stress on anchorage

Figure C11.8: Test results for the effect of confinement reinforcement and compression stress on anchorage

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 431
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

Calibrations at EPFL and at RWTH Aachen performed for laps found mean values of coefficient klb,m = 32 and
klb,m = 33 respectively, and recommended for design klb = 60 and klb = 70, respectively, for a target reliability
factor of about  = 3,8, see Figure C11.9. A coefficient of klb = 60 was retained as recommended values in FprEN
1992-1-1:2022. Note: the coefficient for laps was renamed in FprEN 1992-1-1:2022 to the product (klb kls).

(a) Calibration at EPFL (b) Calibration atRWTH Aachen

Figure C11.9: Test results for laps compared against theory mean values and design values

Table C11.1 summarises the results of the above referenced calibrations for anchorages and laps in terms of
coefficient klb and (klb kls), respectively, as a function of the reliability index .
Mean Characteristic  = 3.8

Anchorages klb 22 35 50

Laps (klb kls) 32 45 60

Table C11.1: Coefficients klb and (klb kls) as a function of reliability index 

The effect of good/poor bond conditions (top cast factor) was recently investigated again, [C11.9] and [C11.12].
Effect of casting position
Main conclusions of this research were that the overall definition of good/poor bond conditions according to EN
1992-1-1:2004, maintained in FprEN 1992-1-1:2022, is reasonable. The coefficient kcp = 1,2 is reasonable or
slightly conservative for low slump concretes, however, slightly non-conservative for high slump concretes.
However, for practical reasons, a single value for kcp is maintained in FprEN 1992-1-1:2022, see Figure C11.10.
Further research by Cairns [C11.14] has demonstrated the top cast effect diminishes as lap length increases.
Results in Fig C11.10 were obtained from relatively short (lb/ϕ = 5) anchorages and are therefore conservative
for practical construction where a minimum lap length of 10ϕ applies.

Top cast factor in low slump concrete (C30, S2) Top cast factor in high slump concrete (C30, S5)

Figure C11.10: Top cast factor as a function of distance from top surface and concrete slump
Note: A comparison of top cast factors between ribbed and indented reinforcing steel bars is given in [C11.10].

The effect of fatigue loading on bond performance has been investigated e.g. in [C11.20]. It has been found that
Effect of fatigue loading the fatigue strength of bond as a function of the number of load cycles follows closely the fatigue strength of
concentrically loaded concrete under compression, see Figure C11.11. This means that no fatigue bond failure
will occur during several million load reversals if for the usual anchorage lengths required for reinforcing steel
bars the upper load (reinforcing steel stress) is smaller than about 50 percent of the static ultimate load of the
specific anchorage length. A previously applied cyclic load does not negatively affect the deformation behaviour
of the anchorage near failure or the ultimate load compared with monotonic load increase in a static bond test
providing failure does not occur during the fatigue test.

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 432
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

Figure C11.11: Fatigue strength of bond as a function of the number of load cycles, [C11.20]
Recent fatigue testing of bond [C11.21] has confirmed the above results and conclusions. Repeated loading
does not negatively affect the ultimate strength, provided that the anchorage does not fail under fatigue loading.
During fatigue loading the peak bond stresses move towards the unloaded end of the bar with increasing number
of load cycles, see Figure C11.12.

Figure C11.12: Distribution of steel stresses and strains over bond length at upper and lower load in fatigue
tests as a function of load cycles acc. to [C11.21].
Based on this favourable behaviour of anchorages, verification of anchorages and laps under fatigue loading
was not specifically covered in FprEN 1992-1-1:2022.

Comparison of lbd for A general trend for the comparison of anchorage lengths between EN 1992-1-1:2004, prEN 1992-1-1:2021 and
anchorages according to FprEN 1992-1-1:2022 can be observed already in Figure C11.3.
EN 1992-1-1:2004, prEN Figures C11.13 and C11.14 show detailed comparisons of design values of anchorage length, lbd, calculated
1992-1-1:2021 and FprEN according to EN 1992-1-1:2004, prEN 1992-1-1:2021 and FprEN 1992-1-1:2022 for klb = 50 in poor and good
1992-1-1:2022 bond conditions, respectively, for two values of cover cd and three different bar diameters.

sd = 200 MPa sd = 300 MPa sd = 435 MPa

Good bond, cd = 1.5

 = 12 mm

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 433
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

 = 16 mm
 = 25 mm

Poor bond, cd = 1.5

 = 12 mm
 = 16 mm
 = 25 mm

Figure C11.13: Comparison of anchorage length lbd between EN 1992-1-1:2004, prEN 1992-1-1:2020-11 and
FprEN 1992-1-1:2022 with coefficient klb = 50 for cd = 1,5 in good and poor bond conditions

sd = 200 MPa sd = 300 MPa sd = 435 MPa

Good bond, cd = 3.75

 = 12 mm
 = 16 mm
 = 25 mm

Poor bond, cd = 3.75

 = 12 mm

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 434
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

 = 16 mm
 = 25 mm

Figure C11.14: Comparison of anchorage length lbd between EN 1992-1-1:2004, prEN 1992-1-1:2020-11 and
FprEN 1992-1-1:2022 with coefficient klb = 50 for cd = 3,75 in good and poor bond conditions

FprEN 1992-1-1:2022 results in larger anchorage lengths compared with EN 1992-1-1:2004 for large diameter
bars and steel stresses in the order of 435 MPa or above, in particular in good bond conditions. Comparable or
smaller anchorage lengths result for steel stresses of sd  300 MPa. Significantly smaller anchorage lengths
are obtained with FprEN 1992-1-1:2022 for small bar diameters, small steel stresses and poor bond conditions
combined with large cover.

A general trend for the comparison of lap lengths between prEN 1992-1-1:2021 and FprEN 1992-1-1:2022 can
Comparison of lsd for
be observed in Figure C11.15: Lap lengths for steel stresses below 300 MPa are smaller due to the effect of the
laps according to EN
1992-1-1:2004, prEN unique exponent n = 1,5, above that value the lap lengths are larger due to the coefficient kls = 1,2. At a steel
1992-1-1:2021 and FprEN stress of 300 MPa the lap lengths of prEN 1992-1-1:2021 and FprEN 1992-1-1:2022 are identical.
1992-1-1:2022

Figure C11.15: Comparison of lap lengths lsd between prEN 1992-1-1:2020-11 and FprEN 1992-1-1:2022
Figures C11.16 shows detailed comparisons of design values of lap length, lsd, calculated according to EN 1992-
1-1:2004, prEN 1992-1-1:2021 and FprEN 1992-1-1:2022 for (klb kls) = 60 in poor and good bond conditions,
respectively, for a cover cd = 1,5 and three bar diameters.

sd = 200 MPa sd = 300 MPa sd = 435 MPa

Good bond, cd = 1.5

 = 12 mm
 = 16 mm
 = 25 mm

Poor bond, cd = 1.5

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 435
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

 = 12 mm
 = 16 mm
 = 25 mm

Figure C11.16: Comparison of lap length lsd between EN 1992-1-1:2004, prEN 1992-1-1:2020-11 and FprEN
1992-1-1:2022 with coefficient (klb x kls) = 60 for cd = 1,5 in good and poor bond conditions
Note: Plots are shown for 6 = 1.15 where up to 1/3rd of bars are lapped at a section and for 6 = 1.5 where 50%
or more of bars are lapped at a section
Except for large bar diameters with high steel stresses sd  435 MPa, lap lengths according to FprEN 1992-1-
1:2022 are smaller than those according to EN 1992-1-1:2004 with 6 = 1,5. For steel stresses sd  300 MPa,
lap lengths according to FprEN 1992-1-1:2022 are generally smaller or comparable to EN 1992-1-1:2004 even
for values 6 < 1,5.

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 436
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

[C11.1] fib, Bond and anchorage of embedded reinforcement: Background to the fib Model Code for
References
Concrete Structures 2010, fib Bulletin 72, 2014, 161 p.
[C11.2] Fernández Ruiz M., Muttoni A., Gambarova P., Analytical modelling of the pre and post-yield
behaviour of bond in reinforced concrete, ASCE Journal of Structural Engineering, Vol. 133, N° 10,
2007, pp. 1364-1372.
[C11.3] Fernández Ruiz M., Muttoni A., Gambarova P., A re-evaluation of test data on bond in R/C by
means of FEM modelling, Studies and researches, V. 27, 2007, pp. 113-134.
[C11.4] Clear C.A., Bonner D. G., Settlement of fresh concrete – an effective stress model, Magazine of
Concrete, 1988, pp. 3-12
[C11.5] Castel A., Vidal T., Viriyametanont K., François R., Effect of Reinforcing Bar Orientation and
Location on Bond with Self-Consolidating Concrete, ACI Structural Journal, 2006, pp. 559-567
[C11.6] Cairns J., Influence of casting position on bond strength, Joint meeting PT.SC2.T1 and AhG
Detailing, Lausanne, 10.1.2017
[C11.7] Metelli G., Marchina E., Plizzari G.A., Effect of casting position on bond, 4th Meeting of fib TG 2.5,
Stuttgart, 25.9.2017.
[C11.8] Metelli G., Cairns J., Plizzari G., The influence of percentage of bars lapped on performance of
splices, Materials and Structures, 2015, pp 2983–2996
[C11.9] Suryanto B., Aitken M., Cairns J., Influence of casting position on anchorage strength of steel
reinforcement, Heriot-Watt University, April 2020
[C11.10] Suryanto, Aitken & Cairns, Influence of Casting Position on Anchorage Strength of Indented Steel
Reinforcement by Heriot-Watt University, December 2020
[C11.11] Amin, Results from database compiled for anchorage at end supports, South Bank University,
London, 2009
[C11.12] Moccia F., Fernandez-Ruiz M., Metelli G., Muttoni A., Plizzari G., Casting position effects on bond
performance of reinforcement bars, fib Structural Concrete, Vol. 22, January 2021, 21 p.
[C11.13] Mancini G., Carbone V.I., Bertagnoli G., Gino D., Reliability-based evaluation of bond strength for
tensed lapped joints and anchorages in new and existing reinforced concrete structures, fib Structural
Concrete, Vol. 19, June 2018, pp. 904-917
[C11.14] Cairns J., Top cast effect: the influence of bond length Structural Concrete 2021;1–14 DOI:
10.1002/suco.202100376
[C11.15] Cairns J. Staggered lap joints of tension reinforcement. Structural Concrete, Vol 15 no. 1. March
2014. DOI 10.1002/suco.201300041
[C11.16] Cairns J., Lap Splices of Bars in Bundles. ACI Structural Journal. Vol 110, no. 2. Mar 2013
[C11.17] fib Task Group 4.5 "Bond models", Splice test database obtained from
http://fibtg45.dii.unile.it/files%20scaricabili/Database_splicetest%20Stuttgart %20sept%202005.xls;
2005
[C11.18] Micallef M, Vollum RL. The behaviour of long reinforcement laps in tension. Mag Concrete Research
2018; 70(14):739–55. https://doi.org/10.1680/jmacr.17.00285
[C11.19] Jones A, Holt D, Design of laps for deformed bars in concrete under bentonite and polymer drilling
fluids, the Structural Engineer, 2004, 82, No. 18, 32-38
[C11.20] Rehm G., Eligehausen R., Bond of ribbed bars under high cycle repeated loads, ACI Journal,
February 1979, pp. 297-309
[C11.21] Lemcherreq Y., Kaufmann W., Vogel Th., Fatigue of bond: Experimental investigation using pull-out
tests with distributed fibre optical sensors, Conference Paper https://doi.org/10.3929/ethz-b-
000465183, Institute of Structural Engineering, ETH Zurich, Switzerland, 2020-11
[C11.22] Hegger J., Camps B., Überprüfung der Bemessungsregeln für Übergreifungsstösse und
Verankerungen (gerade Stäbe, Abbiegungen und Haken) nach dem Entwurf der zweiten Generation
des Eurocode 2, Abschlussbericht, Institutsbericht Nr. 431/2020, IMB, RWTH Aachen, 22.12.2020.
See CEN-TC250-SC2-WG1_N1260_Mr Hegger PRB-PG2_0553
Hegger_Abschlussbericht_PRB_Uebergreifungsstoss-Teil_1_(202... and CEN-TC250-SC2-
WG1_N1258_Mr Hegger 2021-11-04_Presentation Professor Hegger_Anchorages and Laps_final
[C11.23] Camps B., Schmidt M., Hegger J., Anchorages and laps according to prEC2 – Review and
comparison to current EC2, fib Structural Concrete, publication under preparation.
[C11.24] CEN-TC250-SC2-WG1_N1284_Mr Ganz-Mr Muttoni - Latest technical changes - Clauses 11.4 & 11.5
- Joint WG1-SC2 meeting 2022-06-27-rev

List of versions of this document

1.2.2016, A.M. 1st draft (submitted to SC2/WG1, 9.5.2016)

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 437
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4 and 11.5 Background to FprEN 1992-1-1:2023

12.1.2017, A.M. 2nd draft (after discussion with AhG Detailing, 10.1.2017)
13.1.2017, A.M. Average bond strength implemented
30.11.2017, A.M., New template, completed with new provisions in prEN 1992-1-1:2017-10
09.02.2021, HRG Text updated to prEN 1992-1-1:2020-11
12.04.2021, HRG Implemented comments and background information of Ch. Goodchild and J. Cairns, added fatigue, amended Figure C11.6,
corrected anchorage length according to EN 1992-1-1:2004, added comparison of lap length, prEN 1992-1-1:2020-11
15.04.2021, HRG Replaced hand-written examples with printed version, prEN 1992-1-1:2020-11
05.10.2022, HRG Updated to FprEN 1992-1-1:2022
25.01.2023, HRG Implemented comments from J. Cairns
02.03.2023, HRG Added references [C11.23] and [C11.24]
27.03.2023 Peer review

Aurelio Muttoni / Hans Rudolf Ganz et al. 02.03.2023

For citations, copying and referencing see Note on Front page! page 438
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to a new proposal

for clause 11.4.7
Anchorage of headed bars

Aurelio Muttoni EPFL, Ecole Polytechnique Fédérale de Lausanne,

Switzerland
Alejandro Pérez Caldentey FHECOR, Spain
28.03.2022

Page 1/500
For citations, copying and referencing see Note on Front page! page 439
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4.7 Background to FprEN 1992-1-1:2023

Introduction This document describes the development of design formulae for the calculation of the anchorage resistance of
headed bars (clause 11.4.7) in FprEN 1992-1-1:2023.

Reasons for change New provisions on the anchorage of headed bars consistent with those for partially loaded areas have been
drafted in prEN 1992-1-1:2021, but several inconsistencies have been found with respect to the provisions of
EN 1992-4:
- the influence of compressive concrete strength on the anchorage strength is different (providing too
high resistance for high values of fck);
- the group effect (effect of spacing) is different;
- the effect of the distance to edges and corners is different.
For these reasons, a new proposal has been developed based on the same mechanical model which has been
used to draft the provisions for the minimum mandrel diameter ([C11.4.7.2], see also presentations and
discussions within CEN/TC250/SC2/WG1 in 2022). This choice is related to the fact that the two cases are
similar in terms of failure modes and parameters involved. This document describes the developed model as
well as the proposed provisions.

Similarities with the Figure C.11.4.7.1 shows failures of a headed bar near to the free surface and a similar failure of a bent
model developed for reinforcement due to the deviation forces. One can observe that in both cases, failure occurs due to spalling of
verifying the mandrel the concrete cover (so-called concrete blow-out failure).
diameter and
development of a
design formula for
headed bars

a) b)

Figure C11.4.7.1: (a) concrete spalling observed at the anchorage of headed bards (so-called blowout, tests by
Furche and Eligehausen, [C11.4.7.1]); (b) spalling failures due to an insufficient mandrel
diameter of bent bars near to a free edge (tests by Monney et al. [C11.4.7.2]).

This comparison shows that the mechanical model developed for the minimum mandrel diameter (see details in
[C11.4.7.2]) can be adapted to calculate the resistance of anchorages with headed bars. The main ideas and
assumptions are:
- The steel stress which can be activated in the anchored reinforcement bar without accounting for a
bond component corresponds to the triaxial concrete compressive strength fc,3 which can be activated
under the head multiplied by the ratio between the net area of the head and the cross-sectional area
of the reinforcement bar As :
2
A  As  h 
kh, A  h    1 (11.9)
As  
where Ah is the total Area of the head;
- The concrete compressive stress under the head can exceed the uniaxial compressive concrete
strength fc (which should be reduced by the strength reduction factor cc, see clause 5.1.6(1)), due to
the confinement effect of the surrounding concrete. It can be calculated as [C. 11.4.7.3]:

fc ,3  cc  fc  4   c ,1 (C11.4.7.1)

- The confinement concrete compressive stress c,1 in the confined area Ac,c parallel to the free surface
is in equilibrium with a confining tensile stress fct.,eff acting on the confining area Ac,t (see [C11.4.7.2])
and can be calculated as:

A. Muttoni + A. Pérez Caldentey 28.03.2022

For citations, copying and referencing see Note on Front page! page 440
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4.7 Background to FprEN 1992-1-1:2023

Act
 c ,1  f ct , eff (C11.4.7.2)
Acc
- The confined area Ac,c under the head (see blue area in Figure C11.4.7.3c) is assumed to be
proportional to the net head area Ah – As;
- The confining area Act where the effective tensile strength fct.,eff can act (see red area in Figure
C11.4.7.3c) is assumed to be proportional to the distance ad between the bar axis and the free surface
times the nominal head diameter h;
- Similarly to [C11.4.7.2], the effective concrete tensile strength fct.,eff is assumed to be proportional to
the square root of the compressive concrete strength and is reduced with a size effect factor which
depends on the aggregate size parameter ddg (defined in 8.2.1(4)) and the nominal size √∙h with an
exponent of 1/3:
13
 d dg 
f ct , eff  k1 fc   (C11.4.7.3)
   
 h 
where k1 is a calibration factor.
Considering that the force in the bar is equal to the force that can be transmitted to concrete through the head

area before failure (  s
'
 2   c ,1  Ah  As  ), the stress which can be activated in the reinforcement
4
becomes thus:
13
 d dg  a   A  As
  kh , A cc  f c  4  k1
'
fc   k2 d h  h
s
    A h  As
 2
 h  
4 (C11.4.7.4)
56 13
a d  h   d dg 
 kh , A cc  f c  k3 f c    
     
where k3 is the calibration factor replacing 16∙k1∙k2/ which can be determined, for the case of single bars near
to a free surface, on the basis of a comparison with laboratory tests ([C.11.4.7.1], [C.11.4.7.4], [C.11.4.7.5] and
[C.11.4.7.6]). As shown in the comparison of Figure C11.4.7.2 and Table C11.4.7.1, excellent results can be
obtained assuming k3=11,0. Only for the test series by Lee and Breen [C.11.4.7.6], a larger scatter can be
observed. This is due to the fact that for these tests, in some cases, the headed bars were anchored near to a
confinement reinforcement (spiral reinforcement in a pile) or adjacent to very dense longitudinal reinforcement
(both reinforcements not considered in the comparison).

Table C11.4.7.1: Statistical values of the ratio measured / calculated resistance (Rexp / Rcalc ) according to
Formula (C11.4.7.4).

Test series Avg. CoV n

De Vries, 1996, edge bars 1.07 9.3% 52
Furche and Eligehausen, 1991 1.08 8.3% 20
Hasselwander et al., 1977 1.08 14.1% 17
Lee and Breen, 1966 1.07 18.2% 12
all series 1.07 11.2% 101
all series without Lee and Breen 1.08 10.1% 89

A. Muttoni + A. Pérez Caldentey 28.03.2022

For citations, copying and referencing see Note on Front page! page 441
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4.7 Background to FprEN 1992-1-1:2023

2.00
1.80
1.60
1.40

Rexp / Rcalc
1.20
1.00
0.80
0.60
De Vries, 1996
0.40 Furche and Eligehausen, 1991
0.20 Hasselwander et al., 1977
Lee and Breen, 1966
0.00
1.00 1.50 2.00 2.50 3.00 3.50 4.00
a)
h / 
2.00
1.80
1.60
1.40
Rexp / Rcalc

1.20
1.00
0.80
0.60
De Vries, 1996
0.40 Furche and Eligehausen, 1991
0.20 Hasselwander et al., 1977
Lee and Breen, 1966
0.00
0 10 20 30 40 50
b) fc [MPa]

Figure C11.4.7.2: Comparison of Formulae (C11.4.7.4) and (C11.4.7.5) with tests series in ([C.11.4.7.1],
[C.11.4.7.4], [C.11.4.7.5] and [C.11.4.7.6]); (a) influence of head size and (b) influence of
concrete strength.
As shown in Figure C11.4.7.2a, excellent results are obtained also for very small heads despite the fact that
very high concrete stresses are activated under the head (about 15fc in the tests with h=1.24by Furche and
Eligehausen [C.11.4.7.1]).
It has to be noted that formula (C11.4.7.4) covers the failure mode related to spalling of the concrete cover (also
called “blow-out” failure). For large covers, very high stresses can be reached and local failure can occur due to
compression of the matrix in the concrete and collapse of pores [C.11.4.7.7]. For this reason, the concrete stress
fc,3 has to be limited to the maximum value:

f c ,3   part cc  fc (C11.4.7.5)

Based on the tests presented in [C.11.4.7.7], a value part of about 15 can be assumed for uncracked concrete
in the region of the head. Nevertheless, for the sake of simplicity (and despite the fact that the calibration factors
k3 of Formula (C11.4.7.4) and part of Formula (C11.4.7.5) refer to different failure modes and different
calibrations), the same value of part = 11.0 is assumed. In a similar manner as according to EN 1992-4:2018,
coefficient part shall be reduced for concrete cracked in the region of the head (part = 8.0 can be assumed for
this case).

Design value and The design value accounting for material, geometrical and model uncertainties can be determined based on the
design formula (11.8) approach of Annex A of FprEN 1992-1-1:2023 (see [C.11.4.7.8] for details). With respect to the first term in
based on reliability Formula (C11.4.7.4) linearly related to the uniaxial compressive concrete strength, the same partial safety factor
analysis as for uniaxial compression can be used (C = 1.5). For this reason, in the design formula, the expression cc∙fc
can be replaced by the design value fcd according to 5.1.6(1) which contains implicitly the sustained loading
effect. For the other term, which is associated to the concrete tensile strength, the uncertainties defined in Table
C11.4.7.2 below can be considered (see also Table A.2 of FprEN 1992-1-1:2023 and [C.11.4.7.8]). The
coefficient of variation and the bias factor of the resistance can be calculated with following formulae adapted
from Formulae (A.4) and (A.7) of FprEN 1992-1-1:2023:

A. Muttoni + A. Pérez Caldentey 28.03.2022

For citations, copying and referencing see Note on Front page! page 442
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4.7 Background to FprEN 1992-1-1:2023

2 2
 V fc   V ηis 
VR        V d2  V θ2 (C11.4.7.6)
2 
   2 
1
f 2
R   cm  ηis   d  θ (C11.4.7.7)
 fck 

Table C11.4.7.2: Statistical values of the uncertainties assumed to calibrate the partial safety factor

Coefficient of
Uncertainty Bias factor
variation

Compressive strength fc (control

Vfc = 0,100 fcm/fck = exp(1,645Vfc)
specimen)

In-situ factor ηis = fc,ais/fc Vηis = 0,120 μηis = 0,95

Geometrical uncertainty (cover) Vd = 0,100 μd = 1,00

Model uncertainty Vθ = 0,112 μθ = 1,07

Coefficient of variation and bias factor of

VR = 0,169 μR = 1,13
the resistance

The partial safety factor can then be calibrated according to Formula (A.1) of FprEN 1992-1-1:2023 with the
sensitivity factor R =0.8 and the target value for the 50-year reliability index tgt = 3.8:

exp  R   tgt  VR  exp  0.8  3.8  0.169 

R    1.48 (C11.4.7.8)
R 1.13
This means that the same partial safety factor as for concrete in compression (C = 1.50) can be used. The
design formula in FprEN 1992-1-1:2023 (based on Formulae (C11.4.7.4) and (C11.4.7.5)) becomes thus:
56 13
f ck ad  h   d dg 
 sd'  k h , A  f cd   part       k h , A  part  f cd (11.8)
C       

Bars near to a corner As described above, formula (11.8) has been derived for the case of a single headed bar parallel to a free
or group of bars surface (position defined by the distance ad = ay between the bar axis and the free surface) and sufficiently
distant from a corner or another bar. For this case, the failure mode related to spalling (blow-out) is shown in
Figures C11.4.7.3a-b whereas the confined and the confining areas are shown in Figure C11.4.7.3c.
For the case of a bar too near to a corner (Figure C11.4.7.3d), the confining area can be limited by the presence
of the corner, reducing therefore the resistance. An additional detrimental effect is related to the fact that the
failure mechanism is characterized by a possible movement of the spalled concrete portion (red vector toward
the corner in Figure C11.4.7.3d) which has a further reduction of the confining zone as a consequence. To
account for this case, in FprEN 1992-1-1:2023, the nominal value of the distance ad is defined in a different
manner (see Table C11.4.7.3). In addition, for elongated head plates near to a corner, failure can occur first
under the head portion near to the corner. To account for this case, the distance ax shall be defined according
to Figure 11.9c of FprEN 1992-1-1:2023. It has to be noted that for circular and square heads, the formula in
Table C11.4.7.3 refers to the case where ay is the smallest distance to an edge (ax≥ay , otherwise, ax shall be
replaced by ay and vice versa). For the case with an elongated head, both combinations should be verified and
for the case of headed bars near to more than one edge, all edges should be verified separately.
A similar reduction applies also for a group of bars along an edge with a distance sx smaller than 4ay (see Figure
C11.4.7.3e). For this case, the nominal distance ad is determined with Formula (11.10) (see Table C11.4.7.3)
which has been derived assuming a linear transition between the two limits: ad= ay for the upper limit with sx=4∙ay
(no interaction between bars) and the lower limit for sx=x. For the latter case, a safe value of ad can be calculated
assuming that the heads are in contact and act in a similar manner as a strip load parallel to edge (for this case,
since no tests with a large number of headed bars have been found in the literature, the factor k3 in Formula
(C11.4.7.4) has been calibrated on the basis of tests with strip loads parallel to an edge found in the literature).
For the less detrimental case of a group with a limited number of bars, the resistance may be calculated
according to EN 1992-4 which provides a more favourable reduction factor for two and three bars.
Please note that Formula (11.10) in FprEN 1992-1-1:2023 contains an editorial error: coefficient 2,3 should read
0,23 as stated in Table C11.4.7.3 (using a value of 2,3, ad would become larger that ay which is physically not
possible).

A. Muttoni + A. Pérez Caldentey 28.03.2022

For citations, copying and referencing see Note on Front page! page 443
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4.7 Background to FprEN 1992-1-1:2023

1 potential crack associated to spalling (blow-out) or splitting;

2 confined area (with transversal compression);
3 confining area (with transversal tensile stress)
Red vectors: assumed movement of the concrete cover at failure

Figure C11.4.7.3: Failure mechanisms, confined areas and confining areas for: (a-c) single bar; (d) corner bar
and (e) group of bars

Table C11.4.7.3: Definitions of nominal distance ad and limits of applicability

Case Limits of applicability Definition of nominal distance ad

- Single bar
near to an
edge
- Corner bar ax  2a y  1.2h ad  a y
with
- Multiple s x  4a y
bars with
Corner bar a x  2 a y  1.2 h a d  0.5 a y  0.25 a x  0.3 h

Group of h  sx  4 a y s x  h    4a y  s x  1 
ad  a y  0, 23  a y  h  4a    1   (11.10)
bars along
an edge 4a y   h  2  y h   h  
2



Headed bars with a In case of headed bars consisting of ribbed or indented reinforcement, a combination of the force carried by the
bonded length, head and the bond strength developed along the anchorage length may be accounted for according to 11.4.7(3).
11.4.7(3) In this case, the required anchorage length lbd to carry the remaining stress sd-’sd defined in Figure 11.9a may
be calculated according to Formula (11.11):


lbd  1,1 lbd  sd   lbd  sd'   (11.11)

where coefficient 1,1 accounts for:

- the detrimental interaction between the tensile splitting stresses generated by bond and by the head;
- the fact that the maximum bond strength can develop for a bar slip which is smaller than the head
penetration required to activate the head resistance.

A. Muttoni + A. Pérez Caldentey 28.03.2022

For citations, copying and referencing see Note on Front page! page 444
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.4.7 Background to FprEN 1992-1-1:2023

The effect of adjacent In some practical cases, external compressive forces (e.g. from supports) or forces due to opposed active
plates and load headed bars (e.g. corner joint, see for instance Figure C11.4.7.4) can act in the zone of the head providing a
introduction points, beneficial confinement of the head and preventing blow-out failures. In the cases reported in [C.11.4.9] and
11.4.7(5) [C.11.4.7.10] for instance, a significant increase of the resistance has been attained. Since additional research
needs to be conducted to develop general provisions and detailing rules are still missing (position of the external
load with respect to the head, relative positions between heads, etc.), the provision according to 11.4.7(5) opens
the possibility to design these cases using a more refined model without giving detailed design rules.

Figure C11.4.7.4: Confinement effect by opposite plates in case of the reinforcement of a frame corner with a
closing moment anchored with heads (from Marchetto et al. [C. 11.4.7.10])

References [C. 11.4.7.1] Furche, J., Eligehausen, R., “Lateral Blow-out Failure of Headed Studs Near a Free Edge”;
Anchors in Concrete-Design and Behavior, ACI SP-130 81991), p. 235–252.
[C. 11.4.7.2] Monney, F.; Fernández Ruiz, M.; Muttoni, A., “Design against splitting failures in reinforced
concrete due to concentrated forces and minimum bend diameter of reinforcement”,
Engineering Structures, 245 (2021) , p. 21
[C. 11.4.7.3] Nielsen M.P, Hoang L.C., “Limit Analysis and Concrete Plasticity”, CRC Print, 2011, p. 816.
[C. 11.4.7.4] De Vries R.A., “Anchorage of Headed Reinforcement in Concrete”, Dissertation, The University
of Texas at Austin, 1996, p. 314
[C. 11.4.7.5] Hasselwander G.B., Jirsa J.O., Been J.E., Lo K., “Strength and Behavior of Anchor Bolts
embedded near edges of concrete piers”, Research report 29-2F, Center for Highway Reearch,
The University of Texas at Austin, 1977, p. 142
[C. 11.4.7.6] Lee D.W., Been J.E., “Factors Affecting Anchor Bolt Development”, Research report 88-1F,
Center for Highway Reearch, The University of Texas at Austin, 1966, p. 66
[C.11.4.7.7] Lieberum K.H., Reinhardt H.W., “Strength of Concrete on a extremely small bearing area”, ACI
Structural Journal, 1989, p. 67-75
[C.11.4.7.8] Muttoni A., “Background document to clause 4.3.3 and Annex A, Partial safety factors for
materials”, Background Document to FprEN 1992-1-1:2023, CEN/TC 250/SC 2, 23.02.2023, p.
21
[C. 11.4.7.9] Marchetto, F. Use of headed reinforcement bars in construction. A theoretical approach to
determine the dimensions of anchorage plates and experimental tests on knee joints subjected
to closing moment. PhD Thesis. Technical University of Madrid, 2015.
[C. 11.4.7.10] Marchetto, F., Pérez Caldentey, A. and Corres Peiretti, H. (2016), Structural performance of
corner joints subjected to a closing moment using mechanical anchorages: an experimental
study. Structural Concrete, 17: 987–1002. doi:10.1002/suco.201500222.

List of versions of this document

2022.03.06, A.M. + A.P.C. 1st draft

2022.03.22, A.M. + A.P.C. 2nd draft
2023.03.28, A.M. + A.P.C Substantial revision of the document

A. Muttoni + A. Pérez Caldentey 28.03.2022

For citations, copying and referencing see Note on Front page! page 445
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1: 2023

Background document to subsection 11.5.4

Laps using U-bar loops

Henrik B. Joergensen University of Southern Denmark

Linh C. Hoang Technical University Denmark

05.05.2021

For citations, copying and referencing see Note on Front page! page 446
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Introduction
This document describes the theoretical and experimental background of the simple equation proposed for the
next generation of EC2 [C11.2.1] for calculation of the capacity of U-bar loop connections subjected to
combinations of normal forces and bending moments. The proposal is based on an analytical upper bound
plasticity model [C11.2.4, C11.2.5, C11.2.6] which has been verified with a large number of available test
results.
U-bar loop connections (Figure C11.2.1) are frequently used to establish structural continuity between precast
concrete bridge decks. Here, U-bars protruding from each side of the deck elements to be connected are
overlapped in the connection, such that they confine a flattened cylindrical concrete core. After placement of the
precast decks, transverse reinforcement bars are placed within the overlapping U-bars and the connection is
subsequently grouted with concrete or mortar. In many applications, U-bar loop connections will often be
subjected to pure tension or tension in combination with a bending moment.

Figure C11.2.1: Illustration of a U-bar loop connection loaded in combined tension and bending; Plane
illustrations (a, b) and cross sectional illustration (c).

General

The model is based on the upper bound theorem of limit analysis (theory of plasticity). The original formulation
by Joergensen [C11.2.5] was developed following a two-step procedure. The first step involves a method to
design the connection in such a way that premature failure in the concrete at the joint is not governing. The
second step is necessary to ensure that the capacity related to failure in the concrete of the joint is not entirely
depending on the tensile strength of concrete. This eventually leads to the formulation of a procedure to
determine the minimum transverse reinforcement that is required to prevent brittle failure of the concrete. Based
on the analytical results obtained for the case of pure tension, it is possible to extent the model to cover the
case with combination of tension and bending moments.

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 447
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Calculation of the pure tensile capacity related to concrete failure

Symmetric connections
With inspiration from the failure mode observed in tests, a rigid-plastic upper bound model for the pure tensile
capacity of loop connections sensitive to concrete core failure has been developed; see [C11.2.4]. The model
was developed for a so-called symmetric two-to-three connection, i.e. two U-bars overlapped to three U-bars as
used in test specimens, [C11.2.5]. However, the model can be generalised simply by expressing the result as a
load carrying capacity per U-bar.
The assumptions of the model are as follows: The concrete is treated as a perfectly rigid-plastic Modified
Coulomb material that respects the normality condition of plasticity theory. The tensile strength of concrete is
neglected and the reinforcement bars are treated as rigid-plastic components only capable of carrying axial
tension (i.e. dowel action is neglected). Since the aim is to model the case of concrete core failure, it is also
assumed that the U-bars are sufficiently strong such that the ultimate load is not governed by yielding of U-bars.
The failure of the concrete core in a loop connection is complex and involves the development of three-
dimensional fracture surfaces. Therefore, to develop a simple and design-oriented solution, a simplified
mechanism is considered. The mechanism is shown in Figure C11.2.2(a), which shows a connection loaded in
pure tension. The mechanism consists of (four) plane, diagonal sliding yield lines, which develop between the
looped ends of adjacent U-bars and of (three) separation yield lines running from the ends of the U-bars toward
the nearest joint surface. In addition, there are two separation yield lines at the boundaries of the connection.
The system of yield lines divides the joint concrete into a number of segments, ①, ② and ③, which undergo
rigid body displacements as illustrated in Figure C11.2.2(b).
The displacements of the rigid parts may be described as follows: The upper precast element and segment ①
are at rest. The lower precast element moves by the vector uelement = (un, 0). The segments ② and ③ to the
right of the line of symmetry are displaced by u② = (0, 2ut) and u③ = (un, ut), respectively. Here, un and ut refer
to displacements in the n- and t-direction. The displacement field does not require elongation of the U-bars.
On the basis of the assumed failure mechanism, the external work and the internal work can be calculated, see
[C11.2.5, C11.2.6]. This, eventually, leads to the determination of an upper bound for the load carrying capacity.
The optimal upper bound solution is found when the relative displacement in each diagonal yield line forms the
following angle  with the yield line:

 
 2T 
 1
    Arcsin    (C11.2.1)
2 
 a 
   1 
 H 
Here,  = Arctan(a/H), where the parameters a and H are shown in Figure C11.2.2(a). Further, in Eq. (C11.2.1),
 is the effectiveness factor (see Eq. (C11.2.4) below) andT is the mechanical ratio of transverse
reinforcement:

AsT f yT
T  (C11.2.2)
Ac fc
where fyT and AsT are the yield strength and the total cross sectional area of the transverse reinforcement,
respectively. Additionally, fc refers to the uniaxial compressive strength of the joint concrete while Ac refers to the
cross sectional area of the concrete core confined by the overlapping U-bars. Further discussions of Ac will be
provided in the next sub-section.
With  given by Eq. (C11.2.1), the optimal upper bound solution, Nc, appears as shown in Eq. (C11.2.3a), see
[C11.2.5]. It can be noted that the angle determined from Eq. (C11.2.1) is only a valid solution if it is not smaller
than  as well as the internal angle of friction for concrete,  = Arctan(3/4). Otherwise,  must be taken as the
larger of  and . This restriction has been discussed in details in [C11.2.5]. When inserting the two restrictions,
solution (b) and (c) in Eq. (C11.2.3) is derived. The subscript “c” indicates a capacity related to concrete failure:

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 448
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

 2 
 4 T  1  T    a   a  ;if    and    (a)
    
  H   H 
 
  2

 2  T  3  4 a    a   1
    H   H  
Nc      a 3
 nL    ; if    and  (b)
 f c Ac  43
a H 4
 H

a 
2
  a  a 3
  H   1  H  ;if    and  (c)
H 4
   

(C11.2.3)
Here, nL is the number of U-bars available for tension transfer. The above mentioned effectiveness factor, ,
appears in Eq. (C11.2.3) as a reduction factor on fc. This is necessary in order to account for the fact that
concrete is not a perfectly plastic material [C11.2.8]. For the problem treated here,  may be calculated as
follows, [C11.2.5]:

0.88  1 
 1   (f c in MPa , H in m) (C11.2.4)
fc  H
According to Eq. (C11.2.1),  =  + /2 for T = 0. This means that Eq. (C11.2.3a) is valid when T = 0.
Consequently, the model predicts zero capacity when a loop connection does not contain transverse
reinforcement. This result is naturally conservative with respect to reality as, even without transverse bars, loop
connections are capable to carry some level of tension forces.
The tensile capacity of connections without transverse reinforcement is treated in section 2.3.

Figure C11.2.2: Failure mechanism in loop connection loaded in pure tension; yield lines (a) and rigid body
displacements (b).

Non-symmetric connections
U-bar loop connections may also have non-symmetric layout as shown in Fig. C11.2.1b, where the spacing
between overlapping U-bars is not constant but alternates between a1 and a2. It has been shown [C11.2.5] that
the capacity of non-symmetric connections can be determined on the basis of the mechanism illustrated in Fig.
C11.2.3, which is very similar to the one in Fig. C11.2.2. The analysis shows that the capacity of non-symmetric
connections is always larger than the capacity calculated on the basis of the solution for symmetric connections
(i.e. Eq. (C11.2.3)), provided that a  (a1  a2 ) / 2 is inserted. Hence, use of Eq. (C11.2.3) with

a  (a1  a2 ) / 2 will give conservative estimates for the capacity of non-symmetric connections.

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 449
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Figure C11.2.3: Failure mechanism in loop connection loaded in pure tension; yield lines (a) and rigid body
displacements (b).

Theoretical basis to determine the minimum amount of transverse reinforcement

Loop connections without transverse reinforcement are able to carry tension loads because of the tensile
strength of concrete. However, the failure in this case is brittle. Hence, in order to ensure a ductile failure mode,
even in the case where fracture of the joint concrete is governing, it is necessary to have a layout where the
resistance from the transverse reinforcement is larger than that relying on the tensile strength of concrete. This
eventually means that the connection must at least be provided with a minimum amount of transverse
reinforcement. The minimum transverse reinforcement ratio can be determined by requiring that a solution
based on the tensile strength of concrete, but without transverse reinforcement, must not lead to a larger
capacity than Eq. (C11.2.3).
At first glance, it might be tempting simply to include the tensile strength of concrete when developing Eq.
(C11.2.3). This should make the model valid for connections without transverse reinforcement as well.
Unfortunately, the problem is not that simple. In a general solution, it is unsafe to assume yielding in the
transverse reinforcement and at the same time take into account the tensile strength of concrete. The reason is
the pronounced softening behaviour of concrete in tension. Due to softening, the concrete tensile strength in the
yield lines will normally vanish at the time where yielding of the reinforcement occurs. However, if the
connection is lightly reinforced in the transverse direction, the cracking load might be higher than the load
carried by the connection when the transverse bars are at yield. This is analogue to the case of the ultimate
bending moment of concrete cross sections with small reinforcement ratios. In such cases, the cracking moment
might be larger than the moment determined by assuming yielding of the reinforcement.

Figure C11.2.4: Failure mechanism in connection without transverse reinforcement; rigid body displacements
(a), relative displacement in yield lines (b) and concrete areas included in the model (c).

Based on the above discussion, it appears that a solution for connections without transverse reinforcement
should be developed by taking into account the tensile strength of concrete. As for the case of the cracking
moment of a cross section, the solution to be developed in this section may be regarded as a lower limit for the
tensile capacity of loop connections. This result can then be used to calculate the minimum transverse

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 450
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

reinforcement.
Figure C11.2.4 shows a failure mechanism that is similar to the one in Figure C11.2.2. This time, however, there
is no transverse reinforcement. Based on the same displacement field as assumed previously, the internally
dissipated energy and the external work may be calculated, see [C11.2.6]. However, in this calculation, the
separation yield lines which develop parallel to the n-axis should also be considered. The relative displacement
in the yield line between segments ② and ① is u2-1 = u② ‐ u① = (0, 2ut). For the yield line, which coincides
with the line of symmetry, the relative displacement, u3-3, is also (0, 2ut). This is justified because the two
segments ③ both move downwards by un but away from each other by ut. In this case, it is apparent that the
angle between the yield line and the relative displacement vector is /2. These yield lines are therefore called
separation yield lines, while the diagonal yield lines experience a combination of separation and sliding.
Lastly, there are the two yield lines running from the looped end of the outermost U-bars towards the nearest
boundary of the loop connection. However, there are no geometrical restrictions for the inclination of these two
yield lines. Consequently, the yield lines can always be oriented in such a way that they appear as separation
yield lines with a minimum of energy dissipation. The contribution to the load carrying capacity from these two
yield lines is a boundary effect, which can be neglected for loop connections in practice, where the number of
overlapping U-bars is high. Therefore, keeping in mind that the developed solution should be suitable for a
generalization (i.e. the result can be expressed as a capacity per U-bar) the effect of the two boundary yield
lines will not be taken into account in the following.
Now, on the basis of the assumed failure mechanism, the external work and the internal dissipated energy
(including the contribution from the separation yield lines) may be calculated, see [C11.2.6]. This leads to an
upper bound solution for the load carrying capacity. The optimal upper bound solution is found when the relative
displacement in each diagonal yield line forms the following angle with the yield line:

 
 
  2nL 1 Ac*  m  m  1  (C11.2.5)
    Arcsin    l  1  l  
 L n Ac    a
2

   1 
 H 
The parameters l and m are defined as follows:

 t ft sin   t ft 1
l 1 2 , m 1 2 (C11.2.6)
 fc 1  sin   fc 1  sin 
Here, ft is the uniaxial tensile strength of concrete and t denotes the effectiveness factor for concrete stressed
in tension. The purpose of this factor is similar to the purpose of  as mentioned in the above. The following
formulas, which also have been used in other plastic analyses (see [C11.2.8]), are adopted:
0.3
H
ft  0.26 fc2/3 , t  0.6  , (fc in MPa, H in meters) (C11.2.7)
 0.1 
The same restrictions as mentioned for the previous solution apply also to the angle determined from Eq.
(C11.2.5) (i.e. valid only if not smaller than  as well as . With these restrictions and the angle in Eq.
(C11.2.5), the optimal solution for loop connections without transverse reinforcement may now be written as
follows (the subscript “ c,T = 0 ” indicates a tensile capacity related to concrete core failure when T = 0):

  2 2 
  a  2nL  1 Ac*  m  m  a m
l
   H   1     1     ; if    and    a
 nL Ac  l  l  H l 
  

  2n  1  a   m  Ac *  2  
a  m
 l L  3  4   1        1   5  3  
N c , T  0   nL  H  l  Ac
 
H

l 
 a 3
 nL  ; if    and  b 
 f c Ac  43
a H 4
 H
  2 
  a m a  a 3
l  1   H   l H  ; if    and  c
    H 4



(C11.2.8)
Where Ac/cos is the surface area of one diagonal yield line and Ac is the surface area of one separation yield
*

line. Strictly speaking, Ac* should be taken as ½(b  H)t while Ac should be taken as H.t in order to make the
mechanism geometrically possible. However, as discussed in [C11.2.5] on the basis of the observed failure
pattern, it is reasonable for practical purposes to disregard the contribution of the cover. Therefore, as for the

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 451
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

previous solution (Eq. (C11.2.3)), the concrete cover will be neglected when determining the surface areas of
the yield lines. This means that Ac and Ac* are taken as the areas illustrated in Figure C11.2.4(c), i.e.:

 
 D  2L    H  D  2L  D  2L  ,
2
Ac  (a) 
4
 (C11.2.9)

1 
Ac*   b  H  D  2L  , (b) 
2 

Here, the formula for Ac is only correct if H  D  2L , where L is the cross sectional diameter of the U-bar.
This assumption is in accordance with all available experimental data (see [C11.2.5]). On the basis of the
empirical background, the solution in this document is therefore limited to connections with H  D  2L .
Figure C11.2.5 shows the two solutions as a function of the transverse mechanical reinforcement ratio. From
Figure C11.2.5, it appears that the strength is governed by Eq. (C11.2.8) when the mechanical ratio of
transverse reinforcement is low. On the other hand Eq. (C11.2.3) is governing for large mechanical ratios of
transverse reinforcement. Hence, to ensure that the strength is not governed by the tensile strength of concrete,
the mechanical ratio of the transverse reinforcement must be larger than the ratio corresponding to the
intersection of the two curves. Figure C11.2.6 shows the mechanical ratio of the transverse reinforcement that
corresponds to the intersection point of the two curves. The calculations have been carried out by varying a
number of important design parameters.

Figure C11.2.5: Tensile capacity versus the effective mechanical ratio of transverse reinforcement. Parameters
*
used in example: a / H  0.5 , Ac / Ac  0.15 , m / l  0.80 , l  0.77 , nL  2

Figure C11.2.6: Effective mechanical ratio of transverse reinforcement T,min/, corresponding to

N c , T  0  N c

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 452
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Simplification of Equation (C11.2.3)

As mentioned in the general derivation of the design equations, there are some geometrical restrictions to the
angle of relative displacement, . Due to the normality condition,  cannot be lower than the angle of friction, .
In many cases, however, this restriction is not governing. Furthermore, it is in fact on the safe side to completely
disregard this restriction because this would be the same as treating the problem as a plane stress problem.
Hence, by assuming plane stress conditions, the angle  should only be larger than . This lead to the following
simplification of equation (C11.2.3):

  a 
2
 nL f c Ac  4 T   T   a   1
1 
 H  ;if T  (a)
        H  2
  
Nc  
  2
a  T
 a  1
 nL f c Ac  H  1  ;if  (b)
    H  2
   (C11.2.10)
Figure C11.2.7 shows the difference between Eq. (C11.2.3) and the simplified solution, Eq. (C11.2.10). The
difference between the two solutions comes from the contribution to the internal work of the transverse
reinforcement when the mechanical ratio of transverse reinforcement is high. It can be noted that, in practice,
the mechanical degree of transverse reinforcement is seldom so high that the difference between Eq. (C11.2.3)
and Eq. (C11.2.10) becomes significant.

Figure C11.2.7: Tensile capacity of loop connection versus effective mechanical ratio of transverse
reinforcement, T /  .

It appears from the solution (and Figure C11.2.7) that the ratio a/H is very important for the tensile capacity of a
U-bar loop connection. The largest ratio in all tested specimens found in the literature is a/H = 100/170 ≈ 0.6
(see [C11.2.5]). It is therefore recommended to use only the solution in this document for connections with
a / H  0.6 .
Eq. (C11.2.10) can be reformulated into the following format:

 2
a 
 a 
N c  nL f c Ac  r     
 H H
  (C11.2.11)
where r may be called the resistance factor of the transverse reinforcement and given by:

 4 T  T  T 1

  1    ;if   2 (a)
r  
1  1
;if T  (b)
  2 (C11.2.12)
Based on the simplifications introduced, the complete solution for the tensile capacity, Nu, may be formulated as
follows:

 N c (a)
N u  min 
 N y  nL AsL f yL (b)
(C11.2.13)
Where:
Nc is the tensile capacity when concrete failure is critical, see Eq. (C11.2.11)

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 453
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Ny is the yield capacity of one U-bar (both legs)

nL is the number of U-bars available for tension transfer
AsL is the cross sectional area of a U-bar (total area of both legs)
fyL is the yield stress of the U-bars

Simplified calculation of minimum amount of transverse reinforcement

According to the previous sections, a minimum amount of transverse reinforcement within the overlapping U-
bars should be provided to avoid brittle failure. However, determining the minimum reinforcement is tedious as
this requires an iterative process where Eq. (C11.2.3) must give the same result as Eq. (C11.2.8).
From a parametrical study, it can been found that the minimum amount of transverse reinforcement can be
approximated by the following explicit expression:

(C11.2.14)
Where the effective tensile strength ftef = tft is determined by Eq. (C11.2.7). Fig. C11.2.8 shows the comparison
of the minimum reinforcement found by the iterative procedure and found by use of Eq. (C11.2.14) for a range
of parameters that are often used in practice.

Figure C11.2.8: Comparisons of approximate and accurate calculations of the minimum amount of transverse
reinforcement. Parameters used in calculation (when not varied): a  100 mm, H  170 mm, D  110 mm,

 L  20 mm, Ac*  3000 mm2, m / l  0.79 , l  0.76 , nL  2 .

It should be noted that the area Ac* (the area of the separation yield lines outside the concrete core, see Figure
C11.2.4c) plays an important role when determining  min . For the calculations shown in Figure C11.2.8, the
*
area, Ac , corresponds to a width of the connection equal to b  H  2  20 mm.

Extension of model to connections subjected to combinations of tension and bending

Tests with combined tension and bending [C11.2.5 & C11.2.6] show that the yield lines developed in the
concrete core are very similar to those formed in pure tension tests. This suggests that concrete segments in
the core are pulled out (just like in the pure tension tests). The similarity with pure tension tests is further
supported by strain measurements, which show tensile strains in both legs of the U-bars at the ultimate state for
connections loaded in pure bending [C11.2.5]. Based on the observed similarities, a simple model for
connections loaded in combined tension and bending may be established. The model draw on the theoretical
results for the case of pure tension. The model is briefly explained below and can be used to determine the N-M
interaction diagram for loop connections. For a more detailed explanation, see [C11.2.5, C11.2.6].

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 454
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Figure C11.2.9: Loop connection subjected to combined tension and bending; sectional forces (a and b) and
sandwich model for distribution of internal forces (c).
Figures C11.2.9(a) and C11.2.9(b) illustrate a loop connection subjected to a tension force, N, and a bending
moment, M (the figure shows a connection with nL = 2). The set of sectional forces, (N, M), is assumed to be
statically equivalent with the distribution of internal forces in the sandwich model shown in Figure C11.2.9(c).
The sandwich model consists of two layers. The first layer has thickness x and carries a uniaxial compression
stress field with stress magnitude fc. The stress resultant of this stress field is denoted C = fcLx in the figure. The
second layer embeds the U-bars and has the thickness (D+2L). This layer transfers a tensile force, F, which, as
shown in the figure, is assumed to be:
F  C   Nu ,   [0;1] (C11.2.15)

where Nu is the pure tensile strength of the connection as determined by Eq. (C11.2.13). Based on the results
for the case of pure tension, it is possible to state that F can be transferred if F ≤ Nu. Remember that in the pure
tension model, the concrete cover was neglected. The cover can therefore here be used as the layer that
carries the uniaxial compression field. Keeping the picture of a sandwich model in mind, it is clear that the
thickness x of the compression layer cannot be larger than the cover c. Hence, there is an upper limit for the
compression force, namely C ≤ fcLc. The two mentioned conditions, i.e. C ≤ fcLc and F ≤ Nu, can be condensed
into the following condition, when F is replaced by the right hand side of Eq. (C11.2.15):

 N u (1- )
C  Cmax  min  (C11.2.16)
 f c Lc
Now, by requiring static equivalence between (N, M) and (C, F), the following is found:

N   N u ,    0;1 (a)
(C11.2.17)
tx
M  Cz ,z  (b )
2
From Eq. (C11.2.17b) it is seen that maximum bending moment may be obtained when Cmax (see Eq.
(C.11.2.16)) is inserted. To carry Cmax, the compression layer must have a thickness, x, which can be
determined as follows:

 N
 N u (1- N )
 u
x  min  (C11.2.18)
 f c L
c
Cmax and x according to Eq. (C11.2.18) can now be inserted into Eq. (C11.2.17b) to calculate the maximum
bending moment as a function of the tension force, i.e. M(N). The result can be expressed as follows:

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 455
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

   N 
  Nu 1  

 1 Nu 1 N  t   N u 
     
(a)
M (N )  2 f Lc N c f Lc
 min  c  u 
 c
 (C11.2.19)
f c Lc 2  
  

 1  t  1 (b )
 2  c 
Equation (C11.2.19) describes a N-M interaction diagram for U-bar loop connections. The solution is general
and can be used for connections with any number of overlapping U-bars. As a special case, Eq. (C11.2.19) also
provides a solution for the bending capacity of connections loaded in pure bending (when N = 0 is inserted). The
pure bending capacity will in the following be referred to as Mu, i.e. Mu = M(N =0).
Figure C.11.2.10 shows two examples of the calculated N-M interaction diagram. The horizontal cut-off, i.e. Eq.
(C11.2.19b), is only relevant in a small part of the diagram when the concrete core is made sufficiently strong by
increasing the ratio of transverse reinforcement, T. This is illustrated by the diagram with T/= 0.5. For low
values of T and in cases where Nu is governed by yielding of the U-bars (i.e. Eq. (C11.2.13b)), where the U-bar
diameter and spacing correspond a lightly/normally reinforced cross section, the N-M interaction diagram is
entirely described by Eq. (C11.2.19a). In such cases, the N-M diagram is almost linear as shown for T/= 0.2.
This is similar to the N-M diagram for a symmetric reinforced beam/slab cross section subjected tension and
bending.

Figure C11.2.10: N-M interaction diagram. Parameters used in calc.: a/H = 0.50, A*c /Ac = 0.15, nL =2,  = 0.55,
fc =27MPa, Ac = 26000 mm2, L = 600 mm, c = 25 mm.

Simplified calculation of M-N interaction

As mentioned, the N-M interaction diagram obtained for lightly/normally reinforced connections is almost linear
and similar to the N-M diagram for a symmetric beam/slab cross section subjected to tension and bending. This
similarity can be utilised to simplify the analysis, where instead of using the model, a classical cross sectional
analysis may be performed. Such a simplification makes it easy to implement an equation in the Eurocode 2,
which covers both pure tension as well as tension combined with bending (see later). The simplification is
explained below.
Figure C11.2.11a shows a connection subjected to pure tension where the capacity is utilised, i.e. N = Nu. Here
the capacity may be governed by crushing of the concrete core or by yielding of the U-bars. As shown, each of
the two U-bar legs carries Tu = ½Nu. According to the sandwich model, the pure bending capacity of the same
connection should be determined from the distribution of internal forces shown in Figure C11.2.11b, where C =
2Tu in order to be equivalent to the sectional forces (N,M) = (0, Mu). For simplicity, the figure shows a case,
where the thickness of the layer carrying C is equal to the concrete cover, i.e. x = c. By requirering static
equivalent, the model provides the following expression for the pure bending capacity of the connection:

D c
M u  2Tu   L    Tu  D  2L  c  (C11.2.20)
2 2
Now, suppose that the sandwich model is not available and suppose that the only information available is the
capacity of each the U-bar legs, Tu. In that case, a cross sectional analysis may be performed to estimate
bending moment that can be transferred from one side of the connection to the other side. As well known, it is
normal practice to neglect reinforcement bars near the compression zone when calculating the bending capacity
of a cross section. Hence, the internal forces Tu and C shown in Figure C11.2.11c would form the basis of

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 456
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

calculation. Since C = Tu in this case, the compression zone is reduced to x = ½c. Again, by requirering static
equivalent, the sectional analysis provide the following bending capacity:

 3 3 
M u  Tu  D  L  c  (C11.2.21)
 2 4 
It can be seen by comparing Eqs. (C11.2.20) and (C11.2.21) that the bending capacity obtained from the
principle of a cross sectional analysis is smaller than the result of the sandwich model. For typical values of D,
L and c, the results are within app. 10%. As an example where D = 125 mm, L = 16 mm and c = 35 mm, Eq.
(C11.2.21) gives 9% smaller capacity than Eq. (C11.2.20).
The above result shows, that in practice, a sectional analysis can replace the sandwich model in cases with
combined tension and bending. And as the tension force becomes dominant, the results of the two approaches
will converge and be identical for the case of pure tension.

(a)

(b)

(c)
Figure C11.2.11: Internal forces (in U-bars) equivalent to pure tension (a); and internal forces equivalent to pure
bending according to the sandwich model (b) and to a sectional analysis (c)

Comparison with experimental results

Test results for connections subjected to pure tension

In this section the solution given in Eq. (C11.2.13) is compared to all pure tension tests available from the
literature. The comparisons are divided in two groups; the first group contains tests that satisfy the requirement
of the minimum amount of transverse reinforcement given in Eq. (C11.2.14). The second group contains tests
that do not satisfy the requirements of Eq. (C11.2.14). For the latter, the tested tensile capacity is compared to
the solution given in Eq. (C11.2.8) whenever it is required according to theory (i.e. when it gives a tensile
capacity higher than the Eq. (C11.2.11)).

The average value of N u ,test / N u ,calc for all tests is 1.01 with a standard deviation of 0.13.

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 457
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Pure Tension Tests that satisfy the requirement of minimum transverse reinforcement
Tests used in comparison (Fig. C11.2.12):
- Hao (2004 [C11.2.3]): # of tests: 4*
- Gordon (2006 [C11.2.2]): # of tests: 8*
- Ma et al (2012 [C11.2.7]): # of tests: 4
- Joergensen (2014 [C11.2.5]): # of tests: 28
* Where the transverse reinforcement was not anchored at the ends, the capacity of the transverse
reinforcement is reduced according to a linear development curve calculated on the basis of the Eurocode 2
anchorage length; see detailed calculation in [C11.2.5].

The average value of N u ,test / N u ,calc is 1.02 with a standard deviation of 0.12.

Figure C11.2.12: Ratio N u ,test / N u ,calc as a function of different parameters varied in the tests for connections

with    min

Pure Tension Tests that do not satisfy the requirement of minimum transverse reinforcement
Tests used in comparison (Fig. C11.2.13):
- Hao (2004 [C11.2.3]): # of tests: 11*
- Joergensen (2014 [C11.2.5]): # of tests: 10

The average value of N u ,test / N u ,calc is 1.01 with a standard deviation of 0.15.

Figure C11.2.13: Ratio N u ,test / N u ,calc as a function of different parameters varied in the tests for connections

with    min

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 458
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Pure tension tests versus theory using alternative formulation of the effectiveness factor
To keep a consistent formulation of the effectiveness factor in the entire FprEN1992-1-1:2023, the -formula as
stated in Eq. (C11.2.4) will be replaced by the formula below when implemented in FprEN1992-1-1:2023:

 EC 2   fc k3
(C11.2.22)
where:

1
 (C11.2.23)
 fc  min  40 MPa 1/3
 
 fc 

1/3
 d dg 
k3  0.85  
 H  (C11.2.24)
It should be noted that fc accounts for the brittleness of concrete in a similar way as the factor does.
Likewise, k3 as well as the factor in Eq. (C11.2.4) account for the size effect. The parameter ddg is
defined in FprEN1992-1-1:2023.
Figure C11.2.14 shows plots of the ratio between tested and calculated capacity based on the modified
effectiveness factor (Eq. C11.2.22). The average value of N u ,test / N u ,calc is 1.05 and the standard deviation of
0.12. The plots can be directly compared to those in Figure C11.2.12, where Eq. (C11.2.4) was used to
calculate . It can be seen that the results obtained with these modifications are very similar.

Figure C11.2.14: Ratio N u ,test / N u ,calc as a function of different parameters varied in the tests for connections

with    min

Test results for connections subjected to combined tension & bending

Only a limited number of tests with combinations of tension and bending has been published. These tests are
described in details in [C11.2.5, C11.2.6]. Figure C11.2.15 below shows a comparison of the tests with the N-M
diagram determined from the sandwich model. The calculations have been performed by use of the modified -
formula (C.11.2.22). It is noted that the tests in Figure (a) & (b) have    min and therefore, the value of Nu to
be inserted into Equation (C11.2.19) has been based on Equation (C11.2.8). The tests in Figure (c-f) have
   min , which means that Equation (C11.2.11) defines Nu to be inserted into Equation (C.11.2.19).

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 459
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Figure C11.2.15: Comparison of tests with N-M diagram determined by the sandwich model (tests taken from
ref. [C11.2.5])

Proposal in FprEN1992-1-1:2023

The simplified Equation (C11.2.11) is the basis for subsection 11.5.4 (Laps using U-bar loops) of FprEN1992-1-
1:2023.
In addition to change of notation for consistency with FprEN1992-1-1:2023 (which will not be introduced or
explained below) some minor changes have been introduced to make the provision more general (i.e. covering
both pure tension and tension combined with bending) and more consistent with the rest of the code revisions.
These changes are explained below.

Laps versus Connections

Equation (C11.2.11) is formulated in a format that accounts for an entire U-bar connection, i.e. nL is the number
of U-bars that transfer tension to (nL + 1) number of U-bars on the other side of the connection (see Figure
C11.2.11(a) where nL = 1). In this case, the number of developed yield lines (which are accounted for in the
equation) is 2nL. Equation (C11.2.11) has in FprEN1992-1-1:2023 been reformulated in a manner that suits the
problem of a standard unit of U-bar lap, see figure C11.2.15(b). Since only one yield line can develop for a unit
of U-bar lap and since two yield lines develop in a connection with nL = 1, then the factor nL must be replaced by
a factor of 0.5, when Equation (C11.2.11) is used for a unit of U-bar lap. Hence:

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 460
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

 2
a 
1  a
N c   f c Ac  r     
2  H H
  (C11.2.25)

Figure C11.2.15: Number of yield lines formed in U-bar connection with nL = 1 (a) and in a unit of U-bar lap (b).

Pure tension versus combinations of tension and bending

Equation (C11.2.25) has a format that can only deal with pure tension. As shown in the above, combinations of
tension and bending can formally be treated by a simple cross sectional analysis, although the underlying
mechanism is more related to the sandwich model. This leads to slightly conservative results and requires
knowledge about the maximum tension capacity of each U-bar leg, which simply is Tc = ½Nc, i.e.

 2
a 
1  a
Tc   f c Ac  r      (C11.2.26)
4  H H
 
In FprEN1992-1-1:2023, the design compression strength is defined as fcd = fcfck/c. This means that a part of
the -formula in Eq. (C.11.2.22) will be embedded in the definition of fcd. Therefore, when introducing
(C.11.2.22) and fcd into (C.11.2.26) and into the definition of r, the following expression for the design strength of
each U-bar leg in the lap is obtained (except for the difference in notations, the expression has the same
meaning as the Equation of clause 11.5.4(2) in FprEN 1992-1-1:2023):

 d dg   a
1/3 2
  a 
Tcd  0.2 f cd Ac   r      (C11.2.26)
 H   H H

where

 4 T   T   1
  1  ;if T  (a)
     2
r
1  1
;if T  (b)
  2 (C11.2.27)

and

T AsT f ydT (C11.2.28)


  d dg 
1/3

0.85   Ac f cd
 H 
By use of Eq. (C11.2.26), crushing of the concrete within the U bar lap can be prevented when:

max(T1 ; T2 )  Tcd (C11.2.29)

where T1 and T2 are the tension forces in the U-bar legs calculated from a sectional analysis when the lap
transfers combinations of normal forces and bending moments, see Figure C11.2.16. It should though here be
noted that according to the sandwich model, both legs will carry the same tension force and therefore, when
designing U-bar laps using Eq. (C11.2.26), both legs of the U bar must be anchored outside the connection for
the larger of the forces T1 and T2.

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 461
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

Figure C11.2.16: Tension forces in U-bar legs obtained from sectional analysis when lap transfers
combinations of normal forces and bending moments

Minimum transverse reinforcement

Equation (C11.2.14) provides an estimate of the minimum transverse reinforcement that ensures a ductile
behaviour. By adopting Equation (C11.2.7) to determine ftef and by inserting a/H = 0.5 and H = 100 mm (which is
usually the upper and lower limits in practical design), the minimum transverse reinforcement requirement may
conservatively be written as follows:

 min f yT  0.24 f c2/3

(C11.2.30)
which practically gives the same result as the following simpler formulation:

 min f yT  0.5 f c
(C11.2.31)
Equation (C11.2.31) is the one that is adopted in FprEN1992-1-1:2023. Comparison of Equations (C11.2.30)
and (C11.2.31) can be seen in Figure C11.2.17.

Figure C11.2.17: Comparison of Eqs. (C11.2.30) and (C11.2.31) for estimation of minimum transverse
reinforcement.

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 462
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.4 Background to FprEN 1992-1-1: 2023

References
[C11.2.1] CEN European Committee for Standardization, Eurocode 2. Design of concrete structures – general
rules and rules for buildings, EN 1992-1-1, Brussels, Belgium, 2004. 225p.
[C11.2.2] Gordon, S. R., Joints for Precast Decks in Steel Concrete Composite Bridges. PhD thesis, School of
the Built Environment, Heriot Watt University, Edinburgh, UK, 2006
[C11.2.3] Hao, J. B., Structural Behaviour of Precast Component Joints with Loop Connection. PhD thesis,
Department of Civil Engineering, National University of Singapore, Singapore, 2004
[C11.2.4] Joergensen H. B., Hoang L. C., Tests and limit analysis of loop connections between precast
concrete elements loaded in tension, Engineering Structures, Vol. 52, pp. 558-569, 2013
[C11.2.5] Joergensen, H. B. Strength of Loop Connections between Precast Concrete Elements – Part I: U-bar
connections loaded in combined tension and bending, Part II: Wire loop connections loaded in shear.
PhD thesis, Department of Technology and Innovation, University of Southern Denmark, Odense,
Denmark, 2014.
[C11.2.6] Joergensen H. B., Hoang L. C., Strength of Loop Connections between Precast Bridge Decks
Loaded in Combined Tension and Bending, Structural Engineering International, Vol. 25, No 1, pp.
71-80, 2015
[C11.2.7] Ma, Z. J., Lewis, S., Cao, Q., He, Z., Burdette, E. G., and French, C. E. W., Transverse Joint Details
with Tight Bend Diameter U-bars for Accelerated Bridge Construction. Journal of Structural
Engineering, 138(6): 697-707, 2012
[C11.2.8] Nielsen M. P., Hoang L. C., Limit analysis and concrete plasticity. CRC Press, third edition, 2011

List of versions of this document

11.11.2016 1st draft
19.11.2017 2nd draft (section 4 added and title revised)
30.11.2017 new template, small edits
15.04.2021 1) Section about connection subjected to combinations of tension and bending added, including comparison with tests
2) Rearrangement and new titles for some of the existing sections
3) Update of brittleness factor and format of code equation to reflect changes from prEN-1992-1-1: 2017 to prEN-1992-1-1:
2021-01
4) Typo in Eq. (C11.2.7) and (C11.2.8) corrected
01.03.2023 editorial update to FprEN 1992-1-1:2023

H.B. Joergensen, L. C. Hoang 05.05.2021

For citations, copying and referencing see Note on Front page! page 463
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1: 2023

Background document to subsection 11.5.5

Laps using headed bars

Robert Vollum Imperial College London

05.05.2021

For citations, copying and referencing see Note on Front page! page 464
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

Introduction
This document describes the background to the equation proposed for calculation of the tension capacity of short
headed bar laps in the next generation of EC2 [1]. The proposal is based on the analytical upper bound plasticity
model of Joergensen and Hoang [2-4] used to calculate the tensile strength of laps using U-bars in FprEN 1992-
1-1: 2023 [1].
Short headed bar laps can for example be used to establish structural continuity between precast concrete slabs.
Headed bars protruding from each side of the precast elements to be connected are overlapped in the connection
as shown in Figure 1. After placement of the precast units, transverse reinforcement bars are placed over the
overlapping headed bars and the connection is subsequently infilled with concrete. The strength and ductility of
headed bar tension laps is increased by the provision of through thickness tie down reinforcement which provides
confinement. The design recommendations relate to headed bar laps loaded in either pure tension or flexure
induced tension.

Figure 1: Illustration of a headed bar connection loaded in bending.

General
The design approach for headed bar tension laps is based on the upper bound plasticity model proposed by
Joergensen and Hoang [2-4] for U-bar loop splices. This approach is described in detail in the background
document to C11.5.4 of FprEN 1992-1-1:2023 [5] for laps using U-bar loops. This document summarises the key
assumptions implicit in the design method pertinent to headed bar tension laps as well as the validation of the
method with experimental data. A change to the model of Joergensen and Hoang [2-4] for laps of both U-bar loops
and headed bars is that plane stress is considered instead of plane strain. This gives lower and more realistic
strength predictions for tested headed bar joints and is justified by the observation that the anchor plates are
suggestive of a strut and tie system in plane stress. The member depth is assumed to equal to the effective head
depth. Concrete and reinforcement are assumed to behave in a rigid plastic fashion and dowel action is neglected.

Proposal in FprEN1992-1-1:2023

The model applies to short laps of headed bars with head size large enough to anchor the bar force without
additional contribution from bond along the bar. Heads complying with 11.4.7 (1) of FprEN 1992-1-1:2023 [1]
satisfy this requirement for fyd ≤ 435 MPa. The maximum tensile force developed in each bar should be limited by
11.4.7(2).
Overlapping headed bars may be single (Figure 2a)) or multiple (Figure 2b)) and should be anchored for the
design force outside the connection.
The proposed design equation for the resistance of headed bar laps is derived from Eq. (C11.2.25) of the
background document for laps using U-bar loops [5], which is formulated in a manner that suits the problem of a
standard unit of U-bar lap, where only one yield line can develop as shown for a unit of headed bar lap in Figure
2a). The Equation is repeated here:

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 465
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

 2
a
1  a
N c   f c Ac  r     
2  H H
  (1)

For headed bar laps, the dimensions a and H are defined as shown in Figure 2a.

(a) Unit of headed bar lap

(b) Headed bar lap with ns = 3

Figure 2: Number of yield lines formed in (a) a unit of headed bar lap and (b) with ns= 3 headed bars.

Derivation of design expression for headed bar laps

The proposed design equation for headed bar laps is derived from Eq. (1) in similar fashion to the design
expression for U-bar laps [5]. The main differences in the derivation are in the definition of concrete effectiveness
factor and the calculation of minimum transverse reinforcement area.
For headed bar laps, the effectiveness factor  adopted in [5] is replaced by:

 EC 2   fc k3
(2)
where:

1
 (3)
 fc  min  40 MPa 1/3
 
 f c 
and
1/3
 d dg 
k3  1.3  
 lsd  2  (4)
in which H in Eq. (1) is replaced by lsd  2 , where lsd is the lap length measured to the outside of the bar heads
and  is the headed bar diameter which is assumed to equal the head thickness.

In the design expression, the term  fc is absorbed into fcd which is defined in FprEN 1992-1-1:2023 as
f ck (5)
f cd   fc
c
For the assumed range of model applicability 0 ≤ cs ≤ 0.5lsd, the resistance of a single headed bar lap TRd,c may
be calculated as:
/
𝑇 , 0,6𝑓 𝐴 𝑘 (6)
∅ ∅ ∅

where
Ac is the effective concrete area within the heads of the over lapping bars (Fig. 3a)):
Ac = (lsd-2)bh1 (7)
lsd is the overlapping length measured to the outside of the heads which are assumed to be of thickness  where
is the headed bar diameter (Fig. 3a));
ddg is a coefficient that takes into account the concrete type and its aggregate properties according to 8.2.1(4) [1];

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 466
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

bh1 is effective width of the head perpendicular to the plane of the lap (for circular head with diameter 𝜙 :
𝑏 0.5𝜙 √π) (8)
cs is the clear spacing of headed bars (see Fig, 3b));
kst is the resistance factor of the transverse reinforcement, which may be taken equal to kst = 1 when  ≥ 0,5 and
to kst = 4(1 – ) for lower values of 𝜔 /  (9)
. ∅

Ast is the total area of the fully anchored transverse reinforcement positioned within Ac.

a) elevation view
b) plan view of single lap
c) plan view of multiple lap

Figure 3: Laps using headed bars

In case of multiple headed bars as shown in Figure 3c), the resistance of Eq. (6) may be calculated using an
average net spacing cs = 0,5(cs1 + cs2) (see Figure 3c)) and multiplied by (ns – 1) where ns is the total number of
headed bars (ns = 5 in the example of Figure 3c)). Multiple overlaps may also be treated as an assembly of single
overlaps where cs is taken as the smaller of cs1 and cs2.

To avoid brittle behaviour, a minimum amount of transverse reinforcement should be placed:

𝐴st 𝑚𝑎𝑥𝑖𝑚𝑢𝑚 0,75𝐴 ⋅ ck

, 2/8) (10)
yk

where
is the maximum diameter of the lapped headed bars;
To enhance ductility of laps designed for reinforcement yield, tie down reinforcement with total area Astd per single
lap should be provided perpendicular to the plane of the headed bars within Ac (Figure 3a). The tie down
reinforcement should be fully anchored outside Ac and placed symmetrically between the headed bars. The tie
down reinforcement may be provided in the form of either double headed shear studs or links. For lsd ≤ 200 mm,
Astd may be provided by a single double headed shear stud (Figure 3c). The minimum area of tie down
reinforcement per single lap equals:
𝐴std 0,12𝜙 (11)

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 467
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

Minimum transverse reinforcement

The theoretical minimum area of transverse reinforcement is 𝐴st 0,75 𝑓ck ⋅ in which the coefficient of 0.75 is
yk

obtained by multiplying the coefficient of 0.5 in Eq. C11.2.31 of [5] for U-bar loops by 0.6/(2×0.2) = 1.5. The
multiple of 0.6/(2×0.2) results from the increased concrete efficiency factor for headed bars in Eq. (6) compared
to Eq. C11.2.26 of [5] for U-bar loops for which gives the resistance of a single leg of the U-bar. Eq. (10) introduces
an empirical lower limit to the minimum area of transverse reinforcement of 𝐴st 𝜋𝜙 /8 since the theoretical
minimum area of reinforcement is significantly less than that provided in any of the tests within the experimental
database used to calibrate the model. Consequently, based on the minimum area of transverse reinforcement
provided in the tension tests of Vella et al. [6], the minimum area of transverse reinforcement is increased to 𝐴st
𝑚𝑎𝑥𝑖𝑚𝑢𝑚 0,75 𝑓ck ⋅ , 2/8), where  is the headed bar diameter.
yk

Comparison with test data

The model is validated for the experimental data of Vella et al. [6,7], Li et al. [8] and Thompson [9,10]. The
evaluations are carried out with partial factors of 1.0 for steel and concrete.

Tests of Vella et al. [6,7]

Vella et al. [6,7] tested 32 specimens of which 25 had the geometrical arrangement shown in Figure 4. Six further
tests investigated the influence of varying the headed bar spacing and lap length. Other variables were concrete
strength, transverse bar size and arrangement and the presence or absence of through thickness shear studs.
The headed bars were 25mm in diameter with standard 70x70x16mm friction welded square heads. Where
present, two 10mm diameter 125mm long shear studs were placed at the positions shown in Figure 4. The cover
to the stud head cstud was zero, the stud head diameter was 30mm. The 36mm spacing of the transverse
bars shown in Figure 4 was chosen to allow sufficient space for concrete to be placed in contact with the bar
heads and to allow clearance for the friction weld flash. The specimen length was chosen to provide adequate
anchorage for bond failure of the transverse bars not to limit the failure load of the connection. The tests focussed
on concrete controlled failures with a view to determining the critical concrete strength at which bar yield
precedes concrete failure, and to better observe the effect of other variables on joint performance. Table 1
summarises the details of the tested specimens which are divided into five groups, with the test ID describing
the specimens as follows:
For example, G4-39-2H20:TT'-S-100-250:
“ G4 ” – Test group
“ 39 ” – Measured concrete cylinder strength at time of testing
“ 2H20 ” – Number and diameter of transverse bars
“ TT’ ” – Position of transverse bars as indicated in Figure 4
“S” – Shear studs included
“ 100 ” – Lap length of headed bars
“ 250 ” – Spacing of headed bars
Group 1, with through thickness shear studs (see Figure 4), and group 2, without shear studs, investigated the
influence on joint strength of transverse bar diameter and concrete strength. The transverse bars were positioned
at the T and T’ locations illustrated in Figure 4. Groups 3 to 5 respectively varied the spatial arrangement of
transverse bars, the spacing of headed bars and the lap length of headed bars. In group 5, the centreline of the
transverse bars was positioned 32 mm from the inside face of the heads as shown in Figure 4. Consequently, the
spacing of the transverse bars varied with the lap length. In addition, Vella et al. [7] tested 4 slabs with headed
bar laps using the test arrangement shown in Figure 5. The strength of the laps in the tested slabs was significantly
greater than in the tension test specimens as shown in Table 2. Therefore, the proposed design method has been
calibrated using the tension test results.
Figure 6 compares the predictions of the proposed design method with the tension test data of Vella et al [6] with
partial factors of safety s = 1.0 and c = 1.0. Figure 6a shows that the influence of concrete strength is reasonably
predicted with the strength of all specimens safely predicted. Figures 6b and 6c show the influence of /min and
 respectively on predicted strength where  is given by Eq. (9) and min is calculated with Astmin from Eq. (10) for
which 𝐴st 𝜋𝜙 /8 governs. Figures 6b and 6c clearly show that significantly greater areas of transverse
reinforcement were provided in almost all the tested specimens than required by Eq. (10) and that the strength of
the connection is greatest for transverse reinforcement placed in the TT’BB’ arrangement. However, Eq. (6)
provides a conservative lower bound to the strengths of specimens with transverse bars placed in the TT’
arrangement and no through thickness shear studs. Figures 6d and 6e respectively show that Eq. (6) reasonably
captures the influence on lap strength of varying the headed bar lap length and spacing. It should be noted that
the intention of Eq. (6) is to provide a conservative lower bound estimate of lap strength. The actual lap strength

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 468
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

can in some cases be significantly underestimated as shown in Figure 6 since the method is calibrated for
specimens having the TT’ transverse reinforcement arrangement and no through thickness shear studs.
For the 8 specimens in group 2 with no shear studs and transverse bars in the TT’ arrangement the mean and
standard deviation of Ttest/Tcalc are 1.19 and 0.10 respectively. For the 15 specimens with shear studs and
transverse bars in the TT’ arrangement the mean and standard deviation of Ttest/Tcalc are 1.43 and 0.11
respectively.

Figure 4: Typical test specimen

Figure 5: Typical flexural test specimen and section through specimen B2-26-2H20-S-10

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 469
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

Table 1: Tension test specimen details

Ptest
Transverse
sh Hhb (kN)
Test ID fc,cyl (MPa) ft,split (MPa) ∅tr (mm) bar positions
(mm) (mm) {failure
(see Fig. 4)
mode}

G1-39-2H12:TT'-S-100-200 39.2 3.49 12 T T’ 200 100 178 {J}

G1-26-2H16:TT'-S-100-200 25.6 2.38 16 T T’ 200 100 149 {J}
G1-40-2H16:TT'-S-100-200 40.4 3.60 16 T T’ 200 100 232 {J}
G1-54-2H16:TT'-S-100-200 54.3 4.57 16 T T’ 200 100 259 {J}
G1-26-2H20:TT'-S-100-200 25.6 2.38 20 T T’ 200 100 154 {J}
G1-40-2H20:TT'-S-100-200 40.4 3.60 20 T T’ 200 100 242 {J}
G1-54-2H20:TT'-S-100-200 54.3 4.57 20 T T’ 200 100 286 {Y}
G1-48-2H25:TT'-S-100-200 48.0 4.04 25 T T’ 200 100 260 {J}
G1-39-2H25:TT'-S-100-200 39.2 3.49 25 T T’ 200 100 274 {Y}
G1-39-2H25:TT'-10S-100-200 39.2 3.49 25 T T’ 200 100 243 {J}
G2-39-2H12:TT'-100-200 39.2 3.49 12 T T’ 200 100 159 {J}
G2-26-2H16:TT'-100-200 25.6 2.38 16 T T’ 200 100 124 {J}
G2-40-2H16:TT'-100-200 40.4 3.60 16 T T’ 200 100 181 {J}
G2-54-2H16:TT'-100-200 54.3 4.57 16 T T’ 200 100 220 {J}
G2-26-2H20:TT'-100-200 25.6 2.38 20 T T’ 200 100 133 {J}
G2-40-2H20:TT'-100-200 40.4 3.60 20 T T’ 200 100 207 {J}
G2-54-2H20:TT'-100-200 54.3 4.57 20 T T’ 200 100 257 {J}
G2-48-2H25:TT'-100-200 48.0 4.04 25 T T’ 200 100 209 {J}
G3-28-2H20:T'B'-S-100-200 27.7 2.65 20 T’ B’ 200 100 236 {J}
G3-28-4H16:TT'BB'-S-100-200 27.7 2.65 16 T T’ B B’ 200 100 264 {Y}
G3-28-4H20:TT'BB'-S-100-200 27.7 2.65 20 T T’ B B’ 200 100 312 {Y}
G3-46-2H20:T'B'-S-100-200 46.1 3.80 20 T’ B’ 200 100 316 {Y}
G3-46-4H16:TT'BB'-S-100-200 46.1 3.80 16 T T’ B B’ 200 100 318 {Y}
G3-46-2H16:T'B'-S-100-200 46.1 3.80 16 T’ B’ 200 100 290 {Y}
G3-48-1H25:T'-S-100-200 48.0 4.04 25 T’ 200 100 243 {J}
G4-39-2H20:TT'-S-100-150 39.2 3.36 20 T T’ 150 100 288 {Y}
G4-39-2H20:TT'-S-100-250 39.2 3.36 20 T T’ 250 100 190 {J}
G4-39-2H20:TT'-S-100-300 39.2 3.36 20 T T’ 300 100 130 {J}
G5-25-2H20:TT'-S-75-200 24.8 2.13 20 T T’ 200 75 117 {J}
G5-25-2H20:TT'-S-150-200 24.8 2.13 20 T T’ 200 150 213 {J}
G5-25-2H20:TT'-S-200-200 24.8 2.13 20 T T’ 200 200 261 {J}
G5-24-2H20:TT'-150-200 24.1 2.58 20 T T’ 200 150 201 {J}

Note: J depicts joint failure, Y depicts headed bar yield, S depicts strut failure, T depicts tie failure.
ft,split spit cylinder tensile strength; ∅tr diameter of transverse reinforcement bars; H=lsd-2 (lap between inside face
of heads)

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 470
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

Table 2: Comparison of flexural resistance obtained in flexural tests of Vella et al. [6] and flexural resistances
calculated with tensile resistance obtained in corresponding tension test

Specimen M
Test ID
type (kNm)

B2-26-2H20-S-0 (test) Slab 111

G1-26-2H20:TT'-S-100-200 (test) Tension 74
B2-39-2H20-S-0 (test) Slab 129
G1-40-2H20:TT'-S-100-200 (test) Tension 114
B2-24-2H20-/-0 (test) Slab 84
G2-26-2H20:TT'-100-200 (test) Tension 66

a) Influence of concrete strength

3

2,5

2
Ttest/Tcalc

1,5

1
Studs TT' No studs TT'
0,5
StudsT'B' Studs TT'BB'
0
15 20 25 30 35 40 45 50 55 60
fc MPa

b) Influence of /min
3

2,5

2
Ptest/Pcalc

1,5

Studs TT' No studs TT'

0,5
Studs T'B' Studs TT'BB'
0
0 1 2 3 4 5

/min

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 471
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

c) Influence of 

2,5

Ptest/Pcalc
1,5

1
Studs TT' No studs TT'
0,5
Studs T'B' Studs TT'BB'
0
0 0,5 1 1,5 2 2,5 3 3,5 4


d) Influence of lap length H = lsd-2

1,8
1,6
1,4
1,2
Ttest/Tcalc

1
0,8
0,6
0,4
G5 studs
0,2
0
0 25 50 75 100 125 150 175 200 225 250
H mm

e) Influence of cs/H

1,8
1,6
1,4
1,2
Ttest/Tcalc

1
0,8
0,6
0,4
Studs TT'
0,2
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
cs/H

Figure 6: Influence on Ttest/Tcalc of a) concrete strength, b) min, c)  d) lap length H and d) cs/H

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 472
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

Other test data

Apart from the test data of Vella et al. [6,7] there are very few tests in the literature on short tension laps of headed
bars with heads sufficiently large to develop the full headed bar force. Several test data, e.g. Thompson [9,10],
relate to bars with small head sizes. In addition, transverse reinforcement was not provided in most of Thompson’s
tests.

Analysis of Li et al. [8] test data

Li et al. [8] tested 5 flexural specimens with headed bar splices. The headed bars were 16 mm in diameter with
circular heads of 51 mm diameter. Specimens were 610 mm wide. The lap length ranged between 64 mm and
152 mm. The bar spacing was either 102 mm or 152 mm. Transverse reinforcement consisting of 13 mm diameter
headed bars was positioned at the centre of the lap above and below the overlapping headed bars. For these
specimens the transverse bar eccentricity e = 0 (see Figure 3). The headed bar yield strength was assumed to be
470 MPa in the assessments as reported by Li et al. In reality strain hardening occurred in some tests which is
neglected in the calculation of resistance. Figure 7 shows that the proposed design method gives safe strength
predictions for the test specimens of Li et al. with c = 1.0. However, it should be noted that the calculated strength
was limited by headed bar yield in 4 of the 5 tests.

1,6

1,4

1,2

1
Ttest/Tcalc

0,8

0,6

0,4

0,2

0
0 20 40 60 80 100 120 140 160

Lap between heads H = lsd-2 (mm)

a)

1,6

1,4

1,2

1
Ttest/Tcalc

0,8

0,6

0,4

0,2

0
0,0 0,1 0,2 0,3 0,4 0,5 0,6

cs/(H =lsd-2 )

b)
Figure 7: Influence on Ttest/Tcalc for specimens of Li et al. [8] of a) H =lsd-2 and b) cs/H

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 473
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

Analysis of Thompson [9,10] test data

Thompson [9,10] tested 15 slabs in 4 point bending with lapped 1’’ diameter headed bars and headed bars with
net head area to bar area of 4 or more. The bar heads were either rectangular plates measuring 1.5’’× 3’’, for
which the net ratio of head to bar area was 4.70, or 2.25’’ in diameter for which the net ratio of head to bar area
was 4.04. Consequently, these head sizes are insufficient to develop the bar yield strength according to 11.4.5(1).
The lap length between heads ranged between 3.5” and 14”. The bar spacing was either 6” or 10”. With the
exception of specimen LS-08-04.04-12-10(N)-1-TTD none of these specimens included transverse confining
reinforcement within Ac (see Figure 3). Five of the specimens included through thickness confinement
reinforcement in the form of hairpin bent bars (4 specimens) and a tie-down detail (1 specimen) (see Figure 8). A
tie down ratio was defined as the ratio of the total area of vertical legs of confining reinforcement divided by the
sum of cross-sectional areas of all the lapped bars. The provided tie down ratios ranged between 0.20 and 1.01.

Figure 8: Confining details adopted by Thompson [8]

In the absence of transverse reinforcement, the resistance was calculated assuming that the area of transverse
reinforcement equalled 𝐴 0.75 𝑓 since this is the theoretical area at which Eq. 6 gives the same
resistance as calculated in a yield line analysis without transverse reinforcement (see background document for
lapped U-bar loops [5]). The ratio Ttest/Tcalc is plotted against cs/H, kst and the tie down ratio respectively in
Figures 9a to 9c. Interestingly, the proposed design method gives safe results for all the test specimens of
Thomson despite the head size being insufficient to develop the full bar strength. Figure 9b shows that the degree
of conservatism of the design method reduces with the coefficient kst in Eq. 6 which depends on (Eq. 9). It should
be noted that kst was calculated with the theoretical minimum area of transverse reinforcement, which corresponds
to concrete tensile failure, since no transverse reinforcement was provided in these specimens. Figure 9c shows
that the ratio Ttest/Tcalc is almost independent of the tie down ratio. However, the provision of confining tie down
reinforcement increased the deformation at failure in the Thompson tests [9,10] with the transverse tie down detail
in Figure 8 being most effective. In this detail, the area of transverse reinforcement was 28% of the headed bar
cross sectional area and the tie down ratio was 0.32. These ratios compare with a minimum ratio of transverse
bar area to headed bar of 0.46 in the tests of Vella et al. [6,7] and a tie down ratio provided by shear studs of 0.11
per lapped bar (2×78.5/(3×491) = 0.11) or 0.16 per single lap. The presence of the shear studs increased lap
strength in the tests of Vella et al. as well as the deformation at failure in tests where headed bars yielded. The
increase in strength due to the confining studs is given by the ratio of strengths of corresponding specimens in
Group 1 (with studs) and Group 2 (without studs). This ratio varies between 1.12 and 1.28 with an average of
1.18.

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 474
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

The test specimens of Vella et al. [6,7] with only transverse reinforcement exhibited more ductility than those of
Thompson without tie down reinforcement. Additionally, a ductile response was also achieved in the tests of Li et
al. [8] with 152 mm laps despite the specimens having no tie down reinforcement. Consequently, it is proposed
that tie down reinforcement is only necessary when laps are designed for bar yield.
Figure 9d shows Ttest/Tcalc plotted against the minimum area of transverse reinforcement calculated as 𝐴
0.75 𝑓 . The resulting areas of reinforcement are seen to be very small compared with the cross-sectional
area of the 1’’ diameter lapped headed bars. Hence, for practical reasons a minimum area of transverse
reinforcement equal to half the headed bar cross sectional area is proposed. However, Figure 10 shows that the
measured strengths increase with kt in a similar fashion to the calculated strengths with 𝐴 0.75 𝑓 . The
tests of Vella et al. and Thompson suggest that added ductility can be provided with through thickness confining
reinforcement giving a tie down ratio of 0.15 per single lap or more. This leads to Eq. 11.19 for tie down
reinforcement in which the coefficient of 0.12 is given by 0.15 0.25𝜋𝜙 0.12𝜙 .

3,5

2,5
Ttest/Tcalc

1,5

0,5

0
0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0,45
cs/H

a)

3,5

2,5
Ttest/Tcalc

1,5

0,5

0
0,5 0,6 0,7 0,8 0,9 1
kst

b)

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 475
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

4
3,5
3
2,5

Ttest/Tcalc
2
1,5
1
0,5
0
0 0,2 0,4 0,6 0,8 1 1,2
Tie down ratio

c)

4
3,5
3
2,5
Ttest/Tcalc

2 provided
1,5
1
0,5
0
0 20 40 60 80 100 120 140 160
Astmin mm2
d)
Figure 9: Influence on Ttest/Tcalc for specimens of Thompson [9] a) cs/H, b) kst, c) tie down ratio and d) 𝐴
𝑚𝑎𝑥𝑖𝑚𝑢𝑚 0.75 𝑓 𝑡ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙 ; 𝑝𝑟𝑜𝑣𝑖𝑑𝑒𝑑 𝑠𝑡𝑒𝑒𝑙 )

250

200
T per bar kN

150

100

50 Calculated
Test
0
0,5 0,6 0,7 0,8 0,9 1
kst

Figure 10: Influence of kst (calculated with 𝐴 0.75 𝑓 ) on strength of Thompson specimens [9]

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 476
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

Parametric study
The practical implications of the proposed design rules are illustrated in Figure 11. The headed bar arrangement
was assumed to consist of bars spaced at 200 mm centres on each side of the lap. The lap strengths were
calculated assuming 2 single laps per headed bar with 25 mm cover to the bar head which was assumed to be
square with area equal to 10 times the headed bar cross section as in the tests of Vella et al. [6,7]. Partial factors
of 1.5 and 1.15 were assumed for concrete and steel respectively with reinforcement yield strength equal to 500
MPa. The maximum required areas of transverse reinforcement required for the full strength laps shown in
Figure 11 were 145 mm2 for H16 bars, 209 mm2 for H20 bars and 293 mm2 for H25 bars. The cover to the head
was chosen to ensure that attainment of bar yield was not prevented by side blowout according to 11.4.7(2) in [4].
A single 10 mm diameter double headed stud positioned between each headed bar, and midway along the lap,
would be sufficient to satisfy the tie down requirements of Eq. (11).

Influence of lap length (bar spacing 200 mm; 25 mm cover

to head)
1,2

0,8
TRd,c/Tyd

0,6

0,4

0,2

0
100 125 150 175 200 225
Lap length H mm

H16 fck = 35 MPa H20 fck = 40 MPa H25 fck = 45 Mpa

a)

Influence of concrete strength (bar spacing 200 mm; 25 mm

cover to head)
1,2

0,8
TRd,c/Tyd

0,6

0,4

0,2

0
30 40 50 60 70 80
Concrete strength fck MPa

H25 125 mm lap H20 125 mm lap H16 125 mm lap

b)
Figure 11: Influence of a) lap length and b) concrete strength on design lap strength according to proposed design
method

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 477
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.5.5 Background to FprEN 1992-1-1: 2023

References

[1] FprEN 1992-1-1: 2023 Eurocode 2: Design of concrete structures– Part 1-1: General rules – Rules
for buildings, bridges and civil engineering structures
[2] Joergensen H. B., Hoang L. C., Tests and limit analysis of loop connections between precast
concrete elements loaded in tension, Engineering Structures, Vol. 52, pp. 558-569, 2013
[3] Joergensen, H. B. Strength of Loop Connections between Precast Concrete Elements – Part I: U-
bar connections loaded in combined tension and bending, Part II: Wire loop connections loaded in
shear. PhD thesis, Department of Technology and Innovation, University of Southern Denmark,
Odense, Denmark, 2014.
[4] Joergensen H. B., Hoang L. C., Strength of Loop Connections between Precast Bridge Decks
Loaded in Combined Tension and Bending, Structural Engineering International, Vol. 25, No 1, pp.
71-80, 2015
[5] Joergensen H. B., Hoang L. C., Background document to subsection 11.5.4: Laps using U-bar loops.
CEN/TC250/SC2/WG1/TG4, May 2021.
[6] Vella J. P., Vollum R. L., Jackson A., Investigation of headed bar joints between precast concrete
panels, Engineering Structures, Vol: 138, pp. 351-366, 2017.
[7] Vella J. P., Vollum R.L., Jackson A., Flexural Behaviour of Headed Bar Connections between Precast
Concrete Panels, Construction and Building Materials, 154, pp. 236-250, 2017.
[8] Li, L., Ma, Z., Griffey, M. E., and Oesterle, R. G., Improved Longitudinal Joint Details in Decked Bulb
Tees for Accelerated Bridge Construction: Concept Development. Journal of Bridge Engineering,
15(3): pp. 327-336, 2010.
[9] Thompson M.K., The Anchorage Behavior of Headed Reinforcement in CCT Nodes and Lap Splices,
Phd thesis, The University of Texas at Austin, 2002.
[10] Thompson MK, Ledesma A, Jirsa J, Breen J, Lap splices anchored by headed bars, ACI Structural
Journal, 103 (2): p. 271-279, 2006.

List of versions of this document

10‐05‐2021 Number of Vella specimens similar to Figure 4 corrected to 25

01‐03‐2023 Editorial update to FprEN 19192‐1‐1:2023

R.L. Vollum 03.05.2021

For citations, copying and referencing see Note on Front page! page 478
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2022

Background document to clause 11.6.3

Minimum radius of curvature of tendons

Hans Rudolf Ganz Switzerland

30.09.2022

For citations, copying and referencing see Note on Front page! page 479
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.6.3 Background to FprEN 1992-1-1:2022

Introduction
This document has been prepared in order to provide background for the provisions in Clause 11.6.3 on minimum
radius of curvature of tendons. The provisions are mainly based on common practice in post-tensioning industry
and provisions in the European Assessment Document EAD 160004 for prestressing systems, [C11.6-1].

11.6.3(1) Reasons for change

The radius of curvature is important for the detailing of post-tensioning tendons in structures. A small radius of
curvature permits to generate significant deviation forces which can compensate or equilibrate applied actions on
the structure and thus favourably influence the behaviour of the structure both in SLS and ULS. However, small
radii of curvature can in extreme cases cause crushing of concrete on the inside of the tendon curvature or can
reduce the effective tensile strength of the tendon due to the combination of axial stresses with significant
transverse stresses between individual tensile elements of the tendon.
Current EN 1992-1-1:2004 is silent about the minimum radius of curvature of tendons. There is only one reference
to European Technical Approval documents for deviators of external tendons. However, it was felt important for
the 2nd generation EN 1992-1-1 to cover this important detailing aspect to permit a designer to choose appropriate
minimum radii without reverting immediately to prestressing system and hence company specific documentation.

Summary of provisions for minimum radius of tendon curvature

Clause 11.6.3 first states that the minimum radius of curvature and minimum straight length of tendon adjacent to
the anchorage shall comply with the values specified in the technical documentation of the actually used post-
tensioning system.
However, the values in the technical documentation shall not be smaller than the values given in 11.6.3(2) unless
demonstrated by testing in accordance with the relevant standard for prestressing systems. Note: EAD 160004 is
available as relevant standard.
The minimum values of tendon curvature to prevent damage of the unconfined concrete on the inside of the
tendon curvature during prestressing and to avoid a reduction of the axial tensile strength of the tendon may be
taken as:
Rmin = (pd / pRd) Ap (C11.6-1)
where:
pd is the design stress in the tendon at the time of tensioning, including partial factor for prestress (P,unfav)
pRd is the maximum transverse bearing stress on the prestressing tendon or from the tendon on the concrete
on the inside of the tendon curvature which may be taken from Table C11.6-1. These values taken from
FprEN 1992-1-1:2022, Table 11.4 (NDP), are NDPs since they represent design strength.

Table C11.6-1 Maximum transverse bearing stress on the prestressing tendon

Basis of specified values of pRd

(1) Internal tendons:
The relevant criterion used for internal tendons is the limitation of the bearing stress between tendon duct and
surrounding concrete on the inside of tendon curvature to conservatively avoid crushing of unconfined concrete
at the time of stressing of tendons.
EAD 160004 provides the following formula for the minimum radius of curvature of internal tendons consisting of
prestressing strands which may be used in the absence of national provisions:
Rmin = (2 Fpm0 dstrand / pR,max dduct,i)  2,5m (C11.6-2)
where:
Fpm0 is the initial prestressing force of the tendon (not factored)
dstrand is the nominal diameter of the strand
dduct,i is the internal diameter of the tendon duct

Hans Rudolf Ganz 30.09.2022

For citations, copying and referencing see Note on Front page! page 480
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.6.3 Background to FprEN 1992-1-1:2022

pR,max is the recommended maximum allowable pressure which in the absence of national regulations may be
taken as either 130, 150 or 230 kN/m
In the prestressing industry the following simple formula has been used since long for the minimum radius of
tendon curvature both for strands and wires which is also given in fib Model Code, [C11.6-2]:
Rmin = 2,8 (fpk Ap)  2,5m (C11.6-3)
where:
fpk is the specified tensile strength of the prestressing steel; (fpk Ap) expressed in [MN] for Rmin [m]
Ap is the cross sectional area of prestressing steel

Figure C11.6-1 compares the minimum radii of tendon curvature according to Formulae (C11.6-1), (C11.6-2) and
(C11.6-3) as a function of Ap. For the comparison, prestressing steel strand with 15,7mm nominal diameter,
150mm2 cross-section, and 1860 MPa specified tensile strength is considered. It is further assumed that the initial
stress is 0,80 fpk and that Ap represents about 40% of the duct cross section, as typically found in practice. As it
appears from Figure C11.6-1, minimum radii of tendon curvature as permitted in [C11.6-1] and [C11.6-2] are
obtained if the upper limit for the maximum transverse bearing stress in FprEN 1992-1-1:2022, Table 11.4 (NDP),
is set to pRd = 15MPa.

(a) Minimum radii for fck = 20MPa, pRd  12MPa (b) Minimum radii for fck = 30MPa, pRd  12MPa

(c) Minimum radii for fck = 30MPa, pRd  15MPa

Figure C11.6-1: Comparison of minimum radii of curvature for internal tendons

(2) External tendons:

The relevant criterion used for external tendons is the limitation of the transversal bearing stresses between tensile
elements or between tensile elements and duct on the inside of the duct in the tendon curvature / tendon deviator
combined with longitudinal movements of and friction on the tensile elements.
EAD 160004 provides minimum radii of curvature of external tendons consisting of prestressing strands for
deviators away from the anchorage, see Table C11.6-1. These values are based on testing and experience in
France, [C11.6-3].
Tendons with strands Minimum radius of curvature at deviator
19  13mm or 12  15mm 2.5m
31  13mm or 19  15mm 3.0m
55  13mm or 37  15mm 4.0m
61  15mm 5.5m
Table C11.6-1: Minimum radius of curvature for external tendons at deviators away from anchorage
Note: Interpolation between tendon sizes given in table is permitted. Tendons with wires should have the same
minimum radius of curvature as tendons with strands of the same Fpk.
In the prestressing industry about half of the radius given by Formula (C11.6-3) has been used since long for the
minimum radius of curvature of external tendons both for strands and wires. This has also been adopted in fib
Model Code, [C11.6-2]:
Rmin = 1,4 (fpk Ap)  2,0m (C11.6-4)

Hans Rudolf Ganz 30.09.2022

For citations, copying and referencing see Note on Front page! page 481
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.6.3 Background to FprEN 1992-1-1:2022

Figure C11.6-2 compares the minimum radii of tendon curvature according to Formulae (C11.6-1) and (C11.6-4)
with the values in Table C11.6-1 as a function of Ap. For the comparison, prestressing steel strand with 15,7mm
nominal diameter, 150mm2 cross-section, and 1860 MPa specified tensile strength is considered. It is further
assumed that the initial stress is 0,80 fpk. As it appears from Figure C11.6-2, minimum radii of tendon curvature
as permitted in [C11.6-1] and [C11.6-2] are obtained if the upper limit for the maximum transverse bearing stress
in FprEN 1992-1-1:2022, Table 11.4 (NDP), is set to pRd = 30MPa.

(a) Minimum radii for pRd = 24MPa (b) Minimum radii for pRd = 30MPa
Figure C11.6-2: Comparison of minimum radii of curvature for external tendons

(3) Loop in internal tendons:

The relevant criterion used for loops in internal tendons is the limitation of the transversal bearing stresses
between tensile elements or between tensile elements and duct on the inside of the duct in the tendon curvature
without longitudinal movements of and friction on the tensile elements.
EAD 160004 does not provide any rules for minimum radii of curvature of internal tendon loops. In the prestressing
industry the following simple formula, based on system approval testing, has been used since long for the
minimum radius of tendon curvature both for strands and wires which is also given in fib Model Code, [C11.6-2]:
Rmin = 0,6 (fpk Ap)  0,6m (C11.6-5)

Figure C11.6-3 compares the minimum radii of tendon curvature according to Formulae (C11.6-1) and (C11.6-5)
as a function of Ap. For the comparison, prestressing steel strand with 15,7mm nominal diameter, 150mm2 cross-
section, and 1860 MPa specified tensile strength is considered. It is further assumed that the initial stress is 0,80
fpk. As it appears from Figure C11.6-3, minimum radii of tendon curvature as permitted in [C11.6-2] are obtained
if the upper limit for the maximum transverse bearing stress is set to pRd = 70MPa.

(a) Minimum radii for pRd = 55MPa (b) Minimum radii for pRd = 70MPa
Figure C11.6-3: Comparison of minimum radii of curvature for loop in internal tendons
It should be noted that the compressive stress on the concrete inside the loop curvature needs to be checked.
For the parameters of Figure C11.6-3(a) a design concrete strength of fcd = 30MPa, and for C11.6-3(b) of fcd =
40MPa is required. Hence, a sufficiently high concrete strength should be specified or the concrete should be
confined and verified to FprEN 1992-1-1:2022, Clause 8.1.4, or alternatively, a stiff, pre-bent steel tube with
oversized pipe diameter may be chosen.
Figure C11.6.3-4 illustrates a typical detailing for looped tendons used in practice to provide reinforcement
transverse to the plane of the tendon loop and to tie back half of the deviation forces in the plane of the loop to
the concrete on the outside of the loop, in the form of U-shaped bars.

Hans Rudolf Ganz 30.09.2022

For citations, copying and referencing see Note on Front page! page 482
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.6.3 Background to FprEN 1992-1-1:2022

Figure C11.6-4: Typical detailing of reinforcement for looped tendons

Conclusions
‐ The provisions of FprEN 1992-1-1:2022, Clause 11.6.3, may be used to determine minimum radii of
tendon curvature for internal tendons, external tendons and for loops in internal tendons.
‐ In order to arrive at similar radii of curvature according to FprEN 1992-1-1:2022, Clause 11.6.3, as used
in practice and specified in [C11.6-1], [C11.6-2] and [C11.6-3] as relevant, the maximum transverse
bearing stresses of pRd = 15MPa, 30MPa and 70MPa for internal tendons, external tendons and loops
in internal tendons, respectively, as specified in FprEN 1992-1-1:2022, Table 11.4 (NDP) should be
used. However, if NSBs wish to apply more conservative values for tendon curvatures, i.e. larger radii
of tendon curvature, the values of Table 11.4 (NDP) can be reduced.
‐ Whatever values of pRd are chosen in FprEN 1992-1-1:2022, Clause 11.6.3, the actual radii of curvature
should not be smaller than the values specified in the post-tensioning system documentation of the
actual system to be installed in the structure unless verified by testing.

[C11.6-1] EAD 160004-00-0301, European Assessment Document “Post-tensioning kits for prestressing of
References
structures”, European Organisation for Technical Assessement (EOTA), Brussels, September 2016.
[C11.6-2] fib Model Code for concrete structures 2010, International Federation for Structural Concrete (fib),
Lausanne, 2013.
[C11.6-3] External Prestressing, Service d’Etudes Techniques des Routes et Autoroutes (SETRA), Bagneux,
France, February 1990.

List of versions of this document

11.02.2021. 1st draft prEN 1992-1-1:2020-11
25.02.2021 2nd draft (amended objectives of limitation of radii) prEN 1992-1-1:2020-11
30.09.2022 Updated to FprEN 1992-1-1:2022
21.02.2023 Addition of Figure C11.6-4

Hans Rudolf Ganz 30.09.2022

For citations, copying and referencing see Note on Front page! page 483
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to subsection 11.7

Deviation forces due to curved tensile and
compressive chords

Aurelio Muttoni Ecole Polytechnique Fédérale de Lausanne,

Miguel Fernández Ruiz Switzerland

Report EPFL-IBETON 17-01-R5 25.02.2021

For citations, copying and referencing see Note on Front page! page 484
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.7 Background to FprEN 1992-1-1:2023

Introduction This document has been prepared to clarify the background of the design provisions of FprEN 1992-1-1:2023
with respect to Section 11.7 (“Deviation forces due to curved tensile and compressive chords”). These
provisions are new as the topic was not specifically covered by EN 1992-1-1:2004.

Reasons for new Deviation forces due to curved tensile or compressive chords is a common design situation for many members
provisions such as vaulted cut-and-cover tunnels, arches, pipes, shells and similar members (see Figure C11.7.1). Such
deviation forces can be equilibrated by arranging a specific transverse reinforcement (following the general
design procedure of Section 8.5). This solution is however uneconomical (typically in tunnels or pipes) and
unnecessary in many cases, particularly when the deviation forces are low and they can be carried by the
resistance of the concrete cover.
Following this motivation, and also to reduce the probability of accidents (as described for instance in reference
[C11.7.2-3]), a new Subsection has been introduced in the FprEN 1992-1-1:2023 providing explicit rules for
verification of deviation forces of curved reinforcement and compressive chords. The rules are based to a large
extent on Model Code 2010 provisions [C11.7.1], grounded in turn on the works by Fernández Ruiz et al.
[C11.7.2].

Figure C11.7.1: Practical cases where curved reinforcements in tension near can lead to cover spalling

Design approach, A curved reinforcement in tension originates transverse deviation forces following equilibrium conditions, see
background Fig. C11.7.2. When such forces are applied near to the surface and push against the concrete cover, a brittle
failure by cover spalling can occur. This condition is attained when the deviation force (ftr,d = Ftd/r) equals the
effective tensile resistance of the concrete cover (fct,eff,d ∙cu) [C11.7.2].

Figure C11.7.2: Deviation forces of curved reinforcement: (a) deviation forces; and (b) spalling resistance

A. Muttoni, M. Fernández Ruiz 25.2.2021

For citations, copying and referencing see Note on Front page! page 485
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
11.7 Background to FprEN 1992-1-1:2023

The value of the effective tensile resistance depends upon a number of parameters [C11.7.2]:
 Concrete strength (tensile strength)
 Concrete cover
 Spacing of reinforcement
 Level of deformation of the reinforcement (with higher strains in the reinforcement associated to a
reduced strength to spalling due to its interaction with bond near to cracks)
 Size of the member
 Interaction with tensile stresses generated by bond in laps
For design purposes, it is normally sufficient to consider that the reinforcement reaches its yield strength
(maximum strain and force in the reinforcement, due to the strain effect, the effective resistance in case of lower
steel stresses would be higher). In that case, according to [C11.7.2], it results:
fct,eff,d = fctd∙el ≈ 0.38 fctd (C11.7.1)
This expression was slightly rearranged and the coefficient rounded for MC2010 [C11.7.1] as:
fct,eff,d = (1/3)∙fctk,0.05/C (C11.7.2)
For FprEN 1992-1-1:2023, a similar Formula is presented, expressed in terms of the compressive strength of
concrete in the following manner:
fct,eff,d = (1/8)∙fck1/2/C (C11.7.3)
As it can be noted, the expression gives almost identical results for low strength concrete (difference lower than
3% for a C30/37) and slightly more conservative estimates for higher concrete strengths.

Deviation of forces in These provisions define the cases concerned by deviation forces (11.7(1)). In general, Section 8.5 can be
curved reinforcement applied for their design (as a discontinuity region), requiring arrangement of transverse reinforcement (11.7(2)).
near to the free surface Alternatively, the Formula in 11.7(3) can be applied for verification of the resistance against deviation forces for
a simple verification or to avoid transverse reinforcement (Formula according to the previous background
11.7(1-3)
[C11.7.1,2]). The width where the effective tensile strength is considered to develop is calculated according to
the geometric definitions of Fig. 11.3 (same definitions as for verifying anchorages and laps).

Curves laps near to the The presence of curved laps is accounted for in this provision. As already observed by Fernández Ruiz et al
free surface [C11.7.2], laps have a detrimental effect on the resistance against cover spalling of curved reinforcement. This
is attributed to the interaction between bond stresses (laps) and deviation forces (curved reinforcement).
11.7(4)
A detailed procedure for such verification is presented in [C11.7.2] (reducing the value of coefficient el in
Formula (C11.7.1)). In order to provide a simpler approach for design, FprEN 1992-1-1:2023 proposes a linear
interaction between the resistance against deviation forces and the resistance against spalling in laps.

References [C11.7.1] International Federation for Structural Concrete, “fib Model Code for Concrete Structures 2010,” Ernst
& Sohn, Berlin, Germany, 434 p., 2013.
[C11.7.2] Fernández Ruiz, M., Plumey, S. and Muttoni, A., Interaction between Bond and Deviation Forces in
Spalling Failures of Arch-Shaped Members without Transverse Reinforcement, ACI Structural
Journal, V. 107, No. 3, 2010, pp. 346-354
[C11.7.3] Intichar, M.; Ebner, M.; and Sparowitz, L., Deviation Forces in Curved Beams (in German:
Umlenkkräfte in gekrümmten Stahlbetonbalken), Österreichische Ingenieur- und Architekten-
Zeitschrift, V. 149, No. 1, Austria, 2004, pp. 11-16.

List of versions of this document

13.2.2021 1st draft
25.2.2021 some corrections
13.03.2023 review and editorial update

A. Muttoni, M. Fernández Ruiz 25.2.2021

For citations, copying and referencing see Note on Front page! page 486
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/CDG FprEN 1992-1-1:2023

Background document to Section 12

Detailing of members and particular rules

Tony Jones The Concrete Centre

Linh C. Hoang DTU, Technical University of Denmark
Aurelio Muttoni EPFL, Ecole Polytechnique Fédérale de Lausanne,
Switzerland
20.03.2023

For citations, copying and referencing see Note on Front page! page 487
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12 Background to FprEN 1992-1-1:2023

Introduction
This document provides the background to section 12 (Detailing of members and particular rules). The current
Eurocode section 9 (Detailing) has 28 NDPs.

Ease of use
The initial objective of the revision of Section 9 EN1992-1-1:2004 was the reduction of NDPs. However, for the
reasons given below, in many cases some national choice was still required. Therefore, NDPs were
consolidated into clear tables and where helpful some indication of the purpose of the requirements are given. It
is believed that by having the majority of NDP’s in tables they are more readily identified, and this constitutes
ease of use.

A compilation of NDPs in EN1992-1-1:2004 was created as part of the PRB Project, [1]. Its purpose was to
facilitate consensus-building in the reduction of NDPs to enhance ease of use of the Eurocode 2. The following
Figure C12.1 gives on overview of the section 9 NDPs. It can be seen, that in most cases (green boxes) the
recommended values were accepted by the NSBs. However, for every NDP at least some NSBs either adopted
a different value (red) or adopted the recommended value but with further conditions. In many cases the NSB
values/rules are based on local practices and local approaches to safety, not purely on a technical basis. As
such there was a strong demand to keep national choice.

Figure C12.1: Analysis of Section 9 NDP’s in EN1992-1-1.

12.2 Reasons for change

Minimum reinforcement This is a new clause that sets out in greater detail the reasoning behind the minimum reinforcement
rules requirements and signposts the relevant clauses that address the points listed.

12.2 (2) and (3) Bending/axial load reinforcement

Clause 12.2(2) ensures that the resistance of the cracked section is at least as great as the resistance of the
uncracked section. Whilst expression 12.1, for flexure, and 12.2 for pure tension, are not explicitly given in EN
1992-1-1:2004 they do underpin the values that are given. This approach ensures distributed cracking and
enables forces from restrained deformations, where not considered explicitly in the design, to be dissipated. It
also ensures reasonable deformation capacity to enable secondary load paths to form and thereby add to
structural robustness.
In EN 1992-1-1:2004 there is an additional requirement for members with prestressed permanently unbonded
tendons to have minimum reinforcement equal to 1.15 times the cracking moment. It has not been possible to
determine the justification for the 1.15 factor. Post cracking behaviour of an element with unbonded tendons will
be significantly different to that with bonded reinforcement. However, most elements with unbonded tendons
also have some normal reinforcement and the current rule implied a step change in requirement even if the

T Jones, L. C. Hoang, A. Muttoni 20.03.2023

For citations, copying and referencing see Note on Front page! page 488
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12 Background to FprEN 1992-1-1:2023

amount of unbonded tendon was relatively small compared to the bonded steel. It was also noted that the
calculated resistance of an unbonded tendon was dependent on assumptions about the overall deflected shape.
On this basis it seemed to be more logical to remove the 1.15 factor but to be more specific as to the basis
under which the resistance of the unbonded (or external) tendon was calculated. To this end a requirement to
calculate the resistance of the unbonded tendon based on the prestressing stress has been added. This avoids
any reliance on the deflected shape.

“For sections prestressed with internal permanently unbonded tendons or

with external tendons the contribution of the tendons to NEd, MR,min and
Mcr should be calculated using the σp,mt(x) (see Formula (7.33)).”

As NEd affects both the cracking moment and the bending strength and as the same value of prestress is used,
in both cases, for unbonded tendons, it may be thought that little advantage can be taken from these tendons in
satisfying minimum steel requirements. However, the contribution of the prestress is greater when considering
the bending strength. This is because the lever arm between the resultant of the prestressing force and the
centroid of the cracked section is greater than the lever arm in the uncracked case. Therefore, subject to
satisfying other requirements, such as those for tying, it is possible to satisfy the minimum steel requirements
solely with unbonded tendons.
In the case of lightly loaded beams clause 12.2(2) can lead to large amounts of reinforcement. For statically
determinate elements an alternative rule is provided in 12.2(3). This recognises that when the load is
significantly less than the cracking moment, cracking may still occur for other reasons. When such cracking
does occur, the reinforcement must be sufficient to absorb the energy released. This reinforcement can be
calculated by comparing the energy released at cracking with the energy absorbed through the opening of a
single crack, noting that reinforcement less than required by 12.2(2) will not ensure distributed cracking.
The so-called energy balance method is illustrated in Figure C12.2.1, where WI denotes the available internal
dissipation and WE denotes the released potential energy when sudden cracking takes place under constant
external loading, ME.
A more detailed explanation of the model can be found in references [2] & [3]. The model applies to members
where the yield hinge consists of one single crack. This means members where the cross-section yield moment,
My, is smaller than the cracking moment, Mcr, as indicated in Figure C12.2.1. For members where multiple
cracks can develop, i.e. Mcr < My, the model underestimates the deformation capacity and leads to conservative
results.

Figure C12.2.1: Illustration of energy balance concept to verify sufficient strength and deformation capacity
of members containing less reinforcement than “the deem to work” minimum reinforcement
requirement.
It should be noted that the proposed code provision contains a simplification. In the original formulation [3], the
required moment capacity is given as:

T Jones, L. C. Hoang, A. Muttoni 20.03.2023

For citations, copying and referencing see Note on Front page! page 489
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12 Background to FprEN 1992-1-1:2023

2
M Rd ,min  kdc  M Ed ; kdc 
f  1
1  tk  1  
f yk   

Here, MRd,min corresponds to My as seen in Figure C12.2.1 and the factor ξ is calculated as follows:


 u lo     u   y 
f t  fy 
y
fy  

wu  f ct 2 f ct  4
   
wy  fy  
 y  lo  
 2 f ct 4 
Where:
fy = yield stress of reinforcement
ft = tensile strength of reinforcement
fct = tensile strength of concrete
φ= bar diameter
εy = yield strain of reinforcement
εu = strain of reinforcement at stress ft
lo = debonding length at vicinity of crack
The debonding length has an effect on the deformation capacity of the member but no effect on the strength of
the cross section. It is always conservative to disregard the debonding length, i.e. assuming lo = 0 mm. By
doing so, a simplified expression for ξ can be obtained:

    f  
   2 u
 1  t  1  1
 y   f  
    y  
The above equation has been used to calculate the kdc-values in FprEN1992-1-1:2023.
To study the effect of the simplification, calculations of the required yield moment, My, as a function of the
applied moment, ME, have been carried out. The results, normalised with respect to the cracking moment, Mcr,
are shown in Figures C12.2.2 – 4. Table 1 shows the input data used in the calculations. It can be seen that the
model requires more reinforcement when lo = 0 mm. The increase is however not significant.

Compression strength fc 40 N/mm2

Yield stress fy 500 N/mm2

Young’s modulus Es 200000 N/mm2

Bar diameter Ø 12 mm

Cover c 20 mm

Table 1. Input data

T Jones, L. C. Hoang, A. Muttoni 20.03.2023

For citations, copying and referencing see Note on Front page! page 490
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12 Background to FprEN 1992-1-1:2023

Class A Class A

(a) (b)

Figure C12.2.2. Class A steel. Required yield moment, My, as function of applied moment, ME. (a) model
with debonding length lo ≠ 0 and (b) model with debonding length lo = 0.

Class B Class B

(a) (b)

Figure C12.2.3. Class B steel. Required yield moment, My, as function of applied moment, ME. (a) model
with debonding length lo ≠ 0 and (b) model with debonding length lo = 0.

Class C Class C

(a) (b)

Figure 12.2.4. Class C steel. Required yield moment, My, as function of applied moment, ME. (a) model

T Jones, L. C. Hoang, A. Muttoni 20.03.2023

For citations, copying and referencing see Note on Front page! page 491
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12 Background to FprEN 1992-1-1:2023

with debonding length lo ≠ 0 and (b) model with debonding length lo = 0.

12.2 (4) Shear/torsion reinforcement

See separate background document to Cl12.2(4) by Muttoni et al.

12.2 (6) Anchorage and lapping of minimum reinforcement

Generally bars provided to satisfy the minimum reinforcement requirements should be lapped or anchored to
transfer their full capacity. This clause notes clarifies that where additional steel is provided the anchorage
length can be reduced accordingly. It also allows for the specific case of a simply supported element, the
anchorage length to be reduced to 10ϕ past the edge of the support. This relaxation acknowledges the practical
difficulties in providing a full anchorage in these situations and consideration that flexural cracking as
considered in Figure C12.2.1 is very unlikely to occur close to a support in this type of member.

12.3 Reasons for change

Beams
In addition to placing the majority of the NDPs within a Table, clarification has been added on the anchorage of
links. In EN 1992-1-1:2004 no distinction is made between anchorage in the tension or the compression zone. It
is known that the anchorage of some open links is less efficient in the tension zone where flexural cracking can
occur parallel to the link anchorage.

12.3.3 Shear and torsion reinforcement

To address issues around anchorage in tension zones Figure 8.5 on EN1992-1-1:2004 has been expanded into
Figure 12.2 of FprEN1992-1-1:2023. The new figure provides additional cases to highlight that laps in closed
stirrups are still acceptable in tension zones whilst the text notes that cases b) to d) (see Figure C12.3.1) with
open stirrups, are not. It is noted that Figure 12.2 a) is not mentioned with regards to its suitability in a tension
zone, this was because different countries had different practices with regards this detail.

T Jones, L. C. Hoang, A. Muttoni 20.03.2023

For citations, copying and referencing see Note on Front page! page 492
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12 Background to FprEN 1992-1-1:2023

Figure C12.3.1. Anchorage of links a)-e) and closing stirrups f)-h) (Figure 12.2 FprEN1992-1-1:2023)

For torsion links in T-sections a further example of an acceptable link has been added to figure 12.3 of
FprEN1992-1-1:2023. This allows “U-Stirrups” to form the link if the laps are confined within the flange. This
removes the risk of anchorage being compromised by corner spalling which occurs in beams predominantly
loaded in torsion. There is also a requirement to design the lap for 1.2fyd to ensure ductile yielding rather than a
brittle bond failure that may invalidate the design model.
It should be noted that the Figure 12.3d) does not show the full lap length within the depth of the flange and this
is an area where further national guidance may be helpful to reflect national traditions.

12.4 Reasons for change

Slabs
In addition to placing the majority of the NDPs within a Table, clarification has been added in the form of notes
in the table to support the requirements given.

12.5 Reasons for change

Slab-column connections
The punching shear reinforcement requirements required updating to be consistent with the revised punching
and column bases shear models in clause 8.

12.5.1. Punching Shear Reinforcement

To be consistent with the model in Section 8 this section removes the NDP for the position of the outer row of
reinforcement relative to the control perimeter where reinforcement was not required. In addition, the limitations
of shear reinforcement spacing further than 2.0d from the column have been relaxed. Headed studs are now
included explicitly. Finally, the effective area of one leg of shear reinforcement is limited based on the effective
depth of the slab. These revised limits are based on the experimental data used to calibrate the punching shear
rules in clause 8.

See separate paper by Schmidt and Hegger, “Background Document to sub-section 12.5.1(3)”

12.5.2 Integrity reinforcement against progressive collapse of flat slabs

See Separate background document by Muttoni and Ruiz.

12.6 Reasons for change

Columns
In addition to placing the majority of the Nationally Determined Parameters within a Table, clarification has been
added in the form of notes in the table to support the requirements given. The previous rules were also not
easily interpreted for prestressed columns.

Table 12.3(NDP) – Detailing requirements for reinforcement in columns - Row 1

During the revision of the table some effort was made to understand the basis of the requirements for minimum
longitudinal steel. One reason cited for the requirements was to reduce the risk of creep and shrinkage in the
concrete causing yielding of the reinforcement under serviceability conditions. Such yielding would have the
potential to lead to unacceptable cracking. The 0.1NEd/fyd in EN1992-1-1:2004 was shown to be a reasonable
minimum to reduce the risk of such cracking in practical cases for the range of concrete strengths covered.
However, it was considered that the 0.002Ac provided a sensible lower bound for lightly loaded columns. In
addition, as the issue of compression yielding was not relevant for prestressed columns, and the effect of
considering axial load was to penalize the prestressing, the requirements for prestressed columns are relaxed
with just the 0.002Ac minimum applied.

Table 12.3(NDP) – Detailing requirements for reinforcement in columns – Row 2

In EN1992-1-1:2004 there was no limit to the spacing of the longitudinal bars. A limit of ≤400mm has been
added to provide consistency with the requirements for walls. The minimum number of longitudinal bars in
circular columns has increased from 4 to 6 so that the behaviour is less dependent on the orientation of the
reinforcement cage.

Table 12.3(NDP) – Detailing requirements for reinforcement in columns - Row 3

A number of nations chose to vary the limit for link spacing preferring ≤12 to 15 times the longitudinal bar
diameter rather than the recommended value of ≤20. It was agreed that to control bar buckling, when the
concrete cover has spalled, a value of link spacing ≤15 time the diameter of the main bar was required.
However, it was also noted that the recommended values required the link spacing to reduce to 0.6 x 20 x the
longitudinal bar diameter at the ends of columns where such spalling is most likely to occur. It was therefore
decided to leave the recommended values as per the current version of EN1992-1-1:2004. As it is an NDP a
country may decide to use ≤12 or 15 diameters over the full length of the column.
In non-seismic design a limited amount of ductility is required for column members to ensure compatibility with
current analysis methods and to provide a degree of robustness. As concrete strength increases columns

T Jones, L. C. Hoang, A. Muttoni 20.03.2023

For citations, copying and referencing see Note on Front page! page 493
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12 Background to FprEN 1992-1-1:2023

become less ductile, however ductility can be increased by confinement. If only the concrete is considered, a
measure of the plastic rotation capacity can be found by considering the difference between the yield strain of
the concrete and the ultimate strain. In accordance with 5.1.6(3) of FprEN 1992-1-1:2023, this can be presented
as shown in Figure C12.6.1.

Figure C12.6.1. Ultimate Strain compared to Strain at Maximum Stress.

It can be seen that whilst the rotation capacity reduces as fcm increases there is a substantial drop off at fcm =
58MPa. However, the above graph ignores the requirement for requirement for minimum confinement links in all
columns and this increases the ductility for all grades of concrete. The current minimum confinement rules do
not necessarily produce consistent ductility even for the same concrete strength. The column ductility can be
related to the ratio of the confinement stress to the design strength of the concrete. It was found that by using
confinement stress of 0.02fcd the ductility of higher strength concrete columns was similar to that of normal
strength concrete columns using the historical deemed to satisfy link requirements. Thus, for concretes with
fck>50 MPa, transverse reinforcement shall be provided to achieve a minimum confinement of k.fcd, where k is a
NDP with recommended value of 0.02. Away from end regions where ductility is less important the principal
function of the minimum links is to prevent buckling of reinforcement. For prestressed columns this is not
relevant and so the minimum spacing based on bar size can be ignored for columns where all longitudinal bars
are prestressed.

12.7 Reasons for change

Walls and deep beams
In addition to placing the majority of the Nationally Determined Parameters within a Table, the requirements for
walls and deep beams have been combined. It is often difficult in practice to determine when a wall is behaving
as intended for walls and when it is behaving as a deep beam. By including both cases in a revised table and
differentiating between the behavior of elements predominantly axially loaded with some out of plane bending
and elements carrying normal and shear forces it was considered this would improve ease of use. Reference is
made to the design methods in section 8.5 for strut and tie and Annex G for Shell Elements for walls carrying
normal and shear forces and the minimum reinforcement quantities given are consistent with these models.

12.8 Reasons for change

Foundations
It was considered that most of the text in the equivalent clause 9.8 of EN1992-1-1:2004 was related to design
and therefore not appropriate for the new clause 12.8. Instead, clauses are given dealing with the typical strut
and tie assumptions that may be used in the detailing of foundations and particularly pile caps.

In addition, specific rules are given for footings and pile caps without shear reinforcement where the effective
lever arm may be less than that assumed for plane sections remaining plane. Further details are given in [5] and
[6].

12.9 Reasons for change

Tying systems for
robustness of buildings The current version of the EN1992-1-1:2004 was published before EN1991-1-7:2006, which covers accidental
actions. Therefore, there has been a discrepancy between the tie force requirements in EN1992 and EN1991. In
addition to the difference in the force requirements, EN1991-1-7 also considers the consequence class of the
building whereas EN1992-1-1:2004 treats all buildings in the same way.
A comparison of the tie force requirements for EN1992-1-1:2004 and EN1991-1-7:2006 was carried out,
generally it was shown that EN1991-1-7 required the design to consider greater horizontal tie forces than
EN1992-1-1. It was however, noted that EN1991-1-7 did not explicitly include horizontal ties to edge
columns/walls. Overall, recognising that the table will need to be subject to National Choice, it was considered
appropriate to adopt the current EN1991-1-7 approach to avoid the current inconsistency. For clarity, the

T Jones, L. C. Hoang, A. Muttoni 20.03.2023

For citations, copying and referencing see Note on Front page! page 494
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12 Background to FprEN 1992-1-1:2023

existing horizontal tie requirements for walls and columns were maintained.
The tie requirements are summarized in table 12.5(NDP) recognizing that the final values chosen will be NDPs
and ease of use is improved by combining them in a single table.
It should be noted that the second generation version of EN1991-1-7 was not available when FprEN 1992-1-
1:2023 was written, the expectation is that the tie forces will generally align with those given in the first
generation, however as the table is an NDP national bodies can modify if required.

12.10 Reasons for change

Supports bearings and
In EN1992-1-1:2004 section 10 “Additional rules for precast concrete elements and structures” gave a number
expansion joints of detailing rules that could be applicable to situations other than precast concrete. It was considered helpful to
bring these more general rules into clause 12 of FprEN1992-1-1:2023.

EN1992-1-1:2004 section 10 also included several tables giving minimum dimensions and “ineffective” bearing
lengths. These values could not be justified from first principles and in some situations could be either
conservative or unsafe. Therefore, whilst the need to consider ineffective edge distances is stated it is expected
that the Engineer will use both their judgement, the other material Eurocodes and NCCI to determine the
requirements for their design. Similarly, matters of deviation are dealt with in project specifications and whilst the
need to consider them is stated no values are recommended.

12.10(5) Nominal bearing lengths

Whilst Figure 12.12 of FprEN1992-1-1:2023 is largely based on Figure 10.5 of EN1992-1-1:2004, it is

recognized that the requirements for vertical bends can be onerous as they require the start of the bend to be
beyond the end of the bearing. Therefore, where measures are taken to avoid spalling, a relaxation is
introduced to allow the bend to start at the centre line of the support. This is consistent with the strut and tie
modelling for supports.

[1] CEN-TC250-SC2-WG1_N0265_Collection_of_EN_1992-1-1_NDPs_2015-08.pdf
References
[2] Hoang, Linh Cao; Ruiz, Miguel Fernández: Minimum reinforcement rules for robustness –
Performance based calculation of As,min for bending problems. Presentation at the 23rd CEN/TC
250/SC2/WG1 meeting, 7-8 November, 2018, Delft.
[3] Hoang, Linh Cao; Hallgren, Mikael: Minimum reinforcement rules for robustness – Performance
based verification by use of NLFEA and simplified Rigid-plastic modelling. Presentation at the 24th
CEN/TC 250/SC2/WG1 meeting, 25-26 March, 2019, Copenhagen.
[4] Ruiz et al. Post-Punching Behaviour of Flat Slabs, ACI Structural Journal, September-
October 2013.
[5] Simões J.T., Faria D.M.V, Fernández Ruiz M., Muttoni A., Strength of reinforced concrete
footings without transverse reinforcement according to limit analysis, Engineering Structures,
2016
[6] Muttoni A, Fernández Ruiz M, Simões JT. The theoretical principles of the critical shear crack
theory for punching shear failures and derivation of consistent closed-form design
expressions. Structural Concrete, 2017

List of versions of this document

07.01.2021. 1st draft
03.03.2021. 2nd draft
11.03.2021 3rd draft – Additional text on minimum reinforcement for unbonded tendons added
22.11.2022 4th draft – General update to align with FprEN1992-1-1 2022.
12.12.2022 Final – Issued to CDG.
20.03.2023 Updated following Peer reviewed

T Jones, L. C. Hoang, A. Muttoni 20.03.2023

For citations, copying and referencing see Note on Front page! page 495
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to clause 12.2(4)

Minimum shear reinforcement for ductility
classes B and C

Aurelio Muttoni Ecole Polytechnique Fédérale de Lausanne,

Miguel Fernández Ruiz Switzerland
Frédéric Monney
Linh Cao Hoang Technical University of Denmark, Denmark
Report EPFL-IBETON 17-01-R3 24.02.2023

For citations, copying and referencing see Note on Front page! page 496
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.2 (4) Background to FprEN 1992-1-1:2023

Introduction This document has been prepared to clarify the background of the design provisions of FprEN 1992-1-1:2023
with respect to the provisions of 12.2(4) concerning the minimum amount of shear reinforcement for ductility
classes B and C. These provisions are new as the topic was not specifically addressed by EN 1992-1-1:2004.
Since the phenomena governing the minimum reinforcement ratio are the same as those governing the
minimum inclination of the compression field in webs, the values of cotmin defined in 8.2.3(4) are also modified
for class A reinforcement.
The clause on minimum shear reinforcement is also applicable to minimum torsion reinforcement. Since torsion
is resolved into shear, the considerations in this document are equally applicable to torsion.

Reasons for change The shear design of beams and slabs is performed in FprEN 1992-1-1:2023on the basis of a variable-angle
truss model. Such approach, consistent with EN 1992-1-:2004, is an equilibrium-based model for design
grounded on the application of limit analysis [C12.2(4).1-4]. Other than equilibrium, the strength conditions of
the materials have to be respected and the materials shall present sufficient deformation capacity. In particular,
it is considered that reinforcement in tension is sufficiently ductile so that failure in the compression field governs
before rupture of the stirrups.
In usual design situations, the deformation capacity of the reinforcement is not relevant to limit the shear
resistance. This might however be the case for members with relatively low amounts of shear reinforcement (as
for members designed with minimum amounts of shear reinforcement). This is justified in these cases by the
fact that relatively flat angles of the compression field can be required to ensure equilibrium, leading to large
deformation demands in the stirrups. Following the relatively high deformation demand of the stirrups, the
ductility class of the reinforcement (defining its deformation capacity together with its hardening properties) can
be governing. With this respect, it is worth mentioning that according to 7.3.3.1(5), “Plastic analysis shall only be
used for reinforcement steel in Class B or C.”
Theoretically, when failure occurs by rupture of the stirrups, the shear resistance depends on the deformation
capacity of the reinforcement. This is due to the fact that rupture of the reinforcement will limit the minimum
angle of the compression field and thus the number of activated stirrups at ultimate. Consequently, when
rupture of the stirrups governs, a reinforcement class A for the stirrups leads to a lower shear resistance than a
Class B or C reinforcement. Other consequence is that, in order to ensure a certain level of resistance with a
minimum amount of reinforcement (assuming stirrup rupture), the amount of stirrups can be lowered for a Class
B or C reinforcement with respect to a Class A reinforcement. These considerations allow for savings in the
amounts of reinforcement and ensure a more uniform level of safety and enhanced consistency with the values
of the compression field recommended in 8.2.3(4).

Design approach, Remark: The material presented hereafter is a concise version of [C12.2(4).7] which is referred to for further
background details.
The main aim of the minimum amount of shear reinforcement is to allow for a safe application of shear design
provisions (VATM) of Section 8.2.3 (“Members with shear reinforcement”). This implies two aspects:
 Control of cracking: Avoid crack localization and ensure that the strength reduction factors of the
compression field  in 8.2.3 are applicable
 Allow for redistribution of internal stresses (variable orientation of compression struts, sufficient
deformation capacity)
The VATM is based on limit analysis and allows for design based only on equilibrium conditions. Nevertheless,
the state of strains is also implicitely considered by means of the efficiency factors for the compression field
(depending on the principal tensile strains) and the limits of compression field inclination cotmin (8.2.3(4)). An
explicit calculation of such strain state is considered in Formula (8.45), where the maximum tensile strain (1) is
estimated assuming that the compressive strain in concrete is 0,001 and that the strain tensor is parallel to the
stress field (assuming cracks to have the same inclination  as the compression field):
1 = x +( x + 0.001)∙cot2 (C12.2(4).1)
Consistently with that approach, depending on the horizontal strains (x), the vertical strains can also be
determined (see Mohr’s circle in Figure C12.2(4).1):
 y =( x + 0.001)∙cot2- 0.001 (C12.2(4).2)

A. Muttoni, M. Fernández Ruiz, F. Monney, L.C. Hoang 24.02.2023

For citations, copying and referencing see Note on Front page! page 497
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.2 (4) Background to FprEN 1992-1-1:2023

Figure C12.2(4).1: Mohr’s circle of strains

Such vertical average strain increases as the angle of the compression field reduces as can be seen in Figure
C12.2(4).2 and also for higher levels of longitudinal deformation (x).

1.4
x/y [-]
1.2
0.5
1.0 0.375
0.25
w,ave [%]

0.8
0.125
0.6
0
0.4

0.2 wy

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
cot  [°]

Figure C12.2(4).2: Average vertical strain in the compression field according to Formula (C12.2(4).2)
For some of the cases shown in Fig. C12.2(4).2, the vertical strains reach 1.3%, which indicates the risk of
failures in tension for brittle reinforcement when tension-stiffening effects are considered [C12.2(4).5] (roughly
estimated as half of the rebar strain at rupture uk = 2.5%).
When compared to an extensive database of members with varying deformation capacity of the reinforcement,
such effect can be clearly observed. This fact is presented in Figure C12.2(4).3, where the model of FprEN
1992-1-1:2023 has been compared to selected slender shear beams [C12.2(4).6] by considering a constant
value  = 0.5 and cot = 2.5. In that Figure, the value of w,min is adopted as 0.08fc1/2/fy according to Formula
(12.4) of FprEN 1992-1-1:2023 (without modification for other ductility classes). The results show that all tests
with ductility Class C are safely estimated by the code, only one of Class B is (slightly) estimated in an unsafe
manner, but a large number of tests with reinforcement Class A are unsafe, independently of the shear
reinforcement ratio (in some cases, tests with brittle reinforcement and w = 2÷3ꞏw,min show insufficient
resistance). On the other side, tests with ductile reinforcement (classes B and C) and a shear reinforcement
ratio smaller than the minimum value show sufficient resistances.

A. Muttoni, M. Fernández Ruiz, F. Monney, L.C. Hoang 24.02.2023

For citations, copying and referencing see Note on Front page! page 498
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.2 (4) Background to FprEN 1992-1-1:2023

2.5

Class A
2.0
Class B
Class C
1.5

test /R [-]

1.0

0.5
prEN 1992-1-1:2018(D3)
v = 0.5 ; cot = 2.5
0.0
0.0 1.0 2.0 3.0 4.0
ρw /ρw,min [-]

Figure C12.2(4).3: Comparison to test results assuming cotmin = 2.5 for all ductility classes

2.5

Class A
2.0
Class B
Class C
1.5
test /R [-]

1.0

0.5
prEN 1992-1-1:2020-11(D7)
(variable limits cotmin)
0.0
0.0 1.0 2.0 3.0 4.0
ρw /ρw,min [-]

Figure C12.2(4).4: Comparison to test results with a penalty for Class A reinforcement in the definition of
cotmin

Changes in 8.2.3(3) and Following these observations:

12.2(4)  A penalty for Class A stirrups has been introduced in the allowable angles of the compression field for
Class A reinforcement (20% reduction of cotmin, refer to 8.2.3(4)). The results of this consideration
show a more uniform level of safety, as shown in Figure C12.2(4)4.
 It is proposed in 12.2(4) that the minimum amount of shear reinforcement can be reduced to 10% for
Class B reinforcement and to 20% for Class C reinforcement. These limits are in agreement to the
experimental results of Figure C12.2(4)4, where no deviation on the model predictions is observed for
reductions up to 20% of the minimum amount of shear reinforcement.

References [C12.2(4).1] Nielsen M. P., Braestrup M. W. and Bach F. “Rational Analysis of Shear in Reinforced Concrete
Beams.” IABSE Colloquium Proceedings P-15, Vol. 2, 1978, 16 p.
[C12.2(4).2] Thürlimann B. “Plastic Analysis of Reinforced Concrete Beams.” IABSE Colloquium - Plasticity in
Reinforced Concrete, Reports of the working commissions, Vol. 28, 1979, p. 71-90.
[C12.2(4).3] Nielsen M.P., Hoang L.C. Limit analysis and concrete plasticity. 3rd ed. Boca Raton (USA), CRC
Press, 2011, 788 p.

A. Muttoni, M. Fernández Ruiz, F. Monney, L.C. Hoang 24.02.2023

For citations, copying and referencing see Note on Front page! page 499
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.2 (4) Background to FprEN 1992-1-1:2023

[C12.2(4).4] Sigrist, V., Bentz, E., Fernández Ruiz, M., Foster, S.J., Muttoni, A., Background to the Model
Code 2010 Shear Provisions - Part I: Beams and Slabs, Structural Concrete, Vol. 14, No. 3,
2013, pp. 204--214
[C12.2(4).5] Marti, P., Alvarez. M., Kaufmann, W. and Sigrist, V., “Tension chord model for structural
concrete”, Structural Engineering International, IABSE, Vol 8, No 4, 1998, 287–298.
[C12.2(4).6] Test results from: Huber P., Huber T., Kollegger J., 2016; Yoon Y.-S., Cook W. D., Mitchell D.,
1996; Piyamahant S., Shima H., 2002; Angelakos D., 1999; Bresler B., Scordelis A. C., 1963;
Vecchio F.J., Shim W., 2004; Yoshida Y., 2000; Cladera A., Mari A. R., 2005; F. Monney 2019;
Campana S. et al., 2013
[C12.2(4).7] Monney F, Fernandez Ruiz M, Muttoni A. Influence of amount of shear reinforcement and its
post-yield response on the shear resistance of reinforced concrete members. Structural Concrete.
2022. 1–33. https://doi.org/10.1002/suco.202200331

List of versions of this document

15.02.2021, AM, MFR, FM, LCH 1st draft
26.02.2021, AM some improvements
02.03.2021, AM minor changes
22.02.2023, MH peer reviewed and updated to FprEN 1992-1-1:2023 version
24.02.2023, AM final review

A. Muttoni, M. Fernández Ruiz, F. Monney, L.C. Hoang 24.02.2023

For citations, copying and referencing see Note on Front page! page 500
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to
clauses 12.5.1 and 12.4.2
Detailing of punching shear reinforcement -
maximum diameter of punching shear
reinforcement

Dr.-Ing. Philipp Schmidt Institute of structural concrete (IMB)

Univ.-Prof. Dr.-Ing. Josef Hegger RWTH Aachen University

24.01.2023

For citations, copying and referencing see Note on Front page! page 501
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.4.2 and 12.5.1 Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared in order to give a background to the proposed EC2 provisions for the
maximum allowable diameter of punching shear reinforcement Øsw,max in flat slabs and column bases as well as
shear reinforcement assemblies in slabs. The proposal was derived from database evaluations published in [1]
and previously discussed in CEN/TC250/SC2/WG1/TG4.

12.5.1 (3) Reasons for change

(+ 12.4.2 (6)) To obtain the highest possible level of maximum punching strength (Vmax), an upper limit defining the maximum
permissible punching shear reinforcement diameter Øsw,max should be given in design provisions to guarantee
sufficient and uniform activation of shear reinforcement. Few shear reinforcement elements with large diameters
Detailing of punching and thus large spacings to adjacent bars negatively affect the resulting punching load, even if allowable
shear reinforcement – maximum tangential spacings are not exceeded. Instead of pursuing a stepwise definition of Øw,max according to
maximum limit of Øsw,max
Model Code 2010 (MC2010) [2], a continuous limitation of Øw,max depending on the slab’s effective depth d was
required. Consequently, a simple mathematical correlation between d and Øw,max was derived in
CEN/TC250/SC2/WG1/TG4.

Database evaluation
A punching database evaluation was conducted to derive application limits and to assess current detailing
provisions as systematic test series usable for quantifying the influence of Øsw,max are not available. The
database was discussed in CEN/TC250/SC2/WG1/TG4 and is more detailed explained in [1]. All tests failing
with any kind of punching failure inside or outside the shear-reinforced zone or on the level of Vmax were
selected for evaluation to get an overview of experimentally tested values of Øsw. Considering deviating
anchorage and thus activation conditions of different types of shear reinforcement, three different categories are
defined in accordance with the given examples in Figure 12.6 from section 12.5.1 of prEC2. Other shear
reinforcement elements which cannot be clearly classified and specific types such as bent-up bars or I-beam
section cuts were excluded for this database evaluation. However, bent-up bars revealing well anchorage can
generally be assigned to the same category as double headed studs. A summary of all selected punching tests
is given in the appendix in Table A and can also be found in [1].
 Single leg links and open stirrups exhibiting a small anchorage stiffness

 Closed stirrups and bars with similar anchorage exhibiting an average anchorage stiffness

 Bent up bars and headed bars exhibiting a high anchorage stiffness

Evaluating and deriving Øsw,max depending on the slab’s effective depth d turned out to be the most suitable
method as the Øsw/d-ratio strongly influences the bond conditions and thus the activation of punching shear
reinforcement [3–5]. Figure X.1 visualizes the database selection and also depicts corresponding permissible
upper limits according to MC2010, Eurocode 2 in combination with the German National Annex (EC2/NAD) and
prEC2.

J.Hegger, P.Schmidt 24.01.2023

For citations, copying and referencing see Note on Front page! page 502
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.4.2 and 12.5.1 Background to FprEN 1992-1-1:2023

32 MC2010
32 32
(a) (b) (c)

Shear rfct diameter Øsw [mm]

EC2/NAD
prEC2
24 test 24 24
0.05∙d 16 d /200

16 16 16
11 d / 200 10 d /200

8 8 8
MC2010 MC2010
prEC2 prEC2
d >160 mm d >160 mm test d > 160 mm test
0 0 0
0 100 200 300 400 500 600 0 100 200 300 400 500 600 0 100 200 300 400 500 600
Effective depth d [mm] Effective depth d [mm] Effective depth d [mm]

Figure X.1: Database evaluation of punching tests for determining Øw,max

The point cloud evolving from the selected punching tests reveals a degressive trend of the upper limit of tested
Øsw for increasing d and indicates that better anchorage of shear reinforcement may go along with larger
permissible bar diameters. Slabs reinforced with stiffly anchored double headed studs principally featured larger
diameters than slabs with stirrups. The filtered database for shear links is small and represents only a very
limited range of d.

It is now interesting to compare the database with current detailing provisions. The stepwise graduated upper
limit for Øsw,max according MC2010 does not distinguish between the type of shear reinforcement and only
reflects double headed studs adequately. Whereas general EC2 does not mention any upper limit for Øsw at all,
EC2/NAD provides a linear relationship for stirrup-reinforced slabs resulting in overestimating Øsw,max for larger
values of d compared to the experimentally tested values. The prEC2 square root approach, recently defined by
CEN/TC250/SC2/WG1/TG4 using the presented database, represents the upper limit quite well by introducing
simple factors to differentiate between the three shear reinforcement categories (formula shown in Figure X.1).
An additional limitation ensuring proper anchorage of shear reinforcement in slabs is the minimum slab height
specified as hslab ≥ 200 mm according to EC2, EC2/NAD and prEC2 or alternatively defined as dv ≥ 160 mm
according to MC2010 which is also plotted in Figure X.1.

Conclusions
All in all, the presented database evaluation allows for a more profound definition of Øsw,max replacing the roughly
estimated upper limits of MC2010 or EC2/NAD and paves the way for the novel prEC2 approach enabling a
straightforward formula. This continuous definition of Øsw,max distinguishing between three different anchorage
categories can be extended by further categories at any time. Future punching test series with systematically
varying Øsw in combination with different tangential spacings and types of shear reinforcement may allow for
deriving more progressive upper limits.
Although the derivation is based on a punching test database, the proposed limitations for shear links shall also
be applied to shear reinforcement assemblies in slabs (cf. clause 12.4.2. (6)).
Detailed list of database entries can be presented if required.

[1] Schmidt P, Kalus M, Hegger J. Discussion of the level of maximum punching strength of flat slabs
References
and column bases using database evaluations. Structural Concrete 2021 (submitted for publication).
[2] fib Model Code for Concrete Structures 2010, Ernst & Sohn, Berlin, 2013.
[3] Fernández Ruiz M, Muttoni A. Applications of the Critical Shear Crack Theory to Punching of RC
Slabs with Transverse Reinforcement. ACI SJ 2009;106(4):485–94.
[4] Brantschen F. Influence of bond and anchorage conditions of the shear reinforcement on the
punching strength of RC slabs. PhD-Thesis. Lausanne, Switzerland; 2016.
[5] Schmidt P. Punching in Shear-reinforced Flat Slabs and Column Bases. Dissertation. Aachen; 2021.

List of versions of this document

18.03.2021 1st draft
11.06.2021 2nd draft
24.01.2023 Final Draft
06.03.2023 peer reviewed and editorially updated

J.Hegger, P.Schmidt 24.01.2023

For citations, copying and referencing see Note on Front page! page 503
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 AhG Detailing FprEN 1992-1-1:2023

Background document to clause 12.5.2

Integrity reinforcement against progressive
collapse of flat slabs

Aurelio Muttoni Ecole Polytechnique Fédérale de Lausanne,

Miguel Fernández Ruiz Switzerland

Report EPFL-IBETON 17-01-R6 24.02.2023

For citations, copying and referencing see Note on Front page! page 504
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.5.2 Background to FprEN 1992-1-1:2023

This document has been prepared to clarify the background of the design provisions of FprEN 1992-1-1:2023
Introduction
with respect to clause 12.5.2 (“Integrity reinforcement against progressive collapse of flat slabs”). These
provisions extend and complete clause 9.4.1(3) of EN 1992-1-1:2004, providing a consistent design approach
for integrity reinforcement in flat slabs.

Flat slabs are particularly vulnerable to punching shear failures. This is due to the fact that a brittle failure at a
Reasons for change
slab-column connection leads to a significant loss of its load-carrying capacity (unless specific measures are
adopted). Thus, following a first punching failure, neighbouring connections are overloaded, with the risk of also
triggering punching failure at them. As a consequence, a local punching at one connection can propagate
through the whole flat slab, leading to its eventual progressive collapse [C12.5.2-1].
The clause 9.4.1(3) of EN 1992-1-1:2004 provides guidance on the detailing of integrity reinforcement, stating
that at least two well-anchored bars shall be arranged in each orthogonal direction on the compression side of a
slab passing over a column. Some Countries had supplementary rules for these bars that required them to have
a tensile capacity equal to the accidental load reaction at the column. If the aim is that such bars would carry the
whole accidental reaction, greater deformation and ductility would be required than is available. These
provisions give good guidance, but are relatively incomplete in terms of design and detailing of such
reinforcement. Recent state-of-the-art on this topic [C12.5.2-1,2] has shown that rational approaches can be
proposed for design of integrity reinforcement. In addition, these works have shown that previous assumption
(full activation of the yield strength of integrity bars in vertical direction to carry the shear force) is clearly too
optimistic. On that basis, Model Code 2010 [C12.5.2-3] provisions a number of rules for the design of the
integrity reinforcement, accounting for the rebar arrangement and ductility class of the reinforcement. In the
revision of EN 1992-1-1:2004, this state-of-the-art was incorporated in an effort to provide design guidance for
integrity reinforcement of new structures and an assessment method for existing ones.
In addition, since punching shear reinforcement is increasingly used in many Countries, there is a need for
consistent provisions on how to combine integrity and punching shear reinforcement.

Slabs without shear Several strategies can be followed to enhance the post-punching resistance of a slab-column connection after a
reinforcement punching failure. They normally consist of arranging shear reinforcement in the form of vertical or bent-up bars
(enhancing the deformation capacity and toughness of the connection) and/or providing the so-called integrity
12.5.2(1) reinforcement. The latter is typically a horizontal reinforcement provided in the compression zone which can
carry the shear force after punching by catenary action. This reinforcement can be activated in tension following
the significant deformations occurring after a punching shear failure.
The performance of the different solutions is highly variable, particularly with respect to the integrity
reinforcement. This is for instance shown in Figure C12.5.2.1 with reference to a number of tests performed by
[C12.5.2-1] considering different integrity reinforcement details. As can be seen, without any shear
reinforcement or reinforcement arranged on the compression side (Figure C12.5.2.1a), a sudden drop in the
load occurs after punching with a relatively low residual resistance (this residual resistance after punching being
associated to the catenary action of hogging reinforcement which is limited by the dowel resistance associated
to the concrete tensile strength [C12.5.2-1]). When longitudinal reinforcement is placed on the compression side
(Figure C12.5.2.1b) or bent-up bars (Figure C12.5.2.1c) are provided, the performance is significantly better,
and levels of load comparable to those leading to first punching can be attained after a reasonable
displacement. These tests also show (Figure C12.5.2.1d) the significance of the anchorage conditions of the
hogging reinforcement, governing both the punching and post-punching performances.

A. Muttoni, M. Fernández Ruiz 24.02.2023

For citations, copying and referencing see Note on Front page! page 505
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.5.2 Background to FprEN 1992-1-1:2023

Figure C12.5.2.1: Deviation forces of curved reinforcement: (a) deviation forces; and (b) spalling resistance
(from [C12.5.2-1])
Following these experimental findings, a rational model was proposed by [C12.5.2-1,2] for calculating the post-
punching resistance of flat slabs. The model considered notably:
 The contribution of the reinforcement in the compression zone which can be activated by catenary
action after some vertical displacement following punching failure. The performance of such
reinforcement is shown to be highly dependent on its ductility class and the capacity of the concrete
core around the column to anchor it;
 The contribution of bent-up bars intersected by the punching cone, whose angle with the slab plane
makes their contribution more efficient for post-punching resistance than straight bars (the catenary
action being activated for significantly smaller deformations);
 The contribution of the hogging reinforcement. This contribution depends on the tensile strength of the
concrete cover and its anchorage conditions.
On the basis of the rational model of [C12.5.2-1], simple design provisions were implemented in fib Model Code
2010 [C12.5.2-3]. For such provisions, the following considerations were accounted for:
 The contribution of the hogging reinforcement is neglected (as it depends on the tensile strength of
concrete, it only contributes to a relatively low post-punching resistance, Figure C12.5.2.1a and its
efficiency highly depends on the anchorage conditions [C12.5.2-1], Figure C12.5.2.1d);
 The contribution of the reinforcement in the compression zone or of bent-up bars is considered by
means of their strength at rupture and their angle b associated to catenary action (refer to Figure
C12.5.2.2a). For calculation of the angle, the deformation capacity of the reinforcement is governing;
 A condition is introduced limiting the catenary action by rupture (pull-out) of the concrete core around
the column (Figure C12.5.2.2b).

(a) (b)

Figure C12.5.2.2: Post-punching resistance: (a) dowelling of compression reinforcement and resulting angle
(ult); and (b) rupture (pull-out) of concrete core limiting the angle of dowelled bars (from
[C12.5.2-1])
The resulting expression of fib Model Code 2010 reads:

A. Muttoni, M. Fernández Ruiz 24.02.2023

For citations, copying and referencing see Note on Front page! page 506
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.5.2 Background to FprEN 1992-1-1:2023

0.5 fck
VRd ,int   As fyd (ft / fy )k sin b  d res bint (C12.5.2.1)
c

where As refers to the sections of reinforcement in the compression zone suitably anchored over the column or
passing through it (see definition in Figure C12.5.2.3), fyd to their design value of the yield strength, (ft/fy)k to the
characteristic ratio between tensile and yield strength, b to the angle of the integrity reinforcement at rupture
(0° for class A, 20° for class B and 25° for class C and 40° for bent-up bars class BC, fck to the characteristic
value of the compressive strength of concrete, C to its partial safety factor, dres is the core height between layers
(Figure C12.5.2.2b) and bint the integrity perimeter. As it can be noted, in Formula (C.12.5.2.1), the left-hand
side part refers to the capacity of the catenary action of the reinforcement (initially horizontal or bent-up bars)
while the right-hand side refers to the upper limit related to the capacity of the concrete core to carry the shear
force in the integrity reinforcement.
The value of the post-punching resistance provided by the integrity reinforcement (Formula (C.12.5.2.1)) is to be
compared to the action considering an accidental design situation (with reduced partial safety factors for the
actions, refer to commentary to §7.3.5.6 of MC2010). Also, MC2010 indicates that integrity reinforcement
according to Formula (C.12.5.2.1) is not required provided that the deformation capacity of the slab is sufficient
(MC2010 commentary to §7.3.5.1), which can be ensured by arrangement of conventional shear reinforcement
(MC2010 §7.3.5.6, designed to comply with the minimum amount given in Formula (7.3-68) of MC2010).
Based on the Model Code 2010 provisions, a design expression was proposed for FprEN 1992-1-1:2023, by
considering some simplifications:
 The considerations of ductility Class of reinforcement, influencing both ult and (ft/fy)k in Formula
(C.12.5.2.1), can be lumped into one single parameter
 The check for the capacity of the concrete core is only governing in special cases (low concrete
strength and/or shallow slabs). It is thus disregarded for the purpose of ease-of-use.
From these considerations, the following design formula results:

VRd ,int   As ,int  f yd  k int  VEd (12.10)

where As,int refers to the sections of reinforcement suitably anchored over the column or passing through it (see
Figure C12.5.2.3), kint defines the performance of the integrity reinforcement and depends on the ductility class
and VEd refers to the action under accidental situation. The value of coefficient kint, lumping the information on
the angle of the reinforcement related to catenary action and tensile strength, results 0.37 for Class B
reinforcement and 0.49 for Class C reinforcement. As it can be noted, this formulation is comparable to that of
MC2010 (Formula (C.12.5.2.1)). Due to its limited deformation capacity, class A reinforcement is not allowed as
integrity reinforcement (as in fib MC2010).

Slabs with shear As previously stated, in MC2010, no integrity reinforcement is required provided that shear reinforcement is
reinforcement provided (complying with the conditions of minimum amount given in Formula (7.3-68) of MC2010). Different to
that consideration, FprEN 1992-1-1:2023 states that integrity reinforcement is required for shear-reinforced
12.5.2(2)
slabs when the contribution of the shear reinforcement during the post-punching regime (VRd,w,int) is not sufficient
to ensure the resistance of the connection under accidental actions (VEd):

VRd ,w,int   w  f ywd  sin  w  b0 ,5  dv  VEd (12.11)

This Formula considers that, during the post-punching regime, the displacements are sufficiently large to ensure
yielding of the shear reinforcement. This can be considered as a reasonable assumption according to test
results (refer for instance to Figure C12.5.2.1).
When the resistance of the shear reinforcement in the post-punching regime (VRd,w,in) is lower than the shear
demand (VEd), integrity reinforcement has to be arranged according to Formula (12.10) providing a resistance
higher than VEd – VRd,w,int . This considers thus that a part of the shear force is carried by the shear reinforcement
and the remaining part by the integrity reinforcement.
Tests conducted with slabs reinforced with bent-up bars have shown that this type of reinforcement is not only
efficient in increasing the punching shear resistance, but can also be easily activated as integrity reinforcement
[C12.5.2-1] (see also Fig. C12.5.2.1c). For this reason, in fib MC 2010 [C12.5.2-3], bent-up bars are mentioned
as an efficient type of integrity reinforcement. In the approach of FprEN 1992-1-1:2023, such reinforcement can
be considered as a shear reinforcement leading in fact to the same result.

Detailing rules for It is worth mentioning that to ensure the required efficiency, integrity reinforcement should fulfil the detailing
integrity reinforcement rules defined in 12.5.2(1). The bars considered in the design Formula (12.10) should be:
 of ductility class B or C (ductility class A reinforcement does not have sufficiently deformation capacity
12.5.2(1)
to activate catenary action),
 anchored in the column or passing through it (see in Figure C12.5.2.3 the sections which may be
considered) and
 placed on the compression side of the slab within the vertical column reinforcement (ideally, integrity

A. Muttoni, M. Fernández Ruiz 24.02.2023

For citations, copying and referencing see Note on Front page! page 507
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.5.2 Background to FprEN 1992-1-1:2023

reinforcement should be placed as a 1st and 2nd layer of reinforcement to ensure sufficient resistance
against rupture of concrete core (see Figure C12.5.2.2b), otherwise, the verification according the
rigth-hand side part of Formula (C12.5.2.1) of Model Code 2010 becomes governing. Concerning the
requirement of the position with respect to the column reinforcement, it is necessary to avoid local
failures at the edge of the column, which can potentially limit the capacity of the integrity
reinforcement to transmit its force to the column).
In addition, punching shear reinforcement which is considered in the reduction of the required integrity
reinforcement according to Formula (12.11) shall enclose the hogging reinforcement according to 12.4.2(4)
(studs should be arranged so that the upper level of their heads is at the upper level of the hogging
reinforcements, see Figure 12.2e) and fulfil the requirements of 12.5 (particularly with respect to the distance
between the column face and the first shear reinforcement in Figure 12.7).

Figure C12.5.2.3: Example of integrity reinforcement over an edge column with 6 sections of bars which may
be considered

Contribution of hogging As shown in Figure C12.5.2.1a, the contribution of significant flexural reinforcement is not negligible. As already
reinforcement described above, the residual resistance after punching associated to hogging reinforcement results from the
catenary action which is limited by the dowel resistance associated to the concrete tensile strength [C12.5.2-1]).
12.5.2(3)
According to [C12.5.2-1], this contribution can be calculated as:

f ck
VRd ,hog  nhog     bef ,hog (12.12)
C
where nhog is number of bars crossing the control perimeter b0,5 (for the sake of simplicity, compared to [C12.5.2-
1], only the bars inside the control perimeter are considered) and fully anchored at a distance 4d from the
control perimeter (bars crossing two times the control perimeter can be considered twice, whereas bars
anchored within the column are considered only once in a similar manner as shown in Figure C12.5.2.3).
According to [C12.5.2-1], the catenary action of the hogging reinforcement has to be transferred to the concrete
of the slab through dowel action activating the concrete tensile strength (approximated as 0 ,5 f ck ) over an

area 2∙∙ bef,hog for every bar, where bef,hog can be calculated as:

bef ,hog  min  s- ; 6 ;4c  (C12.5.2.2)

where s is the spacing of the hogging reinforcement and c is the cover of the hogging reinforcement.

Alternative forms of Alternatives to the forms of integrity reinforcement described above can consist of special shear reinforcements
integrity reinforcement or shearheads, providing a significant deformation capacity and a sufficient residual resistance after punching
(an instance based on bars ensuring significant dowel action with ductile behaviour and efficient anchorages is
described in [C12.5.2-4]).

Economic considerations With respect to EN 1992-1-1:2004, the provisions of FprEN 1992-1-1:2023 require more integrity reinforcement
in slabs without punching shear reinforcement and with moderate hogging reinforcement. On the other side, the
new provisions allow combining the contributions of shear, hogging and integrity reinforcement. From an
economic perspective, in most cases, it is more efficient to invest in a reinforcement which can be used for both
the accidental as well as for persistent and transient design situations, allowing to reduce the slab thickness
or/and the column dimensions.

[C12.5.2-1] Fernández Ruiz, M., Mirzaei, Y., Muttoni, A., Post-punching behavior of flat slabs, American
References

A. Muttoni, M. Fernández Ruiz 24.02.2023

For citations, copying and referencing see Note on Front page! page 508
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
12.5.2 Background to FprEN 1992-1-1:2023

Concrete Institute, Structural Journal, Vol. 110, No. 5, 2013, p. 801—812

[C12.5.2-2] Habibi, F., Cook, W.D., Mitchell, D., Predicting Post-Punching Shear Response of Slab-Column
Connections, ACI Structural Journal, Vol. 111, No. 1, 2014, p. 123-134
[C12.5.2-3] International Federation for Structural Concrete, “fib Model Code for Concrete Structures 2010,”
Ernst & Sohn, Berlin, Germany, 434 p., 2013.
[C12.5.2-4] Cantone, R., Fernández Ruiz, M., Bujnak, J., Muttoni, A., Enhancing the punching strength and
deformation capacity of flat slabs by arrangement of horizontal dowel studs, American Concrete
Institute, Structural Journal, Vol. 116, No. 5, 2019, pp. 261--274

List of versions of this document

26.02.2021, AM, MFR 1st draft
02.30.2021, AM review, with some additional content
24.02.2023, AM comment to 12.5.2(3) added, references to FprEN 1992-1-1:2023 updated

A. Muttoni, M. Fernández Ruiz 24.02.2023

For citations, copying and referencing see Note on Front page! page 509
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1:2023

Background document to clause 13.5.1

Minimum concrete cover for pre-tensioning
tendons

Dr.-Ing. Philipp Schmidt Institute of structural concrete (IMB)

Univ.-Prof. Dr.-Ing. Josef Hegger RWTH Aachen University
24.01.2023

For citations, copying and referencing see Note on Front page! page 510
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
13.5.1 Background to FprEN 1992-1-1:2023

Introduction
This document was prepared in order to provide background information to the specifications regarding the
minimum concrete cover for pre-tensioning tendons. A sufficient concrete cover (cmin,b) and clear tendon spacing
(s) shall guarantee crack-free transfer of prestressing within the anchorage area and shall ensure the bond
between concrete and prestressed tendon.

13.5.1 Reasons for change

Table 13.1 While in current EC2 the requirements for clear spacing between tendons are given in clause 8.10 and the
Minimum concrete cover specifications regarding the required minimum concrete cover can be found in a Note of clause 4.4.1.2, the
regulations are merged for the new EC2 in clause 13.5 (illustrated by Figure 13.1 and specified by Table 13.1).

Background
In general, it should be differentiated between the required minimum concrete cover for strands and indented
wires due to the different bond mechanisms.
For strands, the specifications of Table 13.1 are in accordance with the work of NITSCH [1] to avoid splitting of
concrete and were also discussed by GEßNER [2]. In contrast to Model Code 2010 provisions [3], no distinction
is made between normal-strength and high-strength concrete since higher concrete strengths are automatically
accompanied by shorter transfer lengths and thus higher bond stresses leading to larger splitting forces [4].
Even if the tests of DEN UIJL [5] seem to allow smaller concrete covers for high-strength concrete, this effect is
omitted to provide minimum concrete covers on the safe side. The illustration of DEN UIJL [5] shown in Figure
13.1, which was also picked up by GEßNER [2], summarizes the interplay of occurring bond-splitting cracks
depending on concrete cover and clear spacing of tendons.

Figure 13.1: Occurrence of bond-splitting cracks in normal- and high-strength concrete as a function of
strand spacing and concrete cover reproduced from DEN UIJL [5]
Note: “1/3” means 1 specimen with splitting cracks out of 3 test specimens. Grey marked
zone for specimens without splitting cracks
Overall, it must be considered that most tests such as of NITSCH [1] and DEN UIJL [5] were conducted with 0.5’’
strands and that it is not clearly verified whether the relative values of cmin,b/Ø and s/Ø can also be directly
applied for larger strand diameters. In case of members with multiple prestressing tendons in one layer and little
or no transverse reinforcement, NITSCH [1] pointed out that splitting can not be totally ruled out regarding the
recommended minimum concrete cover values. The use of the provisions of clause 13.5 should, without further
evidence, be limited to a maximum strand diameter of 15.7 mm.
For indented wires, previous specifications were given in Model Code 2010 [3]. However, an experimental test
campaign by GEßNER [2] revealed that prestressed indented wires failed more often due to longitudinal cracks
compared to tests with strands or smooth wires. Consequently, a concrete cover of cmin,b/Ø = 2.5 is insufficient
for indented wires. Tests with cmin,b/Ø = 3.5 showed better results [2] but anchorage failure could still not be
excluded with certainty. As a result, an increased concrete cover of cmin,b/Ø = 4.0 is proposed for indented wires,
which is nevertheless smaller compared to the specifications of corresponding technical approvals. In the tests
of GEßNER [2], the clear spacing of prestressed tendons was s > 2.5Ø. To account for smaller clear spacings of
s = 2.0Ø, the concrete cover for indented wires should be increased by 0.5Ø in accordance to the approach of
NITSCH [1] for strands.
Furthermore, the values for the minimum concrete cover given in the technical approvals of pre-tensioning
tendons have generally to be considered. For example for special applications such as hollow-core slabs, other
values of tendon spacing and concrete cover may be used for pre-tensioned members to fullfill all requirements
regarding the ultimate limit state as well as serviceability limit state and providing that satisfactory behaviour in
service is demonstrated by testing or permitted by the specific European product standard with factory
production control.

P. Schmidt, J. Hegger 02.06.2021

For citations, copying and referencing see Note on Front page! page 511
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
13.5.1 Background to FprEN 1992-1-1:2023

[1] A. Nitsch, “Spannbetonfertigteile mit teilweiser Vorspannung aus hochfestem Beton“, Dissertation,
References
RWTH Aachen University, Aachen, 2001.
[2] S. Geßner, “Bond and anchorage of pre-tensioning tendons“, Dissertation, RWTH Aachen University,
Aachen, 2019.
[3] International Federation for Structural Concrete, “fib Model Code for Concrete Structures (Model
Code 2010)”, Ernst & Sohn, Berlin, 2013.
[4] G. Bertram, “Zum Verbund- und Querkrafttragverhalten von Spannbetonträgern aus Ultra-hochfestem
Beton“, Dissertation, RWTH Aachen University, Aachen, 2012.
[5] J.A. den Uijl, “Transfer length of prestressing strand in HPC“, Progress in Concrete Research, Vol. 4,
TU Delft, 1995.

List of versions of this document

02.06.2021. 1st draft
01.03.2023 Final Draft

P. Schmidt, J. Hegger 02.06.2021

For citations, copying and referencing see Note on Front page! page 512
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to clauses

13.5.3 and 13.5.4
Transmission length and anchorage length
of pre-tensioning tendons

Hans Rudolf Ganz Switzerland

30.09.2022

For citations, copying and referencing see Note on Front page! page 513
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
13.5.3 and 13.5.4 Background to FprEN 1992-1-1:2022

Introduction
This document has been prepared in order to explain the slight modifications in the presentation of transmission
length and anchorage length of pre-tensioning tendons introduced in FprEN 1992-1-1:2023 when compared
with EN 1992-1-1:2004. These changes are mainly of editorial character with the intent not to significantly
change the actual transmission and anchorage length.

13.5.3 and 13.5.4 Reasons for change

Transfer of prestress and Current provision for the transfer of prestress in EN 1992-1-1:2004 (clause 8.10.2.2) starts with the definition of
anchorage of tensile force the design bond strength of pretensioning tendons, Formula (8.15), and using it then to calculate the
at ULS transmission length in Formula (8.16). In line with the changes introduced for anchorage of reinforcing steel, the
concept of bond strength has been removed in FprEN 1992-1-1:2023 also for the transmission length of
pretensioning tendons. Instead, the two formulae have been merged into one providing the design value of the
transmission length directly, FprEN 1992-1-1:2023, Formula (13.4). This was felt to improve ease of use.
In EN 1992-1-1:2004 the design value of the concrete tensile strength was used for the determination of the
design bond strength. The design value of the concrete tensile strength was determined with a recommended
value of ct = 1,0. In FprEN 1992-1-1:2023 ct was replaced by ktt with a recommended value of 0,8 or 0,7. To
avoid a significant change in the value of the transmission length, it was decided to avoid the use of fctd and
instead, similar to other provisions in Clause 8, use fck instead of fctd.
The transmission length in EN 1992-1-1:2004 includes, through the use of the design value of bond strength,
the partial factor for concrete C. Hence, to give NSBs the opportunity to adapt the safety level, and in order to
use the formula for transmission length also for fire design with different partial factors, the partial factor C was
introduced directly into the formula for the transmission length in FprEN 1992-1-1:2023, with a coefficient
(C/1,5), i.e. setting this coefficient to 1,0 if the recommended value of C = 1,5 is used.
The same approach was used for the presentation of the design anchorage length of pretensioning tendons
under Clause 13.5.4.

Summary of results
The above changes resulted in the formulae (13.4) and (13.9) for transmission length and anchorage length,
respectively, in FprEN 1992-1-1:2023. There were no changes in the coefficient 1 for release of tendons, nor
for the coefficient 1 for bond conditions:

H.R. Ganz 30.09.2022

For citations, copying and referencing see Note on Front page! page 514
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
13.5.3 and 13.5.4 Background to FprEN 1992-1-1:2022

Comparison of results
Figures C13-5.1 and C13-5.2 show a comparison of transmission and anchorage length according to EN 1992-
1-1:2004 and FprEN 1992-1-1:2023 as a function of the concrete compressive strength at the relevant time,
fck(t). For the comparison, good bond conditions and gradual release of tendons were assumed. In addition, for
the anchorage length, a ratio (pd - pm)/pm0 = 0,38 was assumed. The vertical axis has been made
dimensionless by presenting (lpt x 1)/(p x 1) and (lbpd x 1)/(p x 1), respectively.

Figure C13-5.1: Comparison of transmission length between EN 1992-1-1:2004 and FprEN 1992-1-1:2022

Figure C13-5.2: Comparison of anchorage length between EN 1992-1-1:2004 and FprEN 1992-1-1:2023

Conclusions
‐ The two approaches in EN 1992-1-1:2004 and FprEN 1992-1-1:2023 give very similar results for
transmission length at relatively low concrete strength which is considered typical for transmission.
The differences between the two approaches are slightly larger for wires.
‐ The two approaches in EN 1992-1-1:2004 and FprEN 1992-1-1:2023 give very similar results for
anchorage length at medium to higher concrete strength which is considered typical for ULS
verifications. The differences between the two approaches are slightly larger for wires.
‐ According to the knowledge of the author of this background document, there is no testing available
for strands diameters larger than 15.7mm. Hence, the use of provisions 13.5.3 and 13.5.4 should
without further evidence be limited to a maximum strand diameter of 15.7mm.

References

H.R. Ganz 30.09.2022

For citations, copying and referencing see Note on Front page! page 515
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
13.5.3 and 13.5.4 Background to FprEN 1992-1-1:2022

List of versions of this document

26.02.2021. 1st draft prEN 1992-1-1:2020-11
30.09.2022 Updated to FprEN 1992-1-1:2023
22.02.2023 Final draft

H.R. Ganz 30.09.2022

For citations, copying and referencing see Note on Front page! page 516
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to clause 13.5.5

Shear resistance of precast members
without shear reinforcement

Mikael Hallgren Tyréns AB

Josef Hegger RWTH Aachen
Simon Wijte Eindhoven University of Technology
14.02.2023

For citations, copying and referencing see Note on Front page! page 517
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
13.5.5 Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared to explain the changes regarding pure shear tension capacity of prestressed
members without shear reinforcement.

13.5.5 Reasons for change

Shear resistance of In EN 1992-1-1 [2004] the shear resistance of precast members without shear reinforcement is described by
precast members without equation 6.4:
shear reinforcement

Ibw
 fctd 
2
VRd,c =   l cpfctd (1)
S
In equation (1) the shear resistance is verified by checking the principal tensile stress to the design tensile
strength.
This description of the shear resistance is moved to main clause 13 as it applies only to prestressed members
which usually refers to precast elements. It should be noted that the limitation to simply supported members as
contained in EN 1992-1-1:2004 was removed during the preparation of FprEN 1992-1-1:2023 since the verification
method applies to continuous members also.
The provision in this clause is limited to the effective depth d not larger than 500 mm unless the size effect is
considered in a refined analysis. This is to conservatively consider the size effect on the brittle type of failure
caused by exceeding the tensile strength of plain concrete. The Commentary to EC2 [1], i.e. the background to
EN 1992-1-1:2004, specifically refers to prestressed hollow core slabs when exemplifying special cases where
pure shear tension failure can occur. Hollow core slabs typically do not have effective depths exceeding 500 mm.
There are very few tests on prestressed beams failing in pure shear tension and with d exceeding 500 mm that
are reported in the literature. The ACI-DAfStB database of shear tests on prestressed beams without shear
reinforcement contains only a few tests with d ranging from 575 mm to 825 mm, and only one test with d = 1025
mm, see Figure 1.

Figure 1: Number of tested prestressed beams versus the effective depth d, adopted from [2].

Tests by Bertram [3] include tests on prestressed girders with d up to 917 mm, but these beams also contained
steel fibres. However, the test [3] indicate a decreasing load reserve after shear cracking with increase of beam
height, se Figure 2.

Figure 2: Shear test results of prestressed girders of SFRC without shear reinforcement, from [3].
Resent research by Roosen [4] on shear failure of prestressed girders in regions without flexural cracks also
confirms that the shear strength depends on the size of the region subjected to high tensile stresses.

M. Hallgren, J. Hegger, S. Wijte 14.02.2023

For citations, copying and referencing see Note on Front page! page 518
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
13.5.5 Background to FprEN 1992-1-1:2023

The provisions for shear resistance of 13.5.5 based on principal stress may be applied to members with larger
effective depth than 500 mm only if the size effect is considered in a refined analysis. However, since beams
typically have a minimum shear reinforcement to Clause 12, it is suggested to alternatively verify shear capacity
with the refined provisions of 8.2.3 accounting for the provided minimum shear reinforcement and considering the
minimal inclination of the compression field under the design prestressing force.
Equation (1) describes the resistance based on the principal tensile stress in the cross section at the centroidal
axis. In case of irregular shaped cross sections, it is very well possible that the maximum principal tensile stress
is found in another fibre than that at the centroidal axis. Therefore, the new clause 13.5.5(2) introduces a limit of
the tensile stress in concrete based on the general formulation of principal tensile stress using Mohr´s circle:

x,Ed x,Ed
𝜎1,Ed 𝑦 𝜏Ed 𝑦 (2)

where the normal stress σx,Ed(y) and the shear stress τEd(y) in the cross section are assessed in a fibre at distance
y from the centroidal axis. The shear stress is defined by
Ed
𝜏Ed 𝑦 (3)
⋅

where VEd is the design shear force at the cross section, S(y) is the first moment of the cross section area, I is the
second moment of the cross section area, b(y) is the width at y.
With this general formulation, the maximum principal tensile stress can be assessed for a cross section with any
given cross-sectional shape.
The design condition follows
ctk,0,05
𝜎1,Ed (4)

where the design tensile strength is given by fctk0.05/C, and is not affected by a factor for high sustained load for
this design.
Another elaboration of this verifcation can be found in EN 1168 [2008], in which the method for prestressed hollow-
core slabs is given and in which also the shear stress due to the transfer of the prestress is considered.

[1] Eurocode 2 – Commentary. European Concrete Plattform ASBL, June 2008.

References
[2] Dunkelberg D, Sneed LH, Zilch K, Reineck K-H. The 2015 ACI-DAfStb data base of shear tests on
slender prestressed concrete beams without stirrups—Overview and evaluation of current design
approaches. Structural Concrete. 2018;19:1740–1759.
[3] Bertram G. Zum Verbund- und Querkrafttragverhalten von Spannbetonträgern aus Ultra-hochfestem
Beton. Dissertation, RWTH Aachen, 2012.
[4] Roosen M.A. Shear failure of prestressed girders in regions without flexural cracks. PhD dissertation
TU Delft, ISBN 978-94-6421-285-3, 2021.

List of versions of this document

12.05.2021 1st draft
06.02.2023 Peer review
14.02.2023 Final draft

M. Hallgren, J. Hegger, S. Wijte 14.02.2023

For citations, copying and referencing see Note on Front page! page 519
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

CEN/TC250/SC2/WG1/TG6 FprEN 1992-1-1:2023

Additional background information to

Clause 14.4.5.2
Simplified design method for unreinforced
walls and columns

Simon Wijte Eindhoven University of Technology

03.01.2023

For citations, copying and referencing see Note on Front page! page 520
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared in order to provide additional background information to the revision of the
design rule to the determine the ultimate load bearing resistance of a braced unreinforced concrete wall loaded
by an axial force with a maximum first order eccentricity.

Reasons for change

14.4.5.2 In EN 1992-1-1:2004 in clause 12.6.5.2 a simplified design rule, equation 12.11, is given for the determination of
Simplified design method the axial load resistance of a braced unreinforced concrete wall.
for walls and columns
NRd = b hw fcd  (1)

where N Rd is the axial load resistance, b is the width of the cross-section, hw is the height of the cross-section,
fcd is the design value of the compressive strength of unreinforced concrete and  is a factor taking into account
eccentricity, second order effects and normal effects of creep. The factor  is described by formula (2)

1,14(1 – 2 etot/hw) – 0,02 lo/hw  1 – 2 etot/hw (2)

where etot is the sum of the first order eccentricities caused by the loads, the floor wall interaction and additional
eccentricity covering the effects of geometrical imperfections and l0 is the effective length of the wall.

While discussing prEN 1992-1-1:2004/A1:2014 it was proposed to change equation 12.11 to:

 = 0,71(1 – 2 etot/hw) – 0,013 lo/hw  1 – 2 etot/hw (3)

By WG1/N132 the French suggested the following equation:

 = 1,07(1 – 2 etot/hw) – 0,026 lo/hw  1 – 2 etot/hw (4)

Thereafter the several proposals were discussed during the 7th WG1 meeting in Ispra. As a result of that
discussion document WG1/N171 [1] is written. In that document the several proposals were compared and
additional also a comparison with results from structural analyses based on the finite difference method was
made. In these structural analyses a bilinear stress-strain-relation as described in figure 3.4 of
EN 1992-1-1:2004 is used as a starting point for the determination of the moment-curvature relation. From this it
is concluded that:
- formula (3) leads to more conservative results than formula (4);
- formula (4) leads to more conservative results than formula (2);
- formula (2) leads to a capacity which is comparable to the capacity that can be found when a structural
analysis, in this case using the finite difference method, is performed;
- the difference between the theoretical results from the structural analysis and formula (1) is covered by
introducing an additional eccentricity due to creep: e = 0,001 l0.

The proposal from N171 was adapted for EN 1992-1-1:2004/A1:2014. So, in EN 1992-1-1:2004/A1:2014
equation 12.11 [1] was not changed, while the variable e is introduced to contribute to etot. Formula (2) was
maintained in the early drafts of prEN 1992-1-1, where the number of the formula was changed to (14.11).

In WG1/N179 again formula (4) is maintained as French proposal for equation 14.11. In the proposal it is stated
that when using formula (4) the addition of e is not required. Formula 4 is the result of many calculations
performed. Also, it is learned that the effective creep coefficient, calculated from annex B of EN 1992-1-1:2004
is considered in the background of that proposal.

For citations, copying and referencing see Note on Front page! page 521
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

Comparison of several formulas

In [1] a comparison between the formulas (2), (3) and (4) is given without the addition of e. Hereafter additional
to this, a comparison also is made with the m variable in EN 1996-1-1:2005+A1:2012. This variable m is
comparable to the variable  in equation 14.11. It describes the maximum axial load in an unreinforced
masonry wall based on first order eccentricities and a ratio between the effective wall height and the effective
wall thickness.
From figure 14.4.5.2-1 is can be learned that, as already is described in [1], the formula in EN 1992-1-1 (formula
(2)) leads to the highest resistance compared with formula (4) and the proposal for prEN 1992-1-
1:2004/A1:2014 (formula (3)). It also can be learned that approach in EN 1996-1-1:2011 (Eurocode 6 for
masonry structures), for E = 1000 fk leads to an even higher resistance than should follow from the present EN
1992-1-1:2004. So, when these formulas are applied, the capacity of a masonry wall with similar dimensions
and a comparable design value of the compressive strength is higher than that of a unreinforced concrete wall.
In EN 1996-1-1:2022 a revised method to determine m is described in annex F which, when it is compared to
EN 1996-1-1:2012, for smaller ratios of etot/hw and higher ratios of l0/hw even leads to an increase of the m
value, see figure 14.4.5.2-1.
1.0
for 2 (0.05)
for 3 (0.05)
0.9 for 4 (0.05)
EN 1996‐1‐1:2011 (0.05)
EN 1996‐1‐1:2022 (0.05)
0.8
for 2 (0.20)
for 3 (0.20)
0.7 for 4 (0.20)
EN 1996‐1‐1:2011 (0.20)
EN 1996‐1‐1:2022 (0.20)
0.6
for 2 (0.35)
for 3 (0.35)
for 4 (0.35)


0.5
EN 1996‐1‐1:2011 (0.35)
EN 1996‐1‐1:2022 (0.35)
0.4

0.3

0.2

0.1

0.0
0 5 10 15 20 25 30
l0/hw

Figure 14.4.5.2‐1: Comparison of several formulas for F with a variation of l0/hw, for etot/hw ratios of 0,05, 0,20
and 0,35
As a result from the comparison it is learned that when the creep coefficient implictly is taken into account, as
according to the background documentation, is done in formula (4), obvious smaller values for will be found.
Whether the creep coefficient should be applied to a second order analysis for braced unreinforced walls is
discussed between several member states represented in WG1.
It may be noted that from the background information for annex F of prEN 1996-1-1:1999 [2] it can be learned
that in this code for masonry structures creep effects are not considered in the design of vertical loaded walls,
see figure 14.4.5.2-2. In EN 1996-1-1:2005+A1:2012 creep was considered by applying and additional
eccentricity for slender walls (lo/hw  15). In FprEN 1996-1-1:2021 this additional eccentricity is not described.

For citations, copying and referencing see Note on Front page! page 522
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

Figure 14.4.5.2‐2: Figure 2 from SC6/N65 background information to annex F of prEN 1996-1-1 [2]
Comparison of the proposed formula with experimental data and the formula for e/h = 0

Therefore a proposal for formula (14.11) in FprEN 1992-1-1:2023 is derived in which the creep coefficient is a
variabel.

Background FprEN 1992-1-1:2023 formula (14.11)

To derive an alternative formula for formula (14.11) in which the creep coefficient is a variable, an analysis
model is created according the general method for non-linear analysis as described in 7.4.3.3, in which:
 a braced wall loaded by an axial load with a first order eccentricity etot which is constant over the length of the
wall is considered, see figure 14.4.5.2-3;
 the stress-strain relation presented in figure 5.1 of FprEN 1992-1-1:2023, is used where instead of Ecm the
value for Ecd is used with, CE = 1,2 and where the strain is increased by the creep coefficient by multiplying
with (1 + eff), see figure 14.4.5.2-4.
 a relation between moment and curvature is determined based on the assumption that plane section remain
plane and the tensile strength as well as the stiffness of concrete in tension is neglected.

Figure 14.4.5.2‐3: Scheme of wall in calculation model

Assuming a first order eccentricity which is constant over the length of the wall leads to the largest second order
effect for the particular wall. Therefore the method can be used as a conservative approach where in case of
lateral loading the largest eccentricity resulting from the lateral loading is used in the formula.
With the analysis model iterative calculations are performed for an increasing axial force in which both material
and geometrical non-linear behaviour are considered. From the maximum value of the axial force for which an
equilibrium can be found, the relative factor  is derived.

For citations, copying and referencing see Note on Front page! page 523
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

1.1

1.0

0.9

compressive stress / fcd

0.8

0.7

0.6
C20/2
C20/25 5 = e0f f = 0
‐ jeff
0.5

0.4 C20/2
C20/25 5 = e1f f = 1
‐ jeff

0.3 C40/5
C40/50 0 = e0f f = 0
‐ jeff

0.2 C40/5
C40/50 0 = e1f f = 1
‐ jeff
0.1

0.0
0.000 ‐0.001 ‐0.002 ‐0.003 ‐0.004 ‐0.005 ‐0.006 ‐0.007 ‐0.008

strain

Figure 14.4.5.2‐4: Relative stress-strain relation for plain C20/25 en C50/60 with eff equal to 0 and 1

As can be seen from figure 14.4.5.2-4, due to the use of the stress-strain relation as described in figure 5.1 of
FprEN 1992-1-1:2023 the relation between the stress and the strain no longer is linear related to the
compressive strength. Therefore, calculations have to be performed for several concrete grades. For this
C20/25 and C40/50 have been chosen as being a minimum and maximum grade applied in practice of plain
concrete walls.
Results of the several calculations are summarized in table 14.4.5.2-1 and figure 14.4.5.2-5 hereafter.

Table 14.4.5.2‐1: Summary of results from calculation model

For citations, copying and referencing see Note on Front page! page 524
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

0.9

etot/hw fck eff

0.8 G00.05‐20‐0
.05 20 0,0

G00.05‐20‐0.6
.05 20 0,6

G00.05‐20‐1.4
.05 20 1,4
0.7
G00.05‐40‐0
.05 40 0,0

G00.05‐40‐0.6
.05 40 0,6

0.6 G00.05‐40‐1.4
.05 40 1,4

G00.2‐20‐0
.20 20 0,0

G00.2‐20‐0.6
.20 20 0,6

0.5
G00.2‐20‐1.4
.20 20 1,4

G00.2‐40‐0
.20 40 0,0

0.4 G00.2‐40‐0.6
.20 40 0,6

G00.2‐40‐1.4
.20 40 1,4

G00.35‐20‐0
.35 20 0,0
0.3
G00.35‐20‐0.6
.35 20 0,6

G00.35‐20‐1.4
.35 20 1,4

0.2 G00.35‐40‐0
.35 40 0,0

G00.35‐40‐0.6
.35 40 0,6

G00.35‐40‐1.4
.35 40 1,4

0.1

0
0 5 10 15 20 25 30
l0/hw

Figure 14.4.5.2‐5: Maximum value of  related to a variation of eccentricity ratio, concrete strength, creep
coefficient and aspect ratio

Based on [3] and the results presented in table 14.4.5.2-1 a new formula for  has been derived. It is similar to
that in [3], but with slight modifications proposed by the author of [3], Bo Westerberg:"


 l e
1   2,1  0,02 0  tot
 h h
= 2 0,6
(5)
 l0   etot  0,8  eff   fcd,pl 
1     0,9  6   20 
h  h 
 1000  

Where etot is the sum of e0 and ei, e0 is the first order eccentricity caused by the load, ei is the additional
eccentricity covering the effects of geometrical imperfections and fcd,pl is the design value of the compressive
strength of plain concrete.

In the figures 14.4.5.2-6 and -7 hereafter the results of the formula (5) are compared with the results
from the analysis model. Each concrete class considered, is presented in a separate figure.

For citations, copying and referencing see Note on Front page! page 525
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

1.0
e t o t /h w -f c k - eff
e t o t /h w -f c k - eff
0.9
0.05‐20‐0
0.05‐20‐0.6
0.8 0.05‐20‐1.4
0.2‐20‐0
0.7 0.2‐20‐0.6
0.2‐20‐1.4

0.6 0.35‐20‐0
0.35‐20‐0.6
0.35‐20‐1.4
0.5

(5) 0.05‐20‐0
(5) 0.05‐20‐0.6
0.4 (5) 0.05‐20‐1.4
(5) 0.2‐20‐0
0.3 (5) 0.2‐20‐0.6
(5) 0.2‐20‐1.4
0.2 (5) 0.35‐20‐0
(5) 0.35‐20‐0.6
(5) 0.35‐20‐1.4
0.1

0.0
0 5 10 15 20 25 30
l0/hw
Figure 14.4.5.2‐6: Comparison for C20/25 between calculation results (dots) and result from
formula (5) (line)

1.0

e t o t /h w -f c k - eff
0.9
0.05‐40‐0
0.05‐40‐0.6
0.8 0.05‐40‐1.4
0.2‐40‐0
0.7 0.2‐40‐0.6
0.2‐40‐1.4
0.6 0.35‐40‐0
0.35‐40‐0.6
0.35‐40‐1.4
0.5


(5) 0.05‐40‐0
(5) 0.05‐40‐0.6
0.4 (5) 0.05‐40‐1.4
(5) 0.2‐40‐0
0.3 (5) 0.2‐40‐0.6
(5) 0.2‐40‐1.4
0.2 (5) 0.35‐40‐0
(5) 0.35‐40‐0.6
(5) 0.35‐40‐1.4
0.1

0.0
0 5 10 15 20 25 30
l0/hw
Figure 14.4.5.2‐7: Comparison for C40/50 between calculation results and formula (5)

From the figures it can be concluded that the formula (5) results in a reasonable value for the factor  when it is
compared to the results of a more detailed calculation according to the general method to determine second
order effects. Formula (5) is presented in prEN 1992-1-1 as formula 14-11.

Influence of CE
In the previous, the partial factor for the modulus of elasticity of concrete is taken as 1,2. In table 4.3 of the final
version of FprEN 1992-1-1:2023, the partial factor is a NDP for which a value of 1,5 is advised. Hereafter the
influence of the difference between 1,2 and 1,5 for CE is presented. First this is done by comparing some results
of the general calculations for the different values of CE. For C20/25 and for C40/50 this is done in figure
14.4.5.2-8.

For citations, copying and referencing see Note on Front page! page 526
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

1.0 1.0

0.9 0.9

0.8 1.2‐ 0.05‐20‐0 0.8 1.2‐ 0.05‐40‐0

1.2‐ 0.05‐20‐1.4 1.2‐ 0.05‐40‐1.4

0.7 0.7
1.5‐0.05‐20‐0 1.5‐0.05‐40‐0

0.6 1.5‐0.05‐20‐1.4 0.6 1.5‐0.05‐40‐1.4

1.2‐ 0.2‐20‐0 1.2‐ 0.2‐40‐0


0.5


1.2‐ 0.2‐20‐1.4 0.5 1.2‐ 0.2‐40‐1.4

1.5‐0.2‐20‐0 1.5‐0.2‐40‐0
0.4 0.4
1.5‐0.2‐20‐1.4 1.5‐0.2‐40‐1.4

1.2‐ 0.35‐20‐0 1.2‐ 0.35‐40‐0

0.3 0.3
1.2‐ 0.35‐20‐1.4 1.2‐ 0.35‐40‐1.4

0.2 1.5‐0.35‐20‐0 0.2 1.5‐0.35‐40‐0

1.5‐0.35‐20‐1.4 1.5‐0.35‐40‐1.4
0.1 0.1

0.0 0.0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
l0/hw l0/hw

Figure 14.4.5.2‐8: Influence of CE equal to 1,2 or 1,5, for C20/25 (left) and C40/50 (right)

From figure 14.4.5.2-8 it shows, that the influence of the variation of CE is limited, and only significant in case of
higher slenderness’s. Based on this small difference a small adjustment of formula (5) is suggested in case CE
is equal to 1,5, see formula (6) hereafter:

 l e
1   2,1  0,02 0  tot
 h h
= 2 0,6 (6)
 l0   etot  1  eff   fcd,pl 
1     0,9  6    
h  h  1000   20 

In case CE is considered a variable, formula (6) can be expressed as:

 l e
1   2,1  0,02 0  tot
 h h
= 2 0,6 (7)
  
l e  0,67  CE  eff   fcd,pl 
1   0   0,9  6 tot    20 
h
   h  1000  

In the figures 14.4.5.2-9 and -10 hereafter the results of the formula (6) are compared with the results
from the analysis model. Each concrete class considered, is presented in a separate figure.

For citations, copying and referencing see Note on Front page! page 527
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

1.0

0.9
e t o t /h w -f c k - eff

0.8 0.05‐20‐0
0.05‐20‐1.4
0.7
0.2‐20‐0
0.2‐20‐1.4
0.6
0.35‐20‐0
0.35‐20‐1.4
0.5

(6) 0.05‐20‐0
(6) 0.05‐20‐1.4
0.4
(6) 0.2‐20‐0
(6) 0.2‐20‐1.4
0.3
(6) 0.35‐20‐0

0.2 (6) 0.35‐20‐1.4

0.1

0.0
0 5 10 15 20 25 30
l0/hw
Figure 14.4.5.2‐9: Comparison for C20/25 between calculation results (dots) and result from
formula (6) (line)

1.0

0.9
e t o t /h w -f c k - eff

0.8 0.05‐40‐0
0.05‐40‐1.4
0.7 0.2‐40‐0
0.2‐40‐1.4
0.6
0.35‐40‐0
0.35‐40‐1.4
0.5


(6) 0.05‐40‐0
(6) 0.05‐40‐1.4
0.4
(6) 0.2‐40‐0
(6) 0.2‐40‐1.4
0.3
(6) 0.35‐40‐0

0.2 (6) 0.35‐40‐1.4

0.1

0.0
0 5 10 15 20 25 30
l0/hw

Figure 14.4.5.2‐10: Comparison for C40/50 between calculation results and formula (6)
From the figures it can be concluded that the formula (6) results in a reasonable value for the factor  in case
CE is equal to 1,5.

[1] Wijte, S.N.M, 2014, CEN TC250/SC2/WG1/N171, Equation 12.11

References
[2] Barkeer, T., Jäger, W.; 2018, Improvement of the reduction factor for slenderness and eccentricity, part
of SC6/N654 – Background document on prEN 1996-1-1;
[3] Westerberg, B.; 2011, Load bearing capacity of unreinforced walls, Design of concrete structures and
bridges using Eurocodes, Bratislava.

For citations, copying and referencing see Note on Front page! page 528
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
14.4.5.2 Background to FprEN 1992-1-1:2023

List of versions of this document

03.06.2021 1st draft
03.01.2023 2nd draft
28.02.2023 peer reviewed
07.03.2023 final version

For citations, copying and referencing see Note on Front page! page 529
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG7 FprEN 1992-1-1:2023

Background document to Annex B, B.3-B.8

Time dependent behaviour of materials:

Creep, shrinkage and elastic strain of concrete

Harald S. Müller (Conv. TG 7) SMP Ingenieure, Karlsruhe

Fernando Acosta Urrea Ed. Züblin AG, Stuttgart

20.01.2023

For citations, copying and referencing see Note on Front page! page 530
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

Introduction to the background document (BD) to Annex B

The comparison of Annex B in FprEN 1992-1-1:2023 [1] with Annex B in EN 1992-1-1:2004/AC:2010-11 [2]
indicates that the new standard represents a fundamental revision and a large extension in terms of content,
structure and layout. Like Sections 5.1.1 to 5.1.5 of [1], Annex B has been developed by the Task Group 7 (TG 7)
"Time dependent effects" of CEN TC 250 SC 2 WG 1. Current members of TG 7 are F. Acosta (DE, secretary), T.
Kanstad (NO), H.S. Müller (DE, convenor), A. Perez (ES), E. Stack (IE), P. Takacs (GB), J.-M. Torrenti (FR), N.
Tosic (SRB), C. Vogt (SE), R. Wan-Wendner (BE).
Firstly, it has to be noted that Annex B in [1] is normative while Annex B in [2] is informative. While Annex B in [2]
is confined to information on the determination of the creep coefficient (magnitude and time development) and
equations for determining the drying shrinkage strain (basic shrinkage is dealt with in Sec. 3.1.4 in [2], Annex B in
[1] covers a much wider scope of concrete data.
Annex B in [1] indicates above all the development of concrete strength and stiffness with time, the basic formulae
for determining the creep coefficient, the basic formulae for determining the shrinkage strain, information on tests
on elastic deformations, creep and shrinkage as well as equations for the detailed analysis for creep at variable
loading. Further, a section on the relaxation of prestressing steel has been added. Similar information was placed
elsewhere in [2]. Introductory subsections of Annex B in [1] deal with the use of the Annex, the scope and field of
application and general information for design.
In this background document (BD) it is neither possible nor necessary to give detailed background information on
the models on creep and shrinkage. The reason is that both models are nearly identical to the corresponding
models given in fib Model Code 2010 (MC 2010) [3], which are scientifically well-founded in comprehensive
documents, see e.g. [4, 5]. Also, the commentary column (parallel to the guideline column) in MC 2010 gives
additional information.
However, for the new FprEN models on creep and shrinkage minor improvements have been introduced in the
underlying MC 2010 models. These improvements are documented in [6, 7], including various further references.
Nevertheless, the background document at hand picks out and summarizes some information, which is considered
to be valuable for the user of FprEN 1992-1-1:2023. Experts in the field of creep and shrinkage, who are up-to-
date with the scientific literature will scarcely find really new information. Designers who are interested in more
detailed scientific information on creep and shrinkage beyond the information given in [4, 5, 6, 7], may refer e.g. to
[8, 9, 10, 11].

B.3 General Related to Note (1) it should be mentioned that in contrast to the approach for creep in [2], the new creep approach
in [1] separates distinctly between the components basic creep and drying creep. This distinction is also given by
the creep model in [13].
The time development of drying shrinkage (Note (2)) depends on the migration of water molecules through
hardened concrete, which may be perfectly described by Fick’s 2nd law, i.e. drying shrinkage is a diffusion-type
process. This has to be taken into consideration for the modelling of this process.
It should be mentioned that both deformation components, i.e. creep and shrinkage, depend on the compressive
strength of concrete which serves as a substitute parameter for the effect of the composition of concrete (see
background document (BD) to Section 5.1.5. While basic shrinkage increases with the compressive strength, drying
shrinkage decreases. Related to creep, both components, i.e. basic creep and drying creep decrease with
increasing compressive strength.
Referring to Note (3), it has firstly to be stated that investigations carried out in the past show that the average
shrinkage is not significantly affected by short term varying ambient humidities [30]. However, considering seasonal
changes, pronounced differences are observed. In the winter season shrinkage is much lower than in the summer
season [31, 32], what can be considered in design. However, the average shrinkage over years corresponds to the
shrinkage which is obtained for mean constant ambient conditions. This means that reduced shrinkage in winter is
compensated by an increased shrinkage in summer; for more information see also B.6.
Note (4) does not at all mean that the presented models in FprEN 1992-1-1:2023 are not reliable for today’s
concrete. On the contrary, the reliability has been checked carefully even though there are much less experimental
data available from the years 2000 to 2020. However, the reliability is definitively reduced when the concrete
composition deviates from the limiting values defined in EN 206.
Table B.1 indicates the strength development classes of concrete. The related background is indicated in the BD
to Section 5.1.5.
The subsequent Table 1 corresponds very closely to Table B.1 and indicates some editorial improvements and
clarifications. Table 1 was developed by Task Group 7 (TG 7) "Time dependent effects" of CEN TC 250 SC 2 WG
1 during the process of discussing the draft of FprEN 1992-1-1:2023, and all changes compared to Table B.1 were
accepted by WG 1 and SC 2. However, due to an oversight Table B.1 was forgotten to be replaced by Table 1 in
the final document.

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 531
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

Table 1: Strength development classes of concrete

Composition and properties of the cement or the binder of the concrete

Concrete containing only cement Concrete containing binder composition
Class Cement
Type of Composition of the binder
strength class b,
cement a (Portland cement clinker c and SCM d)
e.g.
CEM III Portland cement clinker and more than 65 M.-% of ground granulated
CS 32,5 N; 42,5 N blast furnace slag (ggbs) or more than 35 M.-% of fly ash (fa)
CEM II/B
CEM II/A Portland cement clinker and more than 35 M.-% but less
CN CEM I 32,5 R; 42,5 N than 65 M.-% of ground granulated blast furnace slag (ggbs) or more
than 20 M.-% but less than 35 M.-% of fly ash (fa)
CEM II/A 42,5 R; 52,5 N; Portland cement clinker and less than 35 M.-% of ground granulated
CR blast furnace slag (ggbs)or less than 20 M.-% of SCM other than ggbs
CEM I 52,5 R
a CEM IV, CEM V and special cements may be classed considering Portland cement clinker content and strength class
b according to EN 197-1
c observe that the content of Portland cement clinker in CEM I may be 95 M.-%
d considered as binder for the water-binder-ratio according to EN 206

Clause (3) defines the range of validity of the creep and shrinkage models, if the structural concrete complies with
EN 206. For further information see BD to Section 5.1.1.

B.4 Development of The aging function for the mean compressive strength (Formulae (B.1) and (B.2)) have principally been maintained
from the previous standard [2]. However, the function for the aging coefficient βcc(t) was extended and readjusted,
concrete strength and compare Formula (B.2) in [1] and Expression (3.2) in [2]. The origin of this equation is reference [12].
stiffness with time While the coefficient s according to Expression (3.2) in [2] depends only on the type of cement used, the new
coefficient sC in Formula (B.2) takes into account the influence of the concrete strength and its development over
time, see Tables B.2 and B.1 in [1]. The values for sC have been derived from new experimental data (see BD to
Annex D). This new approach is much more precise and is also up-to-date, as the influencing variables take into
account today's concrete composition.
The relationship for the aging coefficient βcc(t), i.e. Formula (B.2) was extended by the introduction of the reference
age tref, which can deviate from the usual reference age of 28 days; see BD to Section 5.1.3. For tref = 28 days the
original Expression (3.2) in [2] is obtained.
Formula (B.3) in [1] has been updated compared to Expression (3.4) in [2]; for details see BD to Annex D.
Formula (B.4) in [1] is different to the Expression (3.5) in [2]. The reasons are detailed founded in the BD to Section
5.1.4.

B.5 Basic formulae for Concerning the background of the prediction of the creep coefficient, refer to the BD to Section 5.1.5, Clause (1) -
(3). Further, as the new creep model is very closely related to the corresponding model given in [3], detailed
determining the creep background information may be found in [4]. Moreover, additional information is given in [6, 7]. Nevertheless, few
coefficient aspects will be highlighted subsequently.
When comparing the determination of the creep coefficient in [1] with the corresponding determination in [2], it
becomes evident that by means of Formula (B.5) in [1], a distinct separation of the creep components into basic
creep and drying creep has been realized. From a physical point of few this separation is well founded by the fact
that different microstructural mechanisms prevail for the development of the individual strain components. This
effect is well- known since roughly half a century as the so-called Pickett effect, and is well described in textbooks,
see e.g. [8, 11].
From a practical point of view, i.e. from the intention to derive the best possible approach for estimating the creep
coefficient, the separation into the strain components is associated with two fundamental advantages. Firstly, it is
known that the time development of basic creep and drying creep is different, i.e. different types of mathematical
functions have to be used to describe closely the observed time development. Secondly, if the compressive
strength of concrete is used as a substitute parameter to describe the effect of the concrete composition on creep,
what is the case in all creep models given in standards, the fact has to be faced that basic creep and drying creep
are differently affected by the magnitude of the compressive strength. Test results show that the ratio of basic creep
and drying creep increases significantly with increasing strength. Hence, if there is no separation into the both
creep components, as this is the case for the previous standard [2], neither a correct consideration of the effect of
the compressive strength nor of the effect of the ambient conditions on the time development of creep is possible.
Related to the consideration of the parameter compressive strength of concrete, both for creep and shrinkage
modelling, refer to the BD to Section 5.1.5, Clause (1) - (3). The major aspects related to the time development
functions for creep, i.e. the functions given by Formulas (B.8) and (B.13) are subsequently summarized.

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 532
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

Time functions for creep models

The creep of concrete is characterised by two decisive time dependencies, namely the influence of the age at
loading t0 and the influence of the duration of loading (t-t0). Both influencing variables have been known for almost
as long as creep itself. It is therefore not surprising that numerous functions, some of them very different, have
been proposed in the past to describe these influences. The main reason for this is the test results considered in
each case, i.e. for different types of concrete and different boundary conditions with respect to the concrete age
and the ambient climate different tendencies for the time development of creep may be observed. On the contrary,
it has been known for far less time that the mathematical formulation of the two associated time functions and their
coupling, for example in the form of a product as often realised, i.e. εcc ~ f(t0) ꞏ g(t-t0), is associated with some
implications, and has to be carefully considered in order to avoid thermodynamically nonsensical predictions.
Relevant aspects of the new approach in [1], some aspects are summarized in the following.

Time development of creep

When looking to the very comprehensive literature on creep, it becomes evident that it is not possible to find a
single function that is able to describe the development of concrete creep under the various boundary conditions
relevant to practice. From countless experiments, it can be summarised that a power function describes the short-
term creep best, but otherwise clearly overestimates the deformation after a longer loading period (t-t0 > 1 year).
The often chosen hyperbolic functions – numerous different mathematical types are known, see e.g. [2, 13, 14] –
can accurately describe creep over a very long period of time, regardless of whether basic creep or drying creep
occurs. However, a regression on test data shows – this is general scientific consensus – that small deviations
from test data occur in the initial phase, as well as after a very long period of loading.
A purposeful clarification of the open question regarding the "correct" time function is facilitated if a systematic
distinction is made between basic creep and drying creep. This distinction is indicated for physical reasons anyway,
as already mentioned above. For explanation, the test results of Troxell et al [15] on a concrete C12/15 are
subsequently used as an example (see Figure 1). The creep deformations were recorded over a period of approx.
23 years. The trends that can be derived from the curves were later confirmed by other authors for stronger and/or
usual structural concretes submitted to very high load durations [16, 17].

Figure 1: Long-term creep of concrete (specific creep is the ratio of the creep strain and the creep inducing
stress); basic creep (RH = 99 % and 100 %) and drying creep (RH = 50 % and 70 %) data from
Troxell and regression analysis applying different time functions

In Figure 1, the logarithmic function of the type ln [1 + const∙(t-t0)] – this is in principle identical to Formula (B.8) –
is plotted next to the test data, which also visually represents a good regression to the basic creep (measured
values for RH = 99 % and RH = 100 %). In the usual semi-logarithmic representation of creep curves, the time
development changes to a linear development after a slight leftward bend and they have no final value.
If one describes the time development of the basic creep with a logarithmic function and the time development of
the overlapping drying creep with a hyperbolic function according to Formula (B.13), the regression result is
obtained as shown in Figure 1 for the test values with RH = 50 % and RH = 70 %. Although a final value for the
time development is no longer given, a good adaptation to the measured values is achieved even after a very long
period of loading.
Since many of the formulae used today for the design of concrete components require a final creep coefficient, a
calculated final creep coefficient must be defined when using the logarithmic time function for the basic creep. For
this purpose, the service life of the structure can be used as the load duration to be included in the calculation. The

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 533
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

service life, which is usually specified in guidelines, varies between 50 and 200 years depending on the guideline
and/or the type of structure. In FprEN 1992-1-1:2023 [1] it is given as 50 years. For a structural member made of
a usual structural concrete C30/37, which has an effective member size of hn = 200 mm and is loaded at the age
of 28 days in a normal environment (RH = 65 %), the increase in deformation between 50 and 200 years of load
duration amounts to approx. 5-10 % of the creep deformation occurring after 50 years when extrapolating the
superimposed logarithmic and hyperbolic time development functions. For further details see [4].

Aging function of creep

The aging function takes into account the influence of the age at loading t0 on the creep of concrete. In the so-
called product or factorial approaches, i.e. for a presentation of creep as εcc ~ f(t0) ∙ g(t-t0), a hyperbolic function
was usually chosen for f(t0), which is multiplied by the function for the time development g(t-t0), see e.g. the
approach for drying creep in [1] (Formulae (B.12) and (B.13)). This results in the typical affine representation of the
creep, which generally agrees reasonably well with test results, when no higher demands on the prediction quality
is desired. In this kind of product formulation, the age at loading only influences the level of creep or a creep
component, respectively, but not its development over time.
A more detailed analysis of the influence of the age at loading on the creep deformation is provided by the
exemplary test results shown in Figure 2 [18]. They show that the time development of creep is delayed with
increasing age at loading. This effect is represented in the new creep model in [1] analogous to [3], by taking into
account the influence of the age at loading on the drying creep component in the exponent of the time function by
γ(t0), see Formula (B.13). In contrast, the exponent in the EN 1992-1-1:2004 creep model [2] is constant, i.e.
γ(t0) = 0,3 see Expression (B.7). Another advantage associated with the γ(t0) function introduced in [1] is shown in
the subsequent section on the coupling of time functions.

Figure 2: Time development of creep of concrete loaded at different ages; test data from [18] and
predictions from MC 2010 (curves are identical to predictions of FprEN 1992-1-1:2023) and EN
1992-1-1:2004

A likewise non-affine consideration of the age at loading is given in the representation of the basic creep according
to Formula (B.8) in [1], being first introduced in the MC 2010 [3]. There, the influence of the age at loading t0 is
integrated into the logarithmic function. This representation has the effect that the creep deformations decrease
with increasing loading age, but at the same time the creep curves also run parallel in the usual graphical
representation (strain over the logarithm of time) after longer load duration. This is in good agreement with
experimental investigations and can be justified thermodynamically [19].

Coupling of the aging function and the time development function

The typical coupling of aging function and the time development function is given by a product and form the so-
called product-type approach for the creep coefficient, i.e. φ(t,t0) = f(t0) ... ki ... g(t-t0). In this form, the factor f(t0)
describes the influence of the age at loading (aging function) and the function g(t-t0) is the time development
function of creep. The factors ki are functions of other influencing variables, whereby these factors are partially
correlated, as the various influencing parameters are interrelated. This simple form of representation, also known
as "affine creep", has been used since many years [14, 20] and was also chosen for the creep approach in MC
1990 [12], and it corresponds with the approach in EN 1992-1-1:2004 and FprEN 1992-1-1:2023. For sake of
completeness the summation-type approach, also known as "rate-type creep" should be mentioned here. Details
may be found in [4, 5].

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 534
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

Figure 3: Prediction of creep recovery (left) and stress relaxation (right) when applying the principle of
superposition in combination with the creep models given by EN 1992-1-1:2004 and FprEN 1992-
1-1:2023

The product-type approach has the weakness that in combination with the application of the superposition principle
according to Formulae (B.41) or (B.42) under the boundary conditions of full unloading and pure relaxation, it leads
to mechanically nonsensical deformation or stress predictions [5, 21], respectively, see also Figure 3.
Irrespective of the fact that the use of the superposition principle must inevitably lead to erroneous predictions
because concrete is not a linear viscoelastic material as considered by this principle, the strain reversal after
unloading (Figure 3, left) and the change of sign for stress from compression to tension during relaxation (Figure
3, right), both occurring for the creep model in EN 1992-1-1:2004, must be eliminated as far as possible. This can
successfully be realized if the relevant time functions for the age at loading and the time development are matched
in a suitable way, as has been done for MC 2010 [3, 22] and FprEN 1992-1-1:2023, respectively.
Without going into further constitutive details, it may be stated here that a satisfying solution is obtained by the age
dependency of the exponent in Formula (B.13) and the particular combination of aging and time development in
the increment of the logarithmic function of Formula (B.8) in FprEN 1992-1-1:2023. Note, that this is quite different
to the approach in EN 1992-1-1:2004. In this way, the weaknesses of the product-type approach can be overcome
to a large extent. An easier mathematical solution does not exist.

Creep of concrete in existing buildings

Since all normative models for creep take into account the influence of the age at loading, one could assume that
the relationships given can also be applied when the creep behaviour of old concretes in existing buildings has to
be estimated, e.g. when a new loading occurs after rehabilitation. However, this is not the case. Basically, all creep
models were derived for the prediction of concrete creep of previously unloaded concrete. The data on which the
models are based only include creep tests of virgin concrete and, in addition, only a very few investigations had
been carried out applying a high age at loading (t0 >> 5 years).
A long-lasting preloading, as is the case with existing buildings, has a strengthening effect on the concrete if the
acting stresses are below the limit of the service stresses (σc < 0.4 ∙ fcm(t0)). Softening of the concrete (effects of
sustained loading) only occurs at significantly higher stresses. In both cases, an influence on the creep behaviour
can be assumed from material science considerations. However, to the author’s knowledge, the extent to which
long acting preloads actually influence creep has not yet been systematically investigated.
Nevertheless, it cannot be ruled out that the current creep models provide useful estimates of the creep behaviour
of existing concretes in an old structure for an initial rough calculation, but this is by no means certain. Therefore,
for deformation-sensitive constructions, it must be recommended to carry out creep tests on samples taken from
the existing building. How such test results can be integrated into a prediction model like FprEN 1992-1-1:2023 is
shown in Annex B.7 of this standard [1].

Non-linear creep
In the design of concrete structures, concrete is considered to be a linear viscoelastic material, but in fact it has a
complex non-linearity that depends, among other things, on the magnitude of the applied stress for constant

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 535
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

loading, the ambient conditions, and on the loading history for variable stresses. For details refer to [4, 5, 9, 21,
38].
Formula (B.19) in [1] has been developed originally already for MC 1990 [12] in order to allow for estimating creep
when high initial stresses are applied as this is the case for early high prestressing, even though the overall creep
model is in principal a stress-linear model.
The level of stress at the transition from linear to non-linear creep was fixed to kσ = 0,4 = constant. As a matter of
fact, this level depends actually on the strength of the concrete (porosity, age at loading), which can be neglected
for sake of simplicity. The increasing non-linearity can be described appropriately with the help of the exponential
function (Formula (B.19) in [1]). It is a simple approximation for the non-linear creep behaviour of concrete, but
neglects the influence of the age at loading as well as the influence of the duration of loading. In fact, the non-linear
part of the creep deformation decreases pronouncedly with increasing load duration. Experimental evidence for
Formula (B.19) is shown and discussed e.g. in [4].
In Formula (B.19) the stress level (stress-strength ratio) is defined as kσ = σc / fcm(t0), i.e. the creep generating stress
σc is related to the mean value of the concrete strength fcm(t0) at the age t0. This type of definition is also applied in
[3, 12]. Correspondingly, the limit value in Formula (B.19) for the onset of non-linearity is given as kσ,lim = 0,40. In
contrast, in EN 1992-1-1:2004 [2], the definition kσ* = σc / fck(t0) is used and kσ,lim = 0,45 is set. With this definition
and the given limit value, a somewhat lower stress σc for the transition from linear creep to non-linear creep is
calculated for structural concretes < C60/75. Although this is of no severe practical relevance, conceptually the
approach for kσ* in conjunction with Expression (3.7) in [2] is not correct. The linearity limit and the related
exponential equation were derived directly from experimental investigations, whereby the concretes were
characterized by the experimentally determined value for fcm. In contrast, fck is a characteristic value derived under
safety considerations, starting from the known strength distribution. Therefore, it is consistent to apply Formula
(B.19) using the experimentally determined concrete compressive strength fcm. If this value is not known, it can be
estimated from the concrete strength class or fck using fcm = fck + 8 N/mm².
The non-linearity is much more complicated if the creep-inducing stress is variable. Deviations from the predicted
strain by applying the principle of superposition may occur even if the stress-level is low (kσ < 0,40). Details may
be found in [5], and partially also in [6, 7]. The observed effects can only be described by a complex non-linear
model and require a special superposition principle for the mathematical treatment of creep problems with variable
loads [21]. Further information on the non-linearity presented here can also be found in [23].

Lower and upper characteristic values of the creep coefficient

Section B.5, Clause (8) in FprEN 1992-1-1:2023 provides information on lower and upper limit values of the creep
coefficient, taking into account the scatter of the creep or the coefficient of variation in the prediction of the creep
coefficient, respectively. The EN 1992-1-1:2004 does not contain comparable information.
The scatter, i.e. the coefficient of variation for the creep coefficient was found to be approx. V ≈ 30 %. When
assuming a Gaussian distribution, the fractile value for the creep coefficient may be calculated from
φfrac = φ (1 ± ηfrac∙V), where φ is considered as the mean value of the creep coefficient and ηfrac depends on the
fractile (ηfrac = 1,645 for 5 % fractile, ηfrac = 1,282 for 10 % fractile). As an example, φ0,95 is calculated from
φfrac = φ∙(1 ± ηfrac∙V) when applying φ0,95 = φ (1 + 1,645∙0,30) = 1,494 φ ≈ 1,5∙φ.

B.6 Basic formulae for It should be noted that by definition (sign convention) shrinkage is calculated as a positive strain (elongation) when
applying the Formulae (B.23) to (B.33). In fact, however, shrinkage is a contraction, while swelling is an elongation.
determining the
Important background information on the new shrinkage model in FprEN 1992-1-1:2023, Annex B.6 [1] are already
shrinkage strain given in the BD to Section 5.1.5, Clause (4). It is explained and founded there why the new model gives
approximately 30 to 40 % higher values for shrinkage than given by EN 1992-1-1:2004 [2]. Subsequently, some
additional background information is provided.
As mentioned above, the new shrinkage model in Annex B of [1] is closely related to the corresponding model in
MC 2010 [3]. However, slight improvements have been introduced by the introduction of the Formulae (B.29) to
(B.31) and Formula (B.32) and by replacing the effect of the class of cement (N, S, R) and the strength class of
cement (32,5; 42,5; 52,5), see [3] and [2] as well, by the newly introduced strength development classes of concrete
(see Table B.1 in [1]). Background information is given in the BD to Section 5.1.5, Clause (1) - (3).
The basic approach for determining the shrinkage strain, i.e. the separation of the total strain into the components
basic shrinkage and drying shrinkage has been kept (compare Formula (B.23) in [1] with Expression (3.8) in [2]).
The introduction of the coefficients αNDP,b and αNDP,d in Formulae (B.24) and (B.25) is new. The default values for
both coefficients are 1,0, however, a National Annex may give different values.
From a scientific point of view such a deviation of the coefficients αNDP,b and αNDP,d from 1,0 is only justified, if the
composition of the concrete deviates from the average composition used in Europe. For example, if in a particular
country concretes are primarily produced using high strength limestone aggregate and/or with a low binder content
and/or with binders for which low shrinkage values have been experimentally proven, it makes sense to apply
αNDP,b < 1,0 and αNDP,d < 1,0. It is recommended to run test in order to derive reliable values for αNDP,b and αNDP,d.

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 536
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

Otherwise, the αNDP values should be kept as 1,0, even if for example a high ambient average or a low ambient
average humidity prevails in the particular country. The reason is that the effect of the ambient average humidity is
considered correctly by the formulae (B.29) to (B.32).
Formula (B.26) in [1] which gives the magnitude of the notional basic shrinkage differs from the Expression (3.12)
in [2] in relation to the mathematical form, the application of fcm,28 instead of fck, and by considering the early strength
development of concrete. This results in differences for the prediction of basic shrinkage up to approx. 0,04 ‰
within the validity range of the model, which is not that much. Nevertheless, the new approach is physically correct,
whereas the previous approach is not.
Further, Formula (B.26) is identical to the corresponding Formula in MC 2010 [3], and was already introduced in
MC 1990-99 [24]. The choice of a hyperbolic function as given in the MC 2010 and now in FprEN 1992-1-1:2023
has the advantage that no change of sign occurs in the calculation of the shrinkage values for low strength values
and that a final value for the notional basic shrinkage is reached for high strength values. Both effects make physical
sense but are not given by the simple linear relationship of Expression (3.12) in EN 1992-1-1:2004. Hence, TG 7
(see Introduction) decided to switch for FprEN 1992-1-1:2023 to the corresponding equation of the MC 2020.
Related to the use of fcm in [1] instead of fck in [2], refer to the BD to Section 5.1.3, Clause (1) - (6). The use of the
compressive strength as a parameter, which affects shrinkage (and creep) is elaborately explained in the BD to
Section 5.1.5.
The Formulae (B.29) to (B.31) and Formula (B.32) describe the effect of the ambient humidity RH on the magnitude
of the shrinkage. This parameter is related to the reference value RHeq, which indicates the average relative
humidity inside a member in a state of moisture equilibrium. For an average concrete strength of fcm = 35 N/mm²,
RHeq = 99 %, see Formula (B.32). With increasing concrete strength, the value for RHeq decreases continuously
and as an example amounts to RHeq = 91 % for a concrete strength of fcm = 80 N/mm². The physical consequence
is that in the assumed case of an ambient humidity of RH > RHeq = 91 %, this concrete no longer shrinks but swells,
i.e. its volume increases significantly, because it accumulates water molecules from the environment inside its pore
structure.
The described effect was not correctly considered by Expression (B.11) in [2]. The Formulae (B.29) to (B.32) in [1]
correctly defines three ranges without a discontinuity (jump) at the transition between the ranges. The range
20 % ≤ RH ≤ RHeq covers the drying shrinkage and the range RHeq < RH < 100 % the swelling of the concrete. For
RH = 100 %, swelling also occurs, but this results from water absorption when the member surface is exposed to
water in liquid form. In this case, the capillary transport mechanism of the pore system becomes active, which leads
to increased water absorption and, as a consequence, to an increased volume increase.
The Formulae (B.29) to (B.32) describe only those swelling processes that can occur in a concrete if it has not
previously shrunk as a result of a water release. If, for example, a concrete member has reached a shrinkage of
0.5 ‰ after a long drying phase and is subsequently exposed to a different ambient climate resulting in a continuous
uptake of water molecules (e.g. RH = 99…100 %), the occurrence of the resulting swelling deformation cannot be
determined by means of Formula (B.31). In the relevant literature, no relationships are known with which the
swelling occurring under the conditions mentioned could be calculated. As a guideline for practical applications, it
can be assumed that after long periods of high moisture supply, about 50-60 % of the previously occurring
shrinkage deformation is reversible. Thus, swelling after previous shrinkage of -0.5 ‰ is calculated to +0.25 ‰ to
+0.30 ‰, possibly plus the deformation resulting from the summand "+0.25" according to equation (B.31) in
connection with Formula (B.25), if water can be absorbed in liquid form.
When comparing the functions for the time development for drying shrinkage, it is evident that the Expression (3.10)
in [2] is quite different to the Formula (B.33) in [1]. This is visualized in Figure 4, which gives the time development
of drying shrinkage as predicted from EN 1992-1-1:2004 [2], EN 1992-2:2005 [13] and MC 2010 (identical to FprEN
1992-1-1:2023 [1]).
Concerning the time function of FprEN 1992-1-1:2023, Formula (B.33) in [1], it has to be stated that this kind of
hyperbolic function complies with the diffusion theory and may be directly derived from this process which governs
drying shrinkage.
Firstly, this means that the member size hn affects just the time development exactly according to the square root
((t-ts)/(c∙hn2 + (t-ts))0,5, where the parameter c corresponds to the (average) diffusion coefficient. Hence, the member
size hn does not affect the magnitude of drying shrinkage but the rate of its time development, see e.g. [4].
Moreover, this time function meets very well the average time development observed in hundreds of experiments.
The time function according to EN 1992-1-1:2004 is a purely empirical function that does not meet the typical
development of drying shrinkage of structural concretes. The shrinkage model in EN 1992-2:2005 takes into
account correctly the hyperbolic size effect (t-ts)/(c∙hn2 + (t-ts)) but does not consider the square root for the time
development (see above), which is characteristic for diffusion processes.

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 537
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

Figure 4: Comparison of the time development of drying shrinkage according to EN 1992-1-1:2004, EN

1992-2:2005 und MC 2010 [3] (FprEN 1992-1-1:2023 is identical to MC 2010) for a notional
member size of hn = 200 mm

The time development of the FprEN 1992-1-1:2023 curve lies roughly in the middle of the two curves being given
by EN 1992-2:2005 [13], apart from the initial phase, when presented in the selected semi-logarithmic scale (see
Figure 4). The silica fume (SF) added to the concrete mix (criterion only "contained" or "not contained") causes a
considerable delay in the drying shrinkage curve, regardless of the quantity, which is at least surprising.
In contrast, the drying shrinkage according to EN 1992-1-1:2004 [2] develops much faster. Half of the final value is
reached here after approx. 150 days, whereas with FprEN 1992-1-1:2023 it is only reached after approx. 700 days.
The possible reason why the time development curve of the EN 1992-1-1:2004 deviates so much from the curve
given by MC 2010 (FprEN 1992-1-1:2023) is subsequently explained.
As already mentioned above and in the BD to Section 5.1.5, the factors 0.85 and kh were introduced in the shrinkage
model in EN 1992-1-1:2004 [2] in order to lower the final values of the calculated drying shrinkage to a level given
by former guidelines (e.g. MC 1978 [25]). Obviously, it was then realised that the initial shrinkage is as well lowered
so that for small member sizes it turned out to be much smaller than predicted by MC 1978. To compensate this
weakness, it was necessary to change the time function (more rapid development of shrinkage, see Figure 4) in
order to increase the calculated values after a short drying period. Hence, the originally considered time
development function – see function in prEN 1992-1-1:2001 [26], which is identical to Formula (B.33) – had to be
replaced. In retrospect, it can be seen that by introducing two “errors” (factors 0.85 and kh) a further error
(accelerated time development) had to be introduced in order to achieve a sufficient adaptation to the shrinkage
model of MC 1978, which was considered to be correct. This procedure ignored the reasons for the change from
the MC 1978 shrinkage model to the MC 1990 shrinkage model (diffusion-type model).

Size effect upon shrinkage

As mentioned above, the member size hn does only affect the time development of shrinkage but not its magnitude,
see [1, 4]. The argument, which is still repeated today, that the member size could have a certain influence on the
final shrinkage value of structural concretes, cannot be dismissed in principle, see e.g. [39]. The decisive aspect,
however, is whether the conceivable influences occur in an appreciable order of magnitude. In this context, mostly
the influence of carbonation, which is less effective for thick members, and the influence of a long internal curing
of the core area, which occurs for thick members, were cited. However, both arguments do not stand up to scientific
scrutiny.
The contraction resulting from the carbonation shrinkage (up to max. 0;2 ‰ at full carbonation) is limited after, for
example, 20 years only to a surface zone of 10 mm to a maximum of 20 mm depth. In this zone, tensile stresses
will arise as a result of the deformation restraint by the core cross-section, the upper limit of which is limited by the
comparatively low tensile strength of the concrete. The compressive stress caused by this low tensile stress over
a comparatively small area as a result of deformation constraint on the much larger residual cross-section, which
theoretically produces creep (shortening) there, which is perceived as shrinkage, can be neglected.
Good internal curing of the core area of a large cross-section by the concrete layer of the surface zone will tend to
result in reduced porosity in the core area. This will slow down diffusion and thus shrinkage as well, but this does
not necessarily lead to lower deformations after long periods of time, as is assumed. In particular finely porous,
hydraulically bound microstructures with a low porosity may show markedly large hygric deformations. This
observation has also led to the fact that in all known shrinkage models, the use of finely ground high-strength
cements, which quickly form a very fine and dense microstructure in the course of hydration, results in increased
shrinkage values.

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 538
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

Magnitude of final shrinkage

The shrinkage model of the MC 1990 was derived on the basis of an extensive, updated database and taking into
account the diffusion theory, which had proved to be applicable to concrete shrinkage a few years earlier. The CEB
Technical Committee at that time was of course aware that the shrinkage values of the MC 1990 model were
increased by 30 to 40 % compared to the preceding model in the MC 1978, which is also documented in [5]. Various
reasons could be identified for this.
A major reason for the lower shrinkage values in the MC 1978 is that their determination was still largely based on
the compilation of test data by Wagner from 1958 [27]. These "old" concretes were generally produced with a much
coarser Portland cement than was the case in later years and even more so today. Furthermore, these concretes
were also produced with lower cement/binder contents. Both parameters - lower grinding fineness and lower binder
content - cause a significant reduction in shrinkage. Other reasons are probably to be found in certain shifts in the
proportions of clinker phases that have occurred in cement production over the past decades.
In addition, there is another important aspect that is related to the evaluation of shrinkage tests. At that time, it was
common and considered scientifically valid to extrapolate the deformation behaviour or its final value with the help
of Ross' hyperbolic function, which was possible in simple graphical form (pocket calculators did not yet exist). But
especially when investigating test specimens with larger member sizes, where even after 2 to 3 years of observation
in the laboratory no approximation to a final value is apparent, extrapolation with the help of the Ross hyperbola
leads to a clear underestimation of the final values [28]. It can be easily verified mathematically that the final values
extrapolated according to Ross decrease with increasing member size for the same drying time in the test. This
led to the identification of an influence of the member size on the final value of shrinkage, which, however, does
not actually exist, but results solely as an artefact from the evaluation of shrinkage tests. It was originally Wagner
who came up with right the values given in Table 3.3 in EN 1992-1-1:2004, which were first used in the MC 1970
shrinkage model [20]. With the state of knowledge in 1980s, the shrinkage model in MC 1990 was introduced and
later adopted in an extended form in MC 1990-99 [24] in prEN 1992-1-1:2001 [26].
From the CEN records of the early 2000s it can be seen that the increase in the calculated minimum reinforcement
associated with the increase of the shrinkage values was the decisive motivation to again critically examine the
specifications of prEN 1992-1-1:2001, which then finally led to the introduction of the factors 0.85 and kh in EN
1992-1-1:2004.

Shrinkage of laboratory samples in relation to member shrinkage in practice

A recurring question is to what extent shrinkage values determined in the laboratory can be transferred to member
shrinkage in practice. It is true that, for example, the MC 1990 or MC 2010 models, as well as earlier models such
as MC 1978 and MC 1970 and all other known models in foreign guidelines, were derived directly from the results
of tests on unreinforced test specimens. However, very early investigations by Rüsch had already shown that this
is permissible [29]. There is no doubt about this in the scientific community.
Nevertheless, arguments in favour of lower shrinkage in real structures are often put forward by practitioners, e.g.
with respect to the lower applied ambient humidity in the laboratory tests and the smaller dimensions of the test
specimens compared to real building members outdoors, as well as the constantly changing climate (humidity and
temperature) in practice compared to the constant climate in laboratory tests.
First of all, it should be noted that the influences of the parameters ambient humidity and member size on the
shrinkage of concrete have been extensively investigated in laboratory tests for decades. The underlying physical
processes are considered to be fully understood. Therefore, the conversion of the effects of lower ambient
humidities to higher ambient humidities and the estimation of the influence of larger member dimensions are
possible using very well validated functions. The functions given in MC 2010 describe these dependencies
accurately. The effects of variable climate, i.e. cyclical changes in the daily or weekly cycle, have also been studied
in individual publications [30]. However, no significant effects have been proven.
Another argument repeatedly put forward for the occurrence of smaller shrinkage values in concrete members in
practice compared to those determined on laboratory specimens is the reinforcement present in real structural
members. In fact, the reinforcement has the effect that the shrinkage reduction is somewhat smaller than in
unreinforced members, since the reinforcing steel does not shrink and therefore hinders the shrinkage tendency of
the concrete.
In principle, shrinkage values of unreinforced concrete are (correctly) always given in guidelines. The effect of
reinforcement can, if necessary, be easily estimated by a calculation using the theory of composites, provided it is
not implicitly taken into account in the design approaches anyway.
Although the transferability of laboratory tests on creep and shrinkage to practical conditions is theoretically
assured, corresponding experimental evidence is necessary and of high value. For various reasons, this applies in
particular to the seasonal influences and the influence of very large member sizes as well as the interaction of both
influencing parameters.
In a comprehensive study by Kollegger [31, 32], in addition to regional influences (aggregates used, climatic
conditions) and seasonal influences (summer/winter concreting), the influence of the test specimen size on creep

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 539
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

and shrinkage was also investigated (three different square cross-sections, edge length 1000 mm, 500 mm and
250 mm). Parallel to the concreting of large test specimens on the construction site, laboratory specimens were
also produced on site from the same concrete mix. Unlike the large test specimens, they did not remain on the
construction site, but were tested in the laboratory under the typical test conditions (constant 20 °C, 65 % relative
humidity).
The preliminary results obtained show that the MC 2010 model is able to predict the shrinkage deformations
measured on the test specimens in the laboratory as well as on the construction site with mostly good and partly
very good approximation. With regard to the influence of the member size, the tests confirm the validity of Fick's
2nd law for concrete shrinkage.
All in all, the very valuable new tests of Kollegger confirm the fact that the influences of relative humidity and
member size, which are physically correctly taken into account in the diffusion-type shrinkage models, ensure that
laboratory results can on average be accurately transferred to practical conditions. On the other hand, the tests
also show that the seasonal climatic fluctuations exert a pronounced influence with increasing slenderness of the
members that was not previously known or proven in this way and is therefore not represented by shrinkage models
in standards. This applies equally to the influence of the season of the member production on the shrinkage
behaviour of concrete, which was recognised in [31, 33], and which results from a coupled ambient temperature
and moisture effect.

Lower and upper characteristic values of total shrinkage

Section B.6, Clause (8) in FprEN 1992-1-1:2023 provides information on lower and upper limit values of the total
shrinkage, taking into account the scatter of the shrinkage or the coefficient of variation in the prediction of the total
shrinkage, respectively. The EN 1992-1-1:2004 does not contain comparable information. This coefficient of
variation was found to be approx. V ≈ 30 % (same order of magnitude like for creep). The calculation of the lower
and upper characteristic values is indicated in the corresponding section on creep (see above).

B.7 Tests on elastic The FprEN 1992-1-1:2023 [1] specifies in Section B.7 criteria for the necessity of experimental investigations on
concrete creep and shrinkage as a supplement to the sole application of the prediction models for creep and
deformations, creep and shrinkage. With the Formulae (B.34) to (B.39) relationships are given for an improved model that takes into account
shrinkage the test results by means of additional parameters which are obtained from a regression analysis.
Clause (3) gives the relevant test guidelines. Unfortunately, EN 12390-16 denotes the basic shrinkage as
autogenous shrinkage, as this shrinkage component was termed in former times. Today’s understanding is that
the basic shrinkage results from the superposition of the chemical shrinkage and the autogenous shrinkage. The
chemical shrinkage is the volume decrease caused by the chemical reaction of H2O with the cement clinker phases.
The autogenous shrinkage results from the internal drying as a consequence of the formation of the gel
microstructure of the binder matrix. In both cases there is no moisture exchange with the environment.
The consideration of test results for the considered structural concrete allows the coefficient of variation V for the
prediction of creep and shrinkage to be reduced from approx. V ≈ 30 % to V ≈ 10 %. This increases the factors
given in B.7, Clause (2) from 0,6 and 0,5 to 0,87 and 0,84, respectively, and decreases them from 1,4 and 1,5 to
1,13 and 1,16. Thus, as an example, if the creep coefficient is calculated as φ(50y, 7d) = 3, the upper value for
design is φ0,95(50y, 7d) = 3 ∙ 1,5 = 4,5 if no experiments have been carried out, and drops to
φ0,95(50y, 7d) = 3 ∙ 1,16 = 3,5 (23 % less), if experimental results are considered.
The regression parameters ξbc1, ξbc2, ξds1 and ξds2 are to be determined according to the method of the least squares.
For mathematical reasons, the regression parameters ξbc1 and ξbc2 correlate and cannot be determined
independently of each other. However, this does not imply any limitations.
For such creep and shrinkage tests on usual test specimens, a minimum loading period or drying period of 3 months
must be observed in order to enable a reliable regression of the fitting functions. If drying creep or drying shrinkage
is considered, the strain development curves will show an inflection point in the semi-logarithmic scale (y-axis
strain, x-axis log of time, see of Figures 1 and 4). Waiting for this inflection point in the tests is absolutely crucial
for a reliable regression or extrapolation of the shrinkage and creep process.
Since the drying shrinkage and partially also drying creep is a diffusion-controlled process, the specimen size and
the concrete grade play a decisive role in view of the onset of the inflection point. For usual structural concretes
and specimen diameters of 150 mm, this inflection point occurs roughly after 50 ± 10 days, so that a reliable
extrapolation is possible after a test period of between 90 and 120 days. For slimmer specimens, the inflection
point is reached earlier. This can be estimated using Formulae (B.13) and (B.33).
For creep tests in a moist environment in order to determine the basic creep there will be no inflection point.
However, the minimum loading period of approx. 3 months is sufficient to be able to carry out a reliable regression.
For shrinkage tests in a moist environment in order to determine the basic shrinkage there will occur an inflection
point as well in the semi-logarithmic scale (for mathematical reasons) which is achieved roughly after 30 days. After
roughly 20 days, approx. 60 % of the final strain is reached. Formula (B.27) in [1] may be used for a more precise
estimation.

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 540
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

B.8 Detailed analysis for While in Formula (B.40) the creep function J(t,t0) is described by means of the conventional parameters Ec(t0), Ec,28
and φ(t,t0), i.e. with the elastic moduli and the creep coefficient, Bažant [34, 35] gives a different definition of J(t,t0).
creep at variable loading However, both approaches for J(t,t0) are equally correct if the definitions of the respective deformation components
are consistent with each other.
Formula (B.40) applies exclusively to the effect of a constant stress σc(t0). If, on the other hand, the stress is time-
variable, Formula (B.40) shall not simply use σc(t) instead of σc(t0), as this is sometimes found in the literature, but
for constitutive reasons the superposition principle, also called Boltzmann's superposition principle, shall be used.
According to this, the strain caused by the stress curve σc(t) is obtained by decomposing the stress curve into small
stress increments Δσc(τi) applied at time τi and summing up the strains caused by this, which leads to Formula
(B.41).
It is important to mention that Formulae (B.41) and (B.42) are valid only for a linear viscoelastic material. This holds
not true for concrete, which feature a complex non-linearity. Nevertheless, both Formulae may be applied for
concrete, as it is usually done, and they give in most case reasonably good prediction results. For further details,
see [5, 9, 21].
As a general constitutive relation for concrete, Formula (B.42) allows the treatment of arbitrary creep and relaxation
problems if, given a known material law (creep function J(t,t0)), either the stress history σ(t,t0) or the strain history
ε(t,t0) is known. In the latter case, Formula (B.42) mathematically represents a Volterra's integral equation, which
can only be solved numerically for complex creep functions as used in design today.
For the sake of completeness, it should be mentioned that the viscoelastic behaviour of concrete can also be
described by means of a differential equation that is equivalent to Formula (B.42). However, Formula (B.42) can
only be directly converted into a differential equation if the creep function J(t,t0) fulfils certain mathematical criteria.
Further details can be found in [9, 21].
Since the application of Formula (B.42) is mathematically complex and also requires a high storage capacity for
the numerical solution of creep and relaxation problems - in this respect the differential equation mentioned above
would be more favourable - the effective modulus of elasticity ("age-adjusted effective modulus") according to the
subsequent equation is generally used for calculations in practice:

Ec, 28
Ec, 28, eff 
1  ξ  t , t0   φ  t , t0 

This equation uses the "relaxation coefficient" ξ(t,t0) according to Trost [36], ("aging coefficient" according to Bažant
[37]) and thus allows the treatment of creep at variable stresses and strains as a quasi-elastic problem in a simple
algebraic way. Although there are disadvantages associated with this application, such a calculation is possible in
many practical cases with a good approximation because the value for ξ(t,t0) with 0,5 ≤ ξ(t,t0) ≤ 1,0 varies within
narrow limits and can mostly be assumed to be constant as ξ = 0,8. An exact calculation of ξ(t,t0) is complex; details
are given in [10]. The use of equation for Ec,28,eff also has the great advantage that the mathematical type of the
approach for the creep coefficient φ(t,t0) no longer plays a role. Only the accuracy of the prediction of the creep
coefficient is decisive.

[1] FprEN 1992-1-1:2023 (2023) Eurocode 2: Design of concrete structures – Part 1-1: General rules, rules
References for buildings, bridges and civil engineering structures. CEN, Brussels.

[2] EN 1992-1-1:2004-12/AC:2010-11 (2010) Eurocode 2: Design of concrete structures – Part 1-1:

General rules and rules for buildings. CEN, Brussels.

[3] Fédération internationale du béton (fib) (2013) fib Model Code for Concrete Structures 2010. Wiley c/o
Verlag Ernst & Sohn, Berlin.

[4] fib Bulletin 70 (2013) Code-type models for structural behaviour of concrete: Background of the
constitutive relations and material models in the fib Model Code for Concrete Structures 2010. State-
of-art report, Lausanne.

[5] CEB Bulletin 199 (1990) Evaluation of the time dependent behavior of concrete. CEB Bulletin
d'Information, No. 199, Comité Euro-International du Béton, Lausanne.

[6] Müller, H. S.; Acosta, F.; Kvitsel, V. (2021) Models for predicting shrinkage and creep of concrete - Part

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 541
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

1: Analysis of the shrinkage model in DIN EN 1992-1-1:2011 and new approach in Eurocode 2 prEN
1992-1-1:2020 in: Beton- und Stahlbetonbau 116, H. 1, pp. 2-18; Part 2a: Creep - Principles and
analysis of the creep model in DIN EN 1992-1-1:2011 in: Beton- und Stahlbetonbau 116, H. 9, pp. 660-
676; Part 2b: Creep - New approach in Eurocode 2 prEN 1992-1-1:2020 in: Beton- und Stahlbetonbau
116, H. 10, pp. 808-820 (in German).
[7] Müller, H. S.; Acosta F.; Kvitsel, V.; Torrenti, J. M. (2023) Prediction of creep and shrinkage – improved
approaches for fib Model Code 2020 in: Structural Concrete, in preparation.

[8] Neville, A. M.; Dilger, W. H.; Brooks, J. J. (1983) Creep of plain and structural concrete. Construction
Press, London, Great Britain.

[9] Bažant, Z. P. (1988) Mathematical modelling of creep and shrinkage of concrete. John Wiley & Sons
Ltd., Chichester, Great Britain.

[10] CEB Bulletin 215 (1993) Structural effects of time dependent behaviour of concrete. Comité Euro-
International du Béton, Lausanne, Switzerland.

[11] Bažant, Z. P; Jirasek, M. (2018) Creep and Hygrothermal Effects in Concrete Structures. Springer
Science+Business Media, Dordrecht, The Netherlands.

[12] CEB-FIP Model Code 1990 (1993) Design Code. Committée Euro-International du Béton (CEB),
Thomas Telford Services Ltd., London.

[13] EN 1992-2:2005/AC:2008 (2005) Eurocode 2: Design of concrete structures - Concrete bridges -

Design and detailing rules. CEN, Brussels.

[14] ACI Committee 209 (2008) Guide for modeling and calculating shrinkage and creep in hardened
concrete. ACI Report 209.2R-08, Farmington Hills.

[15] Troxell, G. E.; Raphael, J. M.; Davis, R. E. (1958) Long time creep and shrinkage tests of plain and
reinforced concrete. ASTM Proceedings, Vol. 58, pp. 1101-1120.
[16] Stöckl, S. (1981) Experiments on the influence of the load level on the creep of concrete. German
Committee for Reinforced Concrete (Deutscher Ausschuss für Stahlbeton), H. 324, Beuth Verlag
GmbH, Berlin (in German).

[17] Brooks, J. J. (2005) 30-year creep and shrinkage of concrete in: Magazine of Concrete Research, Vol.
57 Issue 9, November 2005, pp. 545-556.

[18] Hilsdorf, H. K.; Rüsch, H.; Kordina, K. (1970) Unpublished experiments at the MPA Munich (in German)

[19] Le Roy, R.; Le Maou, F.; Torrenti, J. M. (2017) Long term basic creep behavior of high performance
concrete: data and modelling in: Materials and Structures 50, H. 1, pp. 1-11.

[20] CEB Bulletin 72/73 (1970) CEB-FIP Model Code 1970: International Recommendations for the design
and construction of concrete structures. Comité Européenne du Béton, Paris, London.

[21] Müller, H. S. (1986) On the prediction of the creep of structural concrete [Dissertation]. University of
Karlsruhe (in German).

[22] Müller, H. S.; Anders, I., Breiner R., Vogel, M. (2013) Concrete: treatment of types and properties in fib
Model Code 2010 in: Structural Concrete, Vol. 14, No 4, pp. 320-334.

[23] Anders, I. (2017) Material law for the description of the creep and relaxation behaviour of young normal-
and high-strength concretes. Karlsruhe Institute of Technology (KIT). Karlsruher Reihe Massivbau,
Baustofftechnologie, Materialprüfung; Heft 73 and Deutscher Ausschuss für Stahlbeton, H. 624, Beuth
Verlag GmbH, Berlin, 2017 (in German).

[24] Müller, H. S.; Küttner, C. H.; Kvitsel V. (1999) Creep and shrinkage models of normal and high-
performance concrete – concept for a unified code-type approach in: Revue française de génie civil,
Vol. 3, No. 3-4, pp. 113-132.

[25] CEB Bulletin 124/125 (1978) CEB-FIP Model Code 1978: International System of Unified Standard –
Codes of Practice for Structures. Comité Européenne du Béton, Paris, Lausanne.

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 542
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.3 to B.8 Background to FprEN 1992-1-1:2023

[26] prEN 1992-1-1:2001-04 (2001) Eurocode 2: Design of concrete structures – Part 1: General rules and
rules for buildings. European Committee for Standardization.

[27] Wagner, O. (1958) The creep of unreinforced concrete. German Committee for Reinforced Concrete
(Deutscher Ausschuss für Stahlbeton), H. 131, Verlag Ernst & Sohn, Berlin (in German).

[28] Hilsdorf, H. K., Müller, H. S. (1983) Discussion of a note "Comments on the use of Ross' hyperbola and
recent comparisons of various practical creep prediction models" by Z. P. Bažant and J. C. Chern in:
Cement and Concrete Research, Vol. 13, pp. 441–443.

[29] Rüsch, H., Jungwirth, D., Hilsdorf, H. K. (1983) Creep and Shrinkage. Their Effect on the Behavior of
Concrete Structures. Springer Verlag, New York.

[30] Müller, H. S. and Pristl, M. (1993) Creep and shrinkage of concrete at variable ambient conditions in:
Bažant, Z. P. und Carol I., [Ed.] Creep and Shrinkage of Concrete, Proceedings of the 5th International
RILEM Symposium on Creep and Shrinkage of Concrete (ConCreep 5). & FN Spon, London, pp. 15–
26.

[31] Kollegger J. (2020) Creep and shrinkage - Development of practical models for the prediction of creep
and shrinkage of concrete. Report to the Austrian Association for Structural Engineering (öbv) (in
German).

[32] Kollegger J. et al. (2023) Creep and shrinkage - Development of practical models for the prediction of
creep and shrinkage of concrete. Publication in Beton- und Stahlbetonbau, in preparation (in German).

[33] Suza, D. (2020) Influence of the scale effect and environmental conditions on the creep and shrinkage
of concrete [Dissertation]. Vienna University of Technology.

[34] Bažant, Z. P.; Panula, L. (1978) Practical prediction of time-dependent deformations of concrete in:
Matériaux and Constructions, Vol. 11, No. 65, 1978; Vol. 11, No. 66, 1978; Vol. 12, No. 69, 1979.

[35] RILEM draft recommendation: TC-242-MDC multi-decade creep and shrinkage of concrete: material
model and structural analysis. Model B4 for creep, drying shrinkage and autogenous shrinkage of
normal and high-strength concretes with multi-decade applicability. RILEM Technical Committee TC-
242-MDC (Zdenek P. Bažant, chair) in: Materials and Structures (2015), Vol. 48, pp. 753-770.

[36] Trost, H. (1967) Effects of the superposition principle on creep and relaxation problems in concrete and
prestressed concrete in: Beton- und Stahlbetonbau, H. 10 und 11 (in German).

[37] Bažant, Z. P. (1972) Prediction of concrete creep effects using age-adjusted effective modulus method
in: ACI Journal, Vol. 69.

[38] Torrenti, J. M. (2018) Basic creep of concrete-coupling between high stresses and elevated
temperatures in: European Journal of Environmental and Civil Engineering, Vol. 22, No. 12, pp. 1419-
1428.

[39] Torrenti, J. M.; Benboudjema, F. (2013) Desiccation shrinkage of large structures: is there a size effect?
in: Proceedings of ConCreep9, MIT, Cambridge, MA, USA.

List of versions of this document

24.08.2021 1st draft
20.01.2023 2nd draft
09.03.2023 peer reviewed and final draft

H. S. Müller (Convenor TG 7) 20.01.2023

For citations, copying and referencing see Note on Front page! page 543
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to Annex B.9

Validation of provisions for relaxation of
prestressing steel

Hans Rudolf Ganz Switzerland

30.09.2022

For citations, copying and referencing see Note on Front page! page 544
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.9 Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared in order to validate the provisions in FprEN 1992-1-1:2023, Annex B.9, for
relaxation of prestressing steel.

B.9 Reasons for change

Relaxation of prestressing Current provisions in EN 1992-1-1:2004 specify three classes of relaxation: two for ordinary and low relaxation
steel of wire and strand, and one for bars. However, in practice in Europe only one class is used and specified for
wire and strand since many years, and one for bars. For the evolution of the relaxation loss over time, EN 1992-
1-1:2004 gives expressions for each of the three classes, (3.28) to (3.30). However, when applying these
expressions and checking relaxation losses at 1000 hours, the calculated relaxation loss at a stress of 70% of
tensile strength is about 1% and thus significantly lower than the maximum value of 2,5% specified for wire and
strand in FprEN 10138:2009 and national specifications commonly referenced for prestressing steel in practice.
Several comments in the Systematic Review of EN 1992-1-1:2004 have requested to update and correct in
particular this last item, i.e. adapt provisions such that the maximum values specified in product standards apply
unless testing of actually used products confirm smaller values.

Determination of relaxation loss according to B.9

B.9(1), (2), (3)
In FprEN 1992-1-1:2023 the approach given in fib Model Code 2010, [B.9-1], for relaxation losses of
prestressing steel has been adopted with some simplifications to improve ease of use. For the evolution of
relaxation losses over time a power function with the loss at 1000 hours (either value specified in product
standard or value measured in tests) is used which corresponds to a straight line when presenting the relaxation
losses over time in a double logarithmic diagram, Formula (B.43). Coefficient kp represents the slope of the
straight line in the double logarithmic scale which, in lieu of actual test results, may be assumed as kp = 0,16
(Note: time t expressed in days as per Clause 3.9).

For the effect of initial stress between 70% and 80% of the tensile strength, relaxation values at 1000 hours may
be linearly interpolated between the specified or tested values at these stress levels. For stresses below 70% of
tensile strength, relaxation values at 1000 hours may be linearly interpolated between the value specified at
stress of 70% and zero relaxation loss assumed at a stress of 50% of the tensile strength. Table B.9-1 shows
the assumed losses at 50% as well as the specified maximum values at 70% and 80% of the tensile strength of
the prestressing steel (corresponding to Table B.4 in FprEN 1992-1-1:2023).

Table B.9-1: Relaxation losses at 1000 hours

FprEN 1992-1-1:2023 provides a two-level approach for determining the evolution of relaxation losses over
time:
‐ In the simplest case (Level 1), without specific information (test results) on the actual properties of
prestressing steel to be used e.g. at the time of design prior to construction, a designer uses the
maximum relaxation values at 1000 hours specified in the product standard and given in Table B.9-1
above together with Formula (B.43) and the assumed value of kp = 0.16 to estimate the evolution of
relaxation losses and in particular the final value or relaxation loss at the end of the design service life
(or a time specified for the project). Assuming an initial stress level of 70% of actual tensile strength,
this results in total relaxation loss of 2,65 1000 for 50 years, and 3,0 1000 for 100 years design service
life, respectively. In Level 1 approach, since no test data is available, fpk needs to be used in lieu of
the actual tensile strength fp.
B.9(4)
‐ When specific information (test results) on the relaxation properties of prestressing steel to be used is
available, either from test certificates of the actual batch of prestressing steel or based on prior tests
performed by the manufacturer on the same type of prestressing steel, Level 2 may be applied. In the
Level 2 approach, the coefficient kp results from the best fit straight line based on the relaxation loss
values measured in the tests up to 1000 hours. Formula (B.43) may then be used with the measured
value 1000 and the coefficient kp from the best fit straight line. In Level 2 approach, since test data is
available, relaxation loss based on the actual strength fp may be used.

H.R. Ganz 13.03.2023

For citations, copying and referencing see Note on Front page! page 545
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.9 Background to FprEN 1992-1-1:2023

Relaxation losses of prestressing steel are sensitive to temperature, see [B.9-1]. Hence, testing of relaxation
B.9(6)
losses at elevated temperature is recommended if the prestressing steel is expected to be exposed over
extended periods of time to temperatures of 30°C or higher. For temperatures below 20°C, the relaxation values
at 20°C may conservatively be assumed. For prestressed members subject to heat treatment, the provisions in
Clause 13 apply regarding relaxation loss during the period of heat treatment.

Summary of the test results

One argument for using Formulae (3.28) to (3.30) in EN 1992-1-1:2004 was that the actual tensile strength of
prestressing steel can significantly exceed the specified tensile strength. In [B.9-2], a difference of the order of
14% was mentioned, see also [B.9-3]. A recent evaluation of 57 test results for 15.7mm strands, from 8 different
production batches, from 4 different manufacturers showed a quite different picture with a ratio of actual /
specified tensile strength for batches ranging from 1,015 to 1,04, see Figure B.9-1. While this evaluation may
not be representative for all types of prestressing steel and all manufacturers, it shows anyway that the ratio of
actual / specified tensile strength of prestressing steel will often be rather close to unity. Therefore, very
significant overstrength will be the exception in practice. Hence, Level 1 approach above, with the assumption of
fp = fpk, will typically result in only a small overestimation of the relaxation loss. Assuming an actual tensile
strength of 1,03 fpk would reduce relaxation loss at 1000 hours to 0,97 1000.

(a) Actual / specified strength per test (b) Actual / specified strength distribution

Figure B.9-1: Actual / specified tensile strength of 7-wire prestressing strands (sample selection, see text above)

Figure B.9-2 illustrates a limited evaluation of the coefficient kp as a function of the initial stress level. As can be
seen, the coefficient varies with initial stress level. However, kp = 0,16 is a fair value for initial stress levels
between 70% and 80% of the tensile strength.

Figure B.9-2: Coefficient kp as a function of initial stress level in prestressing steel

If actual results of relaxation tests are available, Level 2 approach can be used. Figure B.9-3 shows two
examples of relaxation tests, both performed at an initial stress of 70% of the tensile strength up to 1000 hours,
one on 12.9mm diameter strand, the other on 15.7mm strand of 1860 MPa specified tensile strength. The figure
shows the actual test values together with the extrapolated relaxation loss at 50 years design service life
(highlighted with red circle). The formulae for the best fit straight lines are also shown in the figures.

(a) 12.9mm strand stressed to 70% tensile strength (b) 15.7mm strand stressed to 70% tensile strength

Figure B.9-3: Examples of best fit approximation for relaxation loss based on test results (Level 2 approach)

H.R. Ganz 13.03.2023

For citations, copying and referencing see Note on Front page! page 546
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
B.9 Background to FprEN 1992-1-1:2023

The extrapolated relaxation losses at 50 years design service life in Figure B.9-3 are 6.1% and 4.1% for the
12.9mm strand and the 15.7mm strand, respectively. These two samples had actual values of 1000 = 1.37% and
1000 = 1.09% for the 12.9mm and 15.7mm strand, respectively. This can be compared with the estimations
based on Level 1 approach of 6.6% based on the maximum specified value of 1000 = 2.5% A limited evaluation
of another 10 relaxation tests from three manufacturers with strands 12.5mm to 15.7mm, tensile strength of
1770MPa and 1860MPa, and stressed to 70% and 80% of actual tensile strength, gave extrapolated relaxation
values at 50 years design service life between 2.7% and 6.6% with an average value of 4.5%. Hence, while with
the estimated loss of 6.6% based on Level 1 approach some conservatism may be present in some cases, it is
not conservative in other and certainly not in all cases.

Conclusions
‐ The changes introduced for relaxation losses in FprEN 1992-1-1:2023, Annex B.9, are justified based
on the above data and are easy to use.
‐ The Level 1 approach based on maximum specified relaxation loss at 1000 hours from product
standards and Formula (B.43) with an assumed coefficient kp = 0.16 give fair values of final relaxation
loss at the end of the design service life. The slight conservatism in some cases is felt to have a
marginal effect, if any, on the design and economy of prestressed structures, in particular for bonded
tendons where relaxation accounts for about one third of the total long-term losses.
‐ For large projects Level 2 approach based on actual relaxation test data or past relaxation testing of
the same type of prestressing steel from the manufacturer may be appropriate in some cases where
the effective prestress is relevant in design.

[B.9-1] fib Model Code for Concrete Structures 2010, Fédération Internationale du Béton, Lausanne, 2013.
References
[B.9-2] A. W. Beeby, R. S. Narayanan and H. Gulvanessian, “Designers' Guide to Eurocode 2: Design of
Concrete Structures”, Institution of Civil Engineers, London, 2005.
[B.9-1] CEN/TC 250/SC 2/WG 1 N 834.

List of versions of this document

03.03.2021. 1st draft prEN 1992-1-1:2020-11
30.09.2022 Updated to FprEN 1992-1-1:2022
13.03.2023 Peer reviewed

H.R. Ganz 13.03.2023

For citations, copying and referencing see Note on Front page! page 547
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1 Draft:2023

Background to
ANNEX C (NORMATIVE):
REQUIREMENTS FOR MATERIALS

Frank Fingerloos DBV (German Society for Concrete and Construction Technology), DE
Alexander Lindorf DBV (German Society for Concrete and Construction Technology), DE
Lars Meyer DBV (German Society for Concrete and Construction Technology), DE
Tony Jones The Concrete Centre, UK
Werner Fuchs (for C.8) University of Stuttgart, DE
21.02.2023

For citations, copying and referencing see Note on Front page! page 548
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

Introduction
This background document has been prepared to explain the necessity and content of Annex C:
Requirements to Materials.
NOTE This background document is subject to changes according to the ongoing discussion within
CEN/TC250 and its AHG Interface between Eurocodes and hENs. Therefore, the following explanations are
preliminary.

C.1 Use of this annex The design according to Eurocode 2 is based under the assumption that reliable and accurate
C.2 Scope and field of ap- information at a sufficient confidence level is available for the construction materials and construc-
plication tion products used. This information can be obtained from the following sources:
Case 1 from Declarations of Performance (DoP) based on harmonised technical specifications
(hEN/EAD) based on the European Construction Products Regulation (CPR) [1];
Case 2 from other technical provisions (e.g. non-harmonized European standards);
Case 3 from other and, if applicable, project-related specifications.
The required performance of the construction materials and construction products to be fulfilled as
a minimum when used in accordance with Eurocode 2 shall be defined in the form of requirements.
These requirements, which are based on the fitness for use of materials and products, are initially
expected in material or product standards. However, after the changeover from the Construction
Products Directive (CPD) to the CPR in 2013, such requirements may no longer be included in
harmonised specifications (hEN/EAD) based on the CPR. Therefore, they must be specified in
Eurocode 2 for these cases.
The normative Annex C serves this purpose: It is necessary to formulate the requirements for prop-
erties to be fulfilled for those materials and products that are to be used in the design according to
Eurocode 2 under the assumptions given there.
Case 1 provides a technical basis to assess the performance of construction products. The manu-
facturer has to deliver a Declaration of Performance (DoP) in which at least one (essential) char-
acteristic is declared. However, this characteristic can be also irrelevant for design with Eurocode 2.

To ensure a consistent system of parameters needed for the usability of a construction product in
connection with a Eurocode design, the design codes of CEN/TC 250 shall name product charac-
teristics needed for design – if needed in conjunction with threshold values [2]. Therefore, all per-
formance characteristics with necessary requirements (in values or pass criteria of tests) shall be
given in Eurocode 2, Annex C. This is because there is no other document to which the Eurocodes
could refer, and which would contain these necessary requirements and the rules on how to demon-
strate compliance with them. A "voluntary DoP" can be used for those properties that cannot (yet)
be explained via a mandatory DoP because they are not listed in Annex ZA of a hEN or in an EAD,
but which are required for design according to Eurocode 2 [2]. The necessary 'set of properties' is
included in the normative Annex C.

It is expected, that the hENs for reinforcing and prestressing steels will not be finally published in
the Official Journal of EU (OJEU), when this Eurocode 2 is published in 2023. So national product
standards or approvals may be used if compatible with material requirements of Annex C. There-
fore NDP-formulations are established in the code-text as follows:
(…) … steels used for structures designed in accordance with this Eurocode shall comply with
the relevant standards for … steel.
NOTE 1 The National Annex can specify relevant standards for … steel.
NOTE 2 The harmonised product standard EN … for reinforcing steel is currently under development.
Still unsolved is a short-term solution how to implement missing parameters in the scope of product
hENs and therefore in the DoPs of the producers. A general way to proceed is given in document
JIS Action 5, 2018.08.24: CPR-Procedure to develop a standardisation request. Unfortunately
based on this it must be expected that it will last about 5 years until new hENs will be drafted and
published in the OJEU [2].

Because the new Eurocode 2 should be finalised in 2023 a short-term solution with Annex C is
established. Three general case scenarios for construction product properties were detected (see
Figure BD.C.1).

Case 1 contains properties which are covered in harmonised technical specifications (hENs/EADs)
or in non-harmonised documents and do not need any modification. Furthermore, Case 1 is divided
into the sub-cases 1a), 1b) and 1c).
– 1a) Since the required properties in terms of accuracy, reliability and confidence level are fully
covered by the harmonized technical specification based on the CPR, no further action is re-
quired in Case 1a).

22.02.2023
For citations, copying and referencing see Note on Front page! page 549
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

– 1b) In Case 1b) a parameter is covered by the system of the CPR, but it is not listed in an Annex
ZA (and therefore cannot be declared in a DoP). The required level of accuracy, reliability and
the level of confidence have to be established in Eurocode 2.
– 1c) For parameters not addressed in harmonised specifications, a declaration of the existing
performance in a (mandatory) DoP is not given, see case 1c). Here, the information must be
declared otherwise – e.g. in a voluntary DoP. The basis for such a voluntary DoP can also be
non-harmonized rules.
Case 2 describes cases where a required property is in principle addressed in an EN or in another
rule but requires an adaptation rule in the Eurocode regarding accuracy, reliability or confidence
level. A simple declaration of performance in a voluntary DoP would be insufficient here.

Case 3 occurs when no EN or other rule exists for a required property. In this case, requirements
for the property must be defined in the Eurocode – including regulations on accuracy, reliability and
confidence level.

Figure BD.C.1. The different cases for determining the properties necessary for the design [2]

Additional information on the requirements applicable to a product in a specific Member State can
be obtained from the national Product Contact Point for Construction established by each Member
State.

A mechanism for completing the standards and strengthening the DoP should be established in
the EU-Member States (e.g. with the support of industry) by drawing up up-to-date lists of the re-
quired essential characteristics for construction products for certain uses. Member States shall de-
termine under their own responsibility the performance requirements for the essential characteris-
tics in relation to the intended use of a construction product. In the context of the currently ongoing
consultations on the revision of the CPR, corresponding comments from the industry (FIEC, CPE,
SBS) were submitted and presented as "Option F" [3]. The Commission is also aware of this prob-
lem and has launched a so-called acquis process to drive forward the completion of the standards
[4].

Whilst the preceding topics concerns Members of the European Union, the Eurocodes assumes
non-EU states adopting the Eurocodes will have similar procedures.

C.1 (1) Precast concrete Under the umbrella of the non-harmonised EN 13369:2018: Common rules for precast concrete
products products [5] the most of precast concrete products are covered by harmonised standards (hEN).
The manufacturers shall draw up a declaration of performance DoP based on an appropriate hEN.
The difficulty with precast concrete elements is that the rules for the concrete used are not harmo-
nised in the member states and to that extent differ. The same applies currently and until the final
publication of prEN 10080/prEN 10370 for reinforcing steel and of EN 10138 for prestressing steel.
In addition, different national implementations of FprEN 1992-1-1 are to be expected for the design.
Against this background, the following requirement is formulated in C.1 (1) in FprEN 1992-1-1:

(1) …This Normative Annex also applies to precast concrete products according to EN 13369 de-
signed according to this standard.

Precast concrete products that do not take these national differences into account may not be used
in another member state without further specifications. In this respect, compliance with the rules
valid in the place of use is relevant for usability – not compliance with the rules in the place of
manufacture.

22.02.2023
For citations, copying and referencing see Note on Front page! page 550
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

C.3 Concrete Relevant: Case 1c) in Figure BD.C.1.

C.3.1 Normal weight, The main design property concrete compressive strength fck is classified in concrete strength
heavy weight and Light- classes according to EN 206 [6]. The producer confirms (only) the 5%-quantile fck;0,05 of a class
weight Aggregate Con- with initial testing and during factory production control (conformity tests).
crete (LWAC)
It should be noted, that all other concrete properties according to FprEN 1992-1-1-clause 5.1
used in design (like values of Ec, fct, c, cs, c, c,th) are theoretical or empirical benchmark values
and are not covered in EN 206. If such concrete properties are required in design, the designer
shall specify these characteristics with the required values, reliability and confidence level. These
required values shall be confirmed than by the concrete producer in a Technical Documentation.

In EN 206 [6]/EN 12620 [7] the limitation of the aggregate size is designated in terms of lower
(d) and upper (D) sieve sizes expressed as d / D. In the design the values of Dmax shall be specified
with Dupper (for minimum cover, spacing of bars, tendons and couplers) and Dlower (shear and
punching resistance):
– Dlower: smallest value of upper sieve size D for the coarsest fraction of aggregates in the concrete
permitted by the specification of concrete,
– Dupper: largest value of upper sieve size D for the coarsest fraction of aggregates in the concrete
permitted by the specification of concrete
– Dmax: Declared value of upper sieve size D of the coarsest fraction of aggregates which is actu-
ally used in the concrete
The definition of D via Dlower in EN 12620 can lead to a range of aggregate gradings. EN 206 does
not specify a minimum coarse aggregate content.

The shear and punching models are calibrated against tests carried out with typical gradings. Dlower
should be equal or greater than 8 mm. The use of non-typical aggregate gradings where the per-
centage of larger aggregate sizes in relation to Dlower is small can result in different behaviour. This
can be avoided by specifying grading parameters in addition to Dlower.

It is expected that the new Exposure Resistance Classes (ERC) according to FprEN 1992-1-1-
clause 6.4 will be implemented in a new prEN 206-100. In the event EN 206 or prEN 206-100 does
not refer to ERC, a National Application Document to EN 206 (or a National Annex of EN1992-1-1
can provide the necessary advice on how to implement ERC rules in a country.

C.4 Reinforcing steel Relevant: Case 1b) in Figure BD.C.1  Voluntary DoP for strength and ductility classes required.
C.4.1 Carbon reinforcing
steel For FprEN-1992-1-1-design strength classes (B400, B450, B500, B550, B600, B700) and ductil-
ity classes (A, B, C) of reinforcing steel are required. prEN 10080 [10] does not contain these
classes. Therefore, an interface between the defined FprEN-1992-1-1-classes according to Tables
C.1/C.2 and the DoP according to Annex ZA of prEN 10080 is necessary. Table BD.C.1 contains
the interface conditions. See also CEN-TC250-SC2_N1438 [11].

Intermediate strength classes can be defined and used in the National Annex (see 5.2.2, Table 5.4).

It is expected, that reinforcing steel for ULS and SLS-design is produced under Assessment and
Verification of Constancy of Performance (AVCP) system 1+ addressed in prEN 10080, Table ZA.3.

Each reinforcing steel shall bear on one rib or indentation row, a numerical mark identifying the
manufacturer's works. This mark shall be repeated at an interval of not more than 1,5 m. An addi-
tional voluntary marking with max. 6 characters of Re, Rm, Agt or classes (like 500B) with real
numbers directly on the reinforcing steel is expected to reach the assumed design reliability on
execution on site. A combination with the product identification number allocated by the manufac-
turer himself, added by a code/link to declaration of performance on web is possible.
Table BD.C.1. Interface conditions prEN 10080 [10] to FprEN 1992-1-1
prEN 10080: Table ZA.1 – Relevant clauses for FprEN 1992-1-1
reinforcing steel and intended use
Essential Cl. Classes Notes Required classes and/or threshold
characteristics and/or levels
threshold
levels
Elongation at 5.2 – Declared value For specifying ductility classes A, B, C with
maximum load (10% fractile value) elongation at maximum load as a parameter:
Declared value Agt (used as uk) as 10%
quantileb value for production lotc:
Ductility class A: Agt  2,5 %
Ductility class B: Agt  5,0 %
Ductility class C: Agt  7,5 %

22.02.2023
For citations, copying and referencing see Note on Front page! page 551
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

Weldability 5.1 – Declared values Usable in design and execution of welded

(maximum carbon reinforcement with recommended values for
equivalent value the chemical composition in prEN 10080,
Ceq,max and the maxi- Table 2d). For these values it is known that
mum content of the the essential characteristics can be tested
specified individual properly and that test results are valid.
elements Cmax, Smax,
Pmax, Nmax and Cumax )
Sections and 5.8 – Declared values Declared value for max deviation from
tolerances on (nominal value for d; nominal mass of bar or wire:
sizes range of maximum Δm  4,5% 8 mm
deviation from the Δm  6,0% 8 mm
nominal mass per
meter Δm; range of Recommended tolerance ranges
maximum deviation Bars:
Δl, Δw, Δp, Δo for length (L): ± 40 mm if L ≤ 5,0 m,
welded fabric; ± 0,8 %, if L > 5,0 m;
range of maximum height (H1, H2): +1 mm and − 3 mm;
deviation ΔL, ΔH1 width (B1, B2): ± 7,5 mm;
and ΔH2, ΔB1 and pitch (Ps): ± 2,5 mm;
ΔB2, ΔPs for lattice Welded fabric:
girder) length and width: max{± 25 mm; ± 0,5 %}
wire pitch: max{± 15 mm; ± 7,5 %}
overhangs: > 25 mm
Bendability 5.5 – Declared value Usable in design and execution for bended
(maximum mandrel bars, stirrups, links, hooks:
diameter dmax in bend Pass (both bend and re-bend tests)
and/or re-bend test) for production lotc.
No steel cracking is permitted.

Bend test specified in accordance with

EN ISO 15630-1 [12] with mandrels of
 3 for bars 16 mm,
 6 for bars 16 mm,
and the minimum bend-angle of 180°.
Re-bend test specified in accordance with
EN ISO 15630-1 [12] with mandrels of
 5 for bars 16 mm,
 8 for bars 16 mm < 25 mm,
 10 for bars 25 mm,
and the minimum bend-angles of 90° and re-
bend of 20°
Bonding strength 5.9 – Declared values Usable in design with minimum values of
(for fR a 5% fractile relative rib area fR,min for ribbed bars:
value; for br/c a
maximum value; Ribbed reinforcement
for bm and bu mean Declared 5% quantileb value for minimum
values) relative rib area fR,min for production lotc
5 to 6 mm: 0,035
6,5 to 12 mm: 0,040
 > 12 mm: 0,056
Option to supplement values and declare bm
and bu

Indented reinforcement
The indentations shall form an angle of incli-
nation with the longitudinal axis of 35° to 75°.
The indentation flank inclination a shall be
40° to 80°. Requirements of prEN 10080, 7.4,
are accepted.
Shear force 5.3 – Declared values Usable in design and execution of welded
(for lattice girders (minimum values for fabrics and lattice girders
and welded fab- Fw,min, Fs,min and Fd,min) C.4.1(3): In welded fabric the declared mini-
mum shear force value shall be
rics)
Fs ≥ 0,25  Re,nom  An (C.1)
where Re, nom is the nominal characteristic
yield strength and An is the nominal cross-
sectional area of either:
a) the larger wire at the joint in a single wire
welded fabric or
b) one of the twin wires in a twin wire welded
fabric (twin wires in one direction)
Stress ratio 5.2 – Declared value For specifying ductility classes A, B, C with
(ult. tens. strength/ (fractile value) ratio tensile strength/yield strength and ratio
tens. yield actual value of yield strength/nominal value of
yield strength (Rm/Re and Re,act/Re,nom)
strength)
as parameters:

22.02.2023
For citations, copying and referencing see Note on Front page! page 552
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

Declared value Rm/Re as 10% quantileb

value for production lotc:
Ductility class A: Rm/Re  1,05
Ductility class B: Rm/Re  1,08
Ductility class C: Rm/Re  1,15 and < 1,35
Ratio actual value 5.2 – Declared value Declared value Re,act/Re,nom as 10% quantileb
of yield strength/ (10% fractile value) value for production lotc:
nominal value of Ductility classes B and C: Re,act/Re,nom  1,3
yield strength
Tensile yield 5.2 – Declared value For specifying strength classes B with value
strength (5% fractile value) of yield strength Re (used as fyk) as parame-
ter:
Declared value Re as 10% quantileb value for
production lotc:
Strength class B400: Re  400 MPa
Strength class B450: Re  450 MPa
Strength class B500: Re  500 MPa
Strength class B550: Re  550 MPa
Strength class B600: Re  600 MPa
Strength class B700: Re  700 MPa
Tensile strength 5.2 – Declared value Indirectly specified with declared ratio value
(5% fractile value) Rm/Re for ductility classes

Fatiguea 5.4 – Declared values For reinforcing steel under fatigue design:
(N at 2σa and σmax) Declared values as 10% quantileb value for
production lotc:
Fatigue stress range 2a in testing for N  2 ×
106 cycles based on a maximum stress level
of max = 0,6fyk:
 160 MPa for bars/de-coiled bars  ≤ 12 mm
 140 MPa for bars/de-coiled bars 12 mm < 
≤ 16 mm
 130 MPa for bars/de-coiled bars 16 mm < 
≤ 20 mm and for straight bars  > 20 mm
 100 MPa for wire fabrics  ≤ 12 mm
 80 MPa for wire fabrics  > 12 mm
Different fatigue requirements can be set in a
National Annex (see footnote b in Table C.1).
Cyclic load 5.7 – Declared value No requirements
strength for NCLS at εc Not used in FprEN 1992-1-1
Strength at 5.6 – Declared value No requirements for normal design
elevated for RT at T NOTE: In FprEN 1992-1-1, 1.1 is stated, that
temperature the design rules are valid under temperature
conditions between –40 °C and +100 °C gen-
erally.

Assumed parameters for fire design accord-

ing to FprEN 1992-1-2, 5.3.2:
 slope of the linear elastic range Es,θ
 proportional limit fsp,θ
 maximum stress level fsy,θ
Values for hot rolled and cold worked rein-
forcing steel at elevated temperatures could
be obtained from FprEN 1992-1-2, Table 5.3.
Durability 5.10 – Declared values Recommended are the values for the chemi-
(maximum carbon cal composition in prEN 10080, Table 2d). For
equivalent value these values it is known that the essential
Ceq,max and the maxi- characteristics can be tested properly and
mum content of the that test results are valid. Thus, a conclusion
specified individual to the safe application in building construction
elements Cmax, Smax, is possible for those products.
Pmax, Nmax and Cumax)
a
Not applicable for lattice girders.
b
Characteristic value for which the maximum % of test results falling below the characteristic value is given
(min/max values for individual test results).
c
FprEN 1992-1-1 has a Note in C.4.1 (2) which refers to difference between definitions in prEN 10080 and
FprEN 1992-1-1 and considers methods in prEN 10080 a sufficient check for obtaining fyk.
NOTE: While there is a difference between long term quality level of production and value for production lot
considered in prEN1992-1-1, this difference should be taken care by footnote c for all characteristics.
d)
prEN 10080, Table 2:

22.02.2023
For citations, copying and referencing see Note on Front page! page 553
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

C.4.2 Stainless reinforcing Relevant: Case 1b) in Figure BD.C.1  Voluntary DoP for strength classes required.
steel
For FprEN 1992-1-1-design strength classes (B400, B450, B500, B550, B600, B700) and ductil-
ity classes (A, B, C) of stainless reinforcing steel are required. prEN 10370 [13] does not contain
strength classes. Therefore, an interface between the defined FprEN 1992-1-1-classes according
to Tables C.3 and C.4 and the DoP according to Annex ZA of prEN 10370 is necessary. Table
BD.C.2 contains the interface conditions.

Intermediate strength classes can be defined and used in the National Annex (see 5.2.2, Table 5.4).

It is expected that stainless reinforcing steel for ULS and SLS-design is produced under Assess-
ment and Verification of Constancy of Performance (AVCP) system 1+ addressed in prEN 10370,
Table ZA.3.
Table BD.C.2. Interface conditions prEN 10370 [13] to FprEN 1992-1-1
prEN 10370: Tables ZA.1x – Relevant clauses for FprEN 1992-1-1
reinforcing steel (stainless steel) to be used for
reinforcement of concrete
Essential Cl. Classes Notes Required classes and/or threshold
characteristics and/or levels
threshold
levels
Elongation 4.1 Threshold Declared value For specifying ductility classes A, B, C with
elongation at maximum load as a parameter:
Declared value Agt (used as uk) as 10%
quantileb value for production lotc:
Ductility class A: Agt  2,5 %
Ductility class B: Agt  5,0 %
Ductility class C: Agt  7,5 %
Weldability 4.2 Threshold Declared steel Usable in design and execution of welded
designation reinforcement with stainless steels designa-
tions is given in EN 10088-1 or EN 10088-5.
Sections and 4.3 Threshold Declared nominal Declared value for max deviation from nomi-
tolerances on diameter nal mass of bar or wire:
sizes Δm  4,5% 8 mm
Δm  6,0% 8 mm

Recommended tolerance ranges

length (L): ± 40 mm if L ≤ 5,0 m,
± 0,8 %, if L > 5,0 m;
height (H1, H2): +1 mm and −3 mm;
width (B1, B2): ± 7,5 mm;
pitch (Ps): ± 2,5 mm
Bendability 4.4 – Conformity Usable in design and execution for bended
bars, stirrups, links, hooks:
Pass (both bend and re-bend tests)
for production lotc.
No steel cracking is permitted.

Bend test specified in accordance with

EN ISO 15630-1 [12] with mandrels of
 3 for bars 16 mm,
 6 for bars 16 mm,
and the minimum bend-angle of 180°.
Re-bend test specified in accordance with
EN ISO 15630-1 [12] with mandrels of
 5 for bars 16 mm,
 8 for bars 16 mm < 25 mm,
 10 for bars 25 mm,
and the minimum bend-angles of 90° and
re-bend of 20°.

22.02.2023
For citations, copying and referencing see Note on Front page! page 554
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

Bonding strength 4.5 – Declared geometry Usable in design with minimum values of
relative rib area fR,min for ribbed bars:
Ribbed reinforcement
Declared 5% quantileb value for minimum
relative rib area fR,min for production lotc
 ≤ 6 mm: 0,039
6,5 to 8,5 mm: 0,045
9 to 10,5 mm: 0,052
11 to 50 mm: 0,056
Indented reinforcement
The indentations shall form an angle of incli-
nation with the longitudinal axis of 35° to 75°.
Shear force 4.6 – Declared value Usable in design and execution of welded
according to the fabrics and lattice girders
declared test method C.4.1 (3): In welded fabric the declared mini-
mum shear force value shall be
Fs ≥ 0,25  Re,nom  An (C.1)
where Re, nom is the nominal characteristic
yield strength and An is the nominal cross-
sectional area of either:
a) the larger wire at the joint in a single wire
welded fabric or
b) one of the twin wires in a twin wire welded
fabric (twin wires in one direction).
Stress ratio 4.7 Threshold Declared ratio For specifying ductility classes A, B, C with
(maximum ratio tensile strength/0,2 % proof strength
strength/ tensile Rm/Rp0,2 (where Rm should be limited to stress
at 7 % elongation) as parameters:
yield strength)
Declared value Rm/R0,2 as 10% quantileb
value for production lotc:
Ductility class A: Rm/Rp0,2  1,05
Ductility class B: Rm/Rp0,2  1,08
Ductility class C: Rm/Rp0,2  1,15 and < 1,35
Tensile yield 4.8 Threshold Declared value For specifying strength classes B with value
strength of 0,2 % proof strength Rp0,2k (used as fyk) as
parameter:
Declared value Rp0,2 as 10% quantileb value
for production lotc:
Strength class B400: Rp0,2  400 MPa
Strength class B450: Rp0,2  450 MPa
Strength class B500: Rp0,2  500 MPa
Strength class B550: Rp0,2  550 MPa
Strength class B600: Rp0,2  600 MPa
Strength class B700: Rp0,2  700 MPa
Tensile strength 4.9 Threshold Declared value Indirectly specified with declared ratio value
Rm/Rp0,2 for ductility classes
Cyclic load 4.10 – Conformity No requirements
Not used in FprEN 1992-1-1
Fatigue 4.11 – Declared value For stainless reinforcing steel under fatigue
design:
Declared values as 10% quantileb value for
production lotc:
Fatigue stress range 2a in testing for N  5 ×
106 cycles based on a stress ratio min/max =
0,2 for bars and de-coiled products:
 200 MPa for  ≤ 16 mm
 185 MPa for 16 mm <  ≤ 20 mm
 170 MPa for 20 mm <  ≤ 25 mm
 160 MPa for 25 mm <  ≤ 32 mm
 150 MPa for 32 mm <  ≤ 50 mm
Different fatigue requirements can be set in a
National Annex (see footnote b in Table C.3).
Strength at 4.12 – Declared value No requirements for normal design
elevated NOTE: In FprEN 1992-1-1, 1.1 is stated, that
temperature the design rules are valid under temperature
conditions between –40 °C and +100 °C
generally.
Assumed parameters for fire design accord-
ing to FprEN 1992-1-2, 5.1 (6):
 slope of the linear elastic range Es,θ
 proportional limit fsp,θ
 maximum stress level fsy,θ

22.02.2023
For citations, copying and referencing see Note on Front page! page 555
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

Durability 4.13 Threshold Declared composition Durability of the product characteristics is

considered achieved if the chemical composi-
tion of the stainless steel complies with those
of EN 10088-1 or EN 10088-5.
a
Not applicable for lattice girders.
b
Characteristic value for which the maximum % of test results falling below the characteristic value is given
(min/max values for individual test results).
c
FprEN 1992-1-1 has a Note in C.4.2 (1) which refers to difference between definitions in prEN 10370 and
FprEN 1992-1-1 and considers methods in prEN 10370 a sufficient check for obtaining f0,2k.
NOTE: While there is a difference between long term quality level of production and value for production lot
considered in prEN1992-1-1, this difference should be taken care by footnote c for all characteristics.

Table 2 of prEN 10370 [13] contains the appropriate requirements of ductility classes A, B, C
(the upper limit 600 MPa in Note has to be updated). A “voluntary” DoP for the ductility class is
expected.

Fatigue requirements has been included in FprEN 1992-1-1 based on BS 6744:2016 [14]. This
standard adopts the same S-N curves for B500 grade stainless steel reinforcing as for carbon steel
based on a modified regime of testing. The code provision extends this for other grades of stainless
reinforcing steel and provides countries with a national choice as to whether this approach is
adopted or modified.
The minimum relative rib area in prEN 10370 [13] is widely conservative compared with the re-
quirements for carbon steel. Therefore, the design rules for bond in FprEN 1992-1-1 are also valid
for stainless reinforcing steel.

C.5 Prestressing steel Relevant: Case 1b) in Figure BD.C.1  Voluntary DoP for strength classes required.
prEN 10138-3 [16] specifies the requirements and definitions for the essential characteristics of
prestressing bars, which are used for the prestressing of concrete and which are delivered as fin-
ished products in the form of:
– hot rolled bars,
– processed bars,
– plain and threaded bars.

22.02.2023
For citations, copying and referencing see Note on Front page! page 556
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

Part 3 is the first available new draft of prEN 10138. The new parts for wires and strands are still
under preparation.
For FprEN 1992-1-1-design strength classes Y of prestressing steel are required. prEN 10138-3
[16] contains strength classes (grades of prestressing bars) in its Table 4. The steel designation
follows EN 10027 [18]. The FprEN-1992-1-1 classes according to Tables C.5 to C.7 referring to
Table 4 of prEN 10138. Table BD.C.3 contains the interface conditions.

A “voluntary” DoP is expected for the actual tensile strength Rp,act for each individual test-specimen
with these upper limits:
– Y1030: Rp,act  1180 MPa
– Y1050: Rp,act  1210 MPa
– Y1100: Rp,act  1260 MPa
– Y1230: Rp,act  1370 MPa
Intermediate strength classes can be defined and used in the National Annex (see 5.3.2, Table 5.6).

It is expected that prestressing steel for ULS and SLS-design is produced under Assessment and
Verification of Constancy of Performance (AVCP) system 1+ addressed in prEN 10138, Table ZA.3.
Table BD.C.3. Interface conditions prEN 10138-3 [16] to FprEN 1992-1-1
prEN 10138: Tables ZA.1 – Relevant clauses for FprEN 1992-1-1
prestressing steel to be used for
reinforcement of concrete
Essential Cl. Classes Notes Required classes and/or threshold
characteristics and/or levels
threshold
levels
Stress ratio 5.1 – Declared value Rpk/Rp0,1k  1,10
10% fractile value at
90% probability
Tensile strength 5.1 – Declared value For specifying strength classes:
5% fractile value at Declared value Rpk as 5% quantilea value for
90% probability production lotb:
Y1030: Rpk  1030 MPa
Y1050: Rpk  1050 MPa
Y1100: Rpk  1100 MPa
Y1230: Rpk  1230 MPa
Tensile yield 5.1 – Declared value For specifying strength classes:
strength 5% fractile value at Declared value Rp0,1k as 5% quantilea value
90% probability for production lotb:
Y1030: Rpk  835 MPa
Y1050: Rpk  950 MPa
Y1100: Rpk  900 MPa
Y1230: Rpk  1080 MPa
Elongation at 5.1 – Declared value Declared value Agt as 10% quantilea value
maximum force 10% fractile value at for production lotb:
90% probability Agt  3,5 %
Relaxation 5.2 – Declared value Percentage loss of initial force F0 determined
Frt in % of F0 in isothermal relaxation test in accordance
with EN ISO 15630-3 [18].
Relaxation at 1000 h as Frt:
see Table C.7 for required 1000 [%]
Sections and toler- 5.6 – Declared ranges Values of 5.6 accepted.
ances on sizes for m, l and s
Surface geometry 5.5 – Declared ranges No parameters for plain bars. The DoP is
for thread geometry only P.
parameters Full set of the thread parameters of 5.5 for
threaded bars is accepted.
Fatigue 5.3 – Declared value For prestressing steel under fatigue design:
N at a force range Fr Declared values as 10% quantilea value for
and a maximum force production lotb:
Fup Fatigue stress range 2a in testing for N  2 x
106 cycles with an upper limit of Rpk:
 200 MPa for plain p ≤ 40 mm
 150 MPa for plain p > 40 mm
 180 MPa for ribbed p ≤ 40 mm
 120 MPa for plain p > 40 mm
Different fatigue requirements can be set in a
National Annex (see footnote a in Table C.7).
Modulus of 5.1 – Declared value 205 000 MPa
elasticity 10% fractile value at
90% probability

22.02.2023
For citations, copying and referencing see Note on Front page! page 557
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

Durability 5.4 – Declared value Values of Table 4 accepted

minimum time in
hours (T) without
fracture in test solu-
tion; median time in
hours to fracture (tf,m)
in test solution; each
at an initial force F0
declared as percent-
age of Fm
a
Characteristic value for which the maximum % of test results falling below the characteristic value is given
(min/max values for individual test results).
b
FprEN 1992-1-1 has a Note in C.5 (1) which refers to difference between definitions in prEN 10138-3 and
FprEN 1992-1-1 and considers methods in prEN 10138-3 a sufficient check for obtaining fp0,1k.
NOTE: While there is a difference between long term quality level of production and value for production lot
considered in FprEN1992-1-1, this difference should be taken care by footnote b for all characteristics.

The relevant EAD for Post-tensioning systems for prestressing of structures (see Fig. BD.C.2) is
EAD 160004-00-0301 [19] according to FprEN 1992-1-1, 5.4 Prestressing Systems.

Fig. BD.C.2. Installed PT-system/kit – schematic (Fig. 1 in [19])

C.6 Couplers Relevant: Case 1b) in Figure BD.C.1  “Voluntary DoP” is required.
The National Annex can specify relevant standards (or national approvals) for couplers designed
in accordance with FprEN 1992-1-1.

The available EAD for ETA-based DoPs is EAD 160129-00-0301: Couplers for mechanical splices
of reinforcing steel bars [20] (pending for citation in OJEU). The sizes of couplers range from 8 mm
to 50 mm. The load bearing parts of the couplers are completely made of steel or cast steel.
Types of couplers covered by the EAD are standard couplers, position couplers, bridging couplers
or transition couplers (see Fig. BD.C.3).

a) Standard coupler

b) Positional coupler

c) Transition coupler
Fig. BD.C.3. Types of couplers [20]

22.02.2023
For citations, copying and referencing see Note on Front page! page 558
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

Concerning product packaging, transport, storage, maintenance, replacement and repair it is the
responsibility of the manufacturer to undertake the appropriate measures and to advise his clients
on the transport, storage, maintenance, replacement and repair of the product as he considers
necessary [20]. It is assumed that all couplers will be installed according to the manufacturer’s
instructions. Relevant manufacturer’s stipulations having influence on the performance of the prod-
uct covered by the EAD [20] shall be considered for the determination of the performance and
detailed ETA or in the technical product information.
It is expected that couplers for ULS and SLS-design are produced under Assessment and Verifi-
cation of Constancy of Performance (AVCP) system 1+ addressed in EAD [20].
Couplers are intended to be used for mechanical splices of reinforcing steel bars for concrete struc-
tures designed according to FprEN 1992-1-1 and prEN 1998-1 for:
– Transfer of axial tension and/or compression forces of the connected,
– Limitation of slip,
– Resistance to high-cycle fatigue loading according to FprEN 1992-1-1,
– Resistance to low-cycle seismic loading according to prEN 1998-1.
The EAD [20] covers the following specifications of the intended use:
– Connection between reinforcing bars avoiding lapped splicing,
– Mechanical splices of reinforcing steel bars with a nominal yield strength of 400 MPa  Re,nom 
600 MPa and of ductility classes B or C according to FprEN 1992-1-1,
– Mechanical splices of reinforcing steel bars positioned such that the concrete cover complies
with FprEN 1992-1-1.
It shall be ensured that As,nom,bar ꞏ Re,nom,bar ≤ As,nom,coupler ꞏ Re,nom,coupler.
When the EAD is defined as the relevant standard the essential characteristics 1, 2, 3 (see Table
BD.C.4) shall be declared in a DoP addressed in an ETA. The characteristics 4, 5, 6, 7, 8 shall be
declared if relevant in design.

Table BD.C.4. Essential characteristics of couplers [20]

In the DoP, the percentage total elongation at maximum load Agt,act shall also be declared according
to requirements for reinforcing steel in Table BD.C.1.
If couplers are used in fatigue design the requirements of Tables C.1 and C.2 apply. The DoP may
declare other (often lower) values Rsk, k1, k2 for the S-N curve of a coupler. This has to be con-
sidered in an adapted fatigue design and accepted in the project specification. A National Annex
can provide further requirements (e.g. regarding fatigue strength) and advice to couplers.

22.02.2023
For citations, copying and referencing see Note on Front page! page 559
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

C.7 Headed bars The National Annex can specify relevant standards (or national approvals) for headed bars de-
signed in accordance with FprEN 1992-1-1.

The available EADs for ETA-based DoPs are

– EAD 160012-00-0301: Headed reinforcement steel bars [21] and
– EAD 160003-00-0301: Double headed studs for the increase of punching shear resistance of
flat slabs or footings and ground slabs [22].
Concerning product packaging, transport, storage, maintenance, replacement and repair it is the
responsibility of the manufacturer to undertake the appropriate measures and to advise his clients
[21], [22].

Headed bars:
The head of headed bars should have the same steel strength as the reinforcing steel bar. The
head-to-bar connection shall resist to the yield force Fyd of the bar diameter  (to declare in the
DoP).
The bearing surface of heads should be at >80 degrees to the bar unless other justified by testing.
The thickness of the head tH should not normally be smaller than ¼ of the maximum head width/di-
ameter h (see Fig. BD.C.4). The geometry of anchoring heads may be circular, quadratic or rec-
tangular.

h
th
t

reinforcing
steel bar with 

Fig. BD.C.4. Geometries of headed bars (from [21])

When an EAD is used as the relevant standard the essential characteristics 1, 2, 3, 4 (see Table
BD.C.5) shall be declared in a DoP addressed in an ETA. The characteristics 5 and 6 shall be
declared if relevant in design.

It is expected that headed bars for ULS and SLS-design are produced under Assessment and
Verification of Constancy of Performance (AVCP) system 1+ addressed in EAD [21].
Table BD.C.5. Essential characteristics of headed bars [21]

Double-headed studs:
The EAD [22] covers double headed studs made of ribbed or smooth shafts and a head at both
ends as punching shear reinforcement elements (see Fig. BD.C.5). The reinforcement elements
comprise at least two double headed studs with the same diameter and shape (ripped or smooth)
and the following specifications:
– double headed studs with shaft diameter dA of 10 mm to 25 mm
– double headed studs with a head diameter dk 3dA
– double headed studs with shafts made of weldable ribbed reinforcement bars or weldable struc-
tural steel with the following characteristics:
 yield strength: fyk  500 MPa
 ratio of tensile strength over yield strength: (ft/fy)k  1,05

22.02.2023
For citations, copying and referencing see Note on Front page! page 560
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

 elongation: uK  2,5 %

Fig. BD.C.5. Examples of connections of double headed studs (from [22])

When an EAD is used as the relevant standard the essential characteristics 1 and 2 (see Table
BD.C.6) shall be declared in a DoP addressed in an ETA. The required characteristic resistance to
fatigue loading ΔσRsk in testing for N ≥ 2 × 106 cycles based on a maximum stress level of 0,6 fyk
(characteristic 2) is:
– Rsk ≥ 80 MPa
For the load increasing factor for punching shear resistance (characteristic 1), FprEN 1992-1-1,
C.7 contains no material requirements. In design calculations, the load increasing factor is assumed
according to prEN 1992-1-1, 8.4.4 (5).

The reaction to fire (characteristic 3) shall be declared if relevant in design.

It is expected that double headed studs for ULS and SLS-design are produced under Assessment
and Verification of Constancy of Performance (AVCP) system 1+ addressed in the EAD [22].
Table BD.C.6. Essential characteristics of double-headed studs [22]

22.02.2023
For citations, copying and referencing see Note on Front page! page 561
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

C.8 Post-installed rein- FprEN 1992-1-1, clause C.8 provides the requirements for product properties to ensure the safe
forcing steel systems use of post-installed reinforcing steel systems in concrete structures.

C.8 (1) Post-installed steel reinforcing bar connections carried out with anchoring mortars are capable of
developing bond behaviour equivalent to cast-in-place steel reinforcing bars if they are properly
performed with an anchoring mortar sufficiently qualified to fulfil the requirements for safety, ser-
viceability, and durability of concrete structures in service and the product specific installation tools
[23]. Correct installation with these tools will ensure a durable bond in the interfaces between the
reinforcing steel bar and the anchoring mortar as well as the anchoring mortar and the borehole
wall.
Therefore, applicability of the installation tools, suitability and allowable use conditions for the an-
ticipated field of application of a system for post-installation of deformed reinforcing steel bars
should be verified by means of a European Technical Product Specification.
The prequalification process given e.g. in EAD 330087 [24] consists of appropriate prequalification
tests taking into account installation and environmental conditions on site and in service as well as
the range of reinforcing steel bar diameters and concrete strength classes at independent tests
laboratories considering the type of loading and an assessment of test results by an independent
organization. The result of the evaluation corresponding to the prequalification procedure shall be
documented in a European Technical Product Specification. This document should contain all in-
formation for the installation and design of post-installed reinforcing steel bars required by FprEN
1992-1-1.
Then, the design of post-installed reinforcing steel bars can be performed according to the provi-
sions given in FprEN 1992-1-1, clause 11.4.8.

C.8(2) The bond behaviour (bond stiffness and bond strength) of post-installed reinforcing steel bars shall
be at least equal with the bond behaviour of cast-in reinforcing steel bars under the same condi-
tions. Then the required minimum mean bond strength fbm,pi,rqd of Table C.7 for bond efficiency
class CPI-1,0 is achieved (bond efficiency factor kb,pi = 1,0) when tested and assessed in accord-
ance with EAD 330087 [24]. The minimum mean bond strength of post-installed reinforcing steel
bars is evaluated from the results of suitable pre-qualification tests taking into account the param-
eters given in C.7(3). The use of anchoring mortars referring to bond efficiency class CPI-1,0 is
recommended.
However, there are anchoring mortars for post-installed rebar applications with characteristics in-
ferior to cast-in-place reinforcing steel bars possible. Their assessment results in a bond behaviour
with bond stiffness and/or bond strength lower than that of cast-in-place reinforcing steel bars. In
this case the bond efficiency classes CPI-0,7 to CPI-0,9 apply, provided the assessment in accord-
ance with EAD 330087 indicates that the corresponding minimum mean bond strength of Table
C.7 for the relevant bond efficiency class is met. The lower bond efficiency class yields an increase
of the design anchorage length by a factor of up to 1/kb,pi = 1/0,7 = 1,43 which might result in
uneconomic design.
The factor kb,pi is product dependent and shall be published in the relevant European Technical
Product Specification. For post-installed reinforcing steel bars where kb,pi < 0,7 a design according
FprEN 1992-1-1 is not possible.

C.8(3) Clause C.8(3) lists all parameters influencing the performance of post-installed reinforcing steel
bars during installation and in service. They are taken into account in the derivation of the required
minimum mean bond strength fbm,rqd of Table C.7 for the determination of the bond efficiency class
in accordance with EAD 330087.

22.02.2023
For citations, copying and referencing see Note on Front page! page 562
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex C Background to FprEN 1992-1-1:2023

References [1] Construction Products Regulation (CPR). Regulation (EU) No. 305/2011 of the European
Parliament and of the Council of 9th March 2011 laying down harmonised conditions for the
marketing of construction products and repealing Council Directive 89/106/EEC.
[2] CEN-TC250-SC2_N2328 Guidance note on the listing and categorization of material and
product parameters required for Eurocode design (State 2019-10-14).
[3] Oscar Nieto et al: CPR Future – Option F  https://prezi.com/p/rwlxncpqg3rs/cpr-future-
option-f/
[4] European Commission: CPR Acquis  https://ec.europa.eu/growth/sectors/construc-
tion/product-regulation/acquis_en
[5] EN 13369:2018: Common rules for precast concrete products.
[6] EN 206:2021: Concrete: Specification, performance, production and conformity
(EN 206:2013+A2:2021).
[7] EN 12620:2008: Aggregates for concrete.
[8] EN 14889-1:2006: Fibres for concrete – Part 1: Steel fibres – Definitions, specifications and
conformity.
[9] EN 14651:2007: Test method for metallic fibre concrete – Measuring the flexural tensile
strength (limit or proportionality (LOP), residual).
[10] prEN 10080:2020 (new 2023-01): Steel for the reinforcement of concrete – Weldable rein-
forcing steel – General.
[11] CEN-TC250-SC2_N1438_Moersch-Letter_Reinforcing_Steel_EN_10080-2018-09-21
[12] EN ISO 15630-1:2019: Steel for the reinforcement and prestressing of concrete – Test meth-
ods – Part 1: Reinforcing bars, rods and wire.
[13] prEN 10370:2019: Steel for the reinforcement of concrete – Stainless steel.
[14] BS 6744:2016: Stainless steel bars. Reinforcement of concrete – Requirements and test
methods.
[15] EN 10088-1: Stainless steels – Part 1: List of stainless steels.
[16] prEN 10138-3:2020: Steel for the prestressing of concrete – Part 3: Bars
[17] EN ISO 15630-3:2019: Steel for the reinforcement and prestressing of concrete – Test meth-
ods – Part 3: Prestressing steels.
[18] EN 10027: Designation systems for steels.
[19] EAD 160004-00-0301: Post-tensioning systems for prestressing of structures.
[20] EAD 160129-00-0301: Couplers for mechanical splices of reinforcing steel bars.
[21] EAD 160012-00-0301: Headed reinforcement steel bars
[22] EAD 160003-00-0301: Double headed studs for the increase of punching shear resistance
of flat slabs or footings and ground slabs.
[23] Fuchs, W., Hofmann, J.: Post-installed reinforcing bars – Requirements for their reliable use.
Developments in the Built Environment.
DOI link: https://doi.org/10.1016/j.dibe.2020.100040, Elsevier, 2021.
[24] EAD 330087-00-0601: Systems for post-installed rebar connections with mortar.

List of versions of this document

16.06.2021 Draft 1
31.10.2022 Draft 2
10.02.2023 Draft 3  TJ: Peer review and update of references to FprEN 1992-1-1:2023
22.02.2023 Final draft

22.02.2023
For citations, copying and referencing see Note on Front page! page 563
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

CEN/TC250/SC2/WG1/TG7 FprEN 1992-1-1:2023

Background document to Annex D:

Evaluation of Early-Age and Long-Term Cracking
Due to Restraint

Terje Kanstad, Anja Klausen, NTNU (Norway)

Antonia Menga

20.03.2023

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 564
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Introduction
This document has been prepared to explain and validate the text of the new
Annex D and give reference to the most relevant papers and guidelines.

The work on Annex D has been done within CEN TC250/SC2/WG1/TG7: Time-
dependent effects and CEN TC250/SC2/PT1 (The project team for prEN 1992-1-
1).
The objective of the annex is to give normative guidance for cases where it is
agreed to use EC2 to restrict early age cracking. Together with the new “Annex H:
Guidance on design of concrete structures for water-tightness”, it is meant to
replace the content of the present “EN 1992-3. Eurocode 2 Design of concrete
structures. Part 3: Liquid retaining and containment structures.”

The background for the default values chosen in “Annex B: Time dependent
behaviour of materials: Creep, shrinkage and elastic strain of concrete and
relaxation of prestressing steel, Table B.2” is also included in this document since
these parameters are relevant for both Annex B and Annex D.

Background
D.1 General
Cracks can affect the aesthetics, durability, and tightness, whilst their control has
both economy and sustainability consequences. It is therefore important that
Eurocode 2 gives regulations that clearly describe how and when cracking shall be
accounted for in the structural design. Annex D considers the effects of
temperature changes due to the concrete’s hardening process and basic shrinkage
in combination with imposed deformations occurring at later ages.

The main objective of the methods described is to evaluate the cracking risk and
provide guidance on crack width calculations if cracking is expected to occur. The
focus is on through-cracks, which may span over the whole thickness of the
concrete member. This is typically between 2 and 30 days after casting depending
mainly on the dimension of the structural member and the binder type applied.

Possible measures to reduce the amount of through-cracking are to use concretes

with low heat production, concretes with low coefficient of thermal expansion,
cooling pipes in the hardening concrete, heating cables in the restraining concrete,
reduced fresh concrete temperature, or to reduce the degree of restraint for the
hardening concrete member. The measures may be expensive, and as the
significance of early age cracking is case-dependent, it is important that the
measures are reflecting the criticality of cracking.

The typical temperature and stress histories at early ages are illustrated in Figure
1, where Tci is the fresh concrete temperature, Tc,max the maximum concrete
temperature due to hydration heat, T0 the temperature of the restraining structure,
while ΔTmin is an additional temperature drop which might be due to daily or
seasonal temperature variations. The most unfavourable moment in time for early-
age cracking is denoted tcrit, corresponding to the moment when temperature
equilibrium with the restraining structure is achieved (within 2ºC).

Related to Figure 1a) the most decisive parameters are: The fresh concrete
temperature (Tci), ambient temperature history, insulation conditions, climatic
conditions as wind velocity and solar radiation, temperature of the restraining
structure (To), and finally the additional maximum temperature difference due to
daily and seasonal variations (ΔTmin)

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 565
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

a) b)

Figure 1. (a)Typical temperature history for a structural concrete member, and (b)
corresponding stress history.

It might be relevant to consider crack-control in the following three states: (1) At

temperature-equilibrium between the recently cast concrete structural member and the
restraining structure. (2) At commissioning of the structure. (3) During the design service
life. In the two latter cases, the temperature reduction ∆Tmin and drying shrinkage should
be added to the early age effects. In the annex the formulations are particularly directed
towards (1) and (3). If particular demands are related to tightness, durability or
appearance, the cumulative impact of early age effects, load effects and later imposed
deformations should be considered in the crack control. Otherwise, the crack control of
these states can be verified separately. To restrict cracking either crack prevention or
crack limitation may be used. The latter may be achieved either by minimum
reinforcement for crack limitation (FprEN 1992-1-1, Annex S), or by calculations (9.2.3).

D.2 Assessment of Assessment of the temperature history

temperature history
A reliable temperature calculation or estimate is an important prerequisite to achieve
accurate stress/strain or crack width calculation. Computer programs, hand calculations
and diagrams from handbooks and guidelines may be used. See for instance the UK
CIRIA 660-regulations [2], and the French extensive research report CEOS.fr [3] for
handbook solutions, and the Recommendations of RILEM TC 287-CCS: Thermo-chemo-
mechanical modelling of massive concrete structures towards cracking risk assessment
[4] for more general calculation background and finite element analysis guidance.

To provide some background an approach from the CIRIA-code [2] is referred.

Table 1 below gives relatively rough estimates for the cooling temperature
𝑇, 𝑇 ). The data is representing concretes made with ordinary Portland
cements, and multiplication factors for slow hardening cements containing slag or
flyash are also given below the table. The values are meant to be representative
for summer conditions in Western-central Europe, and reductions for winter
conditions are given in the bottom row.
To account for seasonal variation the following values may be used:
Casting under summer conditons: 𝛥𝑇 20 𝐶
Casting under winter conditons: 𝛥𝑇 10 𝐶

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 566
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Table 1. 𝑇 , 𝑇 )-values for different strength class and section thickness. A

fresh concrete temperature of 20oC and mean ambient temperature of 15oC is
assumed [2].

Correction factor for supplementary cement based materials (flyash or slag):

𝑘 1 𝛼 Default: Flyash: 𝛼 0,5 Slag: 𝛼 0,25

Material properties relevant for the temperature development

The total amount of hydration heat (Q∞ (kJ/kg binder) released during hydration,
and its time-function depend on binder fineness and composition and w/b-ratio.
The strength class influences the binder content, and therefore also the hydration
heat.
The hydration heat may be determined in accordance with the standards: EN
12390-14 «Testing hardened concrete - Part 14: Semi-adiabatic method for the
determination of heat released by concrete during its hardening process», or EN
12390-15 «Testing hardened concrete - Part 15: Adiabatic method for the
determination of heat released by concrete during its hardening process”.
Table values and diagrams are available from cement suppliers, in textbooks and
in guidelines [2,3]. The development versus equivalent time can for instance be
described by the commonly used three-parameter formula [6]:
α
Q(t )= Q exp(-(τ/t ) ) (1)
eq ∞ T
or with piecewise linear curves based directly on experimental data. The parameter
tT represents the equivalent time (maturity) and is in Formula B.18 in Annex B
denoted temperature-adjusted concrete age and defined as:

𝑡 ∑ ∆𝑡 ∙ 𝑒𝑥𝑝 13,65 (B.18)

∆

In the research literature the term “4000” is denoted the Arrhenius-constant, and
for accurate approaches it is often replaced by the activation energy (ET)
corresponding to the Arrhenius constant times the Gas constant (8,314 J/moloK))
expressed as [6]:

ET(Ti) =A+B(20-Ti) (2)

where Ti represents the temperature history, and A and B are material parameters.
The parameter B is 0 for Ti>20oC and therefore only relevant (& necessary) for low
fresh concrete temperatures (winter conditions). According to textbooks the
constants A and B depend on cement type, pozzolan content, and the w/b-ratio.
The typical variation range for A is 20000-40000 kJ/mol, while for B it is 1000-1500
kJ/mol.

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 567
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

For additional input parameters needed for the temperature calculations, see for
instance [5]. The values for the heat conductivity may vary within the range 1,2-3,0
∆J/(s moK). A recommended default value is 2.5 J/(s moK).The thermal diffusivity
corresponds to the conductivity divided by the heat capacity and the specific heat
(mm2/s). The surface convectivity may vary in the range from about 3.3J/(s m2oK)
for 18mm plywood formwork and no wind to about 15 J/(s m2oK) for a free concrete
surface with 5 m/s wind.The heat capacity is defined as: Specific heat ꞏ concrete
density. And the specific heat varies typically in the range 0,85-1,15 kJ/(kg oK).

D.3 Stress Stress calculations

calculations
General

Material properties relevant for the stress calculations

In Annex B the time-development of the compressive strength, for t≤91 days, is

described as:

f (t )=ß f (t ) ß= 𝑒𝑥𝑝 𝑠 1 (B.1/B.2)

cm T cm ref

while in Annex D, a more comprehensive expression is used for ß for cases where
stresses and strength at early ages are important:

ß= 𝑒𝑥𝑝 𝑠 1 (D.3)

The value for tref may be taken between 28 and 91 days, default values for the material
parameter sc are presented in Table B.2 (Annex B), while values for tdor are found in
Table D.1. Compared to present EC 2, the approach is reflecting three novel points:
1- Modification of the formulas adding the last square root-sign term including tref,,
makes it possible to express the strength development with reference strength at
tref later than 28 days without changing the sc-parameter.
2- Introduction of the parameter tdor (Figure 1, describing the “end of the dormant
phase”) defining when stiffness and strength start to be significant. This is a
decisive parameter to consistently include thermal dilation and basic shrinkage.
Both effects may develop fast at early ages, and may typically vary between 8
and 13 maturity hours (0,3-0,5 days Table D.1). The parameter may be strongly
affected by admixtures, and it is important that the choice of the parameter is
consistent with basic shrinkage and temperature development.
3- In Annex B, the three “Strength development classes for concrete” CS, CN and
CR are defined including amounts of the most common cement supplementary
materials. In addition, the parameters sc and tdor are made dependent on the
characteristic compressive cylinder strength (fck).
The default values for sc and tdor are given in Table 1 below, while the background is
explained in the following.

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 568
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Table 1: Parameters for strength development (includes content in Table B.2 and D.1)
CS CN CR
Strength sc tdor(d) sc tdor(d) sc tdor(d)
fck ≤ 35 MPa 0.6 0,5 0. 5 0,45 0.3 0,35
35 MPa <fck < 60 MPa 0.5 0,45 0.4 0,4 0.2 0,3
fck ≥ 60 MPa 0.4 0,4 0.3 0,35 0,1 0,3

Assessment of the sc-parameter

Two approaches were used to assess the default parameters, firstly a data base
including 105 mixes, from the references [7-11] plus unpublished data from NTNU was
established, and parameters defined by optimization procedures. Secondly formulas for
“sc” published by Vollpracht, Soutsos and Kanavaris [12], were used. In the end it was
decided to use only one significant number for the parameters, and the findings from the
two approaches were combined. For both approaches Formula B.2 was used (ie tdor=0).

The database consists of 105 test series with a majority of “modern” concretes. The
series are distributed among the strength and strength development classes as shown in
Table 2 below.

Table 2: Distribution of the 105 test series in the data base used to determine the
strength development parameter “sc”.
CS CN CR
fck ≤ 35 MPa 2 13 6
35 MPa <fck < 60 MPa 6 36 17
fck ≥ 60 MPa 6 9 10

Since the parameters refer to characteristic cylinder strength at 28 days, the following
formulas were used to transfer from the strengths reported:

𝑓, 0,80 ∙ 𝑓 , and 𝑓 𝑓 8 𝑀𝑃𝑎

In the paper by Vollpracht et al [12] 861 experimental series were included, many of
them containing slag or fly ash as supplementary SCM’s. The authors established
formulas for the sc-parameter dependent on the w/b-ratio and the amount of SCM (scm)
included:

𝑤 𝑠𝑐𝑚
𝑠 𝑐 𝑐
𝑏 𝑏

For the lower and higher strength classes the following values for the parameters c1 and
c2 were found:

Cements (with lower strengths) : 32.5 N, 32.5 R, 42.5 N: c1=0,528, c2=0,527

Cements (normal and higher strengths): 42.5 R, 52.5 N, 52.5 R: c1=0,481, c2=0,441

In the statistical evaluation the standard deviations are of the same size of order as for
the present data base previously described.

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 569
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

To utilize these formulas to verify, adjust, and increase the general validity of the sc-
values obtained from the data base the well-known Bolomey’s formula [9] was used to
establish a relation between the w/b-ratio and the mean cylinder strength:

𝑓 𝐾∙ 𝛼 with K=29 and α=0,35

/

The transfer from mean to characteristic strength was carried out as described
previously, and the parameters chosen for K and α were shown to give reasonable
agreement with a selection of the results in the data base. It was assumed that w/b=0,7
is representative for fck<35 MPa, w/b=0,35 for fck > 65 MPa, while w/b=0,45 represents
the strength interval in between. It was also assumed that the CS-group of concretes
contains 55% SCM, while the CN-group contains 35% SCM. Based on these
assumptions the sc-values presented in Table 3 below were found.

As seen in Table 3, the finally chosen default values (already reported in Table 1)
correspond approximately to the average values from the data base and the values
determined from Vollpracht et al’s results.

The statistical evaluation based on the data base, showed that for the optimized values
the overall standard deviation was 1,87 MPa. The statistical evaluation reported in [12]
show comparable experience as already mentioned. Combining the two approaches to
the final values gave a standard deviation for the data base on 2,46 MPa.

Table 3: Determination of the strength development parameter sc.

CS CN CR
Data Voll- Final Data Voll- Final Data Voll Final
base pracht value base pracht value base pracht value

fck ≤ 35 0,54 0,66 0,6 0.36 0,55 0,5 0,27 0.34 0,3
35<fck< 60 0.44 0,55 0,5 0.29 0,42 0,4 0,26 0.22 0,2
fck ≥ 60 0.41 0,47 0,4 0.29 0,37 0,3 0,25 0,17 0,1

Assessment of tdor

Accurate determination of this parameter requires compressive stress tests at very low
ages, special setting time tests or tests in special TSTM-equipment. The proposed
default values are based on test results and reflects that the parameter in general is
decreasing with increasing compressive strength, and that it is increasing with increasing
SCM-content, see for instance Kanstad et.al. [7], Klausen [8], and Bjøntegaard and
Kjellsen [13]. Some typical results are shown in Figure 2.

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 570
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Figure 2.(a) Effect of cement type on the parameter tdor and (b) typical experimental
results [13].

Validity of the formula to describe the 91-days strength

Figure 3a) shows the ratio fc(91)/fc(28) for concretes of the data base where the 91-days
strength was available. Although the scatter is large, a tendency with decreasing ratio
with increasing strength can be seen. The scatter is partly explained by a strong
influence of the cement type and pozzolanic content. It should also be noted that the
green cross markers are representing data from a recent project where the aim was to
develop environmentally friendly high strength concretes for the Norwegian “Ferry-free
E39-project” funded by the Norwegian Public Road Administration.

Figure 3b) shows the calculated ratio fc(91)/fc(28) versus the corresponding experimental
ratio using Formula B.2 and the parameters from Table B.2. It is seen that the scatter is
relatively large, but that still the tendency in the equation is reasonable.

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 571
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Figure 3.(a) Ratio fc(91)/fc(28) versus 28 days strength (fc(28)). (b) Calculated versus
experimental ratio fc(91)/fc(28).

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 572
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Coefficient of thermal expansion and basic shrinkage of concrete

The thermal dilation coefficient (αcth) varies with the concrete age, humidity
conditions and aggregate type within a wide range 7-14 E-6 as seen in Figure 4a)
[14]. The recommended default value is as in the present EC2 10E-6, which is a
reasonable choice as shown in Figure 4a).

The basic (autogenous) shrinkage, εcbs(t), may be calculated from Formula B24. It
should, however, be noted that the uncertainty is large, which yields both the final
value and the time function. See the relevant background documents for Annex B.
An important and relevant aspect for Annex D is that both previous and present
European standards for shrinkage measurements (EN 12390-16) prescribe that
the deformation measurements should start one day (24 h) after casting. There
are, however, concretes which develop a considerable basic shrinkage before 1
day. When relevant this should be considered in the calculations. Figure 4b)
shows a typical situation, and illustrates at the same time the need for coordinating
the thermal dilation and the basic shrinkage with the tdor-parameter.
Furthermore, the temperature influence on the basic shrinkage is not yet fully
understood, and is therefore not sufficiently considered in the model in Annex B.
Figure 4c shows the measured basic shrinkage as a function of the temperature
increase 14 days after casting for three concretes with fly ash addition.

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 573
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

a)

b)

c)

Figure 4. a) Experimentally determined coefficients of thermal expansion of

concrete (αcth) vs concrete age for concrete with different w/b-ratios [14]. b)
Illustration of the relevance of introducing the parameter tdor (to in the figure)) in
stress calculations at early ages [14]. c) Influence of the temperature increase due
to the hydration heat on basic shrinkage (autogenous deformation) at 336 hours
maturity (14 days) for concretes with w/b=0,4, and either 17, 33 or 45% flyash [15].

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 574
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Simplified stress calculation method

This section is mainly meant for verification of if cracking will occur or not. Obviously,
nonlinear time-dependent finite element analysis may be used for such stress
calculations, but this is not always convenient, and it is useful with simplified methods.
The present method is based on the effective E-modulus method, and application of the
“degree of restraint” as a key parameter. Such method can give the solution for a
specified position in a structural member at a specific time and may be utilized as part of
a more comprehensive solution regime, see for instance the three-step method
described in [14,16,17]. The simplified approach is based on the commonly used age
adjusted effective E-modulus method (Trost–Bazant method) and has previously been
proven to work well also at early ages [16,17]. The method is based on uncracked
concrete cross sections and applies an effective long-term E-modulus accounting for
aging and creep.

Using parameters described in Table 1 and Figure 1, the tensile stress in a predefined
position, including cumulative impacts, can be determined.

σ1(t)= (ΔεT(t-t2) + εcbs(t-t2))E1 ˑR1 + αcthˑ ΔTmin ˑE2 ˑR2 + εcds(t-t2)ˑE3 ˑR3
(2)
Where t2 is the time when the stress changes sign. For early age problems the time t
should be set equal to tcrit as defined in Figure 1, while if cumulative impacts are included
t is the relevant design service life. ΔεT(t-t2)=kTemp(Tc,max-To)αcth represents the part of the
temperature causing tensile stresses and εcbs(t-t2) is the corresponding part of the basic
shrinkage occurring in the cooling phase and εcds is the drying shrinkage strain. E1
=Ec(t2eq)/(1+k2*φ(t,t2eq)) is the effective E-modulus representing the cooling phase.
Default values based on amongst others Scandinavian practice [6,7] are k1=0,9 and
k2=0,8. The parameters E2 and E3 are corresponding effective E-modules representative
for the additional effects. Three different values for the degree of restraint are used in
formula (2) because this parameter will vary with the concrete age and the time period
considered, and because the structural system may change with time.
The present draft of Annex D [1] includes formulas for the tensile stresses at two
particular stages, expressed by the parameters in Figure 1.
At temperature equilibrium between restrained and restraining member ( 𝑡 ):

𝜎 𝑡 𝑅 , 𝑘 𝛼 𝑇, 𝑇 𝜀 𝑡 𝜀 𝑡 (D.4)

And at end of design service life (long-term):

𝜎 𝑅 , 𝑘 𝛼 𝑇, 𝑇 𝜀 𝑡 𝜀 𝑡
,

𝑅 , 𝐸 𝛼 Δ𝑇 𝑅 , 𝜀 𝑡 𝜀 𝑡 (D.5)
,

t2 is the time when the stress changes sign from compression to tension. For the early
age problem at tcrit, kTemp αcth (Tc,max-To) represents the part of the temperature causing
tensile stresses while [εcbs(tcrit)- εcbs(t2)] is the part of the basic shrinkage occurring in the
cooling phase. Default values based partly on Scandinavian practice [16,17], and verified
below are kTemp=0,9 and χφ 0,55.
For the long-term formula, the temperature term is the same, while the effective E-
modules and the shrinkage terms are adjusted to the relevant time-intervals. The
additional temperature change Δ𝑇 corresponding to the daily temperature variation is
added as a short-time effect. The value X=0,8 is in agreement with the general
application of the Trost-Bazant method. Ec,28 is the tangent modulus of elasticity of

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 575
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

concrete at an age of 28 days.Three different values for the degree of restraint (𝑅 ,

𝑅 , 𝑎𝑛𝑑 𝑅 , ) are used in the latter formulae. The explanation is that the structural
system may change significantly over time. Relevant background for the degree of
restraint are given in many text-books and guidelines, see for instance [2,3]. It may also
easily be determined by linear finite element analysis as described for instance in
[14,16].

Test results used for verification of formula (D.4)

A typical test regime used for the verification presented in the following, consist of
hydration heat development, activation energy for maturity calculation, development of
E-modulus versus time, Temperature-Stress-Testing-Machine (TSTM), and free dilation
test results (dummy specimen for the TSTM). For more thorough explanations, see
[8,15,18]. The realistic temperature histories applied are calculated for a 700 mm thick
wall, with both fresh concrete temperature and equilibrium temperature equal to 20oC.
The predefined degree of restraint is either 0,5 or 0,7. Figure 5 shows typical E-modulus
development, while Figure 6 includes three different sets of so far unpublished TSTM-
test results.

To generalize the measured stress development, a calculated stress development is

determined by curve-fitting the creep model parameters and applying linear
viscoelasticity for aging materials [8,18]. In the verification of Formula (D.4) this
calculated curve replaces the experiment. The parameter tcrit is chosen as the time when
the temperature in the hardening concrete has reached the final temperature To and the
tensile stress therefore has stabilized. The parameter tcrit is highly problem dependent,
but in this verification-task it is set to 9,25 days for all cases since the same wall is
considered, and the various concretes are similar.

Figure 5. Typical E-modulus development (unpublished results from NTNU).

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 576
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

a)

b)

c)

Figure 6. Typical TSTM test results including the realistic temperature curve applied, the
tensile strength development, and the measured and calculated temperature history.
(Unpublished results from NTNU)

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 577
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

In the practical applications of the method, available input may vary, and therefore the
following three approaches are chosen in the verification of the method:

1) -t2, tcrit,Tc,max, To , and E-modulus from the experiment. αcth =10-5 (10E-6) and
basic shrinkage from Formula B.26 and B.27.
2) As 1), but with t2=2 days as proposed default value in Annex D
3) As 1), but with experimental values for the free dilation (sum of basic
shrinkage and thermal dilation)

The results are presented in Table 5, 6 and 7. The number in the parenthesis in the last
column represents the ratio between calculated and experimental values. In some cases
it is expected that the first and the second approach may give large scatter because the
basic shrinkage and the thermal dilation coefficient deviate considerably from the default
values. In particular this is the case when a relatively large basic shrinkage occur before
1 day.

In Figure 7, 8 and 9 the calculated stress versus the corresponding experimental value is
presented. And it is seen that the agreement between the calculated and experimental
values is good for all the considered cases.

Table 5. Verification results,1st approach: t2, tcrit,Tc,max, To , and E-modulus from the
experiment. αcth =10-5 and basic shrinkage from Formula B.26 and B.27.
Test description Ec(t2e) Tmax- εcbs(tcrit)- σ(tcrit) calc σ(tcrit) exp
(MPa) To(oK) εcbs(t2) (MPa) (MPa)
(10-6)
w/b=0,38, Cem I SR, 34176 41,8 16 4,32 4,00(1,08)
0% SCM, R=0,5, t2=3d
tcrit=9,25d, To= 20oC
w/b=0,38, Cem II, incl 32633 36,2 17 3,61 3,75(0,96)
FA, R=0,5, t2=3d
tcrit=9,25d, To= 20oC
w/b=0,38, Cem I, R=0,5, 32835 39,5 18 3,95 3,50(1,13)
t2=3d tcrit=9,25d,
To=20oC
w/b=0,48, Cem I SR, 30844 35,2 13 3,28 2,80(1,17)
R=0,5, t2=3,2d
tcrit=9,25d, To= 20oC
w/b=0,48, Cem II incl 28902 28,8 10 2,51 2,55(0,99)
FA, R=0,5, t2=3d
tcrit=9,25d, To= 20oC

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 578
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Table 6. Verification results, 2nd approach: t2=2 days, tcrit,Tc,max, To , and E-modulus from
the experiment. αcth =10-5 and basic shrinkage from Formula B.26 and B.27.
Test description Ec(t2e) Tmax- εcbs(tcrit)- σ(tcrit) calc σ(tcrit) exp
(MPa) To(oK) εcbs(t2) (MPa) (MPa)
(10-6)
w/b=0,38, Cem I SR, 33253 41,8 21 4,26 4,00(1,06)
0% SCM, R=0,5, t2=3d
tcrit=9,25d, To= 20oC
w/b=0,38, Cem II, incl 31577 36,2 22 3,54 3,75(0,94)
FA, R=0,5, t2=3d
tcrit=9,25d, To= 20oC
w/b=0,38, Cem I, R=0,5, 32077 39,5 23 3,91 3,50(1,12)
t2=3d tcrit=9,25d,
To=20oC
w/b=0,48, Cem I SR, 29510 35,2 18 3,19 2,80(1,14)
R=0,5, t2=3,2d
tcrit=9,25d, To= 20oC
w/b=0,48, Cem II incl 27519 28,8 13 2,42 2,55(0,95)
FA, R=0,5, t2=3d
tcrit=9,25d, To= 20oC

Table 7. Verification results,3rd approach: t2, tcrit,Tc,max, To , and E-modulus from the
experiment. Thermal dilation and basic shrinkage also from the experiment.
Test description Ec(t2e) Tmax- εcbs(tcrit)- σ(tcrit) calc σ(tcrit) exp
(MPa) To(oK) εcbs(t2) (MPa) (MPa)
(10-6)
w/b=0,38, Cem I SR, 34176 41,8 -1,6 4,05 4,00(1,01)
0% SCM, R=0,5, t2=3d
tcrit=9,25d, To= 20oC,
αcth=9,8ꞏ10-6
w/b=0,38, Cem II, incl 32633 36,2 7,1 3,37 3,75(0,90)
FA, R=0,5, t2=3d
tcrit=9,25d, To= 20oC,
αcth=9,6ꞏ10-6
w/b=0,38, Cem I, R=0,5, 32835 39,5 6,0 3,68 3,50(1,05)
t2=3d tcrit=9,25d,
To=20oC, αT=9,6ꞏ10-6
w/b=0,48, Cem I SR, 30844 35,2 -14 2,76 2,80(0,99)
R=0,5, t2=3,2d
tcrit=9,25d, To= 20oC,
αcth=9,2ꞏ10-6
w/b=0,48, Cem II incl 28902 28,8 -11 2,21 2,55(0,87)
FA, R=0,5, t2=3d
tcrit=9,25d, To= 20oC,
αcth=9,6ꞏ10-6

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 579
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Figure 7. Calculated versus experimental results applying formula (D.4), 1st approach.
Statistics: μθ=1,07 , CoV=0,007.

Figure 8. Calculated versus experimental results applying formula (D.4), 2nd approach.
Statistics: μθ=1,04, CoV=0,007

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 580
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

Figure 9. Calculated versus experimental results applying formula (D.4), 3rd approach.
Statistics: μθ=0,96, CoV=0,005

D.4 Crack width

calculations Crack width calculation
While Annex D describes the crack strain due to imposed strains, the crack spacing
to be used is found in the main SLS-chapter as for ordinary loads (9.2.3 Refined
control of cracking):

𝑤 , 𝑘𝑤 ∙ 𝑘1/𝑟 ∙ 𝑆 , ∙ 𝜀 𝜀 (9.8)

wk,cal is the characteristic surface crack width, kw the conversion factor from mean to
characteristic values, Srm,cal the calculated mean crack spacing, while εsm and εcm are
the mean strains in reinforcement and concrete, respectively, and their
difference 𝜀 𝜀 represents the crack strain. The parameter k1/r allows for out
of plane curvature, for uniaxial tension as typically encountered with thermal and
shrinkage cracking this factor may be taken as 1,0.

For elements subjected to imposed strains which are restrained at the edges
(typically walls on foundation slabs), the crack strain (εsm - εcm) in Formula (9.8) may
be replaced by:

𝑓 ,
𝜀 𝜀 𝑅𝑎𝑥 𝜀 𝑘 ∙ (9.13)
𝐸

where kt is a coefficient dependent on the nature and duration of the load, fct,eff the
mean value of the tensile strength of the concrete at the time when the cracks are
expected to occur and Ecm the modulus of elasticity. Raxεfree in Formula (9.13) can be
taken as:

For early age cracking:

𝑅 𝜀 𝑅 , 𝑘 𝛼 𝑇 𝑇 𝜀 𝑡 𝜀 𝑡 (D.6)

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 581
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

For long-term cracking:

𝑅 𝜀 𝑅 , 𝑘 𝛼 𝑇 𝑇 𝜀 𝑡 𝜀 𝑡 𝑅 , 𝛼 ∆𝑇
𝑅 , 𝜀 𝑡 𝜀 𝑡 (D.7)

The parameters are described above in connection with Formulas (D.4) and (D.5).
Elements restrained at the ends can according to 9.2.3(3) be considered in the crack
formation stage and the crack strain in Formula (9.8) may be replaced by Formulae
(9.11):

(9.11)

Relevant background for Formula (D.6) and (D.7) is the UK CIRIA 660 regulations [2]
and the French report CEOS.fr [3] which both are using similar approaches.
An interesting comparison is done for the typical wall on foundation slab system,
Figure 8 [19], ie an investigation of a section one wall-thickness above the
foundation. It is seen that the agreement between the various approaches
investigated is reasonably good. It should be noted that the work was finished in
2018, and therefore based on the literature available at that time. An interesting
publication for verification of the crack width calculation combining 9.2.3 and Annex
D is the comprehensive investigation of wall-on-slab-structures presented in [20].
Furthermore, is the topic highly addressed in the recent RILEM committee TC 287-
CCS Early age and long-term crack width analysis in RC Structures.

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 582
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

a)

b)

Figure 10. Comparison of various crack width calculation methods [19].

References References
[1] FprEN 1992-1-1 2023: Eurocode 2: Design of concrete structures – Part 1-1:
General rules, rules for buildings, bridges and civil engineering structures,
Annex D.
[2] CIRIA 660 (P.B. Bamforth et al) Early-age thermal crack control in concrete.
London 2007 (& draft of 2015-version).
[3] CEOS.fr Cracking and shrinkage. Behaviour and evaluation of special
structures. Recommendations for the control of cracking phenomena
(January 2015).
[4] Azenha, M. et al., Recommendations of RILEM TC 287-CCS: Thermo-chemo-
mechanical modelling of massive concrete structures towards cracking risk
assessment. Materials and Structures (RILEM), Accepted for publication per
April 2021.
[5] Laube, M., Werkstoffmodell zur berechnung von temperaturspannungen in
massigen betonbauteilen im junge alter, Technical University of
Braunschweig, Doctoral thesis, Doctoral thesis, 1990. (In German)
[6] Freiesleben Hansen, P., Hardening technology 2, Decrement method, BKF-
centralen 1978 (In danish)
[7] Kanstad, T.; Hammer, T.A.; Bjøntegaard, Ø.; Sellevold, E. J. Mechanical
Properties of Young Concrete: Part I - Experimental Results related to Test
Methods and Temperature Effects. Materials and Structures 2003 ;Volume
36. p. 218-225

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 583
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex D Background to FprEN 1992-1-1: 2023

[8] Klausen, A.; Early age crack assessment of concrete structures:

experimental investigation of decisive parameters. Trondheim: NTNU,2016
(ISBN 978-82-326-1850-7) PhD-thesis.
[9] Neville, A.M.; Properties of concrete, Pitman, London, 1975.
[10] Graf, O., et al; Die eigenschaften des betons. Springer Verlag, Berlin, 1960.
[11] Ji, Guomin; Cracking risk of concrete structures in the hardening phase:
experiments, material modelling and finite element analysis. Trondheim:
NTNU, 2008 (ISBN 978-82-471-1080-5) PhD-thesis.
[12] Vollpracht, A., Soutsos, M., and Kanavaris, F., Strength development of
GGBS and fly ash concretes and applicability of fib Model Code’s maturity
function – A critical review. Construction and building materials 162 (218) pp
830-846.
[13] Bjøntegaard, Ø., Kjellsen, K., Property development and cracking tendency
in hardening concrete: Effect of cement type and fly ash content. COIN
Project report 40, 2012. https://www.sintef.no/en/projects/coin. ISBN 978-82-
536-1279-9.
[14] Bjøntegaard, Ø., Basis for and practical approaches to stress calculations
and crack risk estimation in hardening concrete structures – State of the art.
COIN project report 31. 2011. ISBN 978-82-536-1236-2 (pdf).
https://www.sintef.no/en/projects/2007/coin/coinp/
[15] Klausen, Anja Birgitta Estensen; Kanstad, Terje; Bjøntegaard, Øyvind;
Sellevold, Erik Johan. The effect of curing temperature on autogenous
deformation of fly ash concretes. Cement & Concrete Composites 2020
;Volume 109. p. -

[16] Kanstad, T. Verification of three different calculation methods for early age
concrete. In: Crack Risk Assessment of Hardening Concrete Structures. The
Nordic Concrete Federation 2006 ISBN 82-91341-97-4. p. 101-110
[17] Larson, M.; Estimation of crack risk in early age concrete, simplified methods
for practical use, Licentiate thesis, Luleå technical university, Sweden
2000:10, ISSN 1402-1757.
[18] Klausen, Anja Birgitta Estensen; Kanstad, Terje; Bjøntegaard, Øyvind.
Hardening Concrete Exposed to Realistic Curing Temperature Regimes and
Restraint Conditions: Advanced Testing and Design Methodology. Advances
in Materials Science and Engineering 2019 ;Volume 2019. p. –
[19] Klausen, Anja; Early age crack assessment: codes, guidelines and calculation
methods. DaCS Project memo. SINTEF Building and infrastructure.
Trondgheim, Norway 2018.
[20] Jedrzejewska, A., Kanavaris, F., Zych, M., Schlicke, D., Azenha,M.;
Experiences on early age cracking of wall-on-slab concrete structures. In
Structures, Volume 27, October 2020, DOI: 10.1016/j.istruc.2020.06.013

Kanstad, Klausen, Menga

For citations, copying and referencing see Note on Front page! page 584
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG6 FprEN 1992-1-1:2023

Background document to Annex F

Safety formats for non-linear analysis

Diego Gino Politecnico di Torino, Turin, Italy

Paolo Castaldo
Giuseppe Mancini

28.11.2022

For citations, copying and referencing see Note on Front page! page 585
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared in order to provide background information related to the FprEN 1992-1-1 Annex F
- “Safety formats for non-linear analysis”. In the last decades, the non-linear analyses (NLAs) have increasingly
become the most common instrument able to simulate the mechanical response of structural systems, such as
reinforced concrete (RC) members, from serviceability to the ultimate limit state (ULS). Concerning the results of such
complex calculations, they need to be properly processed in order to satisfy the target reliability levels required by this
standard (i.e., EN 1990 [1]; FprEN 1992-1-1 [2] – Annex A). To this aim, several methodologies (i.e., safety formats)
have been introduced by scientific literature [3]-[7] and codes as [8]-[9]. All the different safety formats provide a
framework devoted to assess the design value of the structural resistance Rd including the relevant sources of
uncertainties: the aleatory and the epistemic ones [10]-[14].

FprEN 1992-1-1 - The FprEN 1992-1-1 - Annex F - “Safety formats for non-linear analysis” aims to provide a standard to perform
Annex F validation (with reference to general and project specific attempts) and definition (e.g., characterization of the loading
process, selection of the solution strategy and of the modelling hypotheses) of non-linear numerical (NL) models of RC
F.1 Use of this annex structural systems/members and to establish the methods (i.e., safety formats) useful to derive design values of
F.2 Scope and field of structural resistance Rd for ULS verification according to EN 1990 [1].
application

Pros and cons of use non-linear analysis

In general, the use of non-linear analysis has advantages and disadvantages as listed below:

Advantages:

- Analysis is performed with properties and material behaviour closer to reality, therefore the distribution of
internal forces and deformations is more realistic;

- It enables an assessment of structures and structural elements that cannot be easily calculated with the
standard analytical formulas as these are typically valid only for beams, columns or slabs with simple geometrical
shapes without openings;

- It can provide a useful insight in the real structural behaviour and the actual failure mode;

- In some cases when deeper knowledge of the structure and its properties is available, and very well validated
model is used, it can provide higher load carrying capacity then standard methods. This is especially useful for
the assessment of existing structures.

Disadvantages:

- Nonlinear analysis is more time consuming in model preparation, analysis time as well as post-processing;

- Principle of superimposition does not apply to actions and their effects, so the analysis must be performed for each
combination of the actions separately;

- Requires higher knowledge and expertise.

Uncertainties in non-linear analysis of RC systems/members

In structural engineering, the design process is subjected to different sources of uncertainty that, conveniently, can be
classified in two main families: aleatory and epistemic [10]-[14]. The aleatory uncertainties concern the inherent
randomness of the variables that governs a specific structural problem, whereas the epistemic uncertainties are
mainly related to the lack of knowledge in the definition of the structural model [10]-[17] and represented also by
auxiliary non-physical variables/choices (e.g., assumptions for empirical coefficients and/or selection of specific
material models) [12]. The safety assessment of a structural system by means of NLAs should account for explicitly
both these sources of uncertainty.
Concerning the case of RC structures, the aleatory uncertainties relate to the inherent randomness of both materials
(e.g., concrete compressive strength and reinforcement yielding strength) and geometric (e.g., members size,
concrete cover and reinforcement location) properties. This kind of uncertainties may be described by appropriate
probabilistic models [8], [18] that, with particular reference to the case of existing structures, may be updated on the
base of information from tests and inspections [2], [19].
With reference to the epistemic uncertainties, these are mainly related to missing knowledge, simplification and
assumptions performed within the definition of the resistance model itself [10]-[14].The numerical models used for
non-linear analysis, as well as the predicted response, are merely representations, estimations and approximations of
the actual structural behavior of an RC structural member/system. In fact, the prediction of the actual structural
response through NLAs is denoted by a certain level of uncertainty because any model that aims to describe the
actual structural response of a member or system neglect, inevitably, some more or less important aspects. In
general, the definition of a structural model grounds on the basic principles of mechanics as equilibrium of forces,
kinematic compatibility and constitutive laws that, all together, denotes a specific modelling hypothesis [10]-[17].
Specifically, focusing on NLAs, the mentioned above basic principles are met by means iterative solution methods
(e.g. Newton-Raphson algorithm, Arch-length algorithm) that leads, inevitably, to an approximation of the exact
solution for the structural response [20]. Furthermore, the different assumptions related to the equilibrium, kinematic

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 586
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

compatibility and constitutive laws always lead to diverse solutions for the specific structural problem. Then, the
multiplicity of choices that may be performed defining NLN model leads to a further degree of uncertainty having
epistemic nature [10]-[17]. This source of uncertainty can be efficiently represented by the model uncertainty random
variable ϑ, which can be probabilistically characterized comparing experimental and related numerical results. This
probabilistic calibration should be carried out having care to reduce as much as possible the influence of other
epistemic sources of uncertainty as the statistical (i.e., limited sample of observation) and experimental (i.e., deviation
of tests set configuration and/or of measurement devices) ones, that inevitably may affect the probabilistic calibration
of the model uncertainty random variable ϑ.
Next, the background to safety formats implemented in FprEN 1992-1-1 Annex F aimed to account for the different
sources of uncertainty in ULS verifications of RC structures/members is reported.

Verification of ultimate limit states and safety formats for NLAs

In general, refined mechanical and geometrically NLAs are able to account for redistribution of stress fields, of internal
forces and local or global instability issues during the loading process. According to EN 1990 [1], FprEN 1992-1-1 [2]
the verification of ULS is performed by means of comparison of the design value of the effect of the actions (Ed) (e.g.,
bending moment, shear and axial forces) and the related local resistance (Rd) according to the inequality Ed≤Rd
(Section 8.3 - EN 1990 [1]).
In this framework, the effect of external actions Ed are evaluated by means of linear elastic analysis performed
properly combining the actions, while, the associated resistance Rd is evaluated according to the specific resistance
models FprEN 1992-1-1 [2]. This approach to the assessment of the structural safety is defined as local, as it involves
cross-sectional/local verifications of the structural members disregarding from the global structural response.

Figure BG-F.1: Local and global approaches for safety verification of RC structures (Castaldo et al. 2019 [6]).

For instance, in case the ULS verification is performed using refined NLAs, the global capability of the reinforced
F.3 General concrete structures/members to redistribute stress fields and internal forces can not be neglected. In this context, the
use of non-linear numerical models requires the adoption of a global approach for evaluation of the structural safety,
comparing directly the actions combined at ULS according to the selected load case Fd and the associated design
value of structural resistance Rd. This approach is denoted as global resistance format (GRF) in line to [8]. The main
differences between the two approaches are explained in Figure BG-F.1.
In line with the global resistance format [8] and EN 1990 [1], the ULS verification of an RC structural members/system
can be exploited as follow:
Clause (5) Fd  R d (BG-F.1)

where the design value of the actions Fd can be evaluated in line with the EN 1990 [1] according to the proper
combination, while, the design value of structural resistance Rd can be evaluated through NLAs with the application of
safety formats (which will be introduced next in this document) according to the following general expression:

R  X rep ; a rep 
Rd  (BG-F.2)
 R*  R d

In the formula (BG-F.2), the term R{∙} denotes the structural resistance estimated by means of NLA using the
representative values Xrep for material and arep for geometric properties according to the adopted safety format, as
described next. The level of structural reliability adopted for ULS verification is involved by means of two different
partial safety factors:
- the global resistance factor γ*R, which is a partial safety factor that takes into account, at the level of global
structural response, of the influence of the aleatory uncertainties related to the material properties and
geometry. This partial safety factor can be determined in compliance with a specific target level of reliability
according to the methodology described by the selected safety format;

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 587
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

- the model uncertainty factor γRd, which is a partial safety factor that takes into account the uncertainty (of
epistemic nature) related to simplifications, assumptions and choices performed to define of the non-linear
numerical model. This partial safety factor is horizontal and independent from the adopted safety format.

The software used in non-linear verification of the ultimate limit state should be validated by comparison between
Clause (3) numerical and experimental or benchmark results. Similarly, the choices made with respect to the specific numerical
model should be tested by sensitivity analysis.
1) general validation attempts:
- basic material tests;
- structural tests with the characterization of relevant failure modes;
2) project specific attempts:
- mesh sensitivity tests;
- solution method tests.
In general, these issues may be solved partially by software provider and partially by the analyst which is directly
involved in the specific project.
The structural resistance R{∙} that corresponds to the achievement of the ultimate limit state (i.e., structural failure)
shall be identified by the analyst in concomitance of actions level where lack of convergence of numerical procedure
occurs. The analysts should interpret the results paying attention to distinguish and recognize potential numerical
failure and/or physical failure and to characterize the relevant failure mode (see F.5 FprEN 1992-1-1 [2] – Annex F).

The loading process implemented by a non-linear simulation in order to evaluate the structural resistance R{∙} against
the actions shall be defined according to engineering judgement consistently with the considered combination at
Clause (4) ultimate limit state. In particular, the outcomes of the NLA are significantly affected by the loading procedure that, with
particular care to staged construction methods, should be identified accounting for the actual loading sequence and
even including time dependent effects if relevant. In particular, the structural resistance R{∙} may be evaluated
increasing the actions from their initial values (e.g., characteristic values) by incremental steps, such that the
associated design values in the adopted combination are reached in the same step. The incremental process of the
actions should be continued beyond their design values with the same law until structural failure is reached. The
structural resistance R{∙} corresponds to the values of the actions which lead to structural failure. With the adoption of
this approach, the estimated design value of of structural resistance Rd can be directly compared to design value of the
axions Fd with reference to the specific combination at ULS.

The safety formats based on the above described approach, which are introduced by FprEN 1992-1-1 Annex F, are
Clause (6) the following:
i. Partial factor method (PFM);
ii. Global factor method (GFM);
iii. Full probabilistic method (FPM).
In particular cases, where higher degree of refinement for estimation of structural reliability for ULS verification is
required, the (iii.) Full Probabilistic Method (FPM) can be adopted in compliance to EN 1990 [1] and FprEN 1992-1-1
[2] – Annex A .
In the following, the principal features of the safety formats (i., ii., iii.) for NLAs of RC structural systems/members
reported in FprEN 1992-1-1 Annex F are described.

The Partial Factor Method (PFM) has been introduced by [8]. The design value of structural resistance Rd is obtained
F.4 Partial factor method by means of one NLA according to formula (BG-F.3):
(PFM)
R  X d ; ad  (BG-F.3)
Rd 
Clause (1)  Rd

where R{Xd; ad) represents the structural resistance estimated by means of a non-linear numerical analysis using the
design values of the material (Xd) and geometrical (ad) properties; γRd is the partial safety factor which accounts for
model uncertainty related to global NLAs.
With the PFM, the influence of the aleatory uncertainties is accounted for by Xd according to EN 1990 [1]. The partial
safety factors to be used for the evaluation of Xd should consider the statistical variability of the material property and
the geometric uncertainty but should not be inclusive of model uncertainty (which is accounted for trough the model
uncertainty factor γRd related to global resistance estimated by NLA). For instance, the method of EN 1990 [1] should
be implemented according to specifications of Table A.1 line (e) of FprEN 1992-1-1 [2] - Annex A and FprEN 1992-1-1
[2] - Annex A in general. The appropriate target reliability index βtgt can be adopted in the evaluation of the mentioned
above partial safety factors always in compliance to EN 1990 [1] and FprEN 1992-1-1 [2] - Annex A. In case the
geometric uncertainty is not accounted for within the evaluation of the partial safety factors associated to material

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 588
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

properties (as previously mentioned), the design value of geometrical properties (ad) should be used according to EN
1990 [1]. Alternatively, the NLN model should be defined using the nominal values for geometric properties (anom)
(prEN 1990:2020 [1]).

In case of adoption of simplified approach for non-linear analysis according to following assumptions:
Clause (2)
- structural resistance is determined based on numerical model established in a way that the influence of the
concrete tensile strength is neglected;
- and the model follows the assumptions and verification procedures of Section 8 and/or Annex G of FprEN
1992-1-1 [2];
the model uncertainty can be assumed to be equal to the one included in calculation of fcd and fyd according to Section
5 of FprEN 1992-1-1 [2]. In this circumstance, formula (BG-F.3) may be replaced by:

R d  R f cd , f yd ; a nom 

where fcd and fyd are derived from Section 5 of FprEN 1992-1-1 [2] and anom is the nominal value of the geometric
property according to EN 1990 [1].

The Global Factor Method (GFM) is based on the evaluation of the design value of structural resistance Rd by means
F.5 Global factor method of global partial safety factors (i.e., γ*R and γRd) applied to mean value of structural resistance Rm{X; a} which is
(GFM)
estimated throughout NLAs. In particular, the principles of this method descends from global resistance safety formats
F.5.1 General as the “Method of Estimation of Coefficient of Variation” (ECoV) and “Global Resistance Factor Method” (GRF)
Clause (1) reported by [8]-[9], which has been furtherly investigated by [6]-[7], [21].
The Global Factor Method (GFM) allow to estimate the design value Rd assuming that the structural resistance R is
lognormally distributed. The statistical parameters (i.e., mean value Rm and coefficient of variation V*R) should be
evaluated trough probabilistic analysis including materials (X) and geometrical (a) uncertainties using a probabilistic
model for basic variables compliant with EN 1990 [1] and FprEN 1992-1-1 [2] - Annex A. Although the assumption of
lognormal probabilistic distribution turns out to be a simplification, it is widely adopted to describe resistance random
variables and its efficiency is demonstrated by several applications [5]-[7]. In case the lognormal probabilistic model is
not able to represent the structural resistance R the Full Probabilistic Method (FPM) can be adopted to perform
reliability-based verification of the specific ULS.
The Global Factor Method (GFM) can be used to determine the design value of structural resistance Rd according to
the following expression in line to assumption of lognormal distribution:

R m  X ; a (BG-F.4)
Rd 
 R*  R d

The global resistance factor γ*R can be evaluated as follows:

F.3.2 Determination of  R*  e xp  R   tgt  V R*  (BG-F.5)
the global resistance
factor In formula (BG-F.5) the target reliability index βtgt can be adopted in compliance to prEN 1990:2020 [1] and FprEN
Clause (1)
1992-1-1 [2] - Annex A. Note that the same target level of reliability should be adopted to determine the model
uncertainty factor γRd according to FprEN 1992-1-1 [2] - Annex F and as described next in this document. The first-
order reliability method (FORM) factor αR should be set equal to 0.8 in line to the assumption of dominant resistance
variable within the reliability analysis (for more explanation, see the section related to assessment of model
uncertainty for NLAs in this document).
As the Global Factor Method (GFM), in general, requires the execution of probabilistic analysis to define the mean
value Rm and coefficient of variation V*R, the following simplifications can be introduced to enhance the applicability of
the method to analyst and practitioners:
(a) the mean value of structural resistance Rm{X; a} can be assumed equal to the structural resistance
R{Xm; anom} estimated by NLA adopting mean values of material properties (Xm) and nominal values for
geometric ones anom. This simplification correspond to perform the truncation to the first term of the Taylor
series expansion of the structural resistance R around the mean or nominal values of basic variables [7];
(b) the coefficient of variation V*R can be expressed accounting for separately the contribution of the uncertainty
associated to materials VR,M and geometric VR,G properties as follow:

V R*  V R2, M  V R2,G (BG-F.5)

The coefficient of variation VR,M reflect the influence of the statistical variability of material properties on
variability of structural resistance. According to the ECoV method [3],[8] and in line to the assumption of
lognormal probabilistic model for structural resistance R, the coefficient of variation can be estimated with the
following simplified approach:

1  R  X m ; a nom   (BG-F.6)
V R ,M  ln  
1 .6 5  R  X k ; a nom  

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 589
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

where R{Xk; anom} is the structural resistance estimated by means of NLA using characteristic values of
material properties (Xk) in line to FprEN 1992-1-1 [2].

Castaldo et al., 2019 [6]

- SL, SM, SH: RC beams with square openings
(experimentally tested)
Concrete failure - CLX: RC beam with circular openings
(experimentally tested)

VR,M [-]
- T-Beam: RC beam with “T” cross section
designed in compliance to EN1992-1-1
SM SH
T-Beam

SL CLX

Reinforcement yielding

Figure BG-F.2: Coeffiecient of variation related to uncertainties associated to material properties VR,M (Castaldo et
al. 2019 [6]) estimated by means probabilistic analysis using NLA (non-linear finite element
analsysis).
The results from probabilistic analysis of ultimate response of RC beams performed by [6] including material
uncertainties only, highlighted that, in case of missing explicit estimation of VR,M, such value can be set equal
to the upper bound of 0.15 assuming that failure is fully governed by concrete cylinder compressive strength.
This assumption is valid until the coefficient of variation of concrete cylinder compressive strength is VC≤0.15
(as generally assumed by [8],[18]). In case of higher values for VC, reference to the FprEN 1992-1-1 [2]-
Annex A can be performed and the value of VR,M should be estimated appropriately. The coefficient of
variation VR,G account for the contribution of geometric uncertainty to statistical variability of structural
resistance. The value of VR,G can be assumed according to the FprEN 1992-1-1 [2] – Annex A or, as further
simplification on safe side, can be set equal to 0.05 [8],[22] until the governing size of structural component is
bigger than 200 mm.

F.6Full Probabilistic The Full Probabilistic Method (FPM is in compliance to EN 1990 [1] - Annex C. This method allows to quantify directly
Method (FPM)
Clauses (1), (2), (3) the level of structural reliability with reference to considered ULS. In order to fulfill the safety verification, the achieved
level of reliability should be higher than the one associated to the target value EN 1990 [1], FprEN 1992-1-1 [2] –
Annex A. Additional information related to probabilistic methods for structural reliability analysis can be found in [25].
Probabilistic models for model uncertainties (both from actions and resistance sides), actions, material and geometric
properties should be defined in line to EN 1990 [1] and FprEN 1992-1-1 [2] – Annex A. With particular reference to the
probabilistic model for uncertainty associated to the resistance model established using NLA (i.e., model uncertainty),
it should be defined according to FprEN 1992-1-1 [2] – Annex F and as described in the following. More details about
probabilistic models for concrete structures can be acknowledged in [8] and [26].

Assessment of the model uncertainty related to NLAs

F.7 Model uncertainty In this section, background to the characterization of epistemic uncertainty associated to NLAs of RC
structures/members is reported together with the methodology to define the model uncertainty factor γRd. The
interaction between aleatory and epistemic uncertainties is also discussed in order to take into account the shape of
the response surface of structural resistance which can be locally non-convex nearby the design point [6].

The epistemic uncertainty in NLAs is associated to “missing” knowledge, hypotheses and simplifications related to
Clauses (1), (3), (4), (6) constitutive laws, kinematic compatibility and equilibrium of forces and to assumptions concerning auxiliary non-
physical variables or individual choices [11]-[17]. All the mentioned above aspects characterize the selection of a
specific modelling hypothesis [12]-[14], also denoted in literature as “solution strategy” [17]. Moreover, the level of
epistemic uncertainty can be influenced by the nature of the governing failure mode (i.e., brittle, ductile) [16]-[17].
With reference to [11]-[17], the epistemic uncertainty in resistance models may be exhaustively represented by the
resistance model uncertainty random variable denoted as:

R exp , i  X , Y  (BG-F.7)
i 
R num , i  X 
The formula (BG-F.5) relates the structural resistance estimated from an experimental test Rexp,i(X,Y) to the structural
resistance estimated through numerical simulation (i.e., NLA) Rnum,i(X). X is a vector of basic variables included into
the numerical resistance model, while Y is a vector of variables that may affect the resistance mechanism but are
neglected in the model and, their unknown influence, can be indirectly incorporated including the model uncertainty
random variable ϑ within the reliability analysis.
An in-depth probabilistic calibration of the resistance model uncertainty random variable and of the related partial

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 590
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

safety factor γRd needs to be addressed in relation to the selected set of modelling hypotheses and the target reliability
differentiation according to EN 1990 [1] and FprEN 1992-1-1 [2] - Annex A . In line with the approaches of [12]-[17],
the probabilistic calibration of the resistance model uncertainty random variable can be performed in line to the
following steps:
1. Selection of the benchmark experimental tests: the selection of the benchmark set of experimental results
should be performed accounting for geometries, material properties and failure mechanisms inherent to the
investigated problem. In general, all the information about materials and geometry to define the non-linear
numerical models should be derived by the benchmark experimental test. In case of missing information
default values can be assumed according to FprEN 1992-1-1 [2];

2. Definition of the non-linear numerical models: the numerical models defined in order to reproduce by NLAs
of the benchmark experimental tests should be established adopting a specific set of modelling hypotheses
and assumptions. The modelling uncertainty so far estimated will be valid and representative only for the
considered sets of modelling hypotheses;
3. Quantification of the model uncertainty: the model uncertainty should be estimated for each benchmark
experimental test according to formula (BG-F.7);
4. Statistical analysis of model uncertainty: the outcomes of observed model uncertainty should be
appropriately processed from statistical point of view with the aim to determine the best estimate of
statistical parameters (e.g., mean value μϑ and coefficient of variation Vϑ) and the most likely probabilistic
distribution for model uncertainty random variable ϑ. According to [12],[17], in the most of the cases, the
preferred probabilistic distribution able to describe the model uncertainty random variable is the lognormal
one. The methods for statistical analysis should be able to account for the influence of the limited number of
observed samples ϑi on the estimation of mean value μϑ and coefficient of variation Vϑ. Techniques as the
methods of moments, maximum likelihood method (ML) [23] and Bayesian approaches [24] can be adopted.

Once the probabilistic calibration of the model uncertainty random variable ϑ has been performed, it can be possible to
derive the related partial safety factor as follows:

1 (BG-F.8)
 Rd 

F  1     R  tgt  
where F(∙) is the likely cumulative probabilistic distribution function able to describe resistance model uncertainty
random variable; Φ is the standard normal cumulative probabilistic distribution function; αR is the first-order-reliability-
method (FORM) sensitivity factor which is equal to 0.32 in the assumption of non-dominant variable in line to [8]; βtgt is
the target reliability index that should be adopted in compliance to EN 1990 [1] and FprEN 1992-1-1 [2] - Annex A
depending from the considered ULS.

In case the adopted probabilistic distribution is the lognormal one (as it happen in the large part of the cases [12]-[17]),
the formula (BG-F.8) can be rewritten as:

e xp  R  tgtV  (BG-F.9)
 Rd 


In general, the assumption of non-dominant variable for model uncertainty [8] turns out to be verified until the following
expression is satisfied [27]:

V R*  V (BG-F.10)
where V*R is the coefficient of variation of structural resistance inclusive of materials and geometric uncertainties (see
F.3 of FprEN 1992-1-1 [2] - Annex F and related explanation in this document) while, Vϑ is the coefficient of variation
related to model uncertainty.

In cases where the reliability analysis is dominated by the model uncertainty (i.e., the formula BG-F.10 is not fulfilled),
the non-dominant variables becomes the materials and geometric ones (i.e., aleatory). In this case the formula BG-F.9
(and the related one in EN 1992-1-1 [2] - Annex F) should be applied adopting first-order-reliability-method (FORM)
sensitivity factor αR equal to 0.8. On the opposite, the methods described by F.2 and F.3 of EN 1992-1-1 [2] - Annex F
should be applied adopting αR equal to 0.32 to account for the uncertainties related to materials and geometry within
the calculation of related partial safety factors.

Clause (2)
In the FprEN 1992-1-1 [2] - Annex F, fixed values of γRd are also provided in order to cover cases where the
probabilistic calibration of model uncertainty is not carried out by the analysis and/or software produces. These fixed
values are inclusive of the influence of statistical uncertainty and are consistent with the assumption of non-dominant
variable for model uncertainty (i.e., αR=0.32). In case the model uncertainty results to be the dominant variable, these
fixed values are not valid and the probabilistic calibration according to FprEN 1992-1-1 [2] - Annex F and as it is
previously explained in this document have to be performed. In the literature, several investigations reports studies
related model uncertainty quantification (with particular reference to non-linear finite elements analysis) [12]-[17] that
justifies the suggested values assumed on the safe side by FprEN 1992-1-1 [2] - Annex F with reference to 50 years
reference life reliability index βtgt equal to 3.8.
The investigation of Castaldo et al. 2018 [12] considered a selection of 25 structural members from several literature

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 591
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

references including deep beams, shear panels and walls. All the structural members have been realized with
statically determined scheme and tested up to failure with a monotonic incremental loading process. The experimental
tests have been reproduced by non-linear numerical simulations adopting 9 different modelling hypotheses
distinguishing between the software platform and concrete tensile response. A total number of 225 NL simulations has
been performed to estimate the resistance model uncertainty random variable. The results of the investigation are
reported in Table BG-F.2.
The work of Castaldo et al. 2020 [13] propose the investigation related to the epistemic uncertainty in non-linear
numerical analyses using the finite element method in case of cyclic loading process. In particular, 17 shear walls
tested by several Authors under different cyclic loading histories and statically determined static scheme has been
considered. These experimental tests cases have been reproduced by means 18 different modelling hypotheses
differentiating between the adopted software, concrete tensile behavior and shear stiffness of concrete after cracking.
A total number of 306 NLN simulation has been collected in order to characterize the resistance model uncertainty
random variable, with results listed in Table BG-F.2.
Cervenka et al. 2018 [16] reports the investigation of 33 RC structural elements which are inclusive of slabs and
beams which presented both shear and bending failure modes. The experimental tests on these members has been
reproduced by means NL numerical simulations adopting single modelling hypothesis and using monotonic
incremental loading process. The Table BG-F.2 reports the main results in terms of characterization of resistance
model uncertainty random variable distinguishing between bending, shear and punching failure modes.

Table BG-F.2: Characterization of model uncertainty random variable from different literature references with the
related model uncertainty partial safety factor γRd.
Model uncertainty ϑ Partial
Type of RC Coefficient safety
Reference Type of NLA Failure mode Probabilistic Mean value
members of variation factor
distribution μϑ γRd*1
Vϑ
2D finite Different failure
elements modes due to
Castaldo Deep beams,
(plane stress); both concrete
et al. 2018 walls, shear Lognormal 1.01 0.12 1.15
incremental crushing and
[12] panels
monotonic reinforcement
loading yielding
Different failure
2D finite modes due to
Castaldo
elements both concrete
et al. 2020 Shear walls Lognormal 0.88 0.13 1.35
(plane stress); crushing and
[13]
cyclic loading reinforcement
yielding

3D finite Slabs Punching 0.97 0.08 1.16

elements
Cervenka Beams Shear 0.98 0.07 1.13
(brick);
et al. 2018 Lognormal
incremental Beams Bending 1.07 0.05 1.01
[16]
monotonic
loading Slabs, Beams All 0.98 0.08 1.16

Mainly ductile
(failure
characterized
1.04 0.05 1.02
extensive
reinforcements
3D finite yielding)
elements
Engen et Beams, deep Mainly brittle
(brick);
al. 2017 beams, walls, (failure Lognormal
incremental
[17] frames characterized by
monotonic concrete crushing
loading 1.14 0.12 1.02
without or with
limited amount of
reinforcements
yielding)

All 1.10 0.11 1.04

1D finite Buckling (failure
elements characterized by
(beam, fiber- concrete crushing
Gino et al. RC slender without yielding of 0.15 – 1.15 –
beam); Lognormal 1.04
2021 [14] columns reinforcements or 0.18*2 1.19*2
incremental both concrete and
monotonic reinforcements still
loading in elastic field)

*1 the values of γRd are derived considering 50 years reference life reliability index βtgt= 3.8 and αR= 0.32 (non-dominat).
*2 the range bounds the influence of significant/limited level of experimental uncertainty within the tests sets considered for investigation
of model uncertainty [14].

The research of Engen et al. 2017 [17] reports the results of non-linear numerical simulations performed in order to

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 592
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

reproduce 38 RC members experimentally tested under monotonic loading process by several authors. The structural
members were inclusive of beams, deep beams, frames and walls. Also in this case a single modelling hypothesis has
been adopted to realize the NL numerical models of the considered members. The outcomes of the investigation
leaded to the characterization of the resistance model uncertainty random variable distinguishing between ductile and
brittle failure mechanisms. Table BG-F.2 reports the main results of this last investigation.
Finally, the work of Gino et al., 2021 [14] rely to the characterization of the epistemic uncertainty related to NLAs of
slender RC members. A total number of 40 experimental outcomes of tests on reinforced concrete columns having
slenderness ratio between 15 and 275 and different geometries in line with [22] limitations have been considered. A
discussion related to the influence of the experimental uncertainty on estimation of model uncertainty is proposed and
statistical values/model uncertainty factors are derived performing different assumptions. The main results of the
investigation are reported in Table BG-F.2.
Clauses (6), (7)
The investigation of Castaldo et. al, [6] raised the issue of the influence of the probabilistic distribution of aleatory
uncertainties (i.e., materials) on prediction of failure mechanism by means numerical modelling and related
repercussions on the application of safety formats as GFM and PFM (i.e., methods that not require an extensive
probabilistic investigation of structural resistance). In particular, the results of [6] highlights that the possible presence
of concavity in the response surface of structural resistance in proximity of the design point may lead to unsafe
conclusions when the GFM and PFM are used. This particular source of uncertainty is not covered by the model
uncertainty factor γRd alone because it can manifest only in cases where the design value of structural resistance is
evaluated without full probabilistic analysis and concern the interaction between aleatory and epistemic (i.e., model)
uncertainties.

For instance, previously to perform the specific ULS verification, the sensitivity of structural response to probabilistic
distribution of material properties should be investigated [6],[28]. In line to the results of [6] and subsequent discussion
and elaboration in CEN/TC250/ SC2/ WG1/TG6, the preliminary procedure described in FprEN 1992-1-1 [2] - Annex F
have been defined in order to check the shape of response surface of structural resistance in proximity of the design
point.
In case the following relationships between the structural resistances evaluated with different values of material
properties are verified, the numerical model can be considered to be not sensible to probabilistic distribution of
material properties:

R fcd , f yd ; a nom   R fcd , f ym ; a nom  a nd R fcd , f yd ; a nom   R fcm , f yd ; a nom  (BG-F.11)

where fcm and fym are the mean values of concrete cylinder compressive strength and reinforcement yielding strength,
respectively, while fcd and fyd are the related design values determined according to FprEN 1992-1-1 [2].
In case the conditions of (BG-F.10) turn out to be not fulfilled, it means that in the proximity of the design point may
exist possible weak failure mechanisms that give rise to local modification of concavity of the response surface. In this
case, an additional uncertainty should be included within the estimation of the model uncertainty factor γRd that, in line
with [6], can be increased of 15% as safe assumption.

References
[1] EN1990 Eurocode – Basis of structural design. CEN . Brussels
[2] FprEN1992-1-1 Eurocode 2 – Design of concrete structures. CEN 2013, Brussels.
[3] Cervenka V., Global safety formats in fib Model Code 2010 for design of concrete structures, Proceedings of
the 11th Probabilistic Workshop, Brno, 2013.
[4] Cervenka V., Reliability–based non-linear analysis according to fib Model Code 2010, Structures Concrete,
Journalof the fib, vol. 14, March 2013, ISSN1464-4177, p.p.19-28, 2011.
[5] Shlune H, Gylltoft K, Plos M. Safety format for non-linear analysis of concrete structures. Magazine of
Concrete Research 2012; 64(7): 563-574.
[6] P. Castaldo, D. Gino, G. Mancini (2019): Safety formats for non-linear analysis of reinforced concrete
structures: discussion, comparison and proposals, Engineering Structures, 193, 136-153,
https://doi.org/10.1016/j.engstruct.2018.09.041.
[7] Allaix DL, Carbone VI, Mancini G. Global safety format for non-linear analysis of reinforced concrete
structures. Structural Concrete 2013; 14(1): 29-42.
[8] fib Model Code for Concrete Structures 2010. Fib 2013. Lausanne.
[9] EN 1992-2: Eurocode 2 – Design of concrete structures. Part 2: Concrete bridges – Design and detailing
rules. CEN 2005. Brussels.
[10] Kiureghian AD, Ditlevsen O. Aleatory or epistemic? Does it matter?. Structural Safety, 31:105-112, 2009.
[11] Holický M, Retief JV, Sikora M. Assessment of model uncertainties for structural resistance. Probabilistic
Engineering Mechanics 2016; 45: 188-197.
[12] P. Castaldo, D. Gino, G. Bertagnoli, G. Mancini (2018): Partial safety factor for resistance model uncertainties
in 2D non-linear analysis of reinforced concrete structures, Engineering Structures, 176, 746-762.
https://doi.org/10.1016/j.engstruct.2018.09.041.
[13] P. Castaldo, D. Gino, G. Bertagnoli, G. Mancini (2020): Resistance model uncertainty in non-linear finite
element analyses of cyclically loaded reinforced concrete systems, Engineering Structures, 211(2020),
110496, https://doi.org/10.1016/j.engstruct.2020.110496
[14] Gino D, Castaldo P, Giordano L, Mancini G. Model uncertainty in nonlinear numerical analyses of slender
reinforced concrete members. Structural Concrete. 2021;1–26. https://doi.org/10.1002/suco.202000600
[15] P. Castaldo, D. Gino, G. C. Marano, G. Mancini (2022): Aleatory uncertainties with global resistance safety

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 593
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F Background to FprEN 1992-1-1:2023

factors for non-linear analyses of slender reinforced concrete columns, Engineering Structures, 2022; 255;
113920. https://doi.org/10.1016/j.engstruct.2022.113920
[16] Cervenka V, Cervenka J and Kadlek L (2018) Model uncertainties in numerical simulations of reinforced
concrete structures, Structural Concrete, 2018;19:2004–2016.
[17] Engen M, Hendriks MAN, Köhler J, Øverli JA, Åldtstedt E. A quantification of modelling uncertainty for non-
linear finite element analysis of large concrete structures. Structural Safety 2017; 64: 1-8.
[18] JCSS. JCSS Probabilistic Model Code. 2001.
[19] fib Bulletin N°80. Partial factor methods for existing concrete structures, Lausanne, Switzerland; 2016.
[20] fib Bulletin N°45. Practitioner’s guide to finite element modelling of reinforced concrete structures – State of
the art report. Lausanne; 2008.
[21] Cervenka V., Reliability –based non-linear analysis according to fib Model Code 2010, Structures Concrete,
Journalof the fib, vol. 14, March 2013, ISSN1464-4177, p.p.19-28, 2011.
[22] EN 1992-1-1: Eurocode 2 – Design of concrete structures. Part 1-1: general rules and rules for buildings. CEN
2014. Brussels.
[23] Faber, Michael Havbro Statistics and Probability Theory, Springer, 2012.
[24] Gelman A, Carlin JB, Stern HS, Dunson DB, Vehtari A, Rubin DB. Bayesian data analysis. 3rd ed. CRC Press;
2014.
[25] ISO 2394. General Principles on Reliability for Structures; ISO 2394:2015(E); International Organization for
Standartization (ISO): Geneva, Switzerland, 2015; p. 111.
[26] JCSS. JCSS Probabilistic Model Code. 2001.
[27] fib Bulletin N°80. Partial factor methods for existing concrete structures, Lausanne, Switzerland; 2016.
[28] Yu, Q., Valeri, P., Ruiz, M.F. and Muttoni, A. 2021. A consistent safety format and design approach for brittle
systems and application to textile reinforced concrete structures. Engineering Structures, 249: 113306,
https://doi.org/10.1016/j.engstruct.2021.113306.

2023.03.14 reviewed and editorial update

D. Gino, P. Castaldo, G. Mancini 09.02.2021

For citations, copying and referencing see Note on Front page! page 594
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

CEN/TC250/SC2/WG1/TG6 FprEN 1992-1-1:2023

Additional background information to

Annex F
Application examples for the verification of
ultimate limit states by non-linear analysis

Jan Cervenka Cervenka Consulting s.r.o., Czech Republic

28.02.2021

J. Cervenka
For citations, copying and referencing see Note on Front page! page 595
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Introduction
This document has been prepared in order to provide additional background information to the Annex F. It
extends the information provided in the background information document [12]. It summarizes the safety
formats for non-linear analysis as they are proposed in EC2-Annex F. It demonstrates their application and
compares the calculated design values with analytical formulas and experiments. This annex describes three
possible safety formats:
F.4 Partial factor method
F.5 Global factor method
F.6 Full probabilistic method
This documents demonstrates the application of these safety formats to a selected set of benchmark problems
summarizing the most typical elements and failure modes occurring in the design of reinforced concrete
structures.
The first section describes the appropriate choice of input parameters for the application of the safety formats
F.4. to F.6.
The second section describes the benchmark problems used in this study. This section includes also an
important discussion on the choice of model uncertainty partial safety factor.
The third section summarizes and compares the obtained results and some concluding remarks.

Input parameters for safety formats for non-linear analysis

Input parameters for structural assessment are usually defined by the structural geometry, used materials and
other design assumptions, and the corresponding uncertainties shall be considered. It is possible to identify the
following categories of input parameters for the nonlinear analysis:
- Geometry: is defined by the structural shape, reinforcement detailing and uncertainties shell be
selected according to the used safety format F.4, F.5 and F.6
- Material parameters: shall be selected according to the used safety format F.4, F.5 and F.6
- Actions: determined according to EN 1990 and uncertainties are addressed by the corresponding
load factors and load combinations
- Solution and modelling parameters: includes FE mesh, accuracy criteria or other parameters related
to the used numerical method. Associated uncertainties should be evaluated by sensitivity studies.

Input parameters for Partial factor method PFM

Clause F.4
In this case the design value of the structural resistance is obtained as:

R  X d ; a
Rd  (1)
 Rd
where:

R  is the structural resistance based on the numerical simulation;

is the design value of the material property calculating adopting partial safety factors
Xd
according to Table A.1 line (f) accounting for materials and geometric uncertainties, but
excluding model uncertainties, which are considered separately by  Rd . In general case
the design value is calculated as:

Xk ; exp  R  tgt  VRM 

Xd  M  ;
M  RM

 Rd is the partial safety factor, which accounts for the model uncertainty according to F.5;

Index M S for reinforcement and C for concrete in compression;

R Is sensitivity factor for resistance according to Table A.3;

tgt Is the target value for the reliability index according to Table A.3;

Is the coefficient of variation of the resistance, which may be calculated as:

VRM

J. Cervenka
For citations, copying and referencing see Note on Front page! page 596
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

n
VRM  V I
I
2 ;

 RM Is the bias factor for all uncertainties, which may be calculated as a product of all bias
n
factors:  RM    I .
I

For typical situations in accordance to Annex A, the material safety factors for concrete and steel are listed in
Table 1

Table 1: Standard values for material factors for the safety format F.4
persistent and transient design situations

S C
1.2 1.54 (1.46)1

Input parameters for Global factor method (GFM)

Clause F.5
In the case of GFM method, the design strength is calculated as:

Rm  X ; a
Rd  (2)
 R*   Rd
where:

Rm  is the mean value of the structural resistance based on the numerical simulation and
shall be estimated by means of a probabilistic analysis including uncertainties related
to material and geometric properties;

 R* is the global resistance factor for the uncertainties of material properties and geometry
according to F3.2;

 Rd is the partial safety factor, which accounts for the model uncertainty according to F.5;

The global resistance factor may be evaluated as:

 R*  exp  R  tgt  VR* 

Clause F5.2
(3)

where:

1  R  X m ; anom   1  R  X m* ; anom  
(4)
VR*  ln    ln  
1.65  R  X k ; ak   1.65  R  X k* ; anom  
 
In (4) R  X m ; anom  represents an estimate of the structural strength calculated by a nonlinear analysis using
average material parameters and nominal geometric properties. R  X k ; ak  is the analogic estimate of the
characteristic strength that is obtained by a nonlinear analysis using the characteristic values of the material
and geometric parameters. This approached was originally proposed in [8] and [9]. It became part of the
nonlinear safety formats proposed in the fib model code 2010 [17] along with the partial factor method F.4. In
practical cases it would be however rather inconvenient to create a model taking into account the characteristic
values of its dimensions or reinforcement location. It is therefore proposed to keep geometric parameters to

their nominal values and rather define an auxiliary material property X m* , X k* representing the uncertainty in
both material and geometric properties. This can be done using the information in Annex A and the assumption
that structural resistance Rmodel calculated by the numerical model is approximately linearly dependent on the

concrete d c  f c ,is  and reinforcement contribution  d y  f y  .

1
1.46 according to the new version of Annex A (N0267), while 1.54 according to the previous version of Annex A (N 222). This version of the
document was initially prepared according to Annex A version N 222. All examples are calculated according to this Annex A version with the
exception of the last punching example, which was calculated according to Annex A version N0267.

J. Cervenka
For citations, copying and referencing see Note on Front page! page 597
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Rmodel  R  d c  f c ,aIs , d y  f y , ..... (5)

where dc is concrete section area, dy is the effective depth, f c , aIs is the actual in-situ concrete compressive
strength, and f y is the reinforcement yield strength. In general the uncertainties of the input

parameters/variable I can be described by the corresponding bias factor  I  X Im X Ik (ratio of mean to

characteristic value) and coefficient of variation VI .

Then it is possible to express the uncertainties in the steel contribution by its mean value:

f ym
d ym  f ym  d y , nom  d f yk  d y , nom f ym
*
(6)
f yk
and coefficient of variation:

VRS  Vd2  V fy2 (7)

By analogy the uncertainties in the concrete contribution can be expressed by its mean value:

f cm
d cm  f cm  d c ,nom  Ac  Is f ck  d c ,nom f cm* (8)
f ck

where  Is  f c ,aIs f cm and coefficient of variation:

VRC  VAc2  V2Is  V fc2 (9)

The material strengths in (6) and (8) can be described by auxiliary variables:

f ym
*
f ym  d f yk   d exp 1.645V fy  f yk (10)
f yk

f ck   Ac  Is exp 1.645V fc  f ck
f cm
f cm*   Ac  Is (11)
f ck
which represent the material mean strength considering the uncertainties of the material strength, effect of
insitu strength as well as geometric uncertainties.
Using the analogic assumptions the auxiliary variables for characteristic analysis in Eq. (4) can be calculated
as:

f yk*  f ym
*
exp  1.645 VRS   f yk  d exp V fy  VRS  (12)

f ck*  f cm* exp  1.645 VRC   f ck  Ac  Is exp 1.645 V fc  VRC   (13)

These assumptions are in agreement with the Eurocode EN 1992-1-1:2005 approach of partial factors for
materials as shown for instance in the case of F.4.
It should be noted however, that in nonlinear analysis it is usually necessary to define other parameters to
describe the behaviour of reinforced concrete structures. It is desirable to define unique rules how these
additional parameters such as for instance: tensile strength ft , fracture energy GF and others are related to
these basic strength parameters. Then all material parameters can be uniquely derived from the auxiliary steel

yield strength f y* and concrete compressive strength f c* . Alternatively, the mean and characteristic values of
the additional material parameters can be derived using JCSS probabilistic Model Code [16].

J. Cervenka
For citations, copying and referencing see Note on Front page! page 598
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Table 2: Standard values for auxiliary steel and concrete material parameters for the safety format F.5
Parameter persistent and transient design situations
Reinforcement

Yield strength mean

*
f ym
*
f ym  1 .0 2 3 f yk

Yield strength char. f yk* f yk*  0 .9 1 6 f yk

Concrete

Compressive strength mean f cm

* f ck*  1.062 f yk

Compressive strength char. f ck

* f ck*  0 .8 1 5 f ck

Clause F.6 F.6 Full probabilistic method

In this document an analogical approach is taken also for the probabilistic analysis. This means the geometric
uncertainty is taken into account into the corresponding auxiliary material parameters. The basic assumptions
are summarized in

Table 3: Standard values for auxiliary steel and concrete material parameters for the full probabilistic analysis
F.6
Parameter persistent and transient design situations
Reinforcement

Yield strength mean

*
f ym
*
f ym  1 .0 2 3 f yk

Coefficient of variation VRS  Vd2  V fy2 VRS  0.0452  0.052  0.067

Concrete

Compressive strength mean f cm

* f ck*  1.062 f yk

Coefficient of variation VRC  VAc2  V2Is  V fc2 VRC  0.12  0.122  0.042  0.16

Procedure for structural assessment by nonlinear analysis

The procedure for the structural assessment by nonlinear analysis typically involves the following steps:
- Definition of a loading scenario: proper loading sequence should be identified to accurately represent
the investigated load combination.
- Development of a numerical model
- Sensitivity analysis: includes mesh size sensitivity, load step size, different sequences for the
application of actions, etc.

- The evaluation of the structural strength Rd using the developed numerical model with the given

loading scenario. It should be understood that in most cases Rd is not a single number but it rather
represents an overall load factor for the investigated load/action combination.

J. Cervenka
For citations, copying and referencing see Note on Front page! page 599
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Benchmark example problems

The following section contain a set of examples with known experimental results that document the application
of the safety formats for nonlinear analysis from the Annex F to typical cases and failure modes that can be
encountered in the design practice. Each example first contains a description of the geometry and
reinforcement arrangement. Then graphs of the experimental and numerical response is shown. The examples
are analysed using the finite element software ATENA [7], however, the presented concept can be used with
other nonlinear software as well. When using the safety formats from Annex F with a particular analysis
software it is possible to pay special consideration to the following issues as they are listed in the Annex F.
General validation tests should be performed to verify the software behaviour:
- in basic material tests,
- and structural tests with the relevant failure modes.
These tests can be done only once. In most cases they will be provided as standard validation benchmarks by
the software developer. In addition they should be used to provide the calibration of the model reliability factor
 Rd in Eq. (1) and (2). For ATENA software such a validation has been published by Cervenka et al [10].
These factors have been used in the presented example cases to calculate the final values of the design
strength Rd , and they are listed in Table 4. Similar calibration studies have been performed for other programs
and methods such as for instance: [4],[5], [11] and [13].

Table 4: Model uncertainty factors used in this study based on [10]

Failure mode Model uncertainty factor
Punching 1.16
Shear 1.13
Bending 1.01
All failure modes 1.16

Additional important tests that should be performed in each project involving application of nonlinear analysis to
structural assessment are project specific sensitivity studies that typically consist of the following types of
analyses:
- mesh sensitivity tests
- solution method tests such as effects of load steps size, solution method parameters, loading history,
various modelling idealizations, boundary conditions, etc.
For each example a table of used material properties is provided as well as the specific values that are used for
each method. This documents compares the design strength values obtained by PFM (F.4), GFM (F.5) and full
probabilistic method (F.6). The probabilistic method can be considered as the most accurate estimation of the
design strength. These three methods are also compared with the analytical formulas from EC2. The final
comparison is available in Table 9 and Table 10.

J. Cervenka
For citations, copying and referencing see Note on Front page! page 600
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Bending example problem:

Experimental data by Bosco & Debenardi [2]. Beams

Figure 1 : Bosco & Debenardi [2] experiments, geometry and dimensions

Figure 2 : Bosco & Debenardi [2] experiments, load-displacement curves

J. Cervenka
For citations, copying and referencing see Note on Front page! page 601
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Table 5: Overview of material parameters for beams T-8, T-11 from : Bosco & Debenardi [2]
Reinforcement B500H Concrete
Yield strength Strain limit[-] Compressive Tensile Fracture
[MPa] strength [MPa] strength [MPa] energy [Nm]
Test data

f y , (COV)
---
587 (0.04)

ft (COV)
0.07 27.8 – 30.9 --- ---
672 (0.02)

f yk  500 (1)
f ck  25 (1)

EC2 Analytic
(2)
f yd  459 N/A f cd  16.7 --- ---

PFM

f yd  442 f cd  16.2 f ctd  1.17

(2)
0.07 GFd  29.87

GFM

Mean
*
f ym  523 (2)
0.07 f cm*  26.6 *
f ctm  1.91 *
GFm  48.76
(2)
Char. f yk*  478 0.07 f ck*  20.4 *
f ctm  1.47 *
GFk  37.44
Full probabilistic

 1.91
(2)
Mean
*
f ym  523 0.07 f cm*  26.6 *
f ctm *
GFk  37.44

VRC  0.12  0.122  0.042

VRS  0.0452  0.0252
COV ---  0.16 VRC  0.16 VRC  0.16
 0.051

Notes:

(1) Test material properties were matched to standard steel f yk  500 MPa and concrete to f ck  25 MPa .
(2) Geometric uncertainties for steel are modified for the case od d = 565, i.e.

Vd  0.05  200 / d   0.025 ,  d  1  0.05(200 d )  0.975

23 23

J. Cervenka
For citations, copying and referencing see Note on Front page! page 602
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Experimental beam failing in compression by Hallgren [15]

Figure 3 : Geometry of the beam B92B from Hallgren [15]

Figure 4 : Experimental and numerical results for the beam B92B from Hallgren [15]

J. Cervenka
For citations, copying and referencing see Note on Front page! page 603
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Table 6: Overview of material parameters for beam B92B from Hallgren [15]
Reinforcement B500H Concrete
Yield strength Strain limit[-] Compressive Tensile strength Fracture
[MPa] strength [MPa] [MPa] energy [Nm]
Test data

f y , (STD)
---
631 (±5)

ft , (STD)
0.03 42.1 --- ---
749 (±7)
f yk  586 f ck  36 --- ---

EC2 Analytic
f yd  510 N/A f cd  24 --- ---

PFM
f yd  488 0.025 f cd  23.4 f ctd  2.05 GFd  44

GFM
Mean *
f ym  599 0.025 f cm*  38.2 *
f ctm  3.34 *
GFm  72

Char. f yk*  537 0.025 f ck*  29.3 *

f ctm  2.56 *
GFk  55.4

Full probabilistic
Mean *
f ym  599
(1)
0.025 f cm*  38.2 *
f ctm  3.34 *
GFk  72
(1)
COV VRS  0.0452  0.052 VRC  0.12  0.12 2  0.042
--- VRC  0.16 VRC  0.16
 0.067  0.16

Notes:
Geometric uncertainties for steel are not modified d < 200 mm

J. Cervenka
For citations, copying and referencing see Note on Front page! page 604
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Shear problem (Anderson & Ramirez [1])

Figure 5 :Anderson & Ramirez [1] geometry and dimensions, (1 in = 25.4 mm, 1 kip = 4.448 kN,
1 psi = 0.00689 MPa)

Figure 6 :Anderson & Ramirez [1] load displacement curves, (1 in = 25.4 mm, 1 kip = 4.448 kN,
1 psi = 0.00689 MPa)

J. Cervenka
For citations, copying and referencing see Note on Front page! page 605
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Table 7: Overview of material parameters for beam W1-W2 from Anderson & Ramirez [1]
Reinforcement B500H Concrete
Yield strength Strain limit[-] Compressive Tensile strength Fracture
[MPa] strength [MPa] [MPa] energy [Nm]
Test data

f y , (COV)
587 (0.04)

ft , (COV)
0.07 29.3 --- ---
672 (0.02)

f yk  500 (1)
f ck  25 (1)

EC2 Analytic

f yd  455 (2)
N/A f cd  16.7 --- ---

PFM

f yd  439 (2)
0.07 f cd  16.2 f ctd  1.17 GFd  29.87
GFM

 523 f cm*  26.5

(2)
Mean
*
f ym 0.07
*
f ctm  1.91 *
GFm  48.76

(2)
Char. f yk*  476 0.07 f ck*  20.4 *
f ctm  1.47 *
GFk  37.44

Full probabilistic
*
 523 f cm*  26.5 *
 1.91
(2)
Mean f ym 0.07 f ctm *
GFk  37.44
(2)
VRS  0.0452  0.0352 VRC  0.12  0.122  0.042
COV --- VRC  0.16 VRC  0.16
 0.057  0.16

Notes:

(1) Test material properties were matched to standard steel f yk  500 MPa and concrete to f ck  25 MPa .
(2) Geometric uncertainties for steel are modified for the case od d = 344, i.e.

Vd  0.05  200 / d 
23
 0.035 , d  1  0.05(200 d ) 2 3  0.965

J. Cervenka
For citations, copying and referencing see Note on Front page! page 606
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Punching problem :
Punching example taken from the experiments by Guandalini et al. [14]. This example is selected because it is
a slab with large thickness, i.e. representing a large structure where nonlinear analysis can be used in practice.

Figure 7 : Punching experiment by Guandalini et al. [14], PG3 experiment with dimensions B = 6 m,
rg = 2.85 m, c = 0.52 m, h = 0.5 m, reinforcement Ø16@135 mm, cover 28 mm.

Figure 8 : Results of punching experiment of Guandalini et al. [14] and numerical simulation from the
uncertainty factor calibration [10].

J. Cervenka
For citations, copying and referencing see Note on Front page! page 607
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Table 8: Overview of material parameters for slab PG3 Guandalini et al. [14]
Reinforcement B500H Concrete
Yield strength Strain Compressive Tensile strength Fracture
[MPa] strength [MPa] [MPa] energy [Nm]
limit[‐]
Test data

fy , 520

0.025 42.1 2.6 65

ft , 607

f ym 0.025 f ck 
f cm
f yk 
exp 1.645V fy  exp 1.645V fc 
 520  483  32.4
1.077 1.18
 27.5

EC2 Analytic

f yd  420 N/A f cd  18.3 --- ---

PFM

f yd  403 0.025 f cd  18.8 f ctd  1.45 G Fd  37

GFM

Mean
*
f ym  494 0.025 f cm*  30.8 *
f ctm  2.47 *
GFm  61.75

Char. f yk*  442 0.025 f ck*  23.6 *

fctm  2.12 *
GFk  53.1
Full probabilistic

Mean
*
f ym  494 0.025 f cm*  30.8 *
f ctm  2.47 *
GFk  61.75

VRC  0.12  0.122  0.042

VRS  0.0452  0.052
COV ---  0.16 VRC  0.16 VRC  0.16
 0.067

J. Cervenka
For citations, copying and referencing see Note on Front page! page 608
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Conclusions and comparison of the design strength values calculated by the methods from
Annex F
This section contains the comparison of design strength values calculated for the investigated examples using
the three safety formats for nonlinear analysis from Annex F. They are:
F.4 Partial factor method,
F.5 Global factor method,
F.6 Full probabilistic method.
The three methods are also compared with the standard analytical formula from the corresponding section of
the Eurocode 2. The analytical formulas are taken from a previous draft 4 of Eurocode prEN 1992-1-1.
The results are compared in two tables: Table 9 and Table 10. The first table contains the comparison of the
actual strength values in terms of the maximal loads, while the second table contains the comparison of relative
values that are normalized with respect to the analytical formula from the Eurocode.
From the result tables, it is possible to draw the following conclusions:
- for cases with clear analytical solution, which is mainly the case of bending with steel failure the nonlinear
analysis provides same results as the analytical solutions,
- for other situations the nonlinear analysis provides slightly more conservative results than the analytical
formulas. This is however a desirable situation since it can be expected that nonlinear analysis will be
mainly used in situations when analytical formulas are not valid or can be applied only with strong
simplifying assumptions,
- GFM method provides always results on the safe side with the exception of the punching case. However,
after the application of the clause F.7(8) even for this case conservative results are obtained,
- it can be observed that more conservative results from the nonlinear analysis are obtained mainly in the
cases when concrete crushing is observed, i.e. bending with crushing, shear with concrete crushing and
punching example. This results is strongly related to the used software where the default software
settings assume rather conservative brittle behaviour of concrete during compressive softening, i.e.
crushing. This shows the importance of the sensitivity and validation studies that should be performed for
each software or material model to be used in structural assessments according to Annex F as they are
described in clauses: F.3(3) and F.7.
It should be noted that currently the results are slightly inconsistent as the material parameters for first 4
examples were calculated with the previous version of Annex A (N 222) while in the case of the last punching
example the Annex A version (N 0267) was used as is explained in the footnote on page 3. The changes in
Annex A are however minor and it should not significantly affect the main results and conclusions of this
document.

Table 9 : Comparison of design strength obtained by various methods

J. Cervenka
For citations, copying and referencing see Note on Front page! page 609
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

Example Exp. EC2 - Draft 4 PFM GFM Full

Analytic [kN] [kN] [kN]
[kN]
Bending
Bending
41 40 40 41
steel failure 50
 Rd  1.01  Rd  1.01  Rd  1.01
T8-A1 [2]
Bending
concrete 21  21   1.01 23   1.01
34.2 Bending - 22  1.01 Rd Rd
crushing Rd

B92B1-42 [14]
Shear
Bending/shear
Bending - 296 214   1.01 229   1.01 244   1.01
concrete Rd Rd Rd
380 Shear - 223
crushing 191  204   1.13 218   1.13
 1.13
[  38 ] Rd Rd Rd
T11-A1 [2]
Shear 316 [   30 ]
W1-W2 460-
217 [   40 ] 228   1.13 211   1.13
Rd
240   1.13
Rd
Anderson & 549 Rd

Ramirez [1] 182 [   45 ]

Punching
Punching 1382  1392  (1)
1432 
2153 1661  1.16  1.31  1.16
Guandalini [10] Rd Rd Rd

Table 10 : Comparison of design strength obtained by various methods

Example Exp. EC2 - Draft 4 PFM GFM Full
Analytic [kN] [kN] [kN]
[kN]
Bending
Bending
steel failure 1.22 1 0.98 0.98 1.0
T8-A1 [2]
Bending
concrete crushing 1.55 1 0.95 0.95 1.05
B92B1-42 [14]
Shear
Bending/shear
concrete crushing 1.70 1 0.85 0.91 0.98
T11-A1 [2]
Shear
W1-W2 Anderson & 2.32 1 1.05 0.97 1.10
Ramirez [1]
Punching
Punching Guandalini [10] 1.30 1 0.83 0.84 0.86

Notes :
(1) Model uncertainty factor is increased by 15% according to the clause F.7(8)

[1] Anderson, N.S. & Ramirez, J.A. 1989. Detailing of stirrup rein-forcement. ACI Structural Journal 86(5):
References

J. Cervenka
For citations, copying and referencing see Note on Front page! page 610
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex F - Examples Background to FprEN 1992-1-1:2023

507-515.
[2] Bosco C. & Debenardi, P. G. 1993. Influence of some basic parameters on the plastic rotation of
reinforced elements. Ductility requirements for structural Concrete - Reinforcement. Comite Eruo-
international du Breton. CEB bulletin 218. pp25-44
[3] CEB-FIP (Comité Euro International du Béton; Fédération International de la Précontraint). CEB-FIB
Model Code 1990: Design Code. London: Thomas Telford.
[4] P. Castaldo, D. Gino, G. Bertagnoli, G. Mancini (2018): Partial safety factor for resistance model
uncertainties in 2D non-linear analysis of reinforced concrete structures, Engineering Structures, 176,
746-762. https://doi.org/10.1016/j.engstruct.2018.09.041.
[5] P. Castaldo, D. Gino, G. Bertagnoli, G. Mancini (2020): Resistance model uncertainty in non-linear
finite element analyses of cyclically loaded reinforced concrete systems, Engineering Structures,
211(2020), 110496, https://doi.org/10.1016/j.engstruct.2020.110496
[6] Cervenka J, Pappanikolaou V. Three dimensional combined fracture-plastic material model for
concrete. Int J Plast 2008;24(12):2192–220. https://doi.org/10, 1016/j.ijplas.2008.01.004.
[7] Cervenka, V., Cervenka, J., Jendele, L., ATENA Program Documentation, Part 1: Theory, Cervenka
Consulting s.r.o., 2020, www.cervenka.cz
[8] Cervenka, V.: Global Safety Format for Nonlinear Calculation of Reinforced Concrete. Beton- und
Stahlbetonbau 103, (2008), special edition, Ernst & Sohn, pp. 37–42.
[9] Cervenka, V., Reliability-based non-linear analysis according to Model Code 2010, Journal of FIB,
Structural Concrete 01/2013, pp.19-28
[10] Cervenka V, Cervenka J, Kadlec L. Model uncertainties in numerical simulations of reinforced concrete
structures. Structural Concrete 2018;19(6):2004–16. https://doi.org/10.1002/suco.201700287.
[11] Engen M, Hendriks MAN, Köhler J, Øverli JA, Åldtstedt E. A quantification of modelling uncertainty for
non-linear finite element analysis of large concrete structures. Structural Safety 2017; 64: 1-8.
[12] Gino D., Castaldo, P., Mancini, G., Background document to Annex F, Safety formats for nonlinear
analysis, CEN/TC250/SC2/WG1/TG6, 28. 11. 2022
[13] Gino, D., Castaldo, P., Giordano, L., Mancini, G., Model uncertainty in nonlinear numerical analyses of
slender reinforced concrete members. Structural Concrete. 2021;1–26.
https://doi.org/10.1002/suco.202000600
[14] Guandalini S, Burdet OL, Muttoni A. Punching tests of slabs with low reinforcement ratios. ACI Struct J.
2009;106:87–95.
[15] Hallgren M., Flexural and Shear Capacity of Reinforced High Strength Concrete Beams without
Stirrups, TRITA-BKN, Bulletin 9, 1994, ISSN 1103-4270
[16] JCSS, (2001), JCSS Probabilistic Model Code, Joint Committee on Structural Safety, ISBN 978-3-
909386-79-6
[17] ModelCode 2010, fib Lausanne, Ernst & Sohn: Switzerland, 2011

List of versions of this document

05. 02. 2020. 1st draft
23. 02. 2020 2nd draft
28. 02. 2020 3rd draft
21. 03. 2023 editorial update to FprEN 1992-1-1:2023

J. Cervenka
For citations, copying and referencing see Note on Front page! page 611
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG4 FprEN 1992-1-1: 2023

Background Document to Annex G:

Design of membrane-, shell- and slab
elements

Linh C. Hoang Technical University of Denmark

22.12.2022

For citations, copying and referencing see Note on Front page! page 612
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

Introduction
This document has been prepared to clarify the background for the provisions introduced in subsections G.3 -
G.5 of Annex G in FprEN-1992-1-1:2023. The subsections provide methods for ULS design of membrane-,
shell-, and slab elements and SLS analysis of membrane elements.

Reasons for change

In the 1st generation of EC2 (EN 1992-1-1:2004) the provisions for ULS design of membrane elements is found
in Annex F: Tension reinforcement expression for in-plane stress conditions. However, Annex F is incomplete
(only formulas for the case Edx > Edy are provided) and provisions to calculate the efficiency factor  were not
provided. Furthermore, it does not appear clearly (e.g. by means of cross references) in EN 1992-1-1:2004 that
the method of Annex F and the methods for shear- and torsion in subsections 6.2.3 and 6.3 in fact are based on
the same approach grounded on limit analysis. For users in practice, this lack of information has had the
consequence that it is not obvious that e.g. problems with shear in members with large compressive axial
forces could be treated by combining subsection 6.2.3 with Annex F. Finally, EN 1992-1-1:2004 does not have
provisions for ULS design of shell- and slab elements.
In FprEN-1992-1-1:2023 the Annex F of EN 1992-1-1:2004 has been renamed Annex G and the content has
been expanded. In the process of developing this new Annex G, the main aims of Task Group TG4 have been:
1) To include methods for design of shell- and slab elements, which is fully consistent with the
theoretical background for the existing formulas for design of membranes
2) To have a general Annex G, which could serve as a reference point for the specific provisions for
member design in Section 8
With regard to aim 2), TG4 has worked with the roadmap illustrated in Figure 1 showing the framework for ULS
design of members with adequate deformation capacity. The general rules in Annex G can be related to specific
cases treated in Section 8 as shown by the roadmap. Topics written in red font are new and not treated in EN
1992-1-1:2004. The provisions for the new topics have largely been formulated to be consistent with Annex G.
Furthermore, cross references from various parts in Section 8 to Annex G will make it much more clear
(compared to EN 1992-1-1:2004), that the various provisions are based on the same lower bound approach of
plasticity theory.

Figure 1: Roadmap to link the various provisions in Section 8 with Annex G. Topics written in red font are new
and not included in EN 1992-1-1:2004.

Consistency with provisions for shear in beams requiring shear reinforcement

To demonstrate the link between Annex G and the provisions in Section 8 as illustrated in Figure 1, a simple
design example is provided in the following. Figure 2 shows a cantilevered beam subjected to a point load,
which is assumed to be transferred to the end of the beam by mean of a strong diaphragm (to make the
boundary condition simpler). Design of this beam to carry the shear force V can be performed by use of
Equations (G.3) – (G.5) in Annex G. The procedure is shown in Figure 3, where free body diagrams of the
beam web and the tension- and compression chord are shown. Here, the web can be considered as a
membrane element subjected to pure shear  = V/bz. By using Eqs. (G.3) – (G.5), it is possible to calculate the
required distributed reinforcement in the x- and in the y- direction as well as the compression stress c. The
calculations appear in Figure 3. For shallow beams, it is not common practice to supply a distributed

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 613
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

reinforcement in the x-direction. The total amount of required reinforcement in this direction, which carries a
tension force of the magnitude F = Vcot, can however be re-allocated to the positions of the chords. Hence,
half of the horizontal reinforcement (i.e. ½Vcotis allocated to the tension chord while the other half is
allocated to the compression chord as a reduction of the compression force. The re-allocation is illustrated in
Figure 4. As can be seen in Figure 4, the re-allocation of the horizontal reinforcement leads to the same design,
as if the beam was designed using subsection 8.2.3, where the shift-rule required ½Vcotto be added
(subtracted) to the tension (compression) chord.

Figure 2: Example, cantilever beam with point load

Figure 3: Free body diagrams and design scheme for cantilever beam using the method of Annex G

Figure 4: Free body diagram and design scheme for cantilever beam using the method of Subsection 8.2.3.

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 614
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

Subsection G.3 – Table G.1

In contrast to Annex F in EN 1992-1-1:2004, the full set of equations to determine the optimum reinforcement in
membrane elements is now given in Table G.1. This enhances the ease-of-use and furthermore reduces the
risk of misinterpretation when used for design in combination with FEM analysis. The theoretical background for
the equations in Table G.1 have been explained in details in e.g. reference [2]. It should be noted that these
equations have been implemented in many commercial Finite Element software packages.

Clause G.3(3), Expressions (G.6) and (G.7)

When the reinforcement is designed by use of Expressions (G.3 – G.5) and depending on the choice of cot,
redistribution of stresses and rotation of strut angle will take place as the load increases to the ultimate load. To
limit cracking in SLS and to ensure deformation capacity without explicit check, limitations have to be imposed
on the extent of stress redistributions. Expressions (G.6) and (G.7) provide these limitations. These limits are
more liberal than those found in EN 1992-1-1:2004 Annex F, and are introduced to be consistent with the
limitations provided in Subsections 8.2.3 and 8.3. For example in the case of a membrane subjected to pure
shear, limiting ftdx to 2,5f*tdx and ftdy to 0,4f*tdy would be the same as adopting the limiting value cot = 2,5 (1/cot
= 0,4), which is exactly the limit imposed in Clause 8.2.3(3) for shear design and in Clause 8.3.4(4) for torsion
design.

Clause G.3(5) - G.3(7)

Expression (G.9) provides a refined way to calculate the efficiency factor , which accounts for the extent of
transverse cracking. This expression is identical to the one given in clause 8.2.3(6) for beam shear design.
Reference is made to the background document to subsection 8.2.3 for further explanations, [1].
Instead of expression (G.9), a default value of  = 0,4 (expression (G.8)) may be adopted. This value is lower
than the default value of 0,5 proposed in Clause 8.2.3(5) for beam shear, however, higher than provided by the
Expression (G.9) when cot = 2,5 is used in conjunction with the simple estimates/guidelines provided in clause
G.3(7). The main reason for this difference is that when calculating  for beams, the gradient of normal strains
in the cross section may be taken into account allowing the use of only half of the tensile strain in the
longitudinal reinforcement to estimate  (see [1]). It is noted that for beam shear, the simple approach
recommends to use  = 0,4 when verifying the compression field in the vicinity of a plastic hinges. This
guideline appears in the commentary text to clause 8.2.3(5) in the draft D6 of prEN-1992-1-1:2020.
Because of the dependency on the transverse strain, 1, Expression (G.9) is mostly suitable for implementation
in a numerical framework (non-linear finite element modelling) where calculation of deformations is an
integrated part of the analysis. However, to enhance the ease-of-use in manual calculations, clause G.3(7)
provides Expressions (G.10) and (G.11) to estimate 1. These expressions are based on the condition of strain
compatibility and on the assumption of smeared cracks, see also [1]. Furthermore, in these expressions, the
minor principal strain, i.e. the compression strain in the concrete strut, has as an approximation been taken as a
constant c = 0,001.
In order to estimate 1, the designer/user has to know the value of either x or y (the normal strains in the
direction of the reinforcement). In a design situation, these values can be obtained by following the guidelines
(a) – (d) in clause G.3(7). Basically, once the value of cot has been chosen for design, then either x or y can
be estimated depending on how the stress redistribution takes place when the load increases from zero to the
ultimate state. Three scenarios for stress redistributions exists:
1) No stress redistributions at all
2) Stress redistributions take place by rotation of strut toward the x-axis.
3) Stress redistributions take place by rotation of strut toward the y-axis.
Scenario 1) is relevant when reinforcement is not required in one of the two directions (i.e. for shear combined
with large axial compression in one of the two directions) – see e.g. the second and the third row in Table G.1.
Hence, assuming a fully utilized reinforcement design, the reinforcement stress will reach the yield stress when
the load increases to the ultimate state. This means that in this case the reinforcement strain (either x or y)
may be taken as the yield strain when inserting into Expression (G.10) or (G.11). This scenario is covered by
the guideline (a) in clause G.3(7). If more reinforcement than needed is supplied, the strain may be reduced
accordingly, which is covered by the guideline (d). Scenario 1) is also relevant when Table G.1 is used for
cases where reinforcement is required in both directions (i.e. when the membrane is loaded in pure shear or
shear-tension). In this case, the optimal stress field inclination is cot = 1, which will not change in a
proportional loading scheme. Again, if the reinforcement is fully utilized, both direction of reinforcement will
reach the yield stress at the same time when the load increases to the ultimate state. This means that x and y
may be taken as the yield strain. In this case, Expressions (G.10) and (G.11) will lead to identical result. This
case is covered by guideline (b) in clause G.3(7).
The guideline (b) also covers scenarios 2) and 3). Consider for example a case with pure shear, where the
chosen strut inclination corresponds to, say, cot= 2. This will lead to a design (see Expression (G.3) and
(G.4)), where the reinforcement content in the y-direction is only 25% of the amount supplied in the x-direction,

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 615
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

i.e. x = 4y. With this design, and in a proportional loading scheme, the weak reinforcement (i.e. y) will start to
yield long before the load reaches the ultimate state. Once the weak reinforcement yields, redistribution of
stresses takes place by rotation of the strut to a flatter angle (scenario 2) while the concrete stress, c,
increases and while the stress in the strong reinforcement (x) increases as well. At the final state,  = 26,6o
corresponding to cot= 2 and the strong reinforcement reaches the yield stress. In this case, x can therefore
be taken as the yield strain while y is larger, but unknown. The opposite will be the case, if instead the design
is based on cot= ½, which will lead to scenario 3).
In order to avoid a situation where the approach leads to too conservative (small) values of , guideline (c) has
been included. Guideline (c) in Clause G.3(7) covers the important situations, where the assumption of zero
tensile strength in the concrete (which is the basis for the Expressions (G.3) – (G.5)) simply lead to a large
estimate of the transverse strain and thereby too small a -value. This issue can be illustrated by the following
example:
Example:
Assume: fyk = 500 MPa, fck = 30 MPa and fctk = 2 MPa (no partial coefficient)

Case 1:  x;  y ; xy    0;0;9  MPa

Case 2:  x;  y ; xy    8; 8;9.5  MPa

The calculations are performed without partial coefficients. Both cases require according to table G.1 tension
reinforcement (ftdx = ftdy = 9 MPa for case 1 and ftdx = ftdy = 1.5 MPa for case 2) and cot = 1. Since the tensile
strength of concrete is neglected, then reinforcement has to yield in both cases and therefore, 1 and thus also
will in both cases assume the same value. This gives when guideline (b) together with Expressions
(G.9) –(G. 11) are used. The consequence is that in case 1, we can verify the compressive strength (c = 2xy =
18 MPa < fc = 0.61*30 = 18.3 MPa) while in case 2 we cannot (c = 2xy = 19 MPa > 18.3 MPa). The fact is,
however, that in case 2, we actually have a higher efficiency factor because the required tension reinforcement
provides a tensile strength (ftdx = ftdy = 1.5 MPa) smaller than the tensile strength of the concrete. The member is
therefore un-cracked. Therefore, in this case 1 to be inserted into Expression (G.9) should not be determined
by taking x and y as the yield strain for of the reinforcement. The most appropriate value forx and y would in
this case be fct/Ec. This would lead to an increase of to ca. 0.89 leading to c < fc.
It should here be noted, that Section 12 (Table 12.4) of FprEN-1992-1-1:2023 provides rules for minimum
reinforcement in walls designed according to Annex G. The minimum requirements may in some situations lead
to higher reinforcement demands than what is necessary according to Annex G. The reason for this is that the
method of Annex G requires the member to have an appropriate deformation capacity ensured through the
possibility to have a distributed crack pattern (smeared cracks). The suggested minimum reinforcement in Table
12.4 (0,25Ac*fctm/fyk on each surface) corresponds well with tests, see e.g. ref. [13].

Verification of expression (G.9)

To demonstrate the validity of Expression (G.9), test results for reinforced membranes subjected to pure shear
have been compared to the calculated strength. When performing the calculations, the yield condition for
membranes have been used, see Figure 5. It is noted that the yield condition and the design formulas in Table
G.1 are based on the same theoretical model. While the yield condition predicts the capacity based on known
reinforcement content, the design formulas predict the required reinforcement content based on known actions.
The so-called tension conical yield surface (protruding from 1st quadrant) in Figure 5 can in fact be obtained by
manipulations with the Expressions in Table G.1.
The calculation procedure used to compare predicted and tested strength are described below. Reference is
made to [2] for detailed explanations of the capacity expressions to be presented.

Figure 5: Yield condition for reinforced concrete membrane elements (from [2])
Due to the fact that the -factor (G.9) depends on the transverse strain and thereby the stress field inclination,
an iterative procedure must be used to calculate the shear capacity of membranes. In the following calculation

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 616
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

scheme, it is assumed that the reinforcement capacity in the x-direction is larger than the one in the y-direction,
i.e.  x f y , x   y f y , y (notation of Annex G). This assumption means that in case there is redistribution of
stresses, it will be the weak reinforcement (in the y-direction) that undergoes strains larger than the yield strain
while the strain in the strong reinforcement (in the x-direction) will be uniquely defined at the ultimate state. The
procedure includes the following steps:
----------ooOOoo----------
Step 1:
Make an initial guess of 

Step 2:
Calculate the mechanical reinforcement degrees:

 x f y,x  y f y, y
x  ; y 
 f c*  f c*
where

1

f c*  cc f c & cc  min  40 MPa
 
1/3
 fc
Here, cc is the brittleness factor as defined in clause 5.1.6(1) of prEN-1992-1-1:2021.

Step 3:
Calculate the shear capacity, u, the strut inclination, , and reinforcement strain, x, according to following
scheme:

IF: x  y 1:

x f y,x
 u   f c*  x  y ; cot   ;  sx  f y , x  x 
y Es

IF: x   y  1:

 *
 f c  y 1   y   ;  y  0.5

u  
1
  f c* ;  y  0.5
2

 1 
  y  ;  y  0.5
cot    y

1 ;  y  0.5

 f c* cos 2    sx
 sx   x 
x Es

Step 4:

Calculate and updated the value of :

1

1  1101
where:

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 617
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

1   x  ( x  0,001)cot 2 

Step 5:

Repeat steps 1 – 4 until the value in step 4 equals the value in step 1.

----------ooOOoo----------
Calculations have been carried out for benchmark test series published by Vecchio & Collins; Yamaguchi &
Naganuma; Vecchio et al.; Pang & Hsu and Zhang & Hsu [3 –7].

In order to evaluate Expression (G.9), only test specimens where the predicted strength is governed by
concrete crushing, combined with yielding of the reinforcement in the y-direction or combined with no rebar
yielding at all, are interesting. Specimens where the predicted strength is governed by yielding of reinforcement
in both directions (see solution above for the case  x   y  1 ) are not interesting, since this case is not
critical to concrete crushing.

The obtained results can be seen in Figures 6 – 10 where the ratio between tested and calculated shear
strength has been plotted versus strength parameters (y and fc) and calculated parameters (sx, cotand 1),
each one having and influence on the -value. Note that in the figures, the term “orthotropic reinforcement”
means  x f y , x   y f y, y While “isotropic reinforcement” means  x f y , x   y f y , y .For all tests, the mean

value of test/cal is 0.96 with a standard deviation of 13%.

Figure 6: test /cal. vs. mechanical reinforcement degree, y, for the weak reinforcement direction

Figure 7: test /cal. vs. calculated level of tension stress in reinforcement along the x-axis (not yielding)

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 618
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

Figure 8: test /cal. vs. calculated 1st transverse strain in the concrete

Figure 9: test /cal. vs. cylinder compresion strength of test specimens

Figure 10: test /cal. vs. calculated value of cot

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 619
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

Clause G.4(1)-(2)
To have an approach for design of shell elements, which is consistent with the provisions for membrane
elements, subsection G.4 adopt a classical sandwich model. This makes it possible to establish a set of
membrane stresses in the two layers of the sandwich model, which are statically equivalent with the in-plane
forces and the torsional and bending moments acting on the shell elements. Calculation of the reinforcement in
the two layers can then be performed according to subsection G.3.

Clause G.4(5)
For slabs, i.e. shells without in-plane forces, explicit design formulas are provided in Expressions (G.18-G.21)
as well as in Tables G.2 and G.3. The formulas and the rules for limitation of redistribution of moments are
analog to those applied to membrane elements. The theoretical background for these formulas may be found in
e.g. [2]. It is noted that the formulas in Tables G.2 and G.3 have been implemented in a number of commercial
finite element software packages. Finally, it is noted that there is a limitation on the torsional moment, |MEdxy | <
0,07d2fcd, which has to be fulfilled when using the Expressions (G.18 – G.21) to calculate the required
reinforcement. This is necessary because these expressions assume that the reinforcement required to carry
bending as well as torsion can be obtained from a classical sectional analysis. However, the torsional capacity
for slabs can be significantly lower than the flexural capacity if the reinforcement content is high. This is so
because there is a different in the internal lever arm for the two cases. Furthermore, in the case of torsion there
is a reduction of the concrete compression strength due to transverse cracking at the top as well as at the
bottom of the slab. The procedure to calculate the torsional capacity of slabs in a more accurate manner can be
found in e.g. Section 7.3 of Ref. [2]. Figure 11 shows calculated flexural capacity (mR) and calculated torsional
capacity (tR) of an isotropic reinforced slab element. As can be seen, for relatively low reinforcement content, mR
and tR are practically the same. This tendency continues up to a capacity of approximately mR = tR = 0,07d2fcd.
As the reinforcement content increases, much smaller torsional capacity is obtained compared to the
corresponding flexural capacity. Hence, for torsional moments larger than approximately 0,07d2fcd, it is unsafe to
use a flexural sectional analysis to calculate the reinforcement needed to carry combinations of bending and
torsion.

Figure 11: Calculated flexural capacity (mR) and torsional capacity (tR) of isotropic reinforced slab element. The
calculation of tR follows the procedure in [2], where As denotes the reinforcement area per unit length of the slab
cross section. The results are valid for  = 0.6 and d/h = 0,84, where h is the total depth.

Clause G.5
Clause G.5 provides guidelines for calculation of crack widths in membrane elements with skew cracks. The
provisions apply when the skew cracks have an inclination, , which is larger than 15o with the reinforcement
axes.

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 620
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

Figure 12: Fully rotational model showing normal strains in reinforced membrane element (averaged over
distance containing several skew cracks) and normal strains in the concrete between two cracks.

The provisions of Clause G.5 are primarily based on the Cracked Membrane Model [8], where the following
main assumptions should be mentioned:
1) The skew cracks are the result of a stress state that includes a diagonal compression field in the
concrete and tension in the reinforcement bars.
2) The reinforcement content is sufficient to ensure a smeared crack system (i.e. formation of more than
just one skew crack)
3) The direction of the compression field coincides with the direction of the cracks (rotating crack model
with stress free cracks)
4) Average total strains in the cracked membrane element = sum of average strains in concrete and
average strains due to crack opening.
5) Effect of tension stiffening is estimated based on a rigid-plastic bond model (bond between rebars and
concrete)
Figure 12 illustrates a cracked membrane element, which fulfills the above-mentioned assumptions. Here, 1
and 2 are the average principal strains (total strains) of the cracked element when measured over several
cracks while c1 and c2 are the average principal concrete strains between cracks. Due to the assumption 3), 2
= c2. Based on assumptions 3) and 4) it is possible to estimate the crack width, w, from the relationship: w/s =
(1 - c1), where s is the spacing between the skew cracks. Transformation from mean crack width to calculated
crack width, wk,cal, takes place by means of the factor kw defined in Clause 9.2.3(2) of FprEN-1992-1-1:2023.
Annex G provides Equation (G.22) to estimate the mean spacing between the skew cracks. This equation is
identical to Equation (5.12) in [8] as well as the one introduced in the fib model code 2010 [9]. The above-
mentioned strain difference (1 - c1) is in Annex G denoted as (sm - cm) and given by Equation (G.23). It is
noted that the right-hand side of Equation (G.23), with input parameters given in Equations (G.24) – (G.26), has
been deliberately written in a format that as much as possible resembles the format of Equation (9.11) in FprEN-
1992-1-1:2023 for the case of uniaxial tension. However, these equations are in fact the result of a
decomposition and rearrangement of the Equation (5.22) in [8] derived from the Cracked Membrane Model, in
which the Poisson’s ratio of concrete is assumed to be zero and where (1- ) has been approximate with (1- )
~ 1 (assuming the reinforcement ratio to be much smaller than unity). For details, the reader is referred to [8]. It
should be noted that in many practical situations, the term |2| in (G.23) is small compared to the other two
terms and may therefore be neglected. Finally, it is noted that the upper limit (1-kt)s/Es in Equations (G.24-25)
has been introduced to achieve consistency with Equation (9.11) and with experimental results. However, unlike
the case of uniaxial tension (i.e. Equation (9.11)), where the effective tensile strength of concrete is taken as fctm,
Equations (G.24-25) assume fct,eff = 0,5fctm. This value has been recommended here because many membrane
shear test series (see e.g. [13]) have shown that transition from un-cracked to cracked element takes place
already when the major principal stress 1 is in the order of 0,5 – 0,6fctm. The introduction of fct,eff = 0,5fctm in
Annex G is consistent with Table 12.4, where the rules for minimum reinforcement in walls subjected to in-plane
normal and shear stresses are based on fct,eff = 0,5fctm.
The inclination,, of the skew cracks may be determined by use of Equation (G.27). This equation simply
expresses the compatibility condition of a fully cracked membrane element (concrete tensile strength neglected)
subjected in-plane stresses x, y and xy. The equation can e.g. be derived by minimizing the complementary
elastic energy of the membrane element.
Calculated crack widths according to Clause G.5 (and with kw = 1.7, see Equation (9.8)) have been compared
with the maximum observed crack widths in membrane tests extracted from refs. [6, 10 - 12]. The results are
shown in Figure 13. When assuming a lognormal distribution, the mean value of the test-to-calculation ratio is
0,86 with a standard deviation of 0,41.

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 621
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

Figure 13: Calculated crack width compared with observed maximum crack widths in selected membrane tests

References

[1] Muttoni A, Ruiz MF, Pejatovic M: Background document to subsection 8.2.3 – Members requiring
design shear reinforcement. CEN/TC250/SC2/WG1/TG4, February 2021.
[2] Nielsen, MP, and Hoang, LC.: Limit Analysis and Concrete Plasticity. Third edition, CRC Press, Boca
Raton, FL, 796 pp.
[3] Vecchio FJ, Collins MP: The response of reinforced concrete to in plane shear and normal stresses.
The Department of Civil Engineering, University of Toronto, Canada, 1982.
[4] Yamaguchi T, Naganuma K.: Experimental study on mechanical characteristics of reinforced
concrete panels subjected to in-plane shear force . Journal of Struct. Constr. Engng., AIJ, No. 419,
1991.
[5] Vecchio FJ, Collins MP, Aspiotos J: High-Strength Concrete Elements Subjected to Shear. ACI
Structural Journal, Vol. 91, No. 4, July - August 1994. pp. 423-433.
[6] Pang XB, Hsu TTC: Behaviour of Reinforced Concrete Membrane Elements in Shear. ACI Structural
Journal, Vol. 92, No. 6, Nov - December 1995. pp. 665-679.
[7] Zhang LX, Hsu TTC : Behaviour and Analysis of 100 MPa Concrete Membrane Elements. Journal of
Structural Engineering, ASCE, Vol. 124, No. 1, Jan. 1998, pp. 24 – 34
[8] Kaufmann W: Strength and deformations of structural concrete subjected to in-plane shear and
normal forces. Doctoral thesis, Swiss Federal Institute of Technology Zürich, Institute of Structural
Engineering, 1998.
[9] Federation international du beton: fib Model Code for Concrete Structures 2010. Ernst & Sohn, 2013.
[10] Proestos, GT: Influence of high-strength reinforcing bars on the behaviour of reinforced concrete
nuclear containment structures subjected to shear. MSc-thesis, Department of Civil Engineering,
University of Toronto; 2014.
[11] Marti P, Meyboom J: Response of prestressed concrete elements to in-plane shear forces. ACI Struct
J 1992;89(5):503–14.
[12] Dyngeland T: Behaviour of reinforced concrete panels. PhD-thesis, Department of Structural
Engineering. Trondheim, Norway: Norwegian University of Science and Technology; 1989.
[13] Beck A: Paradigms of shear in structural concrete: Theoretical and experimental investigation.
Doctoral thesis, Swiss Federal Institute of Technology Zürich, Institute of Structural Engineering,
2021.

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 622
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex G Background to FprEN 1992-1-1: 2023

List of versions of this document

05.05.2021 1st draft, only background for ULS design
01.06. 2021 The 2nd draft includes background for Clause G.5(1) as well as guidelines for how to calculate the strain difference (sm--cm).
Title has been slightly modified.
22.12. 2022 The 3rd draft includes an update of the background for Clauses G.4(5) and G.5 in order to reflex the changes that were
introduced in the Formal Vote version as compared to the Enquiry version of the code. Professor Walter Kaufmann (ETH
Zurich) has kindly provided the theoretical input to G.5 and Dr. Reignad Tan (MultiConsult) has put great efforts in
performing the comparison of model with test results.
22.03.2023 editorial update to FprEN 1992-1-1:2023

L. C. Hoang 22.12.2022
For citations, copying and referencing see Note on Front page! page 623
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1 FprEN 1992-1-1:2023

Background document to Annex H

Guidance on design of concrete structures
for water-tightness

Alejandro Pérez Caldentey FHECOR, Spain

10.03.2023

For citations, copying and referencing see Note on Front page! page 624
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex H Background to FprEN 1992-1-1:2023

Introduction
The content of Annex H is mainly taken from EN 1992-3:2006. The content of this standard relating to water-
retaining structures has been incorporated into FprEN 1992-1-1:2023, part of in into Annex H, and part of it
into Section 9.2.3. In this background document this rearrangement of content will be documented. Parts of
EN 1992-3:2006 dealing with containment structures were discarded as CEN-TC250/SC2 considered that
this content was not concrete-specific and design issues of such structures built using concrete could be dealt
with using the content of the general part. For this reason, the scope of Annex H is limited to water-tightness
of concrete structures, although it may also be used for other liquids if it can be shown that their behaviour is
similar to that of water.
Some new content is provided by suggestion of a systematic review comment to EN 1992-1-1:2004 by the
Belgian Mirror Group. The suggestion was to reinstate content for ENV 1992-4:1998 regarding the calculation
of the leakage rate through cracks. The derivation of this proposal is also documented and discussed in this
BD.

Annex H Reasons for change

The integration of EN 1992-3:2006 into FprEN 1992-1-1:2023 is intended to increase ease-of-use and
consistency of Eurocode formulations on cracking. This allows a reduction in the document’s volume. For
instance, sections 2. Basis of Design, 3. Materials, 4. Durability, 5. Analysis, 8. Detailing provisions, 9.
Detailing of members and particular rules are not needed. Likewise, section 7.2.3, 7.2.4 and Annex M are
already covered by the general cracking model in Section 9.2.3.
Other paragraphs have been deleted due to the reduction in the scope. This is the case of Section 6, which
deals with dust explosions (and is not concrete-specific). Annex K which deals with the effects of low and
high temperatures on concrete is also deleted. If its contents were to be retained, they should be incorporated
into Section 5.
Annex L, dealing with simplified values of the restraint factor, has been deleted. The reasons for this are the
following:
– The Annex provides figures which take significant space and have a limited scope of application.
Such figures are better suited to be included in application guides.
– The restraint factor can be readily obtained from linear elastic FEM without significant effort, given
today’s computation means.

Enhancement of ease of use

As stated above, the integration of the most significant aspects of EN 1992-3:2006 into FprEN 1992:2023
increases ease of use by integrating into Section 9.2.3 important concepts related with imposed restrained
deformations such as how the element is restrained (at the edges or at the ends, or a mixture of both). This
makes the formulation for cracking more general and also serves as an early warning to engineers that such
concepts are to be accounted for. This integration is in line with the integration of the part on bridges into the
main text with a short annex for bridge particularities decided within TC 250/SC2 and finally approved by TC
250. The intention is to foster object-independent rules for design of concrete structures.

H.4.1 Classification Water Tightness Classes

Table H.1, which defines tightness classes is the same as Table 7.105 of EN 1992-3:2006. It provides a
useful classification of the expected performance of the structure with respect to water-leakage and allows to
regulate the corresponding design:
 For Tightness Class 0, normal crack width limits according to Section 9.2.1 are sufficient:
 For Tightness Class 1, for which a small amount of leakage as well as surface staining or damp
patches are accepted, through cracks are allowed to form, but the crack width is limited according
to the ratio of the hydraulic head to the thickness of the wall of the retaining structure. Also, a new
formulation to estimate water flow has been proposed in order to establish a quantitative measure
of what “a small amount of leakage” entails (see next section).
 For Tightness Class 2, for which appearance is not to be impaired by leakage, no through cracks
are allowed, unless other measures such as a lining or a water barrier are incorporated into design
 Finally, for Tightness Class 3, no leakage is permitted. This requires the use of a liner or the
avoidance of cracking by prestressing.

H.4.2 Tightness New formulation for leakage rate

Requirements
The new formulation for leakage rate was proposed by the Belgian Mirror Group and is based on assuming
a laminar flow through a crack that is assumed to have a rectangular section whose width is equal to an
equivalent crack width (wk,cal,e) and whose height is equal to the length of the crack (lw,p).

Pérez Caldentey 10.03.2023

For citations, copying and referencing see Note on Front page! page 625
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex H Background to FprEN 1992-1-1:2023

In order to derive the proposed expression (for the basic equation see (Vennard, 1947)), consider the
equilibium condition of a mass of water centered around the crack axis, as shown in Fig. 1. h is the depth of
the wall, and p1 and p2 are the water pressures at either end of the wall.


2 p1lw, p y x 2 p2 l w , p y


y
h

Fig. 1 Forces acting on a cylinder of water flowing through the crack

The equilibrium of forces acting on the water prism is as follows, where y represents half the height of the
water volume, and p= p2-p1 is the pressure difference:

y y
2 lw, p h  2plw, p y    p  p (1)
h h
This equation shows that the shear stress is proportional to the distance from the centre of the crack.

v=0

wk,cal,e vc
x

Fig. 2 Variation of velocity of flow as a function of the distance form an edge (taken from ref. [1])
Fig. 2 shows that the velocity of flow increases as the distance from an edge increases. At the edge the
velocity is zero. This variation in velocity is due to the existence of a shear stress between fluid layers. It has
been found, by experimental evidence, that this shear stress is proportional to the gradient of velocity in the
direction perpendicular to the flow. The constant of proportionality is known as the coefficient of viscosity (or
just viscosity for short) and is represented by symbol .

dv dv
     (2)
dy dy
By combining expressions (1) and (2) and applying them to a thin rectangular pipe of width wk,cal,e the following
differential equation is obtained

y dv
  p   
h dy
y v
y p
 h 0
p dy  dv  ydy    dv 
h vc
(3)
p y 2
 vc  v  v  vc  Ky 2
h 2
wk ,cal ,e 4  y2 
y v0K  vc  v  vc 1  4 2 
2 wk2,cal ,e  wk ,cal ,e 

The mean velocity of water in the pipe, can be determined from the definition of flow rate:

Pérez Caldentey 10.03.2023

For citations, copying and referencing see Note on Front page! page 626
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex H Background to FprEN 1992-1-1:2023

wk ,cal ,e

2
y2 
q  wk ,cal ,elw, p vm   vdA  lw, p  vc 1  4 2  dy
 wk ,cal ,e 
A

wk ,cal ,e 
2
wk ,cal ,e
 4 y3  2 2 (4)
 lw, p vc  y  2   vc wk ,cal ,elw, p
 3 wk ,cal ,e   wk ,cal ,e 3
2

3 2
 vc  vm  vm  vc
2 3
From Eq. (3), by using the formula for the velocity at y=wk,cal,e/2, where v=0, an expression can be obtained
for the flow rate, accounting for the definition of q as the section area times the mean velocity:
2
p wk ,cal ,e 3 3 q
 vm  
h 8 2 2 l p , w wk ,cal ,e
(5)
1 l 1
q  wk3,cal ,e p p , w
12 h 
This equation, which is the classical equation for Poiseuille flow between two parallel planes, is similar to Eq.
(A107.1) of ENV 1992-4:1998. However, the coefficient proposed by this Standard is 1/50 (instead of 1/12),
with a value that can be increased from 1/50 to 0 in cracks with a calculated crack width smaller than 0.2 mm
where self-healing occurs.
The difference in the coefficient is in that the derivation does not account for the tortuosity and roughness of
the cracks. For small cracks, such roughness would represent a significant part of the channel width.
(Chivers, 2002) [1] presents a study of the influence of roughness of the crack on the flow rate. Based on
(Jones, Wooten, & Kaluza, T.J., 1988) [2], (Button, Grogan, Chivers, & Manning, 1978) [2] and (Gardiner &
Tyrrell, 1968) [4], (Chivers, 2002) fixes an upper bound for the friction coefficient (f) in Darcy’s equation for
laminar flow equal to 96/Re, where Re is Reynolds number.
Based on this value, an alternative derivation can be made directly from Darcy’s empirical equation, adapted
for the case of a thin rectangular section (see Eq. (6)).

p h vm2
hL   f (6)
 water g wk ,cal ,e g
where:
f is Darcy’s friction coefficient
g is the acceleration of gravity on earth
Reynolds’ number can be expressed as follows for the case of a narrow rectangular pipe:

 water
Re  v w (7)
 m k ,cal ,e
Substituting the upper bound value of f in Eq.(6), and accounting for the expression of Reynold’s number, the
proposal of FprEN 1992-1-1:2023 can be readily derived, as shown in Eq. (8).

p 96 h vm2 96 h vm2

 
 water g Re wk ,cal ,e g  water vm wk ,cal ,e wk ,cal ,e g
1 2 p 1  (8)
 vm  wk ,cal ,e l 1
h    q  wk3,cal ,e p w, p
1
96
 96 h 
q  vm wk ,cal ,elw, p 
Expression (H.2), regarding the equivalent crack width comes directly from Eq. (A107.2) of ENV 1992-4:1998.

Pérez Caldentey 10.03.2023

For citations, copying and referencing see Note on Front page! page 627
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex H Background to FprEN 1992-1-1:2023

References Bibliografía
[1] Button, B., Grogan, A., Chivers, T., & Manning, P. (1978). Gas Flow Through Cracks. ASME J. Fluids
Engng, 100, 453-458.
[2] Chivers, T. (2002). The influence of surface roughness on the fluid growth trhough cracks. Fatigue Fract
Engng Mater Struct, 25, 1095-1102.
[3] Gardiner, G., & Tyrrell, R. (1968). The flow resistance of experimental models of naturally ocurring cracks.
Proc. Inst Mech Engrs., 200, No. C4.
[4] Jones, T., Wooten, S., & Kaluza, T.J. (1988). Single Phase Flow Through Cracks. 63rd Annual Technical
Conference of the Society of Petroleum Engineers (pág. Paper SPE 18175). Houston, USA: SPE.
doi:https://doi.org/10.2118/18175-MS
[5] Vennard, J. K. (1947). Elementary Fluid Dynamics. New York: John Wiley & Sons, Inc.

List of versions of this document

27.05.2021, APC. 1st draft
03.01.2023, APC 2nd draft – Adapted to final version of FprEN 1992-1-1:2023
10.03.2023, APC Editorial update following Peer review.

Pérez Caldentey 10.03.2023

For citations, copying and referencing see Note on Front page! page 628
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

CEN/TC 250/SC 2 FprEN 1992-1-1:2023

Background document to Annex I:

Assessment of Existing Structures

Fabrizio Palmisano PPV Consulting & Politecnico di Torino,

(all sections) Italy

John Cairns Heriot-Watt University,

(I.11.4.1(7)) Edinburgh, UK

Craig Giaccio Arcadis Consulting Ltd

(deteriorated structures) UK

Rita Greco Politecnico di Bari,

(I.11.4.1(3)-(5) and I.11.4.1(7)) Italy

Giuseppe Carlo Marano Politecnico di Torino,

(I.11.4.1(7)) Italy

Aurelio Muttoni EPFL, Ecole Polytechnique Fédérale de

Lausanne
(I.5.2.1(3))
Switzwerland

27.03.2023

For citations, copying and referencing see Note on Front page! page 629
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

CEN/TC 250/SC 2 Annex I to FprEN 1992-1-1:2023

Background document to clause I.5.2.1(3)

Evaluation of fck from in-situ testing

Fabrizio Palmisano PPV Consulting & Politecnico di Torino,

Italy

Aurelio Muttoni EPFL, Ecole Polytechnique Fédérale de

Lausanne
Switzwerland

27.03.2023

For citations, copying and referencing see Note on Front page! page 630
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Introduction
When concrete strength is assessed in an existing structure, according to clause I.5.2.1(1) of FprEN 1992-1-
1:2023 the characteristic value 𝑓 , of the in-situ compressive strength of concrete cores should be evaluated
according to EN 13791:2019. However, in the main part of FprEN 1992-1-1:2023, all formulations that include
concrete strength are expressed in terms of 𝑓 . This is why there is the necessity to include a formula for the
evaluation of 𝑓 on the basis of the assessed 𝑓 , .

Definitions
𝑓 Compressive strength of concrete determined from samples of concrete taken in accordance with EN
12350-1:2019 made into cylinder or cube specimens and cured in accordance with EN 12390-2:2019
and tested in accordance with EN 12390-3:2019.
𝑓 Characteristic value of 𝑓 .
𝑓, Actual in-situ concrete compressive strength in the structure.
𝑓, Compressive strength of concrete cores taken at test location within a structural member, expressed
in terms of the strength of a 2:1 core of diameter ≥ 75 mm (5% fractile) and assessed according to
Section 8 of EN 13791:2019.
𝑓 , Characteristic value of 𝑓 , .

Basis of the formulation

Following the approach included in Annex A of FprEN 1992-1-1:2023, 𝑓 , can be defined as:
𝑓, 𝜂 ∙𝑓, 𝜂 ∙𝜂 ∙𝑓 (1)
where
 𝜂 is a reduction factor which accounts for the damage sustained during core extraction;
 𝜂 is a coefficient which covers the following variabilities and uncertainties:
(i) effect of curing;
(ii) effect of consolidation of fresh concrete (settlement and bleeding) leading to a significant geometric
variability within the structure;
(iii) anisotropy (e.g. difference between vertical and horizontal strength resulting from bleeding);
 𝜂 , 𝜂 , 𝑓 are assumed as random variables, log-normally distributed, stochastically independent.
In Annex A of FprEN 1992-1-1:2023 the average value and the coefficient of variation for 𝜂 are 𝜇 0.95
and 𝑉 0.12, respectively, whilst the coefficient of variation 𝑉 of 𝑓 is equal to 0.10. According to several
works (see e.g. Moccia at. al. 2020) the mean value 𝜇 of 𝜂 is about 0.95 and the relevant
coefficient of variation 𝑉 is negligible with respect to the coefficients of variation of the other variables
that have influence on concrete strength.
Following the above-mentioned assumptions on the statistical distributions of all variables involved and on the
relevant parameters, 𝑓 is equal to

. ∙
,
𝑓 𝑓 , (2)
∙ .

that is similar to that given in EN 13791:2019:

,
𝑓 (3)
.

In this scenario, in Annex I the formulation of EN 13791:2019 has been assumed if the cores are extracted
from the region governing for the verification (see case c) of table I.2).
If the cores are extracted from regions not necessarily representing the conditions of whole structural member
nor the region governing for the verification, the effect on 𝜂 of the position of the cores, within the structural
member, should be taken into account.
Figure 1, taken from Moccia et al. (2020) and based on several test results present in the literature, shows the
influence of the position of cores taken from columns on 𝜂 and in particular
 top layer (upper 20% of the depth, red markers in the figure): 40 values, 𝜇 0.95, 𝑉 0.095;

 bottom layer (bottom 20% of the depth, black markers in the figure): 23 values, 𝜇 1.05, 𝑉 0.105.

Thus, if the cores have been extracted from regions not necessarily representing the conditions of the whole
structural member nor the region governing for the verification, there is a risk that the region with the highest
μ is assessed (for instance μ 1.05), whereas the governing verification region could require e.g. μ
0.95. In this case, a safe estimate of 𝑓 from 𝑓 , would be
. , ,
𝑓 ∙ (4)
. . .

For citations, copying and referencing see Note on Front page! page 631
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

that is case a) of table I.2.

An intermediate case can be found when cores are extracted from different regions representing all conditions
in the structural member. In this case
,
𝑓 (5)
.

can be assumed (i.e. case b) of table I.2)

Figure 1. Coefficient 𝜂 in bottom and top layers of columns related to concrete cylinder strength 𝑓 ,
(Moccia et al. 2020)

References
Moccia, F., Kubski, X., Fernandez Ruiz, M., Muttoni, A. (2020). The influence of casting position and
disturbance induced by reinforcement on the structural concrete strength, Structural Concrete, 2020, 1-28.

For citations, copying and referencing see Note on Front page! page 632
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

CEN/TC 250/SC 2 Annex I to FprEN 1992-1-1:2023

Background document to some clauses of

sections I.4, I.5, I.8, I.9

Fabrizio Palmisano PPV Consulting & Politecnico di Torino,

Italy

27.03.2023

For citations, copying and referencing see Note on Front page! page 633
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

I.4.2.1(2) The approach proposed in this clause is consistent with that included in Annex A of FprEN 1992-1-1:2023. If
the coefficient of variation of concrete core strength is greater than the values given un table I.1, the relevant
Adjusted partial factors
partial factor for concrete c, evaluated by using the procedure given in Annex A, is greater than 1.5. This is
why, in this case, it is recommended to adjust c, according to the procedure given in Annex A. Similarly, If the
coefficient of variation of reinforcement yield strength is greater than the values given un table I.2, the relevant
partial factor for reinforcement s should be adjusted by using the procedure given in Annex A since it is greater
than 1.15.

I.5.2.2(1) Refer to:

Effect of stress level Tasevski D., Fernández Ruiz M., Muttoni A. 2019. ‘Assessing the compressive strength of concrete under
and duration of loading sustained actions: From refined models to simple design expressions.’ Structural Concrete, 20, pp. 971–985.
on concrete strength

I.5.3.(2) The formulation included in the clause is consistent with that given in D.7.2 of prEN 1990 draft April 2022, for
log-normal distribution. The clause includes also the table relevant to 10% characteristic value (not included in
Characteristic value of
prEN 1990:2022) for the evaluation of ductility properties of reinforcement.
the properties of
reinforcing steel Table I.5 has been obtained by using the so-called ‘prediction method’ that is the same as that used in Annex
D of prEN 1990:2020 (see Appendix C of Gulvanessian, Calgaro, Holický 2002, ‘Designers’ Guide to Eurocode:
Basis of Structural Design: EN 1990’, Thomas Telford Publishing, London).

I.8.3.2(3) The clause gives a conservative and simplified formulation for the evaluation of shear strength of beams and
slabs having transverse spacing of shear legs greater than that given (as maximum value) in clause 12 of
Shear strength in case
FprEN 1992-1-1:2023. The starting point of this clause is the experimental study by Lubell et al. 2009 (i.e.
of substandard
Lubell A.S., Bentz E.C., Collins M.P. 2009, ‘Shear Reinforcement Spacing in Wide Members’, ACI Structural
transverse spacing of
Journal, Vol. 106, No. 2).
shear legs
The formulation considers that, in the transverse direction, the concrete part of beam/slab that contributes to
shear strength is not greater, per each side of the leg, than half the maximum leg distance given in clause 12
of FprEN 1992-1-1:2023.

I.8.3.3 The clause gives a conservative and simplified formulation that follows the same approach as that of clause
I.8.3.2(3).
Shear at interfaces in
case of substandard
spacing between
interface reinforcement
in a composite slab

I.9.1(2) In the assessment, low values of the target reliability index 𝛽 for non-seismic ultimate limit states can be
allowed. In this scenario fulfilling ULS requirements can also imply that high stresses occur under service load
Limits on reinforcement
situations. This is why stresses in the serviceability limit state should be limited to avoid yielding of the
and concrete stress for
reinforcement and development of micro cracks and cracks in concrete.
low target reliability
index at ULS The stress limits in table I.6 are the same as those in tables 9.1 and 9.2 of FprEN 1992-1-1:2023. However,
explicit reference to these tables has not been made since the stress verifications in table 9.2 are limited to
some exposure classes while in I.9.1(2), for the above-mentioned reasons, they don’t depend on the exposure
classes.

For citations, copying and referencing see Note on Front page! page 634
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

CEN/TC 250/SC 2 Annex I to FprEN 1992-1-1:2023

Background document to clauses

I.11.4.1(3)-(5)

Anchorage and lap length of plain surface

bars

Fabrizio Palmisano PPV Consulting & Politecnico di Torino,

Italy

Rita Greco Politecnico di Bari,

Italy

27.03.2023

For citations, copying and referencing see Note on Front page! page 635
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Introduction
This section provides the background to clauses (3)-(5) of I.11.4.1 and details the derivation of the closed-form
equation for the design anchorage and lap length of plain surface bars of round and square cross section.

Basis of the formulation

The formulation included in Annex I is based on the semi-empirical one proposed by Palmisano et al. (2020).
This, by using as starting point the proposal by Cairns and Feldman (2018) and a reference test database,
proposed a novel step-by-step procedure for the statistical evaluation and optimization of the contribution of each
parameter affecting anchorage and lap strength. The reference database includes 101 test results on plain bars
of which 54 were conducted on bottom cast bars and 47 on top cast bars (i.e. ‘good’ and ‘poor’ bond casting
position respectively, denoted GP and PP, respectively in the following). 17 test results are relevant to square
section bars and the remainder to round section bars.
The equations proposed by Palmisano et al. (2020) are:
. . .
GP 𝑓 11.66MPa ∙ ∙ ∙ (1)

. . .
PP 𝑓 4.13MPa ∙ ∙ ∙ (2)

where 𝑓 is the mean value of the axial stress 𝑓 (called anchorage/lap strength in the following) developed in
the bar by bond (MPa); 𝑙 is the anchorage/lap length; 𝑓 is the mean cylinder concrete compressive strength
(MPa); 𝑐 is the concrete cover; 𝜙 is the equivalent bar diameter assumed as the bar diameter for round section
bars and as

𝜙 𝐴 (3)

for square section bars (𝐴 is the area of the cross section of the square section bar).
The approach proposed by Palmisano et al. (2020) is based on the statistical evaluation of test results. To this
aim model uncertainties, that account for random effects that are neglected in the models and simplifications in
the mathematical relations, have to be considered. According to JCSS (2001), the most common way to introduce
model uncertainties in the calculation model is to use an additive equation or a multiplicative equation or a
combination of both. In this case the following multiplicative relationship has been chosen as proposed by Mancini
et al. (2018):
𝑦 𝜃 ∙𝑦 , (4)
where 𝜃 is the model uncertainty of the ith test, 𝑦 is the ith test result and 𝑦 , represents the relevant value
resulting from the assumed model.
Moreover, taking into account that the proposed equations have a nonlinear form, the coefficient of determination
𝑅 is evaluated by using the following equation:
∑ ,
𝑅 (5)
∑

where 𝜇 is the mean value of the natural logarithm of test results.

In table 1, the mean value 𝜇 and the coefficient of variation 𝑉 of model uncertainties are reported together with
the coefficient of determination 𝑅 . In this table 𝑅 and 𝜇 are equal to 1 as consequence of the statistical
optimisation procedure performed by Palmisano et al (2020).
Figure 1 compares the bar stress 𝑓 , , predicted by equations (1) and (2), with that (𝑓 , ) measured in
tests. The figure shows a good fit of the proposed equations to test results that is mainly the consequence of the
fact that the values of 𝑅 and 𝜇 are equal to 1.
All the details about the derivation of formulae (1) and (2) as well as the comparison with those proposed by
Cairns and Feldman (2018) are included in Palmisano et al. (2020).

Table 1. Fit of equations (1) and (2) to test data.

Equation Casting position 𝝁𝜽 𝑽𝜽 𝑹𝟐
1 GP 1.00 0.18 1.00
2 PP 1.00 0.36 1.00

For citations, copying and referencing see Note on Front page! page 636
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

(a)
400
Good bond
350 casting
position
300

250

fstm,TEST [MPa]
200
Round bars
150 Square bars

100

50

0
0 50 100 150 200 250 300 350 400
fstm,MOD [MPa]

(b)
400
Poor bond
350 casting
position
300

250
fstm,TEST [MPa]

200 Round bars

Square bars
150

100

50

0
0 50 100 150 200 250 300 350 400
fstm,MOD [MPa]

Figure 1. Comparison between bar stress estimated by equations (1) and (2) and that measured in tests. (a)
equation (1) for GP. (b) equation (2) for PP.

Design anchorage length of plain bars

In this paragraph, starting from the equations (1) and (2), the reliability-based design formulations of the
anchorage strength of plain bars are obtained.
The calculation methodology adopted in the following is the one proposed by Taerwe (1993) and furtherly applied
to anchorage strength of ribbed bars by Mancini et al. (2018).
The following assumptions have been made:
 the concrete compressive strength 𝑓 has lognormal distribution with mean value 𝑓 according to
Diamantidis et al. (2016), JCSS (2001), Mancini et al. (2018); its coefficient of variation 𝑉 has been
assumed equal to 0.10 according to table A.2 of FprEN 1992-1-1:2023;
 a lognormal distribution, with 𝑉 = 0.18 and 𝑉 = 0.36 for GP and PP, respectively, and 𝜇 1 is assumed
for model uncertainties according to Mancini et al. (2018) and table 1.
According to Taerwe (1993) and taking into account the format of equations (1) and (2), the overall anchorage
strength random variable 𝑓 may be defined by the following equation:

𝑓 𝜃∙𝑓 ∙𝐴 (6)
where A represents the multiplicative term relevant to all deterministic parameters of equations (1) and (2)

𝐴 𝛿 𝑀𝑃𝑎 ∙ ∙ ∙ (7)

and 𝛿 , 𝛿 , 𝛿 , 𝛿 are the coefficient/exponents of equations (1) and (2), summarised in table 2.

For citations, copying and referencing see Note on Front page! page 637
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Casting
Equation 𝜹𝟏 𝜹𝟐 𝜹𝟑 𝜹𝟒
position

1 GP 11.66 0.45 0.77 0.78

2 PP 4.13 0.35 0.94 0.89

Table 2. Coefficient/exponents of equations (1) and (2).

By applying the procedure shown in Taerwe (1993) to equation (6), the fractiles 𝑓 of the variable 𝑓 can be
evaluated:

𝑓 𝐴∙ ∙𝑓 2
∙ exp 𝛿2 ∙ ℎ ln 1 𝑉2 ℎ ln 1 𝑉 2
𝛿2 2 𝑙𝑛 1 𝑉2 (8)
1 2

where j = m, k, d with:
 m = mean value;
 k = characteristic value assumed as the 5% fractile;
 d = design value relevant to a specific reliability index β;
and coefficients hj are:
 hm = 0 for the mean value;
 hk = 1.645 for the characteristic value;
 hd = αRβ for the design value, where αR is the FORM (First-Order Reliability Method) correction factor
assumed equal to 0.8 for dominant resistance variables (JCSS 2001).
Equation (8) can be rewritten to express directly the design anchorage length 𝑙 as a function of the design axial
stress 𝜎 in the reinforcement and other relevant parameters in a format similar to that given FprEN 1992-1-
1:2023 for ribbed bars:

. ∙
𝛥 ∙ ∙ ∙ (9)

where

∙
𝛥 (10)
.

𝜁 ∙ exp 𝛿 ∙ ℎ ln 1 𝑉 ℎ ln 1 𝑉 𝛿 𝑙𝑛 1 𝑉 (11)

If hd,=3.8 is the value of hd evaluated for  = 3.8, equation (9) can be rewritten, for practical application, as a function
of the concrete partial factor C:

. ∙
𝛥 , . ∙ ∙ ∙ ∙ (12)
.

where c = 1.5 for  = 3.8, and

∙ , .
𝛥 , . (13)
.

𝜁 , . ∙ exp 𝛿 ∙ ℎ ln 1 𝑉 ℎ , . ln 1 𝑉 𝛿 𝑙𝑛 1 𝑉 (14)

The exponent 5 has been obtained by evaluating lbd, by equation (9), for values of  from 3.3 to 5.2 (i.e. from low
consequences of failure to a failure probability equal to 10-7 according to EN 1990:2002) and imposing a ratio of
equation (12) to equation (9) in the range 0.96-1.04 to make negligible the difference between the rigorous
approach (i.e. equation (9)) and the practical one (i.e. equation (12)). To this aim the formulation of C, included
in annex A of FprEN 1992-1-1:2023 (with the parameters included in table A.3), has been assumed.
By using this approach and the coefficients/exponents of equations (1) and (2), equation (12) yields equations
(15) and (16) for GP and PP, respectively.
. . . ∙ . .
132.9 ∙ ∙ ∙ ∙ (15)
.

. . . ∙ . .
414.8 ∙ ∙ ∙ ∙ (16)
.

For citations, copying and referencing see Note on Front page! page 638
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

As the original equations have a semi-empirical origin, it is reasonable to round the coefficient/exponents in order
to have the final design equations. Hence equations (15) and (16) can be rewritten and compacted in the final
equation (17) to be used for practical application.

. ∙
130 ∙ 𝜂 ∙ ∙ ∙ ∙ (17)
.

where
 1 = 2 = 3 = 4 = 1 for GP;
 1 = 3.1, 2 = 1.6, 3 = 0.9, 4 = 3/5 for PP.
The following limitations should be applied to equations (15)-(17):
 the ratio 1.5eq/c should not be taken smaller than 0.5 in view of absence of test data in the reference
database beyond c/eq = 3.0;
 the application of the equations should be limited to sd ≤ 300 MPa since in the reference database most
of the tests have fstm lower than 280 MPa with just a few having fstm of about 350-400 MPa;
 the equations are valid for c/eq ≥ 1.0 in view of absence of test data in the reference database below
c/eq = 1.0;

 the ratio should be taken not lower than 10, since the relevant minimum value of the reference test
database is about 12.

Comparison with provisions of CEB-FIP Model Code 90

The current Eurocode 2 (EN 1992-1-1:2004) does not provide bond length for plain bars. The CEB-FIP Model
Code 90 was the precursor to the current Eurocode 2 and contained the following equation 41 for anchorage
length of plain bars lbd,MC90.
∙
𝑙 , (18)
∙

where
. /
𝑓 1.0 ∙ 𝜂 ∙ 𝜂 ∙ (19)

 η2 is equal to 1.0 and 0.7 for good and poor bond casting position, respectively;
 η3 = 1.0 for ϕ ≤ 32 mm.
In the following the differences between the proposed practical approach (i.e. equation (17)) and that of CEB-FIP
Model Code 90 are studied by analysing the ratio lbd/lbd,MC90 for c = 1.5.
Figures 2 and 3 show the influence of c/ for sd = 200 MPa on lbd/lbd,MC90 whilst figures 4 and 5 show the influence
of sd for fck = 25 MPa on lbd/lbd,MC90.
Figures 3 and 5 highlight that for PP the equation of CEB-FIP Model Code 90 significantly underestimates the
anchorage length. On the other side, for good bond casting position (see figures 2 and 4) this underestimation is
only relevant to low c/ (i.e. lower than 2) while for c/ > 2.0 the equation of CEB-FIP Model Code 90 is
conservative.

For citations, copying and referencing see Note on Front page! page 639
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

2.0 c/=
c/f=1 1.0
c/=
c/fi=1.5 1.5
lbd/lbd,MC90 c/= 2.0
c/fi=2
c/= 2.5
c/fi=2.5

1.5 c/= 3.0

c/fi=3

1.0

0.5

0.0
0 10 20 30 40 50 60
fck [MPa]

Figure 2. Influence of c/ on lbd/lbd,MC90 in case of c = 1.5, sd = 200 MPa and GP.

5.5 c/=
c/f=1 1.0
lbd/lbd,MC90 c/=
c/fi=1.5 1.5
5.0
c/= 2.0
c/fi=2
4.5 c/= 2.5
c/fi=2.5
c/= 3.0
c/fi=3
4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0
0 10 20 30 40 50 60
fck [MPa]

Figure 3. Influence of c/ on lbd/lbd,MC90 in case of c = 1.5, sd = 200 MPa and PP.

For citations, copying and referencing see Note on Front page! page 640
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

2.5 c/=
c/f=1 1.0
lbd/lbd,MC90 c/=
c/fi=1.5 1.5
c/= 2.0
c/fi=2

2.0 c/= 2.5

c/fi=2.5

c/= 3.0
c/fi=3

1.5

1.0

0.5

0.0
0 50 100 150 200 250 300 350 400
sd [MPa]

Figure 4. Influence of sd on lbd/lbd,MC90 for fck = 25 MPa and GP.

5.0 c/=
c/f=1 1.0
lbd/lbd,MC90 c/=
c/fi=1.5 1.5
4.5
c/= 2.0
c/fi=2

4.0 c/= 2.5

c/fi=2.5
c/= 3.0
c/fi=3
3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0
0 50 100 150 200 250 300 350 400
sd [MPa]

Figure 5. Influence of sd on lbd/lbd,MC90 for fck = 25 MPa and PP.

References
Cairns, J., Feldman, L. (2018). Strength of laps and anchorages of plain bars. Structural Concrete, 19(6), 1782-
1791.
CEB-FIP Model Code 90. 1993, Lausanne.
EN 1990:2002. Eurocode: Basis of Structural Design.
EN 1992-1-1:2004. Eurocode 2: Design of Concrete Structures - Part 1-1: General Rules and Rules for Buildings.
Diamantidis, D., Holický, M., Sýkora, M. (2016). Reliability and risk acceptance criteria for civil engineering
structures. Transactions of the VŠB – Technical University of Ostrava, Civil Engineering Series, 16 (2),
1–10.
fib (2010). fib bulletin 52. Structural Concrete Textbook on behaviour, design and performance, Second edition
Volume 2: Basis of design. 2010. Lausanne: International Federation for Structural Concrete (fib).
JCSS (2001). Probabilistic Model Code. Lyngby, Denmark: Joint Committee on Structural Safety.

For citations, copying and referencing see Note on Front page! page 641
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Mancini, G., Carbone, V.I., Bertagnoli, G., Gino, D. (2018). Reliability-based evaluation of bond strength for tensed
lapped joints and anchorages in new and existing reinforced concrete structures. Structural Concrete,19,
904–917.
Palmisano, F., Greco, R., Biasi, M., Tondolo, F., Cairns, J. (2020). Anchorage and laps of plain surface bars in
R.C. structures. Engineering Structures, Vol. 213, 110603/1-12.
Taerwe, R.L. (1993). Towards a consistent treatment of model uncertainties in reliability formats for concrete
structures. CEB Bulletin d’Information No 219, Safety and Performance concepts (pp. 5-61), Lausanne:
CEB.

For citations, copying and referencing see Note on Front page! page 642
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

CEN/TC 250/SC 2 Annex I to FprEN 1992-1-1:2023

Background document to clause I.11.4.1(7)

Anchorage and laps in case of substandard

cover

Fabrizio Palmisano PPV Consulting & Politecnico di Torino,

Italy

John Cairns Heriot-Watt University,

Edinburgh, UK

Rita Greco Politecnico di Bari,

Italy

Giuseppe Carlo Marano Politecnico di Torino,

Italy

27.03.2023

For citations, copying and referencing see Note on Front page! page 643
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Introduction
This section provides the background to clause I.11.4.1(7) and details the derivation of parameters included in
table I.7. Further details are included in Palmisano et al. (2020a, 2021).

Basis of the formulation

The approach proposed by Palmisano et al. (2020a), that yields the formulation included in Annex I, draws
inspiration from that of Metelli and Plizzari (2014) included in SIA (2011). In particular Zwicky (2013) takes account
of the influence of low concrete cover thickness by studying the model proposed by Schenkel (1998) and referring
it to the concrete cover limitations given in SIA (2003).
For the approach by Palmisano et al. (2020a), the more recent model included in fib (2010), that is the basis for
the evaluation of anchorage length in FprEN 1992-1-1:2023, is assumed as reference.
The reference model included in fib (2014), based on the proposal by Canbay and Frosch (2005), is the strut-and-
tie model (STM) shown in figure 1.
The model of figure 1 is then modified by Palmisano et al (2020a), as shown in figure 2, in case of concrete cover
loss. In the modified STM the tensile action that causes bond failure is doubled (𝐹 𝑡𝑎𝑛𝛽 in Fig. 1 whilst
𝐹 𝐹 𝑡𝑎𝑛𝛽 in Fig. 2).
Assuming cd as the nominal concrete cover taken as the minimum between cx, cy, cs/2 according to figure 3, in
the proposed approach the following assumptions are made:
(i) the models in figures 1 and 2 are assumed as reference;
(ii) the design anchorage strength fstd (axial stress in the bar) is reached when cd is at least equal to the bar
diameter ;
(iii) if cd <  the anchorage strength decreases proportionally to the reduction of cd.
Hence, in the proposed approach, if cd <  the anchorage strength of ribbed and plain bars is evaluated replacing
cd by the bar diameter and decreasing fstd by a factor k’lbs,c that is defined in the following.

Figure 1. Reference STM for the evaluation of bond strength (fib 2014).

Figure 2. Reference STM for the evaluation of bond strength in case of concrete cover loss.

Figure 3. Definition of cs, cx, cy, corner bar (1), edge bar (2).

For citations, copying and referencing see Note on Front page! page 644
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

The lower limit of the concrete cover for the application of formulae 11.3 and I.26 of FprEN 1992-1-1:2023 without
any modification is chosen equal to  for the following reasons:
 to be consistent with the limit given in clause 6.5.2.3 of FprEN 1992-1-1:2023;
 to be consistent with the minimum value of concrete cover in the test database for anchorage length of
plain bars assumed as reference for formula I.26 of prEN 1992-1-1:2023;
 even if the minimum value of concrete cover c in the reference test database (fib 2014) for the anchorage
length of ribbed bars is 0.5, the number of tests performed with c = 0.5, is too low to have a statistical
validity; as stated in fib (2014), tests with c ≤ 0.5 were excluded from the statistical summary;
 differently from other approaches present in the literature, the approach here proposed gives a general
procedure valid for all cases of low concrete cover thickness.
Tables 1 and 2 give the vales of k’lbs,c according to Palmisano et al. (2020a) for edge and corner bars, respectively,
as defined in figure 3.

Table 1. k’lbs,c for edge bars.

Case no. Description k'lbs,c

𝑐
1 0.5 0.5
2𝜙

cs/2 < cy  

𝑐
2 0.5 0.5
𝜙

cs/2  cy < 

𝑐
3* 0.5 0.5
𝜙

cs/2 < cy < 

* In case No. 3, bars having cs/2 <  shall be considered as they are in bundle.

For citations, copying and referencing see Note on Front page! page 645
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Table 2. k’lbs,c for corner bars.

Case no. Description k'lbs,c

𝑐
1 0.5 0.5
𝜙

cs/2  cx  cy < 

𝑐
2 0.5 0.5
𝜙

cs/2  cx < cy  

if 1/3  cmin,xy/ < 1

,

3*
if cmin,xy/ < 1/3
,
0.1 0.7

cs/2  cx < cy < 

𝑐
4 ** 0.5 0.5
𝜙

cs/2 < cx  cy < 

𝑐
5 ** 0.5 0.5
𝜙

cs/2 < cx < cy  

* In case No. 3 cmin,xy = min (cx; cy)
** In cases No. 4 and 5, bars having cs/2 <  shall be considered as they are in bundle.

Validation of the formulation

The formulation described in the previous paragraph and proposed by Palmisano et al. (2020a) is merely
theoretical and based on the assumptions that the reduction of anchorage strength is proportional to the reduction
of the concrete cover. From a general point of view, taking account of the way the formulae for anchorage strength
by fib (2014) and Palmisano et. al (2020b) (that are the bases of formulae 11.3 and I.26 of FprEN 1992-1-1:2023,
respectively) consider the effect of concrete cover, a linear reduction seems to be conservative. However, this
consideration needs to be experimentally verified. To this aim, in the following, tests present in the literature, that
have studied the influence of low concrete cover on bond, are firstly selected and, then, used for the validation of
the proposed formulation (Palmisano et al. 2021).
It is worth highlighting that the validation included in this section, due to the type of tests present in the literature,
is made only with reference to case 2 of table 1 and to case 1 of table 2 even if, the relevant results can be
extended to the other cases.
In order to validate the equations proposed by Palmisano et al. (2020a), it is necessary to select a reference
database in which the influence of low concrete cover thickness on bond has been studied. The data used in the
validation have been selected among the investigations carried out by Schenkel (1998), Regan and Reid (2009),
Cairns and Goodchild (2016) and Darby et al. (2016).
The database is composed of both beam-end anchorage and lap splice tests on plain and ribbed bars.
Pullout tests have not been considered because, having the bar in tension and the surrounding concrete
compressed, they are unrepresentative of practical conditions. Tests carried out with corroded bars have been
neglected as well as tests with bars that reached stress well in excess of yield (Cairns and Feldman 2018). This
has been made since the influence of corrosion and high stress on anchorage/laps lies outside the scope of this
note.

For citations, copying and referencing see Note on Front page! page 646
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

The selected tests taken from Schenkel’s experimental program included a series of four lap splice tests. All
specimens had four pairs of 16 mm diameter ribbed bars, lapped in the same section, in bottom position, within
the constant moment zone (CMZ) of a four-point bending scheme. In these series, all beams had the same
dimensions and the bars were cast with a concrete cover equal to 20, 10, 5 or 0 mm.
A four-point bending scheme was also used in the two series of laps investigated by Regan and Reid (2009). The
first series included a pair of 25 mm corner bottom cast ribbed bars, lapped within the CMZ. In the second series,
there were two pairs of edge bottom cast bars. In the investigations, the concrete cover was varied between 33
to 0 mm; some cases with negative concrete equal to -8 mm were also tested. These last tests have been ignored
in the present note, because the influence of negative concrete cover (i.e. c/ < 0) on bond lies outside the scope
of this validation.
Cairns and Goodchild (2016) analysed four sets of tests. Each set comprised three individual specimens, one
with a concrete cover equal to 30 mm and two with concrete cover equal to 15 mm. Sets from 1 to 3 were
composed of two pairs of 25 mm diameter lapped bars. Set 4 was composed of two pairs of 16 mm diameter bars
lapped to 25 mm diameter bars. The latter test has been excluded from this study not only because relevant to
laps of different bar diameter (i.e. a situation, even if very common in practice, not consistent with the other tests
included in the reference database) but also because the concrete cover (i.e. 15 mm) was barely less than the
minimum bar diameter (i.e. 16 mm). Each set was relevant to bottom cast ribbed bars lapped, with different offsets,
in the CMZ of a four-point bending scheme.
Tests on plain bars have been taken from Darby et al. (2016) investigations. In particular, two series of beam-end
specimens, one with and the other without links, have been studied. Each series was composed of two specimens
of a 16 mm diameter bar in bottom cast position with a concrete cover of 37 or 0 mm.
To validate the equations proposed by Palmisano et al. (2020a), each test series has been considered individually.
Since some test results are in terms of fst (anchorage strength) while other are in terms of fb (bond strength), the
following equation is used to derive fb from fst:

𝑓 (1)

where lb is the anchorage length.

By assuming k as the ratio of fst or fb evaluated for c/ < 1 (called fst,c/<1 and fb,c/ <1, respectively, in the following)
to that evaluated for c/ ≥ 1 (called fst,c/>1 and fb,c/ >1, respectively, in the following), the relevant value evaluated
from tests (kTEST or kTEST,rel as defined in the following) is compared with that obtained from the proposed
formulation (kMOD = k’lbs,c)
It is worth noting that the factor k’lbs,c proposed by Palmisano et al. (2020a) is the ratio of the anchorage/lap
strength evaluated for c/ < 1 to that evaluated for c/ = 1 (fst,c/ =1 in the following). However, in the reference
database, no tests with c/ = 1 are included. Hence, fst,c/ =1 is calculated from fst,c/ >1 by using the following formula
,
𝑓 , (2)

where:
 c/ in formula (2) is relevant to that of the test where c/ > 1;
  = 0.25 for ribbed bars according to the formulation of fstm included in fib (2014);
 for plain bars  = 0.77 applies according Palmisano et al. (2020b).
If in a set of investigations there is more than one test with c/ > 1, the mean value of the anchorage/lap strength
of the tests in that set having c/ > 1 is assumed as fst,c/ >1 in formula (2).
The parameter k is evaluated, by using the test results, with reference to both fst,c/ >1 and fst,c/ =1 according to the
following formulae (3) and (4).
,
𝑘 (3)
,

,
𝑘 , (4)
,

As above mentioned, the type of tests included in the reference database are relevant to case 2 of table 1 and to
case 1 of table 2; thus, the following formulation for kMOD according to Palmisano et al. (2020a) should be used:
𝑘 0.5 0.5 (5)

The influence of c/≤ 1 on kTEST, kTEST,rel and kMOD is plotted in the figures 4-6 for ribbed bars and in figure 7 for
plain bars.
Figures 4-7 clearly show that equation (5) is conservative for both ribbed and plain bars since the points relevant
to kTEST and kTEST,rel are always above the line of kMOD.

For citations, copying and referencing see Note on Front page! page 647
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

1.5 Cairns
and
Goodchild
(2016)

kTEST or kTEST, rel or kMOD

1.0

kTEST
Serie1

0.5 kTEST,rel
Serie2

kMOD
Serie3

0.0
0.0 0.2 0.4 0.6 0.8 1.0
c/φ
Figure 4. Influence of c/ on kTEST, kTEST,rel and kMOD for Cairns and Goodchild’s investigations (ribbed bars).

Regan
1.5 and Reid
(2009)
kTEST or kTEST, rel or kMOD

1.0

kTEST
Serie1

0.5 kTEST,rel
Serie2

kMOD
Serie3

0.0
0.0 0.2 0.4 0.6 0.8 1.0
c/φ

Figure 5. Influence of c/ on kTEST, kTEST,rel and kMOD for Regan and Reid’s investigations (ribbed bars).

For citations, copying and referencing see Note on Front page! page 648
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Schenkel
1.5 (1998)

kTEST or kTEST, rel or kMOD

1.0

kTEST
Serie1

0.5 kTEST,rel
Serie2

kMOD
Serie3

0.0
0.0 0.2 0.4 0.6 0.8 1.0
c/φ

Figure 6. Influence of c/ on kTEST, kTEST,rel and kMOD for Schenkel’s investigations (ribbed bars).

Darby et
1.5 al. (2016)
kTEST or kTEST, rel or kMOD

1.0

kMOD 1
kMOD,

0.5 kTEST
kTEST

kTEST,rel
Ktest, rel 2

0.0
0.0 0.2 0.4 0.6 0.8 1.0
c/φ
Figure 7. Influence of c/on kTEST, kTEST,rel and kMOD for the investigations relevant to plain bars included in Darby
et al. (2016).

References
Cairns, J., Goodchild, C. (2016). Effect of reduced concrete cover on strength of lapped joints. Proceedings of the
Institution of Civil Engineers: Structures and Buildings, 169, 34-45.
Canbay, E., Frosch. R. (2005). Bond Strength of Lap-Spliced Bars. ACI Structural Journal, 102(4), 605-614.
Darby, A., Orr, J., Ibell, T. (2016). Half Joint Assessment Programme Research. Topic 1. Residual bond
characteristics of prestressing strand. Final Report. University of Bath, Department of Architecture and
Civil Engineering.
fib (2014). fib Bulletin 72. Bond and Anchorage of Embedded Reinforcement: Background to the fib Model Code
2010. Lausanne: International Federation for Structural Concrete (fib).
Metelli, G., Plizzari, G.A. (2014). Influence of the relative rib area on bond behaviour. Magazine of Concrete
Research, 66(6), 277-294.

For citations, copying and referencing see Note on Front page! page 649
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Palmisano, F., Biasi, M., Greco, M., Marano, G.C. (2020a). Effect of low ribbing and concrete cover thickness on
the anchorage of bars in existing R.C. structures. Proc. of the IABSE 2020 Wroclaw Poland Symposium
– Synergy of Culture and Civil Engineering - History and Challenges, Wroclaw, Poland, 7-9 October 2020,
617-624.
Palmisano, F., Greco, R., Biasi, M., Tondolo, F., Cairns, J. (2020b). Anchorage and laps of plain surface bars in
R.C. structures”. Engineering Structures, Vol. 213, 110603/1-12.
Palmisano, F., Cairns, J., Menga, A. (2021). Anchorage/lap strength of bars in R.C. structures in case of low
concrete cover thickness. Proc. of the IABSE Congress Ghent 2021 – Structural Engineering for Future
Societal Needs, Ghent, Belgium, 22-24 September 2021, 1057-1065.
Regan, P.E. and Reid, I.L.K. (2009). Assessment of concrete structures affected by cover delamination. Part 1 –
Effect of bond loss. Graduate School in Concrete Structures – Fratelli Pesenti, Politecnico di Milano, Italy,
Studies and researches – V.29.
Schenkel, M. (1998). Zum Verbundverhalten von Bewehrung bei kleiner Betondeckung (On the bond behaviour
of reinforcement with little concrete cover), Report No. 237, Zurich: Swiss Federal Institute of Technology,
Institute of Structural Engineering IBK.
SIA (2003). SIA 262, Concrete structures. Zurich: Swiss Society of Engineers and Architects (SIA).
SIA (2011). SIA 269/2, Existing structures – concrete structures. Zurich: Swiss Society of Engineers and Architects
(SIA).
Zwicky, D. (2013). Bond and ductility: a theoretical study on the impact of construction details – part 2: structure-
specific features. Advances in Concrete Construction, 1(2), 137-149.

For citations, copying and referencing see Note on Front page! page 650
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

CEN/TC 250/SC 2 Annex I to FprEN 1992-1-1:2023

Background document to some clauses of

sections I.11 and I.12

Fabrizio Palmisano PPV Consulting & Politecnico di Torino,

Italy

27.03.2023

For citations, copying and referencing see Note on Front page! page 651
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

I.11.2.(2) The minimum of bar spacing (𝑐 ) in 11.2(2) of FprEN 1992-1-1:2023 is a practical rule, the aim of which is
clearly explained in 11.2.(1). This limit of 𝑐 is not referred to bonding. In fact, considered the general model
Minimum bar spacing
for bond that is the basis of the anchorage length in FprEN 1992-1-1:2023 a limit equal to 𝑐 , should be also applied
to (i.e. for bonding should be not lower than the bar diameter). In this scenario I.11.2(2) has been added.

I.11.4.1(2) The minimum value for 𝑙 (i.e. 5𝜙) given in this clause is equal to the minimum value in the database
considered in fib bulletin 72 (see e.g. Fig. 3.8 of fib bulletin 72) that gives the expression for bond strength
Minimum value for lbd
assumed as basis of formula (11.3) of FprEN 1992-1-1:2023.
In particular, about 100 tests out of about 800 have a bond length lower than 15𝜙. However, as stated at page
14 of fib bulletin 72, tests with short bond length of less than 15𝜙 for laps or 10𝜙 for anchorages were excluded
from the statistical summary as unrepresentative of practice.
Clause I.11.4.2(2) includes the limit of 5𝜙 as the result of the discussion of Turin meeting (25th October 2018)
between TC 250/SC 2/SC2.T3, TC 250/SC 2/WG 1/AhG Detailing & fib TG 2.5. In that meeting, it appeared to
be a general consensus on the necessity to remove, in case of existing structures, practical limits of bond
lengths if there is evidence that the area where anchorage develops is uncracked.

I.11.4.1(11) Refer to section 6.1.7.3 of fib Model Code 2010.

Residual bond capacity

of ribbed bars not
confined by links and
of plain bars in freeze-
thaw damaged concrete

I.11.4.2(1) The effect of bends and hooks has been taken from fib bulletin 72 (2014) par. 3.3. It can be used also for plain
bars according to par. 6.1.3.5 of MC 2010 (version of fib bulletin 65, 2012).
Anchorage of bars with
bends and hooks The effect of bar diameter and confinement has been neglected. Moreover, to have the design values, from
the mean value given in fib bulletin 72, by using the same approach described in the ‘Background document
to clauses I.11.4.1(3)-(5)’ (i.e. the statistical approach proposed by Taerwe 1993, ‘Towards a consistent
treatment of model uncertainties in reliability formats for concrete structures’, CEB Bulletin d’Information No.
219, Safety and Performance concepts, pp. 5-61, Lausanne: CEB), has been used by using the following
assumptions:
 Mean value and coefficient of variation of model uncertainties for ribbed bars have been taken as 1.04
and 0.18, respectively, according to Mancini et al. 2018 (i.e. Mancini, G., Carbone, V.I., Bertagnoli, G.,
Gino, D. 2018, ‘Reliability-based evaluation of bond strength for tensed lapped joints and anchorages in
new and existing reinforced concrete structures’, Structural Concrete,19, 904–917).
 Mean value and coefficient of variation of model uncertainties for plain bars in good bond casting position
have been taken as 1.00 and 0.18, respectively, according to the ‘Background document to clauses
I.11.4.1(3)-(5)’.
 Mean value and coefficient of variation of model uncertainties for plain bars in poor bond casting position
have been taken as 1.00 and 0.36, respectively, according to the ‘Background document to clauses
I.11.4.1(3)-(5)’.
 The first coefficient of formula of fib bulletin 72 has been multiplied by 0.8 and 0.5 for ribbed and plain
bars, respectively, to take account of poor bond casting condition. For ribbed bars this factor is constituent
with that of FprEN 1992-1-1:2023. For plain bars this factor is equal to the mean value of the ratio of the
anchorage strength of bars in poor bond casting position to that of bars in good bond casting position in
the reference test database referred in the ‘Background document to clauses I.11.4.1(3)-(5)’.
 The contribution of c has been evaluated by using the same procedure followed in the ‘Background
document to clauses I.11.4.1(3)-(5)’ for the design anchorage/lap length of plain bars.
It is worth noting that 1 in formula I.30 is lower than 0.5 for plain bars because of the high coefficient of variation
of model uncertainties for plain bars in poor bond casting position.

I.12.(1) Clause 12.1(6) of FprEN 1992-1-1:2023 is a requirement (“shall” clause); it obliges the designer to consider a
member as plain or lightly reinforced concrete member if the longitudinal reinforcement is lower than 𝐴 , .
Members with less
longitudinal Many existing structures do not fulfil the minimum requirements on detailing of new structures. If these are
reinforcement than considered as lightly reinforced concrete structures most of the checks will not be satisfied. Thus, even if
As,min 12.1(6) of FprEN 1992-1-1:2023 is suitable for a new structure, it is very restrictive for an existing one. This is
why I.12(1) gives the possibility to ignore 12.1(6) of FprEN 1992-1-1:2023 for existing structures.

For citations, copying and referencing see Note on Front page! page 652
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

CEN/TC 250/SC 2 Annex I to FprEN 1992-1-1:2023

Additional background on deteriorated

concrete structures

Fabrizio Palmisano PPV Consulting & Politecnico di Torino,

Italy

Craig Giaccio Arcadis Consulting Ltd

UK

27.03.2023

For citations, copying and referencing see Note on Front page! page 653
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Introduction
This section gives additional clauses and background on deteriorated concrete structures. This section is intended
only to provide information outlining what are the obvious signs and what should be taken into account in
assessment.
This section does not intend to provide models for predicting degradation of structures for the different
mechanisms that prevail.
This section is mainly based on the input received from CEN/TC 250/SC 2/WG 1/TG 3 (draft V15 February 2019).
A list of recent articles on corrosion is included in the last paragraph.

Reinforcement corrosion

General
(1) Structural damage in structures with corroded reinforcement, to be considered at ULS verifications, are due
to:
- reinforcement deterioration;
- concrete deterioration;
- bond deterioration.
The main consequences of corrosion are:
- loss of bar cross-section;
- loss of steel ductility;
- cracking of concrete cover with the consequential loss in concrete cross section;
- loss of bond between reinforcement and concrete.
(2) The relevant parameters for the assessment are:
- Mechanical strength of concrete.
- Crack width, crack pattern and delaminated areas.
- Concrete cover thickness.
- Mechanical properties of reinforcement.
- Residual cross-sectional area of reinforcement.
Note: In the case of presence of cracks parallel to the reinforcement, typical of the corrosion process, their pattern
is a qualitative indication of the degree of damage and of the affected surface. Its actual value is a key parameter
for the assessment. The cracks and delamination are necessary to estimate the residual concrete section.
Although the reinforcement maximum strength may remain unaffected, the corrosion may induce a reduction of
steel ductility and the embrittlement of the reinforcement core when the loss in cross section is higher than 15-
20%. This has to be taken into account in the case of seismic risks and other dynamic loads and when considering
non-linear or plastic analysis.

Considerations for verification at ULS

(1) For the verification at ULS after corrosion has initiated, the following factors need to be considered:
- nature of corrosion (generalized or localized);
- representative corrosion rate;
- reinforcement detailing;
- consequences of failure;
- structural redundancy.
(2) The assessment of resistance should account for:
- any reduction in the reinforcement area and in its ductility due to corrosion;
- any reduction in the member dimensions due to concrete spalling;
- any reduction in bond strength caused by corrosion or by the spalling of the concrete cover.
(3) A conservative value of the ultimate moment resistance may be achieved by reducing the reinforcement cross-
section and the member cross-section (due to concrete cover spalling or cracking). A possible effect of occurred
bond deterioration should be considered and taken into account when affecting the strength.
(4) In case of corrosion several factors may induce a premature shear failure, such as:
- For stirrups the effect of corrosion may be higher due to:

For citations, copying and referencing see Note on Front page! page 654
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

 initial small diameters of stirrups and therefore relevant large cross-section loss;
 lower cover for stirrups leading to earlier initiation of corrosion compared to main bars.
- Spalling of cover and reduction of the concrete compression zone.
- Loss of bond of longitudinal reinforcement due to corrosion.
- Induced stresses in stirrups due to corrosion of the main bars.
In members with corrosion at the tensile side and without stirrups, the deterioration of bond at longitudinal bars
can reduce shear strength because of the relevant reduction of dowelling action.
When evaluating the shear resistance of members requiring design shear reinforcement, a further limitation of the
minimal inclination of the compression field should be considered due to bond deterioration.
(5) In the assessment of the ultimate axial load of a column:
- the concrete section should be reduced in case of spalling;
- if the stirrups have been broken due to the corrosion, a reduction in the longitudinal bars subjected to
compression due to risk of buckling should be taken into account.

Loss of bond
(1) The effects of corrosion in hardened concrete differ from those associated with corrosion prior to concreting.
Small amounts of corrosion, up to the level required to induce longitudinal cracking, do not cause loss of bond
resistance, and can even augment bond strength by a modest degree, particularly where the bar is in a ‘poor’
bond casting position.
At greater levels of corrosion, residual bond strength is strongly influenced by the degree of confinement provided
by transverse reinforcement in the form of links and by the surrounding structure. Transverse pressure from
support reactions augments bond. Links play a valuable role in maintaining residual strength of anchorages and
lapped joints.
(2) The residual capacity of anchorages and lapped splices should be checked at the ultimate limit state at
locations of high reinforcement stress where longitudinal cracking develops. Away from anchorages and laps, a
substantial loss of bond may be tolerated without ultimate strength being affected.
(3) The magnitude of the reduction in residual bond strength is highly dependent on the confinement to the bar
and is also affected by concrete quality and environment. It may be taken into account by increasing the design
value of the anchorage and lap length.
Note: The increase of lbd can be evaluated by using Table 6.1-4 of fib MC 2010, for plain and ribbed bars, or
section F.2.3 of Contecvect Manual (‘CONTECVET. A validated Users Manual for assessing the residual service
life of concrete structures. Manual for assessing corrosion-affected concrete structures’, EC Innovation
Programme IN30902I, preparation of document led by Geocisa and Torroja Institue, 2000), for ribbed bars.
However, in the literature there are proposals by different researchers that can be taken as reference. A new
proposal will be included in the forthcoming fib MC 2020, chapter 20.
(4) In addition to (3), the effect on bond of concrete cover spalling should be considered and it may be taken into
account by using the reduced cover thickness in the formulae that give lbd.

External sulphate attack

(1) There are several sources of external sulphates: seawater, ground water with sulphides, bacterial action in
sewers, etc. Expansion and damage occur when the sulphates penetrate from outside the concrete and the
expansion is generated in the surface concrete layers by the formation of sulphate-bearing compounds.
(2) The process can be represented by an aggressive front advancing from the concrete surface and inducing
periodically cracking and spalling. The cracks formed due to differential expansion in the concrete surface layers
allow the sulphates to penetrate deeper producing a discontinuous process, which however can be linearized.
(3) The effects on the structure are:
- loss of elastic modulus in the deteriorated zone;
- loss in bond between aggregate/concrete with decrease of strength and disaggregation of external concrete
layer leading into cracking and spalling;
- reduced bearing concrete section due to that delamination and spalling.
These effects should also be taken account of in the cross section checks for the structural consideration of
external sulphate attack. It may be needed to differentiate the external deteriorated layer from the sound interior.
(4) The rate of expansion and generation of cracking depends on the concentration of the external solution and
the length of the contact period. This rate should be monitored if needed for prediction through the measurement
of dimensional changes or of the evolution of the dynamic elastic modulus.
(5) ULS should be verified with the reduced properties and section dimensions.

For citations, copying and referencing see Note on Front page! page 655
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Alkali-aggregate and other expansive reactions

(1) AAR (Alkali-Aggregate Reaction) and DEF (Delayed Ettringite Formation) are expansive processes that may
be treated in a similar manner. In both cases, the concrete integrity is affected in its mechanical parameters and
in the composite steel/concrete function.
Note:
- AAR is an internal chemical reaction between the alkaline components in the cement and active silica-based
mineral or other constituents of some aggregates, when a certain degree of moisture is present. The reaction
results in the formation of a gel that absorbs water, expands, and therefore induces internal pressure which
sometimes can be far in excess of what the concrete can sustain, thereby causing the formation of micro-cracks.
In the case of the attack to other aggregate constituents, as Mg-bearing, the expansion is due to the solution-
precipitation of new compounds that have more volume than the initial ones.
- DEF is also an internal cracking process produced by the formation of expansive ettringite (‘secondary’ or
‘delayed’) when the concrete is already hardened. The risk has been identified when hydration heat generation
produces temperatures in the concrete mass above 60-70°C. It is aggravated if the cements are of high alkaline
(sodium and potassium) content.
(2) The progress of the process (expansion or crack generation) can be represented by a curve where a maximum
is achieved when the expansive agents are exhausted. This trend can also be linearized resembling the two-
period mode of incubation and propagation. In AAR the incubation can last decades because it depends on the
aggregate reactivity. In DEF however the expansion appears in much shorter terms.
Note: Expansion can be measured indirectly through reduction in E-modulus on drilled cores, or specific SDT
(stiffness damage test) parameters.
(3) Prediction of expansion can be made if it is monitored on site as for sulphate attack through the monitoring of
the local volume changes or the dynamic elastic modulus.
(4) The structural effects of expansion induced by AAR are:
- generation of additional internal actions;
- decrease in the compressive strength mainly due to the generation of internal cracking;
- decrease in the axial tensile strength;
- decrease in the modulus of elasticity;
- increase (due to the confinement induced by the expansion) or decrease of steel/concrete bond due to cracking.
Note: When laps and anchorages are restrained by links, expansion within the limits 0,45‰ (restrained) and 4,5‰
(unrestrained) induced by AAR has been shown to have no significant effect on bond strength.
Where transverse reinforcement is not present then bond strength may be reduced by up to 50%.
Care is necessary in the assessment of residual resistance of anchorage zones of prestressed elements,
particularly of pre-tensioned members in which the tendons are not contained by links.
(Reference: section 6.1.7.2 of fib MC 2010).
(5) ULS should be verified with the reduced properties and sections.

Frost attack
(1) Frost attack is produced due to the fact that ice has a higher volume than liquid water. Not all the pore water
solution freezes at the same temperature because as smaller are the pore size lower temperature is needed to
produce ice. The expansion can induce surface delamination of concrete or an internal microcracking when the
whole section is frozen. Its progress depends on the degree of saturation of the concrete and on the temperature
and its cycling regime. Freezing temperatures, if they are maintained constant, do not necessarily impair the
concrete which has shown to last at these temperatures maintaining its bearing properties.
(2) The progress of the process has not been systematically recorded. It can be monitored by following the amount
of delaminated concrete or the crack pattern and width evolution.
(3) The structural consequences are:
- the external frost attack can affect the performance similarly as sulphate attack due to:
 loss in cross section;
 loss of strength due to the loss in aggregate/concrete bond;
- the internal frost attack:
 reduces (locally) the strength due to the internal cracking;
 causes loss on modulus of elasticity;
 causes loss in strength;
 causes loss in steel/concrete bond.

For citations, copying and referencing see Note on Front page! page 656
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

Leaching and acid attack

(1) Leaching and acid attack consists of a diffusive process leading to neutralisation of the alkaline pore solution
and consequent dissolution of hydrated cement phases. It is an external attack that progresses as a front from
the external concrete surface. The mechanism consists first in the neutralization of the calcium and alkaline
hydroxides of the pore solution. Further, the Calcium-Silicate-Hydrate gel decalcifies and consequently softens,
to end with the dissolution of the rest of aluminate phases and the complete disaggregation of the material and
its dissolution.
(2) The progress of the neutralized front follows the square root of time and its advance is similar to carbonation
when the front is sharp or producing a smoother profile where the deterioration is progressive. The neutralization
front can be more advanced than the softening or attack to the cement hydrated phases.
(3) The rate of progress depends on the nature and concentration as well as of the convection regime of the
external solution. For the sake of prediction, it may be monitored on site.
(4) The structural consequences are:
- loss of bearing cross section;
- loss of strength;
- loss of modulus of elasticity.
(5) ULS should be verified with the reduced properties and sections.

Selected references on corrosion

Andrade, C. (2017). Reliability analysis of corrosion onset: initiation limit state. Journal of Structural Integrity and
Maintenance, 2(4), 200-208.
Andrade, C., Anaya, P. (2020). Advances in the description of corrosion induced cracking. Proceedings of the fib
CACRCS DAYS 2020, Edited by B. Belletti & D. Coronelli, 39-48.
Belleti, B., Rodríguez, J., Andrade, C., Franceschini, L., Sánchez, J., Vecchi, F. (2020). Experimental tests on
shear capacity of naturally corroded prestressed beams. Structural Concrete, 21 (5), 1777-1793.
Belletti, B., Vecchi, F., Bandini, C., Andrade, C., Sanchez, J. (2020). Numerical evaluation of the corrosion effects
in prestressed concrete beams without shear reinforcement. Structural Concrete, 21 (5), 1794-1809.
Blomfors, M., Zandi, K., Lundgren, K., Coronelli, D. (2018). Engineering bond model for corroded reinforcement.
Engineering Structures, 156, 394-410.
Coronelli, D. (2020). Resistance of corroded RC beams: Extending fib Model Code 2010 models. Structural
Concrete, 21(5), 1747–1762.
Coronelli, D., Gambarova, P. (2004). Structural assessment of corroded reinforced concrete beams: Modeling
guidelines. Journal of Structural Engineering, 130(8), 1214-1224.
Coronelli, D., François, R., Dang, H., Zhu, W. (2019). Strength of corroded R/C beams with bond
deterioration. Journal of Structural Engineering, 145(10), 04019097.
Fernández, I., Bairán, J.M., Marí, A. (2015). Corrosion effects on the mechanical properties of reinforcing steel
bars. Fatigue and σ-ε behaviour. Construction and building materials, 101, 772-783
Pedrosa, F., Andrade, C. (2017). Corrosion induced cracking: Effect of different corrosion rates on crack width
evolution. Construction and Building Materials, 133, 525-533.
Prieto, M., Tanner, P., Andrade, C. (2016). Multiple linear regression model for the assessment of bond strength
in corroded and non-corroded steel bars in structural concrete. Materials and Structures, 49, 4749-4763.
Rodriguez, J., Anaya, P. (2020). Some thoughts on the structural performance of corroded concrete structures
arising from past experimental results. Proceedings of the fib CACRCS DAYS 2020, Edited by B. Belletti
& D. Coronelli, 187-196.
Sykora, M., Holicky, M., Prieto, M., Tanner, P. (2015). Uncertainties in resistance models for sound and corrosion-
damaged RC structures according to EN 1992-1-1. Materials and Structures, 48(10), 3415-3430.
Walraven, J. (2020). Significance of reinforcement corrosion for modelling the behaviour of existing concrete
structures. Proceedings of the fib CACRCS DAYS 2020, Edited by B. Belletti & D. Coronelli, 239-249.

For citations, copying and referencing see Note on Front page! page 657
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex I Background to FprEN 1992-1-1:2023

List of versions of this document

05.03.2021 1st draft
20.05.2021 2nd draft: editorial improvements; paragraph “Selected references on corrosion” added to “Additional background on
deteriorated concrete structures”
13.01.2022 3rd draft: update to FprEN 1992-1-1:2022
27.03.2023 Peer review

For citations, copying and referencing see Note on Front page! page 658
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG1 FprEN 1992-1-1:2023

Background document to Annex J:

Validation of expressions for Strengthening
of Existing Concrete Structures with CFRP

This document was prepared by

CEN/TC250/SC2/WG1/TG1.
For authors of the subsections see content list
below.
21.12.2022

For citations, copying and referencing see Note on Front page! page 659
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

Authors
Konrad Zilch Technische Universität München (DE)
Craig Giaccio Arcadis Consulting (UK)
Roland Niedermeier Technische Universität München (DE)
Wolfgang Finckh Ostbayerische Technische Hochschule Regensburg (DE)
Anthony Darby Bath University (UK)
Renata Kotynia Lodz University of Technology (PL)
Eva Oller Universitat Politècnica de Catalunya BarcelonaTech (ESP)
Thorsten Leusmann Technische Universität Braunschweig (DE)
Rolf Alex Deutsches Institut für Bautechnik (DIBt) (DE)
Anett Ignatiadis Deutscher Ausschuss für Stahlbeton (DAfStb) (DE)

Stijn Matthys Ghent University (BE)

Vanessa Buchin (FR)

Content and main contributors

J.2 Scope and field of application K. Zilch, R. Niedermeier, W. Finckh, S. Matthys
J.3 General K. Zilch, R. Niedermeier, W. Finckh
J.4.1 Partial factors for materials R. Niedermeier, R. Alex, K. Zilch, W. Finckh
J.5 Materials R. Alex, A. Ignatiadis, V. Buchin
J.8.1.5 Concrete columns confined with fully wrapped CFRP C. Giaccio, A. Darby
J.8.2 Shear R. Kotynia, E. Oller, A. Darby
J.9 Serviceability Limit States (ULS) W. Finckh, S. Matthys
J.10.1, J.10.2 Fatigue analysis for externally bonded CFRP systems T. Leusmann
J.10.3 Fatigue analysis for Near surface mounted CFRP strips W. Finckh, K. Zilch, R. Niedermeier
J.11.1.1.1 Basic anchorage resistance – simplified method W. Finckh, E. Oller, S. Matthys
J.11.1.1.2 Basic anchorage resistance – refined method R. Niedermeier, W. Finckh, K. Zilch
J.11.1.2 EBR CFRP anchorage requirements – flexure R. Niedermeier, W. Finckh, K. Zilch
J.11.1.3 Basic anchorage resistance – CFRP to concrete for NSM CFRP strengthening W. Finckh
J.12 Detailing of members and particular rules R. Kotynia, W. Finckh, A. Darby, C. Giaccio

Introduction
This document has been prepared in order to explain the proposed provisions for strengthening of existing
concrete structures with CFRP in FprEN 1992-1-1, Annex J. The main basis for the proposed provisions were
national guidelines [1], [2], fib-bulletin 90 [3] and national approvals.

CEN/TC250/SC2/WG1/TG1 21.12.2022
For citations, copying and referencing see Note on Front page! page 660
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

J.2 Scope and field of application

Annex J applies to members with plain, reinforced and/or prestressed normal weight concrete.
Strengthening of members with special concretes (e.g. lightweight aggregate concrete, steel fibre reinforced
concrete, concrete with recycled aggregates) is not covered by Annex J as there is no significant experience. The
basic principles apply to special concretes, however some parameters, especially for bond have to be adjusted.
Annex J covers only strengthening with carbon fibre polymers as the main research has been done on this field.
For other fibre materials experience and background is missing to validate or adjust the models for different E-
Modulus and different behaviour under long-term loading. For glass fibre polymers there is experience for
confinement of columns, but it is used for seismic retrofitting mainly, which is not in the scope of FprEN 1992-1-1.
FprEN1992-1-1 does not cover resistance to fire, consequently this is also not covered in Annex J. In 1.1 (4) of
the main text reference to EN 1992-1-2 is made. However, current prEN 1992-1-2 does not contain specific rules
for strengthening with CFRP.
Because of this the following information should be considered.
Fire design of reinforced concrete members strengthened with CFRP will consider evaluation of the fire resistance
of the individual strengthened member, as well as reflect possible effects on the fire resistance of the whole
structure.
For low to moderate strengthening degrees, the strengthening is not needed in an accidental design situation. In
such cases, fire resistance verification of the member in the accidental design situation must be performed,
considering that the bonded CFRP is no longer active and considering the relevant acting loads during fire. The
latter correspond to the fact that the member is strengthened at the start of the fire scenario. The fire resistance
verification is done to prove that no special measures (like protective insulating linings) are needed, or to establish
the required thermal insulation to protect the concrete and internal steel.
For moderate to high strengthening degrees, the strengthening needs to remain functional during the targeted fire
endurance. The design of the fire protection system mainly focuses to limit temperatures to ensure bond
interaction between the concrete and the strengthening system in the event of fire.
NSM FRP technique provides better fire resistance than EBR FRP one. However using a higher glass temperature
epoxy is the most important impact for NSM FRP applications in the internal structural reinforced concrete
members requiring high fire resistance. If the thickness of the cement composite layer (ta) has the largest possible
depth of the most exterior edge of the strip (Zfe), the highest fire resistance is obtained (Fig. BD-J.12.1 c,d) [45].

J.3 General
Annex J contains rules for strengthening of existing concrete structures with carbon fibres. Adhesively bonded
reinforcement (ABR) is used as umbrella term for different application methods and products. Externally bonded
reinforcement (EBR) is glued to the prepared concrete surface. Near surface mounted reinforcement (NSM) is
applied in slots cut in the concrete cover. Figure BD-J.1.1 illustrates the terms, the application methods and the
corresponding products as developed for this Eurocode. Figure BD-J.1.2 shows some examples of strengthening
methods considered in annex J.

Adhesivly bonded reinforcement

ABR

Externally bonded reinforcement Near surface mounted reinforcement

EBR NSM

CF sheet/pre‐
CFRP strips impregnated CFRP strips CFRP bars
laminates

Figure BD-J.1.1: Terms, application methods and corresponding products

K. Zilch, R. Niedermeier, W. Finckh, S. Matthys 21.12.2022 Page 3/34

For citations, copying and referencing see Note on Front page! page 661
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

Figure BD-J.1.2: Examples for strengthening: a) Ilustration EBR and NSM, b) Strengthening a beam for
bending, c) Strengthening of a column by wrapping

The design of FRP strengthened concrete members follow in principle the mechanics of reinforced concrete.
However, some peculiarities have to be taken into account and may lead to different limit states as explained in
the following chapters.
For example, the bond behaviour of strips applied in EBR and NSM shows a distinct difference (comp. Figure
BD-J.1.3).

Figure BD-J.1.3: Principle of the bond force transfer of adhesively bonded CFRP-strips [4]
In NSM the total force capacity of a CFRP strip can be transferred by bond on a single crack in an acceptable
bond length (as with ribbed reinforcing steel). The strength of the adhesive will be governing the design.
In EBR the total force of a strip cannot be transferred by bond on a single crack. To get acceptable utilization of
the strip capacity strength increase of the force over intermediate crack elements has to be considered! As the
tensile strength of the adhesive is in general much larger than the tensile surface strength of the concrete it is
assumed in the models for EBR that the concrete will fail first and the tension strength of the concrete is governing
bond design. So, the surface tension strength of the concrete will be a very important parameter in design.
Furthermore, it has to be considered that the strength parameters of the CFRP and adhesive may depend on the
temperature and environmental conditions.

K. Zilch, R. Niedermeier, W. Finckh, S. Matthys 21.12.2022 Page 4/34

For citations, copying and referencing see Note on Front page! page 662
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

J.4 (1) Partial factors for materials - Description

Table J.1 contains, depending on the design situation, partial safety factors for prefabricated CFRP strips and
bars and CF sheets/pre-impregnated laminates applied in-situ as well as for the bond of ABR with regard to the
failure of near-surface concrete layers or the adhesive. Since, due to construction, only tensile forces may be
transferred to the ABR, the design of strengthened reinforced and prestressed concrete components is possible
for all relevant types of loading in addition to the partial safety factors given for the other materials in FprEN 1992-
1-1, Section 4.3.3, Table 4.3.

Background
The safety concept is based on the regulations of prEN 1990:2020 [47]. According to section 8.3.5 the design
value of resistance 𝑅 for a specific design situation should be calculated from the equation (BD-J.4.1)
corresponding to equation (8.18) in [47]:
∙
𝑅 ∙𝑅 ;𝑎 ;∑𝐹 (BD-J.4.1)

where
𝛾 is a partial factor associated with the uncertainty of the resistance model, and for geometric
deviations, if these are not modelled explicitly (Remark: numerical values may be obtained
from e.g. [48]);
𝑅 … denotes the output of the resistance model;
𝜂 is a conversion factor accounting for scale effects, effects of moisture and temperature,
effects of ageing of materials, and any other relevant parameters;
𝑋 represents the characteristic values of material or product properties;
𝛾 is a partial factor for a material property accounting for:
- unfavourable deviation of the material or product properties from their characteristic
values;
- the random part of the conversion factor 𝜂;
𝑎 denotes the design values of geometrical property;
𝐹 denotes design values of actions used in the assessment of 𝐸 ;
When applying partial factors to resistance, its design value 𝑅 should be calculated from Formula (BD-J.4.3)
where partial factors 𝛾 and 𝛾 are combined into a single partial factor 𝛾 for resistance:
𝛾 𝛾 𝛾 ∙𝛾 (BD-J.4.2)
∙ ; ;∑
𝑅 ∙ (BD-J.4.3)

Where relevant for the application-specific failure modes, the conversion factors are explicitly stated. They have
to be determined by experimental testing in accordance with an European Technical Product Specification unless
otherwise noted. The effects of geometric deviations are limited by design specifications, e.g. regarding the
number of layers of bonded reinforcement or the slot geometry, and are taken into account in the partial factor
𝛾 with a specific magnitude depending on the procedure. For bonded reinforcement, a thickness of adhesive
layer between 1 and 5 mm is assumed.
For the determination of the partial safety factors given in Table J.1, products were considered that were also
used in the structural tests, on the basis of which the design approaches in sections J.8 to J.11 were determined.
In accordance with the specifications in section J.5 for the design, the axial tensile strength is considered. For the
prefabricated products CFRP-strips and bars, data from the factory production control and the external continuous
surveillance were made available for two representative product groups A and B, which had been fabricated with
different production processes.
It must be taken into account that the product groups considered, as it is the case for pre-stressing steel strands,
are not only due to various production parameters, such as fibre type or fibre content, but are also based on a
targeted allocation to manufacturer-specific product classes on the basis of the mechanical properties determined
in the factory production control (FPC). The data from the FPC therefore only represent a certain range of
parameters, so that a suitable distribution function cannot be determined. Furthermore, it must be taken into
account that the data from the FPC may also contain values that are not applicable to the design situation of the
component due to the type of fracture, but are nevertheless accepted within the scope of the FPC if the strength
is sufficient. The data of the external surveillance, on the other hand, representing the entire parameter range of
the respective manufacturer-specific product class, only contain tests with applicable fracture types and show a
normal distribution of the strength.

R. Niedermeier, R. Alex, K. Zilch, W. Finckh 21.12.2022 Page 5/34

For citations, copying and referencing see Note on Front page! page 663
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

Table BD-J.4.1: Statistical data of tensile strength from FPC and the external continuous surveillance for two
representative product groups A and B of prefabricated CFRP strips (statistical data courtesy
of different German cerfication bodies)

Product
Data source Parameter A B
Number of numerical test results 𝑛 185 617
Factory production control Coefficient of variation 𝑉 0,07 0,05
Kolmogorov–Smirnov statistic for a normal
𝐷 0,09 0,11
distribution function
by notified Number of numerical test results 𝑛 77 122
product Coefficient of variation 𝑉 0,06 0,09
Audit-testing in certification Kolmogorov–Smirnov statistic for a normal
the context of body 𝐷 0,06 0,07
distribution function
continuous by external Number of numerical test results 𝑛 61
surveillance product Coefficient of variation 𝑉 0,07 -
certification Kolmogorov–Smirnov statistic for a normal
body 𝐷 0,12
distribution function

For CF sheets applied by the wet lay up method, the partial factor for tensile strength is based on data collected
by Occhiuzzi et al. [5]. They found a lognormal distribution of tensile strength for the single and three layer
application of a typical CF sheet based on the evaluation of tests carried out at eight university laboratories and
give the following parameters.
Table BD-J.4.2: Statistical data of tensile strength of CF sheets according to Occhiuzzi et al. [5]

Number of layers
Parameter 1 3
Number of numerical test results 𝑛 270 288
Mean value of logarithmised tensile strengths 𝜇 8,20 8,19
Standard deviation of logarithmised tensile strengths 𝜎 0,15 0,14
Kolmogorov–Smirnov statistic for a lognormal distribution function 𝐷 0,061 0,058
Coefficient of variation of tensile strengths 𝑉 0,15 0,14

If the maximum coefficients of variation included in tables BD-J.4.1 and BD-J.4.2 are considered representative
of the population, the following partial factors 𝛾 for the tensile strength of ABR are obtained.
For the CFRP strips, the partial factor 𝛾 is determined for a normal distribution with the reliability index 𝛽 3,8
for residential and office buildings and the sensitivity factor (FORM) 𝛼 0,8 for resistance as follows:
, ∙ , ∙ ,
𝛾 1,17 (BD-J.4.4)
∙ ∙ , ∙ , ∙ ,

For the CF sheets applied in wet lay-up procedure, the partial factor 𝛾 is determined for the lognormal distribution:
∙ ∙ , ∙ , ∙ , ∙ , , ∙ ,
𝛾 𝑒 𝑒 1,23 (BD-J.4.5)
The application of these values for the design implies a corresponding confirmation of the applied coefficients of
variation and distribution functions on the basis of experimental investigations of the tensile strength and the
constancy of performance according to a European Assessment Document.
The partial safety factor for resistance 𝛾 for persistent and transient design situations is a combination of 𝛾 and
the partial factor 𝛾 associated with the uncertainty of the resistance model and for geometric deviations:
𝛾 𝛾 ∙𝛾 (BD-J.4.6)
The index “f” is used in prEN1992-1-1 for the material (FRP), while according to EN 1990 𝛾 is a partial factor that
takes account of unfavourable deviation of an action from its representative value.
Depending on the application technique, the partial factor 𝛾 has different magnitudes for the persistent and
transient design situation, the following partial safety factors are obtained, which are determined for the other
design situations analogously to prEN 1992-1-1, Section 4.3.3, Table 4.3. Although the same models are used
for both applications, a higher 𝛾 is applicable for CF sheets amongst others because of higher degree of
geometric deviations with this application technique.
Table BD-J.4.3: Derivation of partial factors for materials

Design situation Partial factor CFRP strips and bars In-situ lay-up CF sheets
𝛾 1,17 1,23
𝛾 1,10 1,15
Persistent and transient 1,30 1,40
Accidental 1,10 1,15
𝛾
Serviceability 1,00 1,00
Fatigue 1,30 1,40

R. Niedermeier, R. Alex, K. Zilch, W. Finckh 21.12.2022 Page 6/34

For citations, copying and referencing see Note on Front page! page 664
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

The safety concept for bond is based on design assisted by testing according to EN 1990, Annex D (cf. section
J.15) by assessment via the characteristic value with 𝛾 𝛾 . The conversion factors 𝜂 ∙ 𝑘 and 𝑘 for the
concrete strengths are considered according to FprEN 1992-1-1.
For the determination of the partial safety factor for bond 𝛾 , taken from fib Bulletin 90 [3], failure of near-surface
concrete layers or the adhesive failure was assumed. For the persistent and transient design situation, 𝛾
corresponds to the partial safety factor for the concrete bearing component 𝛾 for design of shear at interfaces
according to Section 8.2.6 (5) of the main part of FprEN 1992-1-1.
Table BD-J.4.4: Partial factors for bond strength
Partial factor 𝛾 for
Design situation
bond strength
Persistent and transient 1,50
Accidental 1,15
Serviceability 1,00
Fatigue 1,50

The design concept for near surface mounted CFRP reinforcement distinguishes between concrete failure and
adhesive failure (see section 11). For reasons of simplification, the same safety factor was used here. A closer
look could possibly apply a lower safety factor to the adhesive failure due to its probably lower scatter. However,
the long-term strength has to be considered.

R. Niedermeier, R. Alex, K. Zilch, W. Finckh 21.12.2022 Page 7/34

For citations, copying and referencing see Note on Front page! page 665
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

The limits provided in this section have been selected to reflect to testing ranges used in the development of
J.5.1 (3)
relationships used in this annex. They also correlate with products widely available on the market.
Establishing a minimum value for the characteristic tensile strength of adhesive ensures that the basic assumption
of design of externally bonded reinforcement, namely that the concrete layer close to the surface will fail, has
been satisfied.

The material properties are product-specific and have to be given by the supplier. There are some standard test
J.5.2
procedures for the CFRP in ISO 10406 but not for all relevant properties and not for the entire strengthening
system. Further some of the design parameters may not be directly derived from the test results. In this case or
where test procedures are not standardized yet, recommended values are given in the corresponding sections of
Annex J.
In the product standard EN 1504, which provides with definitions, requirements, quality control and evaluation of
conformity, Part 2 handles surface protection systems for concrete and Part 4 give rules for structural bonding.
It has to be considered that all properties of the single components (e.g. CFRP strips, adhesive) have to be
ensured to be valid in the entire strengthening system.
For further information see “Requirements to Materials – CFRP reinforcement for strengthening” (Background
document to Annex J.15) [42].
According to the Construction Products Regulation (Regulation (EU) No 305/2011 (CPR)) on EU level test
procedures for the essential characteristics of construction products may be included in two types of technical
specifications for construction products:
- European harmonized product standards or
- European Assessment Documents (EAD).
At the time of this writing, none of the two types of technical specification was available for CFRP Strengthening
Systems. EADs are under preparation. That’s why [42] contains not only methods for determining the design
parameters for Annex J, but also the descriptions of possible test procedures, where necessary.

The factor 𝜂 is a reduction factor applied to the tensile strength of the ABR CFRP. It considers long-term effects
J.5.3 (1) and (2)
as creep under sustained load and influences of the alkaline environment and the temperature.
For further information see “Requirements to Materials – CFRP reinforcement for strengthening” (Background
document to Annex J.15) [42].
In approvals, where tests had been performed values between 0,65 and 1,0 can be found. That’s why the
recommended value 0,7 has been chosen. In approvals where no tests have been performed lower values may
be found.

The permissible environmental conditions as well as the resulting measures will be regulated in the European
J.6.3
Assessment Documents (or a subsequent execution standard) of the strengthening System. For further
information see Background document to Annex J.15 [42].
Presently, the supplied system dictates the exposure classification. This document was prepared on the basis
that the designer will specify an exposure classification as part of a design, along with the design life to enable a
suitable product to be supplied for the strengthening application.
The concrete cover must be measured (in execution) so that the reinforcement is not endangered when cutting
the slots. Also, the filling and protection must both ensure fire performance and bond of both existing reinforcement
and NSM reinforcement in slots is acceptable.

R. Alex, A. Ignatiadis, V. Buchin 21.12.2022 Page 8/34

For citations, copying and referencing see Note on Front page! page 666
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

This provision aligns with fib 90 [3], Section 3.8.

J.7
This provisions align with fib 90 [3], Section 2 and Section 6.
J.8
The provisions in this code should not be applied to concrete with 𝑓 12 MPa or 𝑓 50 MPa, due to limited
surface tensile strength in the lower concrete classes and less background for higher strength classes.

The flexural capacity of the strengthened member is assessed by means of traditional cross-sectional analysis,
J.8.1.1
based on strain compatibility assuming full composite action between the CFRP and the concrete substrate,
horizontal equilibrium of forces and moment equilibrium. The load level during strengthening is considered through
the acting strains in the cross section when the moment has come for ABR application.
Debonding checks are undertaken to validate this approach.

J.8.1.2 General
The provisions for the confining of concrete columns in Annex J consider only the confining effect due to the
adhesively bonded CFRP. The confining effect of the internal reinforcement has not been considered for reason
of simplification.
An approach for considering the internal reinforcement can be found in [1], [43], [44].
The application of FprEN 1992-1-1, Section 8.1.4 in addition to Annex J, J.8.1.2 is not on the safe side because
of different behaviour of the internal steel reinforcement and the strengthening material.

J.8.1.2 (2) Limits of Application

The characteristic concrete strength fck is less than 50 MPa;
All models have been validated against a large database of test specimens (306 circular cross section
specimens and 216 rectangular/square specimens). For both circular and rectangular/square columns,
the highest concrete strength studied is 55 MPa (very few at this strength, with the average of
approximately 36 MPa). Given that the expressions are largely empirical, the equations may not be
valid outside this range. Furthermore, the stress-strain behavior of higher strength concrete (above 50
MPa) is known to be different to that of normal strength concrete (particularly in relation to triaxial state
of stress) so the same confinement relationships do not apply outside of this range.

The diameter of a circular column, D, or effective diameter of a rectangular or square column, Deq ≥ 150 mm;
The majority of testing has been carried out on relatively small-scale columns, the vast majority with a
cross-sectional dimension of 150 mm. Some larger specimens have been tested to demonstrate
applicability of models and behaviour at more realistic scale (Darby et al. [7] for rectangular columns,
Youssef [52]). The models have therefore been verified for columns of effective diameter 150 mm and
above.

The first order eccentricity satisfies the condition e0/Deq < 0,20;
While there is some debate around an appropriate limit concerning eccentricity of loading, Song et al.
[6] show that for FRP confined square columns tested with eccentricities greater than e/D of 0,25, the
capacity increase is negligible compared to the unconfined equivalent. While there are a few other
studies on eccentrically loaded columns, there is limited data to support a less onerous condition or
modifications to the model.
The basis of the confinement model assumes confinement around the full column perimeter, full bond
to the substrate and failure governed by rupture of the FRP. However, where significantly eccentric
loads are applied a tensile stress in the FRP does not exist around the full perimeter, altering
confinement within the concrete. For more extreme eccentricities debonding of the FRP can occur due
to shear stresses at the FRP-concrete interface (particularly in rectangular columns), reducing
confinement stiffness. There is also a possibility that FRP rupture does not govern failure, but rather,
lateral deformations become gross, P-Delta effects start to influence behavior and steel can rupture in
tension (Darby et al. [7]). To avoid these issues, an e/D value of 0,20 ensures that the whole cross
section is in compression (varying linearly going from zero axial strain on one side to maximum axial
compressive strain on the opposite face) and thus some significant level of confinement exists around
the whole column perimeter.

J.8.1.2 (3) Derivation of provisions

For the design of CFRP confined concrete cross sections the following simplified bilinear stress-strain relationship
may be used:
When calculating the capacity of a column with some moment (applied or due to first order eccentricity) it is
necessary to use a full stress strain model of the concrete in compression. The common model is to assume a
parabolic curve transitioning into a linear ascending line (e.g. Lam and Teng [53], [54]), but a simplified bilinear

C. Giaccio, A. Darby 21.12.2022 Page 9/34

For citations, copying and referencing see Note on Front page! page 667
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

model (e.g. Darby et al. [7], Binici [55]) is much simpler to apply, is conservative, and gives perfectly acceptable
results, transitioning at a strain of 0,00175.
A rectangular stress block is not suitable for confined concrete, due to the significant variation in additional
strength that can be provided.

The value of 𝜀 can be taken as 0,006 unless more accurate information is available.

Not only is ultimate strength enhanced by confinement, but ultimate strain is also significantly increased. The
ultimate compressive strain for FRP confined concrete varies significantly and models to predict this strain are
highly empirical with a large scatter. It is therefore suggested that a lower bound (based on the database of
experimental results used to validate the models) of 0,006 be used for the ultimate strain for the purposes of
developing a conservative stress-strain model.

The provisions in FprEN 1992-1-1, Annex J for the increase in compressive strength of concrete due to
J.8.1.2 (4)
confinement from FRP are adapted from fib-90 (2019) [3] as outlined below.
The aim of the amendments to fib-90 [3] provisions is to implement an increment increase in compressive strength
outlined in clause 8.1.4.
Rectangular Columns:
According to fib-90 [3], Equation (6.55):
, , ,
1 3,3 ∙ ∙𝛼 ∙ ∗
∙ 𝑓𝑜𝑟 ∙𝛼 ∙ ∗
∙ 0,07 (BD-J.8.1)

, ,
1 𝑓𝑜𝑟 .𝛼 . ∗
. 0,07 (BD-J.8.2)

fib-90 [3] defines 𝛼 as follows:

𝛼 𝛼 .𝛼 .𝛼 (BD-J.8.3)
Expanding these terms definitions provided in fib-90 [3], and using the symbols defined in FprEN 1992-1-1, the
following is obtained:
∙ ∙
𝛼 1 ∙ 1 ∙ 1 ∙ (BD-J.8.4)
∙ ∙ ∙ ∙

Where bf, sf, and bf are defined in Figure J.2 (b), while h, b and rc are defined in Figure J.2 (a), in FprEN 1992-1-1.
For ease of use, it was noted that the 2nd – 4th components of af will not always be used in practice. These terms
only apply to helical or intermittently wrapped systems, therefore the terms was re-written as follows:
𝛼 𝑘 ∙𝑘 (BD-J.8.5)
where:
∙ ∙
𝑘 1 (BD-J.8.6)
∙ ∙

𝑘 1 1 ∙ (BD-J.8.7)
∙ ∙

Replacing fib 90 [3] symbols with FprEN1992-1-1 symbols and re-arranging, the following is obtained:

𝑓 , 𝑓 𝑘 ∙ ∙𝑘 ∙ ∙𝑘 ∙𝑓 𝑓𝑜𝑟 ∙𝑘 ∙ ∙𝑘 ∙ 0,07 (BD-J.8.8)

𝑓 , 𝑓 𝑓𝑜𝑟 ∙𝑘 ∙ ∙𝑘 ∙ 0,07 (BD-J.8.9)

The term kh is not explicitly incorporated into the formulae as it is the exception rather than the rule. However
FprEN1992-1-1 allows 𝑓 to be factored by kh if helical/intermittent wrapping is used.
The intention in the provisions in FprEN1992-1-1 was to retain the same approach as Section 8.1.4 in the main
text, the term fcd was introduced such that:
𝑓 , 𝑓 ∆𝑓 (BD-J.8.10)

Substituting formula (BD-J.8.5) into formulae (BD-J.8.3) and (BD.J.8.4), and simplifying, the following Formulae
is used in FprEN 1992-1-1.

∆𝑓 𝑘 ∙ ∙𝑘 ∙ ∙𝑘 ∙𝑓 𝑓𝑜𝑟 ∙𝑘 ∙ ∙𝑘 ∙ 0,07 (BD-J.8.11)

∆𝑓 0 𝑓𝑜𝑟 ∙𝑘 ∙ ∙𝑘 ∙ 0,07 (BD-J.8.12)

C. Giaccio, A. Darby 21.12.2022 Page 10/34

For citations, copying and referencing see Note on Front page! page 668
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

The value of 3,3 in the fib 90, equation (BD-J.8.1), has been substituted by kcc. Comparing the model with the
database of tests, a value of kcc of 1,5 is recommended as a safe upper limit. Due to the complexity of the model
and numerous factors involved, there is considerable scatter of results from the model compared to the measured
results as shown in the figure below, resulting in a high mean value for the ratio of measured/predicted strength
increase and a high standard deviation. A value of kcc = 1,5 ensures that there is 95% confidence in the
measured/predicted strength increase > 1 (with material partial safety factors set to unity) for tests which meet
the criteria in equation (BD-J.8.11).

The term kr, defined as

0,5 ∙ ∙ 2 for 𝑟 50 mm
𝑘 (BD-J.8.13)
0,5 for 𝑟 50 mm
accounts for the reduction in strength observed for FRP wrapped around corners. This is consistent with fib 90 [3]
and is used for both confinement and shear strengthening. Corner radius above 50 mm is rare in tests, due to the
need to maintain concrete cover to internal reinforcement, but the upper limit on kr is 0,5. This is consistent with
the strength reduction factor for the FRP for confinement of circular columns, based on a lower limit observed in
tests.

Circular Columns
Similar substitutions outlined for rectangular columns were applied to arrive at provisions for circular columns in
FprEN 1992-1-1. Resulting in the following formulae:
∙ ∙ , ∙
∆𝑓 0 for 0,07 (BD-J.8.14)
∙

∙ ∙ ∙ , ∙
∆𝑓 𝑘cc ∙ ∙ 0,5 ∙ 𝑓 for 0,07 (BD-J.8.15)
∙

The term kr is omitted as it is always taken as 0,5 (i.e. radius > 50mm) and, as such, is directly included in the
formulae.
The term kh for helical/intermittent FRP confinement of circular columns is not provided in fib 90, but has been
derived on the same basis as the equation for kh given for rectangular columns and is applied in the same way.

𝑘 1 ∙ (BD-J.8.16)
∙

Due to the relative simplicity of the model, there is considerably less scatter in results when comparing the
measured strength increase from the experimental database with the predictions from the model, as shown in the
figure below.

C. Giaccio, A. Darby 21.12.2022 Page 11/34

For citations, copying and referencing see Note on Front page! page 669
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

As a result of the lower mean and standard deviation for this model, the factor kcc = 2,5 ensures that there is 95%
confidence that the measured/predicted strength increase > 1 (with material partial safety factors set to unity) for
tests which meet the criteria in formula (BD-J.8.15). This is lower than the 3,3 suggested in fib 90 [3] but gives a
safe upper bound to the value.

C. Giaccio, A. Darby 21.12.2022 Page 12/34

For citations, copying and referencing see Note on Front page! page 670
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

J.8.2.1 General verification procedure

The reinforced concrete members requiring strengthening in shear (beams or columns) can be strengthened by
using CFRP materials with orthogonally applied fibres in the direction of the axis of the member. There are several
strengthening systems for shear strengthening of beams containing: closed wrapped CFRP systems and opened
CFRP systems, with discrete or continuous application of the strengthening system (see FprEN 1992-1-1, Figure
J.3).

Shear strengthening for columns contains the only fully closed systems (Fig. BD-J.8.1).

(a) (b)
Figure BD-J.8.1: Shear strengthening of columns with
(a) closed – continuous application of the strengthening system and
(b) closed – discrete application of the strengthening system. [3]
Shear strengthening is required if the average shear strength over the cross-section Ed (Eq. (8.18)) is larger than
Rd,c (Eq. (8.27)) in case the unstrengthened element does not have shear reinforcement, or when the average
shear strength Ed is larger than Rd,sy (Eq. (8.42)) or Eq. (8.44) is not fulfilled if the unstrengthened element has
shear reinforcement. That is, in cases where the internal steel reinforcement is inadequate following the detailing
rules of EN 1992-1-1 the difference may be covered using ABR CFRP in accordance with J.8.2.3. The method
described in Annex J allows superposition with the basic section shear capacity and is similar to that given in [3].
𝜏 , 𝜏 𝜏 , (BD-J.8.17)
Where:
 , 𝜌 𝑓 cot 𝜃 cot  sin  (BD-J.8.18)

𝜌 (BD-J.8.19)
∙

∙ 𝑠𝑖𝑛𝛼 for discrete CFRP strips or CF sheets

(BD-J.8.20)
2𝑡 ∙ 𝑠𝑖𝑛𝛼 for continous CF sheets
𝜌 is the externally applied FRP reinforcement ratio
𝛼 is the angle formed between the CFRP system;
𝑓 is design shear strength of the CFRP system intersected by the shear crack in the ultimate limit
state and calculated in accordance with formulae (J.18) to (J.23), depending on the CFRP configuration.
This design shear strength depends on the shear strengthening configuration (fully wrapped or open
CFRP systems). Hereby, it is assumed that the steel stirrups have sufficient deformation capacity, so
that the FRP can reach its design strength 𝑓 . This might be verified by comparing 𝑓 ⁄𝐸 with the
deformation capacity of the steel (with a given ductility class).
𝐴 is the area of externally bonded shear FRP reinforcement measured perpendicular to the direction
of the fibres at the spacing 𝑠 (Fig. J.4)
𝑏 and 𝑡 are the width and effective thickness of the FRP materials
An analysis of the capacity of the compression struts in accordance with Chapter 8.2 shall also be furnished for
shear strengthening with the selected angle θ of the compression strut, as follows:
 
 𝑓 
(BD-J.8.21)

Unless more rigorous analysis is undertaken,  should be taken as 45 degrees for the calculation of Rd and Rd,f.

R. Kotynia, E. Oller, A. Darby 21.12.2022 Page 13/34

For citations, copying and referencing see Note on Front page! page 671
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

If the structure has not the shear strengthening, at least a fraction equal to 3 = 0,5 of the necessary shear
reinforcement should be covered by stirrups. If this condition is not satisfied, the difference in the shear strength
shall be covered by the externally applied CFRP reinforcement.
For the effective thickness of CF sheets Eq. (J.1) has to be considered, so the reduced effectiveness of a jacket
with many layers is taken into account.
In members with flexural strengthening in accordance with J.8.1 and with Ed > Rd,c , the shear verification of
J.8.2.3 is applicable.
Additional Information on shear strengthening in relation to flexural debonding can be found in J.11.1.2.3.

J.8.2.3 (5) Closed wrapped CFRP systems

The design shear strength of closed wrapped FRP systems (see FprEN 1992-1-1, Figure J.3c) is defined as:
𝑓 𝑓 , 0,8 𝑘 ∙ 𝑓 (BD-J.8.22)
where: 0,8 represents a long-term loading factor, 𝑘 is the reduction factor, which accounts for the non-uniform
distribution of stress in the FRP intersecting the shear crack and for the reduction of FRP strength due to bending
of the fibres at the corners of the cross section. Assuming that the radius at the corners is R (mm), 𝑘 is obtained
as follows (see also FprEN 1992-1-1, formula (J.13)):

0,5 2 𝑅 50 𝑚𝑚
(BD-J.8.23)
0,5 𝑅 50 𝑚𝑚
𝑓 can be obtained from Eq. (J.2)

J.8.2.3 (6) Open CFRP systems without anchorages

The shear capacity is determined by the minimum strength of the FRP reinforcement as if it is closed, 𝑓 , and
the bond strength, 𝑓 , 𝑏 Eq. (BD-J.8.25 to BD-J.8.27). The bond strength is obtained as a function of the
geometric assumptions shown in (Fig. BD-J.8.2) and a bilinear bond law shown in (Fig. BD-J.8.3).
𝑓 min 𝑓 , ,𝑓 , (BD-J.8.24)

Figure BD-J.8.2: Strips crossed by the shear crack. (le corresponds to lbf in FprEN1992-1-1)

Figure BD-J.8.3: Generic bilinear b-s constitutive law for bond in EBR FRP systems.

(a) for ℎ /sin 𝑙 , , and 𝑙 , , 𝑠 / cot𝜃 cot sin ℎ /sin , if all strips are intersected by the
shear crack the bond length is ≥ 𝑙 , , and 𝑓 is calculated according to:

𝑓 , (BD-J.8.25)

(b) for ℎ /sin 𝑙 , , and 𝑠 / cot𝜃 cot sin 𝑙 , , , if all strips are intersected by the shear crack,
the bond length is < 𝑙 , , , and 𝑓 is calculated according to:

R. Kotynia, E. Oller, A. Darby 21.12.2022 Page 14/34

For citations, copying and referencing see Note on Front page! page 672
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

𝑓 , 1 1 (BD-J.8.26)

(c) for ℎ /𝑠𝑖𝑛𝛼 𝑙 , , and 𝑠 / 𝑐𝑜𝑡𝜃 𝑐𝑜𝑡𝛼 𝑠𝑖𝑛𝛼 ℎ /𝑠𝑖𝑛𝛼, if all strips are intersected by the shear crack,
the bond length < 𝑙 , , , and 𝑓 is calculated according to:

𝑓 , (BD-J.8.27)
, ,

where: n is the number of strips crossed by the shear crack:

𝑛 ℎ cot𝜃 cot𝛼 /𝑠 (BD-J.8.28)
m is the number of strips for which the bond length is less than integer quotient of:
𝑙 , , 𝑐𝑜𝑡𝜃 𝑐𝑜𝑡𝛼 𝑠𝑖𝑛𝛼/𝑠 .
𝑙 , , is maximum bond length in accordance with J.11.1. In dependence of the chosen method,
simplified method in J.11.1.2 or refined method in J.11.1.3 formulae (J.37) resp. (J.39) have to be
applied.
𝑓 is characteristic maximum bond strength:

⎧ ∙ 2 𝑓𝑜𝑟 𝑠 𝑙 , ,
𝑓 𝑠 (BD-J.8.29)
⎨ 𝑓𝑜𝑟 𝑠 𝑙 , ,
⎩
The characteristic values of the shear strength 𝜏 and the ultimate slip 𝑠 can be taken from Table BD-J.8.1.
Table BD-J.8.1 - Parameters of the bilinear 𝜏 - s constitutive law for bond.
Type τb1 s1 (mm) s0 (mm)
Mean value 0.53 fcm fctm 0.0063 0.21
CFRP strips
5% characteristic value 0.37 fcm fctm - 0.20
Mean value 0.72 fcm fctm 0.0107 0.24
CFRP sheets
5% characteristic value 0.44 fcm fctm - 0.23

J.8.2.3 (7) Full area bond

Continuous CFRP sheets may be treated as a special case of CFRP strips with 𝑠 𝑏 /𝑠𝑖𝑛𝛼.
, ,
For ℎ /𝑠𝑖𝑛𝛼 𝑙 , , : 𝑓 , 1 (BD-J.8.30)
/

⁄
For ℎ /𝑠𝑖𝑛𝛼 𝑙 , , : 𝑓 , (BD-J.8.31)
, ,

Open CFRP systems with anchorage in the compression zone

Anchorage systems are not covered in FprEN 1992-1-1. For application the technical product specification of the
anchorage system has to be considered.
In case of three-sided CFRP strips/sheets anchored in the compression zone (Fig. BD-J.8.4), the effectiveness
coefficient is equal to 𝑘 , so the strength 𝑓 is:
𝑓 𝑘 𝑓 , (BD-J.8.32)

with 𝑘 ≤ 0,9.
If no anchors are used 𝑘 𝑓 , /𝑓 , .

(a) (b)
Figure BD-J.8.4: Anchorage of three-sided CFRPs:
(a) Use of spike anchors in combination with CFRP sheets;

(b) anchorage of individual strips inside the slab.

Punching is excluded due to insufficient data.

J.8.4

R. Kotynia, E. Oller, A. Darby 21.12.2022 Page 15/34

For citations, copying and referencing see Note on Front page! page 673
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

At the serviceability limit state, the strain in the strengthening system must be limited for reasons of the durability
J.9
of the bond. Here is good experience with a strain limitation for rare load combinations of f < 2 mm/m for members
with existing reinforcing steel of fyk = 500 MPa. This corresponds to the more general proposal for a stress limit in
dependence of the existing reinforcing steel strength of:

𝜎 0,8𝑓 (BD-J.9.1)

The intention behind limiting strain resp. stress for rare load combinations is to prevent significant damage to the
bond of the externally bonded reinforcement. Furthermore, yielding of the reinforcing steel under rare load
combinations is ruled out in the main part, likewise to prevent high bond stresses and irreversible deformations of
the structure.
Limiting the reinforcing steel stresses alone is not sufficient, as members without reinforcing steel reinforcement
can also be strengthened.

J.10 Analysis Concept

The proposal for fatigue analysis of externally bonded CFRP in FprEN 1992-1-1, Annex J is based on the German
guideline for the strengthening of concrete members with adhesively bonded reinforcement [1].
The fatigue analysis of externally bonded CFRP proposed in Annex J.10 is divided into two parts. The first part of
the proof is the basic fatigue analysis in J.10.1. Here, the upper load level is used to check whether damage due
to cyclic loading can occur. The second part is the refined fatigue analysis in J.10.2, where the fatigue range S is
used for the determination of a tolerable number of load cycles N on the basis of an S-N curve.

J.10.1 Basic fatigue analysis for externally bonded CFRP systems

The first proof requires that the upper load remains below the load level that leads to the maximum bond stress
𝜏f1k and the corresponding slip 𝑠f1 from the bilinear composite stress-slip relationship. This ensures that the
deformations are in the elastic range and that no damage occurs. Inserting the elastic energy leads to the formula
used in the proof (J.25). Hankers [8] determines the basic fatigue resistance to:
/
∆𝐹 , 0,348 ∙ 𝑓 , ∙𝑓 ∙𝑏 ∙𝑡 (BD-J.10.1)

J.10.2 Refined fatigue analysis for externally bonded CFRP systems

The refined analyses uses the S-N curve determined from the experimental test data described in Bizindavyi [9],
Carloni [10], Budelmann [11], Leusmann [12], Dai [13] and Ferrier [14]. For determining the number of load cycles
N30 needed for reaching a debonded length of 30 mm, a linear increase of the debonded length is assumed. The
number of load cycles needed for reaching a decoupled length of 30 mm is calculated for the tests from Bizindavyi
[9], Carloni [10], Dai [13] and Ferrier [14] by linear interpolation using the bond length and the number of load
cycles until full decoupling. For the tests from Budelmann [11] and Leusmann [12] the number of load cycles is
taken from the strain measurements. A decoupling length of 30 mm is reached when the strain measured with
the strain gauge placed 30 mm from the loaded end exceeds the free strain measured with the strain gauge
located in the unbounded region of the CFRP-plate.
For fitting the S-N curve unified related load ranges S0,i at a lower load level of 0 and the corresponding number
of load cycles N30 is needed. The unified load range S0,i is determined in a projection analysis using the Goodman
relation and Equation (BD-J.10.2), see [15].
∆𝐹𝑓,𝑚𝑎𝑥 ∆𝐹𝑓,𝑚𝑖𝑛
∆𝐹𝑓,0,𝑖 ∆𝐹𝑏𝑓,𝑚 ∆𝐹𝑓,𝑚𝑎𝑥 ∆𝐹𝑓,𝑚𝑖𝑛
𝑆0,𝑖 (BD-J.10.2)
∆𝐹𝑓𝑅𝑑 ∆𝐹𝑓,𝑚𝑖𝑛 ∆𝐹𝑏𝑓,𝑚 ∆𝐹𝑓,𝑚𝑖𝑛
1
∆𝐹𝑏𝑓,𝑚

The Procedure of the Goodman-projection is depicted in Figure BD-J.10.1.

Experimental data from Bizindavyi [9], Carloni [10], Budelmann [11], Leusmann [12], Dai [13] and Ferrier [14] with
the calculated number of load cycles N30 and the corresponding load range S0, as well as the width bf, the
thickness tf, the number of layers nf and Young’s modulus Ef of the tested FRP can be found in Table BD-J.10.1
and Table BD-J.10.2.
Figure BD-J.10.2 shows the S-N curve fitted to the experimental data. The load range S0 is the difference between
maximum and minimum load - related to the monotonic quasi-static load-carrying capacity of the interface.

W. Finckh, T. Leusmann, K. Zilch, R. Niedermeier 21.12.2022 Page 16/34

For citations, copying and referencing see Note on Front page! page 674
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

Figure BD-J.10.1: Goodman projection for determining the load range S0,i

Figure BD-J.10.2: S-N curve and experimental data from Bizindavyi [9], Carloni [10], Budelmann [11], Leusmann
[12], Dai [13] and Ferrier [14]

W. Finckh, T. Leusmann, K. Zilch, R. Niedermeier 21.12.2022 Page 17/34

For citations, copying and referencing see Note on Front page! page 675
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

Table BD-J.10.1 : Experimental data from Bizindavyi [9].

Ref. Name bf tf Ef lbf F fU F fO N30 S0

[mm] [mm] [N/mm²] [mm] [N] [N] [-] [-]
Biz. [9] C1-1a 25,4 0,33 29200 300 0 6480 30 0,76
Biz. [9] C1-2a 25,4 0,33 29200 300 0 5029 17323 0,59
Biz. [9] C1-2b 25,4 0,33 29200 300 0 5029 28656 0,59
Biz. [9] C1-3a 25,4 0,33 29200 300 0 4572 21025 0,54
Biz. [9] C1-3b 25,4 0,33 29200 300 0 4572 29790 0,54
Biz. [9] C1-3c 25,4 0,33 29200 300 0 4572 34739 0,54
Biz. [9] C1-3d 25,4 0,33 29200 300 0 4572 67608 0,54
Biz. [9] C1-4a 25,4 0,33 29200 300 0 3505 186222 0,41
Biz. [9] C1-4b 25,4 0,33 29200 300 0 3505 202950 0,41
Biz. [9] C2-1a 50,8 0,33 29200 160 0 11135 624 0,66
Biz. [9] C2-1b 50,8 0,33 29200 160 0 11135 788 0,66
Biz. [9] C2-2a 50,8 0,33 29200 160 0 9185 6526 0,54
Biz. [9] C2-2b 50,8 0,33 29200 160 0 9185 26832 0,54
Biz. [9] C2-3a 50,8 0,33 29200 160 0 7071 34674 0,42
Biz. [9] C2-3b 50,8 0,33 29200 160 0 7071 93133 0,42
Biz. [9] C3-1a 50,8 0,33 29200 160 0 11460 29 0,67
Biz. [9] C3-1b 50,8 0,33 29200 160 0 11460 60 0,67
Biz. [9] C3-2a 50,8 0,33 29200 160 0 9453 113 0,56
Biz. [9] C3-2b 50,8 0,33 29200 160 0 9453 407 0,56
Biz. [9] C3-3a 50,8 0,33 29200 160 0 7453 40022 0,44
Biz. [9] C3-3b 50,8 0,33 29200 160 0 7453 40352 0,44
Biz. [9] C4-1a 50,8 0,66 29200 300 0 17496 37 0,58
Biz. [9] C4-1b 50,8 0,66 29200 300 0 17496 39 0,58
Biz. [9] C4-2a 50,8 0,66 29200 300 0 14402 190 0,48
Biz. [9] C4-2b 50,8 0,66 29200 300 0 14402 1062 0,48
Biz. [9] C4-3a 50,8 0,66 29200 300 0 11232 7650 0,37
Biz. [9] C4-3b 50,8 0,66 29200 300 0 11232 26687 0,37
Biz. [9] C5-1a 50,8 0,66 29200 300 0 18882 38 0,63
Biz. [9] C5-1b 50,8 0,66 29200 300 0 18882 97 0,63
Biz. [9] C4-2a 50,8 0,66 29200 300 0 16825 195 0,56
Biz. [9] C5-2b 50,8 0,66 29200 300 0 16825 1430 0,56
Biz. [9] C5-3a 50,8 0,66 29200 300 0 15728 18682 0,52
Biz. [9] G1-1a 25,4 1 29200 220 0 6426 5523 0,56
Biz. [9] G1-1b 25,4 1 29200 220 0 6426 6830 0,56
Biz. [9] G1-2a 25,4 1 29200 220 0 4890 8768 0,43
Biz. [9] G1-2b 25,4 1 29200 220 0 4890 9627 0,43
Biz. [9] G1-3a 25,4 1 29200 220 0 4543 112227 0,4
Biz. [9] G1-3b 25,4 1 29200 220 0 4543 135273 0,4
Biz. [9] G2-1a 25,4 1 29200 300 0 6477 2390 0,57
Biz. [9] G2-1b 25,4 1 29200 300 0 6477 7832 0,57
Biz. [9] G2-2a 25,4 1 29200 300 0 4953 21714 0,43
Biz. [9] G2-2b 25,4 1 29200 300 0 4953 35949 0,43
Biz. [9] G2-2c 25,4 1 29200 300 0 4953 42308 0,43
Biz. [9] G2-3a 25,4 1 29200 300 0 4420 88609 0,39
Biz. [9] G2-3b 25,4 1 29200 300 0 4420 245893 0,39

W. Finckh, T. Leusmann, K. Zilch, R. Niedermeier 21.12.2022 Page 18/34

For citations, copying and referencing see Note on Front page! page 676
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

Table BD-J.10.2 : Experimental data from Carloni [10], Budelmann [11], Leusmann [12], Dai [13] and Ferrier
[14].

Ref. Name bf tf Ef lbf F fU F fO N30 S0

[mm] [mm] [N/mm²] [mm] [N] [N] [-] [-]
Car. [10] DSF1 25 0,167 230000 152 1250 7590 255 0,75
Car. [10] DSF2 25 0,167 230000 152 1100 7590 2604 0,61
Car. [10] DSF3 25 0,167 230000 152 1100 7590 23091 0,53
Bud. [11] C201 50 1,4 170000 1150 3820 28382 127862 0,40
Bud. [11] C202 50 1,4 170000 1150 5117 28382 415550 0,55
Bud. [11] C203 50 1,4 170000 1150 5819 28382 1196372 0,46
Bud. [11] C204 50 1,4 170000 1150 9032 30190 156751 0,34
Bud. [11] C205 50 1,4 170000 1150 15292 34296 188462 0,38
Bud. [11] C206 50 1,4 170000 1150 21729 30190 462744 0,38
Bud. [11] C207 50 1,4 170000 1150 22884 30190 462744 0,45
Bud. [11] C208 50 1,4 170000 1150 5236 30952 581 0,60
Bud. [11] C209 50 1,4 170000 1150 10436 30952 428569 0,39
Bud. [11] C2010 50 1,4 170000 1150 10175 30952 428569 0,39
Bud. [11] C401 50 1,4 170000 1150 3249 24847 572408 0,49
Bud. [11] C402 50 1,4 170000 1150 11662 24847 2000000 0,45
Bud. [11] C501 50 1,4 170000 1150 25811 34950 61811 0,44
Bud. [11] C502 50 1,4 170000 1150 25478 34950 61811 0,43
Bud. [11] C503 50 1,4 170000 1150 5284 34950 746055 0,37
Bud. [11] C504 50 1,4 170000 1150 5986 34950 269850 0,39
Leu. [12] 12A 50 5 95800 1150 5500 36000 326600 0,43
Leu. [12] 12B 50 5 95800 1150 5700 36000 679003 0,45
Leu. [12] 13A 50 5 95800 1150 5300 36000 19130 0,59
Leu. [12] 13B 50 5 95800 1150 6200 36000 21821 0,62
Leu. [12] 14B 50 5 95800 1150 6600 36000 1084 0,72
Leu. [12] 22A1 50 5 95800 1150 15200 36000 33780 0,45
Leu. [12] 22B 50 5 95800 1150 14700 36000 405678 0,50
Leu. [12] 22A2 50 5 95800 1150 15200 36000 53531 0,45
Leu. [12] 22A3 50 5 95800 1150 15200 36000 107046 0,45
Leu. [12] 24B 50 5 95800 1150 15300 36000 3440 0,84
Dai [13] S1-1 60 0,11 230000 150 0 17958 57 0,73
Dai [13] S1-2 60 0,11 230000 150 0 15744 4615 0,64
Dai [13] S1-3 60 0,11 230000 150 0 12546 299162 0,51
Fer. [14] A1 16,2 1,38 67900 200 0 10284 3 0,86
Fer. [14] A2 16,1 1,38 67500 200 0 10220 27 0,86
Fer. [14] A3 16,2 1,38 67300 200 0 8942 59 0,75
Fer. [14] A4 16,1 1,38 67200 200 0 8887 118 0,75
Fer. [14] A5 16,2 1,38 66800 200 0 8048 720 0,67
Fer. [14] A6 16,2 1,38 66800 200 0 8048 542 0,67
Fer. [14] A7 16,2 1,38 66300 200 0 7154 6001 0,59
Fer. [14] A8 16,2 1,38 66200 200 0 6707 9653 0,56
Fer. [14] A9 16,2 1,38 65900 200 0 6148 47948 0,51
Fer. [14] A10 16,2 1,38 65700 200 0 5813 120251 0,48
Fer. [14] B1 12,8 2,57 54400 200 0 17797 2425 0,64
Fer. [14] B2 12,8 2,9 50700 200 0 18968 5210 0,6
Fer. [14] B3 12,7 2,65 51600 200 0 16222 74385 0,56
Fer. [14] B4 12,7 2,61 48900 200 0 14982 207728 0,53
Fer. [14] B5 12,7 2,68 - 200 0 13342 573517 0,46

W. Finckh, T. Leusmann, K. Zilch, R. Niedermeier 21.12.2022 Page 19/34

For citations, copying and referencing see Note on Front page! page 677
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

J.10.3 Fatigue analysis for Near surface mounted CFRP strips

The proposal for the fatigue analysis for Near surface mounted CFRP strips is based on the DAfStb Guideline
“Strengthening of Concrete Members with Adhesively Bonded Reinforcement” [1].

Description
When checking fatigue for non-static loads, the Annex J can be used to verify the bond of flexural strengthening
in the form of near-surface-mounted CFRP strips. As the carbon fibres exhibit virtually no signs of fatigue, only
the bond needs to be checked for fatigue when using CFRP strips. Besides the fatigue of the strengthening
system, the concrete, reinforcing steel and prestressing steel must also be checked according to EN 1992.
In contrast to externally bonded CFRP strips, however, there is no comprehensive analysis concept available for
near-surface-mounted strips. Owing to the low number of fatigue tests involving near-surface-mounted CFRP
strips (see [16]), a quasi-fatigue strength analysis is the only option here. With so few test results available, it is
not possible to specify an S-N curve for near-surface-mounted reinforcement. And as an S-N curve is unavailable,
it is not possible to extrapolate for a number of load cycles greater than that given in the test results. Therefore,
the analysis can only assume sufficient fatigue resistance for max. 2ˑ106 load cycles. Design methods for numbers
of load cycles > 2ˑ106 are not covered in the DAfStb guideline [1].
In this analysis, adequate resistance to fatigue for near-surface-mounted CFRP strips may be assumed for up to
2 ꞏ 106 load cycles provided the end anchorage force for a frequent cyclic action to EN 1992 and taking into
account the “shift rule”, does not exceed the value 0,6 𝑓 and the strip stress range does not exceed a value
given by Eq. (J.32). The strip thickness 𝑡 in mm should be used here so that the result is an admissible stress
range in N/mm².

Background
The analysis of the fatigue resistance of the bond of near surface mounted CFRP strips is based on a small
number of tests described in [16] and [3]. Hence it is not possible to provide a S-N-curve for the bond of near
surface mounted CFRP strips. Owing to the lack of a S-N-curve it is not possible to extrapolate to a greater number
of cycles than investigated in the tests. In this analysis it can therefore only be assumed that adequate fatigue
resistance is provided up to 2 ∙ 106 load cycles ).
The few fatigue tests were carried out on individual products. It should therefore be additionally checked for each
product that the following conditions are fulfilled:
 The upper load should be less than 60% of the composite failure load at the point of adhesive cohesion
failure (transition to friction plateau):
𝐹, , 0,6 ⋅ 𝑓 ⋅ 𝑏 ⋅ 𝑡 with 𝑓 according to Formula (J.56)

 The amplitude should be less than 15% of the composite failure load at the point of adhesive cohesion
failure (transition to friction plateau):
Δσ 0,15 ⋅ 𝑓 with 𝑓 according to Formula (J.56)

W. Finckh, T. Leusmann, K. Zilch, R. Niedermeier 21.12.2022 Page 20/34

For citations, copying and referencing see Note on Front page! page 678
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

The fctm,surf formulation in FprEN 1992-1-1, Eq. (J.34) is based on the lower bound of the possible formulations.
J.11.1.1.1
Other possible formulations taking into consideration the dependence of concreting position on the surface tensile
strength by the factor k are possible [18], [19], [20]. Alternatively, usual ranges are shown in figure BD-J.11.3 for
the different concreting positions so that the designer gets some feeling for possible problems. The values
included in this diagram are based on the arithmetic mean from at least four pull-off tests according to EN 1542
which had been performed on concrete specimens made from globular limestone aggregates from the Munich
area.

Figure BD-J.11.3 Usual ranges for the factor k based on [18]

J.11.1.1.2 Basic anchorage resistance – Simplified method

The Simplified method is based on the following conservative simplification of the refined method in Chapter
J.11.1.1.3.
Equation J.38 can be reformulated as follows
⋅ ⋅
𝑓 ⋅𝛽 ⋅𝑓 , (BD-J.11.1)

Where:

2 1 if 𝑙 𝑙 , ,
𝛽 , , , , (BD-J.11.2)
1 if 𝑙 𝑙 , ,

Equation J.40 can be reformulated with the default values of J.11.1.1.3 𝜏 0,37 ⋅ 1,0 ⋅ 𝑓 , ⋅𝑓 and ∙
𝑠 0,2 (ksys = 1,0) as follows

𝑓 , 0.272 ⋅ 𝑓 , ⋅𝑓 (BD-J.11.3)

The front part of the equation (BD-J.11.1) can be conservative reformulated with ktc = 0.85 , kcc=0.7 and 𝜂
/
0.93
.
𝑓 ⋅𝛽 ⋅𝑓 , (BD-J.11.4)

.
𝑓 ⋅ 𝛽 ⋅ 0.272 ⋅ 𝑓 , ⋅𝑓 (BD-J.11.5)

Equation J.39 can be reformulated with the default values of J.11.1.1.3 𝜏 0,37 ⋅ 1,0 ⋅ 𝑓 , ⋅𝑓 and ∙
𝑠 0,2 (ksys = 1,0) as follows

∙ ∙ ,
𝑙 , ∙ (BD-J.11.6)
, ⋅ , ⋅

𝐸𝑓 ∙𝑡𝑓
𝑙 , 𝑙𝑏𝑓,𝑚𝑎𝑥 1,5 ∙ (BD-J.1175)
𝑓𝑐𝑡𝑚,𝑠𝑢𝑟𝑓 ⋅𝑓𝑐𝑚

W. Finckh, E. Oller, S. Matthys 21.12.2022 Page 21/34

For citations, copying and referencing see Note on Front page! page 679
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

J.11.1.1.3 Basic anchorage resistance - Refined method

This section explains the Intermediate crack debonding. The proposal is based on the DAfStb Guideline
“Strengthening of Concrete Members with Adhesively Bonded Reinforcement” [1] and the fib-bulletin 90 [3] with
some adjustments.

Figure BD-J.11.2 Determination of the bond forces at the end anchorage using the differential equation for bond
and the bilinear bond approach, based on [18], [21], [22]
The basic anchorage resistance - CFRP to concrete corresponds to the tensile stress that can be anchored by
the EBR CFRP at a single crack. The refined method for determining this tensile stress is based on a slip-bond
stress relationship that can be derived from force and slip measurements in bond tests with EBR CFRP. The slip-
bond stress relationship of EBR CFRP can be idealised using a bilinear approach as shown in Figure BD-J.11.2.
Up the maximum bond stress f1, a linear relationship between the slip and the bond stress applies. After
exceeding the maximum bond stress, progressive softening occurs until complete debonding of the EBR CFRP
occurs when the critical slip sf0 is reached. If the friction occurring in practical applications at even larger slip values
after debonding is neglected, the idealised bilinear bond approach provides a good approximate
phenomenological description of the complex processes in the composite system. The size of the area under the
bilinear slip bond stress curve is of the unit Nmm/mm2 and corresponds to an areic energy Gf, the bond fracture
energy which must be expended to completely detach a unit area. From this, it can be seen that a fracture energy
limitation must be considered when anchoring EBR CFRP at a single crack. An increase in the tensile stress
prevailing in the crack cross-section is only possible up to a certain maximum effective bond length lbf,max. Larger
bond lengths do not lead to higher tensile stresses that can be anchored at the single crack due to the fracture-
energy limitation. Therefore, additional cracks are required to build up higher tensile stresses, since the bond
fracture energy defined by the bilinear bond approach can then be activated at each intermediate crack element
located between adjacent cracks.
However, for the differential equation that describes the relationship between relative displacement (slip) sf x and
stress f x , neglecting the concrete deformations for low concrete stress c, no analytical solution suitable for
manual calculation can be found for the section-wise defined bilinear bond approach with the corresponding
boundary and transition conditions at a single crack that allows a determination of the tensile stress that can be
anchored as a function of the bond length lbf. In contrast, a simplified solution of the differential equation can be
obtained by the substitute approach of a linear, plastically softening slip-composite stress relation that results from
the bilinear approach with sf1 = 0 and has the same magnitude of composite fracture energy.

1 ∙𝜏 𝑤ℎ𝑒𝑟𝑒 𝑠 𝑠
𝜏 𝑠 (BD-J.11.10)
𝑤ℎ𝑒𝑟𝑒 𝑠 𝑠
0
For this simplified slip-bond stress relationship, the tensile stress f that can be anchored at the single crack may
be determined as a function of the bond length x = lbf:
∙ ∙
𝜎 𝑙 ∙ sin ∙𝑙 (BD-J.11.11)
∙ ∙

On basis of this equation the maximum anchorable tensile stress fbf,max is determined for:
∙ ∙
𝑙 , ∙ . (BD-J.11.12)

∙ ∙
𝑓 , (BD-J.11.13)

However, since the linear elastic branch of the bilinear bond approach was neglected for the approximate solution,
equation (BD-J.11.12) determines anchorage lengths that are too short compared to calculations that consider
both the linear elastic and the plastic softening branch. For a design approach, equation (BD-J.11.12) must
therefore be multiplied by the factor 4⁄ 𝜋 ∙ 𝑘 , , which leads to equation (BD-J.11.14) and (J.33).

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 22/34

For citations, copying and referencing see Note on Front page! page 680
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

∙ ∙
𝑙 , ∙ . (BD-J.11.14)
,

If the sinusoidal course of the tensile stress 𝜎 𝑙 that can be anchored as a function of the bond length 𝑙 is
replaced by a quadratic parabola, which approximately also describes the exact course on the basis of the bilinear
bond approach with sufficient accuracy, the following equation is obtained

𝑓 , ∙ ∙ 2 𝑤ℎ𝑒𝑟𝑒 𝑙 𝑙 ,
, ,
𝜎 𝑙 (BD-J.11.15)
𝑓 , 𝑤ℎ𝑒𝑟𝑒 𝑙 𝑙 ,

For the design at ultimate limit state, the characteristic values of the variables are applied and the equations are
amended with the partial safety factor BA for bond of adhesively bonded CFRP reinforcement. In addition, with
regard to the concrete strengths, the factor cc to account for the difference between the undisturbed compressive
strength of a cylinder and the effective compressive strength that can be developed in the structural member as
well as the factors ktc and ktt considering the effect of high sustained loads and of time of loading on are added to
the equations.

J.11.1.2.1 EBR CFRP anchorage requirements – flexure – General

The design of a strengthening measure must be carried out in such a way that the potential modes of failure are
avoided. The various failure modes and their critical loads must be known in order to carry out the design with a
margin corresponding to the level of safety for the critical load of the respective governing mode of failure. Figure
BD-J.11.4 provides an overview of the failure modes that can occur.

Figure BD-J.11.4: Examples of failure modes for a reinforced concrete beam with flexural strengthening in the
form of externally bonded reinforcement [23]
First of all, modes of failure related to the function of the CFRP strip can be added to those familiar from
conventional reinforced concrete:
 Failure of concrete in compression zone
 Yielding of internal reinforcement followed by concrete failure
 Yielding of internal reinforcement followed by failure of the adhesively bonded strip
 Shear failure
 Yielding of externally bonded steel plate
Besides these modes of failure well known from conventional reinforced concrete and relatively easy to describe,
there are other modes specific to strengthening measures with externally bonded reinforcement. The first of these
that should be mentioned is concrete cover separation failure, where the concrete cover becomes detached at
the end of a strip. This occurs due to the additional, vertical offset between shear link and strip because the tensile
stresses from the strip cannot be tied back to the compression zone of the beam. This mode of failure therefore

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 23/34

For citations, copying and referencing see Note on Front page! page 681
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

corresponds to a horizontal shear failure in the area between the externally bonded reinforcement and the internal
reinforcement. (End Cover Separation as described in J.11.1.2.4)
The bond between the adhesive and the concrete often fails when using externally bonded reinforcement. In such
a bond failure the layers of concrete near the surface break away once the tensile strength of the concrete has
been exceeded. Owing to the only moderate tensile strength of the layers of concrete near the surface, following
local debonding of externally bonded reinforcement, the result is mostly a total failure of the bond between the
externally bonded reinforcement and the concrete as the load rises further because the forces involved cannot
normally be carried by any remaining areas of intact bonding (unzipping effect) This behaviour means it is
necessary to consider the bond of externally bonded reinforcement very carefully.
In conventional reinforced concrete usually, bond checks are done by end anchorage verifications, which are
based on bond values from pull-out tests. If we introduce such verifications in similar form to structural elements
with externally bonded reinforcement, the full tensile strength of the CFRP-strip cannot be anchored, because
after a specific length, the bond forces cannot be increased anymore. (See also Figure BD-J.11.5). However, the
full-scale tests have shown that much higher forces in the externally bonded reinforcement are reached at the
place of the maximum bending moment, as it would be possible looking only at the end anchorage.
On CFRP strips with their high tensile strength, the sole consideration of the end anchorage would be highly
uneconomical. Thus, the bond forces of the CFRP-strips must be transferred at the point where the forces occur.
For this reason, two areas are distinguished for bond verifications, the end anchorage zone and the rest of the
structural element.
At the end anchorage zone, the forces of the CFRP-strip must be anchored, which occur at outermost bending
crack. The bearable bond forces at the end anchorage zone can be determined by so-called idealized end
anchorage tests, in which the externally bonded reinforcement will be pulled off in the longitudinal direction.
(End Anchorage as described in J.11.1.2.2)
In the remaining area of the structural element, the bond force can be transferred by elements, which are
separated by bending cracks, the so-called intermediate crack element.
On such an intermediate crack there is always a basic force in the strip at lower stressed crack and at the higher
stressed crack there is this basic force plus an additional force. This additional force must be transferred by bond
to the structural element. (Intermediate Crack Debonding as described in J.11.1.2.3)

Figure BD-J.11.5: Principle of the bond force transfer of externally bonded CFRP-strips [24]
In addition to the failure modes described above a premature uncoupling of the externally bonded reinforcement
at shear cracks due to high shear force may occur. (End Anchorage as described in J.11.1.2.2)

J.11.1.2.2 End anchorage

As was already explained in section J.11.1.2.1, an analysis of the end anchorage is necessary in addition to
checking the bond at the element between cracks. The end anchorage analysis can be performed in two different
ways depending on requirements:

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 24/34

For citations, copying and referencing see Note on Front page! page 682
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

 The analysis at the flexural crack closest to the point of contraflexure represents the standard case. In
this case the FRP force acting at this flexural crack must be lower than the resistance of the cross-
section taking into account the “shift rule”
 Where strengthening is done in a small region of the member, the CFRP strengthening shall a distance
of 𝑙bf + ℎ beyond the area of need.

Additional methods for end anchorage

There are three additional methods for an end anchorage known in several codes:
 The first method is a more refined method to the concept of J.11.1.2.2. The resistance of the cross-
section is determined based on bond according to the concept [25], [26], which considers a
redistribution between the externally bonded and the internal reinforcement. Owing to the different bond
behaviour and depending on the strain state of the bonded reinforcement, a different distribution of the
forces between the various lines of reinforcement occurs, which is described via the slip of the strip.
 In the section option – analysis of end anchorage with shear wrapping – the shear wrapping, mostly
provided as a result of the design for shear or to avoid a concrete cover separation failure, may also be
employed to increase the bond force. The concept of increasing the bond force as a consequence of
shear wrapping has been borrowed from Husemann [27], [28]
 The third method is using mechanical end anchorage plates. [29]

J.11.1.2.3 Intermediate crack debonding

As was already explained in section J.11.1.2.1 an additional bond check is needed to ensure the bond force
transfer over the entire structural member.
The proposal is based on the DAfStb Guideline “Strengthening of Concrete Members with Adhesively Bonded
Reinforcement” [1] and the fib-bulletin 90 [3] with several simplifications.
This verification must be performed in two steps.
1. Calculation of the crack pattern
2. Verification of each intermediate crack element
The verification of each intermediate crack element is a limitation of the force increase of the CFRP-strip through
a constant value. The concept of a fixed constant force plate change has the advantage that it is both very clear
and easy to understand, and limits the computational effort. The concept is illustrated in figure BD-J.11.6. In this
figure it is evidence, that the slope of the force of the strip may not exceed a certain value.

Figure BD-J.11.6 limitation of the force increase [23]

The force, which can be transferred by bond by each intermediate crack element, can be calculated with equation
(J.39) and consists of three parts, one part which is caused by the basic of the bilinear bond approach, a second
part caused by bond friction as well as third part caused by the curvature of the structural element. The foundation
of the different part is lead back to [22], [24].

Background simplified model in Annex J.11.1.2.3

The analysis of the change in the strip force at the intercrack elements is based on a simplification of the detailed
analysis of the change in the strip force at the intercrack elements. The simplification was performed in [26]. The
main feature of the simplified analysis is the constant resistance of the intercrack elements to the change in the

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 25/34

For citations, copying and referencing see Note on Front page! page 683
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

strip force which does not depend on the base force in the strip. The change in the tensile force of the strip is
verified along the member. As the analysis does not depend on the base force in the strip, it can be performed on
superimposed load combinations, unlike the detailed analysis. Several boundary conditions were specified in [26]
with the aim of simplifying the detailed analysis. In addition to the boundary conditions listed in Equation (J.50), it
was also necessary to specify an ultimate strain of 10 mm/m for the strips which must not be exceeded at any
point on the member. It is therefore specified in FprEN1992-1-1, J.11.1.2.3 (3) that the ultimate strain and the
design tensile strength of the strips must not be exceeded.
In [1],[3],[22],[23],[34] the equation for the simplified analysis is given below:

, ⋅ . ⋅√ , ⋅ , ⋅ ⋅
𝛥𝐹 ⋅𝑏 (BD-J.11.16)

This Eq. can be transformed to the Eq. (J.39) to (J.42):

, ⋅ . ⋅ , , ⋅ . ⋅ , ⋅ ,
𝛥𝑓 (BD-J.11.17)
⋅

,
. ⋅ , ⋅ ⋅ , ⋅ . ⋅ , . ⋅ ⋅ . ⋅ , ⋅ ,
𝛥𝑓 (BD-J.11.18)
⋅

, , ,
, ⋅ , ⋅ ⋅ , ⋅ ⋅ ⋅
𝛥𝑓 (BD-J.11.19)

, , ,
𝛥𝑓 (BD-J.11.20)

For durability reasons the adhesive bond resistance is reduced as in Equation (J.38)
⋅ ⋅ ∙ , , ,
𝛥𝑓 (BD-J.11.21)

with:
√𝑠
𝛥𝑓 , 0,85 ⋅ 𝑘 , ⋅ 𝑓 ⋅𝑓 , ⋅
𝑡

, 𝑠 ,
𝛥𝑓 , 𝑓 ⋅
𝑡

𝜅 𝑠 ,
𝛥𝑓 , ⋅
ℎ 𝑡

Background refined model

As described above, the model in Annex J.11.1.2.2 is a simplified approach of a more refined model. This refined
model can be found in [1], [3]. The model is based on a mechanical approach. It was also evaluated with a huge
test database. The mechanical background and the evaluation can be found in [22], [24].

Crack Pattern
The proposal is based on the DAfStb Guideline “Strengthening of Concrete Members with Adhesively Bonded
Reinforcement” [1] and the fib-bulletin 90 [3].
The crack spacing is determined on the safe side over 1,5 times the insertion length of the inner reinforcement.
Thus, an initial crack pattern is assumed there and only the already existing reinforcing steel is considered.
In Equation (J.45) 𝜙 is the diameter of a conventional (unstressed) steel reinforcement bar and 𝑛 , is the
number of bars of equal diameter in the tension zone of a cracked section. The summation refer to addition of the
term 𝜙 ⋅ 𝑛 , ⋅ 𝑓 for each individual bar in the tension zone of a crack concrete section.
More accurate and complex calculations are required for prestressed concrete.

J.11.1.2.4 End cover separation

The proposal is based on the DAfStb Guideline “Strengthening of Concrete Members with Adhesively Bonded
Reinforcement” [1] and the fib-bulletin 90 [3].
When using externally bonded or near-surface-mounted reinforcement, the layer of concrete directly beneath the
reinforcement can become detached near the supports. This is known as a concrete cover separation failure. The
additional, vertical offset between internal shear links and external strip leads to the build-up of tensile forces
between the internal and the externally bonded reinforcement. This problem is similar to the situation with dapped

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 26/34

For citations, copying and referencing see Note on Front page! page 684
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

supports, which are frequently encountered in precast concrete construction. The acting forces can be determined
approximately with a truss model, as shown in figure BD-J.11.7.

Figure BD-J.11.7 Cause of transverse tensile stresses at end of strip [23]

To do this, the shear force is distributed in line with the truss over the strut in the concrete and the tie in the form
of the internal shear link for the unstrengthened cross-section. A similar truss is now set up for the component of
the tensile force in the surface-mounted reinforcement and the two systems are superimposed to create one total
system. A concrete cover separation failure occurs when the tie in the concrete, which results from the force in
the strip, can no longer be carried by the tensile strength of the concrete.
In order to prevent this type of failure, a shear strap must be positioned at the end of the strip according to (J.51)
when the acting shear force at the end support, or end of the beam, is greater than the shear capacity. In Eq.
(J.51), Rd,c is determined according to EN 1992. The variable aL is the distance of the strip from the end support,
or end of beam, in mm.
The critical value of the shear stress comes from a conversion of the model by Jansze [30] to the model of DIN
EN 1992-1-1 [31] in conjunction with its German National Annex [32]. This conversion and a validation can be
found in [33].
Since the formulation for Rd,c in prEN 1992-1-1 differs significantly from the basis of the calibration of the formula,
the value according to DIN EN 1992-1-1/NA should be used here again:

𝑉Rd,c 0,1 ⋅ 𝑘 ⋅ 100𝜌 𝑓ck 0,12 ⋅ 𝜎cp ⋅ 𝑏 ⋅ 𝑑 (BD-J.11.26)

For the usual components of building construction, k = 1+ (200/d)0,5 can be set approximately to 1,5 on the safe
side. From this follows approximately when neglecting the normal force:
,
1
𝑉Rd,cfE 0,75 ⋅ 1 19,6 ⋅ , ⋅ 0,1 ⋅ 1,5 ⋅ 100𝜌 𝑓ck ⋅ 𝑏w ⋅ 𝑑 (BD-J.11.27)

This results in equation (BD-J.11.28) resp. FprEN1992-1-1, (J.51):

.
1
𝑉Rd,cfE 0,11 2,2 ⋅ . ⋅ 100𝜌 𝑓ck ⋅ 𝑏w ⋅ 𝑑 (BD-J.11.28)

Where a shear strap is necessary, it should be designed according to Figure BD-J.11.8 for the design value of
the acting tensile force according to Eq. (J.52).

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 27/34

For citations, copying and referencing see Note on Front page! page 685
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

Figure BD-J.11.8 Calculation of force in shear wrapping due to tensile force in bending reinforcement
When using externally bonded reinforcement, an approach that lies on the safe side is to design the force at the
end strap for the maximum force in the flexural strengthening that can be accommodated by the end anchorage
according to Eq. (J.40). This is because a larger force cannot occur in the strip at the end strap except when the
shear strap counts towards increasing the bond force of the flexural strengthening. The force in the shear strap
is, however, superposed on the maximum force that can be accommodated at the end anchorage according to
J.11.1.1.3.

J.11.1.2.5 Shear Induced Separation

The proposal is based on the DAfStb Guideline “Strengthening of Concrete Members with Adhesively Bonded
Reinforcement” [1] and the fib-bulletin 90 [3].
Delamination caused by movement of the crack edges due to the shear force may occur in members with
externally bonded bending reinforcement if the tie and strut of the truss on which the transfer of shear forces is
based are subject to a high degree of stress. This relationship and delamination are shown in Figure BD-J.11.9.
The limits for the type of delamination shown here are based on a model with a subsequent parameter study in
[34],[35]. The stirrup stress included in Equation (J.53) can be calculated by transforming Equation (8.42).

Figure BD-J.11.9 Movement of the crack edges due to the shear force [21]

Beyond a certain level of shear stress, adhesively bonded stirrups are also required in order to ensure that the
tensile forces from the adhesively bonded reinforcement can also be transferred to the flexural compression zone
of the member by the truss shown in Figure BD-J.11.10. The limit 02 according to DIN 1045:1988 [36] has proven
to be a suitable parameter in such cases (cf. Z-36.12-73:2009 [37]). Equation (J.54) describes a similar limit [33].

Figure BD-J.11.10 Mechanism for transferring the tensile forces from the adhesively bonded reinforcement to
the flexural compression zone of the member using the truss model [21]
The design value of the applied shear force of the externally bonded shear reinforcement is determined as the
maximum of two components as it is in existing approvals. The first component in Equation (J.55) is obtained by
distributing the design value of the total applied shear force over the elastic stiffnesses and the second component
is the difference between the design value of the total applied shear force and the design value of the shear
resistance of the steel reinforcing links.

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 28/34

For citations, copying and referencing see Note on Front page! page 686
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

J.11.1.3 Basic anchorage resistance – CFRP to concrete for NSM CFRP strengthening
The proposal for the basic anchorage resistance – CFRP to concrete for NSM CFRP strengthening is based on
the DAfStb Guideline “Strengthening of Concrete Members with Adhesively Bonded Reinforcement” [1] and the
fib-bulletin 90 [3].
Essentially, the concept proposed by Blaschko [16], [17] is used for verifying the bond. In this concept it is
assumed that the CFRP strip makes a full contribution and there is good composite action between strip and
concrete, with the strength of the adhesive usually governing this composite action. As the composite action very
effective, the full tensile strength of the CFRP strip can be anchored within a very short length – similar to
conventional steel reinforcing bars. It is therefore sufficient to check the end anchorage at the point at which the
strip is no longer required for the load-carrying capacity, very similar to anchoring steel reinforcing bars. This
concept to describe the way in which strips in slits work has proved worthwhile over the past 10 years in the former
national technical approvals and was therefore included in the DAfStb guideline [1].
As with conventional reinforced concrete construction with curtailed reinforcing bars, checking the bond requires
verification of the curtailment taking into account the end anchorage of the CFRP strip. This involves verifying that
the design value of the member resistance is greater than the design value of the acting internal forces in the
strengthened condition for every cross-section of the strengthened member. The partial tensile forces assigned
to the lines of reinforcement can be determined in a simplified way by assuming a planar strain distribution. Figure
BD-J.11.11 provides an overview of the curtailment verification.

Figure BD-J.11.11 test results – fatigue resistance nsm-bond

As is apparent in Figure BD-J.11.11, verifying the anchorage of a near-surface-mounted CFRP strip can therefore
be carried out at the point at which the CFRP strip is first required for loadbearing purposes (point A). The
anchorage length lbf of the strip in this analysis is the distance between point A and the end of the strip. To verify
the anchorage, the resulting design value of the bond capacity per CFRP strip depends on this anchorage length
lbf and the distance of the longitudinal axis of the strip from the free edge of the member ar, which may not be >
150 mm, according to Eq. (J.56) and (J.57).
To determine the bond capacity, Eq. (J.56),(J.57) also requires the bond strengthbAd of the near-surface-mounted
CFRP strip, which according to Eq. (J.59) is found from the result of the minimum of the concrete bond strength
and the adhesive bond strength.
Equation (J.59) distinguishes between concrete failure and adhesive failure. For reasons of simplification, the
same safety factor was used here. A closer look could possibly apply a lower safety factor to the adhesive due to
its probably lower scatter.
Usually, failure will occur in the adhesive of the near surface mounted CFRP strips. In this case, the maximum
bond strength in Equation (J.60) is determined in accordance with [16], [17] calculating the shear strength from
the axial tensile strength and the axial compressive strength of the adhesive by use of the Mohr-Coulomb failure
criterion. By way of departure from [16], [17], a system factor 0,6 in accordance with [38], [39] was also introduced
into Equation (J.60) in order to cover effects such as the surface structure of the CFRP strips. Furthermore, a

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 29/34

For citations, copying and referencing see Note on Front page! page 687
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

reduction factor to take account of long-term effects of the adhesive bond, bA, was introduced in Equation (J.59).
A value of 0,5 was proposed for this factor although a different value may be specified in the system approval.
Failure will occur in the concrete if the concrete strength is very low, i.e. below C20/25. In this case, a concrete
wedge will break away on either one side or both sides of the strip, with the base of the wedge in the concrete
member corresponding to the base of the slit. This can be calculated with Equation (J.61). In this Equation this
bond strength is calculated from the square root of the concrete compressive strength and a calibration factor kbck.
The system coefficient for the bond failure of the concrete can be taken from the national technical approval for
the system. Tests carried out at the Technische Universität München (Technical University of Munich) established
a characteristic value kbck = 4,5 [40], [41].
The factors bC is also introduced into Eq. (J.59) to take account of the long-term durability behaviour of the
materials involved. As these are also coefficients specific to particular products, they can again be obtained from
the national technical approvals.

Range of application
The concept in J.11.1.3 is based on bond test with CFRP-strips. For this concept the CFRP-strips width shall be
between 10 mm and 30 mm and the CFRP-strips thickness shall be between 1 mm and 3 mm.
For other CFRP elements such as round or square bars the bond behaviour is different. So for the determination
of FbfRd further considerations are necessary. This may be bond tests to determine FbfRd directly or to recalibrate
the factor 0,6 in equation (J.60) and the factor 4,5 in equation (J.61).

R. Niedermeier, W. Finckh, K. Zilch 21.12.2022 Page 30/34

For citations, copying and referencing see Note on Front page! page 688
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

J.12.1 Detailing of CFRP - Flexural strengthening with eternally bonded CFRP

The limits for the centre-to-centre spacing are based on experiences. The limit of sf ≤ 3 times the slab thickness
≤ 400 mm is in line with the provisions for reinforcing steel

J.12.2 Detailing of CFRP - Flexural strengthening with NSM CFRP

The geometrical limits for NSM CFRP strips in FprEN1992-1-1, Table J.2 are known from experiences and based
on literature [1], [16].

The slot width and depth limits ensure that there is 1 to 3 mm of adhesive between the NSM bar or strip and the
concrete. This ensures that there is full bond around the perimeter of the slot. The 3 mm upper limit is consistent
with adhesive thickness limit for external FRP plate bonding and the range between 1 mm and 3 mm is in line
with testing which has been carried out on NSM flexural strengthening of beams.
The centre to centre spacing ensures that the width of concrete between slots is at least equal to the length of the
sides of each slot so that there is not a propensity for concrete cover separation across the width of the beam in
preference to debonding around the concrete slot perimeter. This also prevents overlapping of confining stress
zones around each NSM bar, maximising bond stresses ([50],[51].
The minimum edge distance equal to 4 times the depth of the slot is based on findings of Hassan and Rizkalla
[50], for round bars, and Rashid, et al. [51] for strips. This is to prevent detachment of the concrete cover along
the edges of the beam, and the NSM reinforcement being able to carry the same force as a bar/strip far from the
edges.

The NSM technique can be used for the flexural and shear strengthening of deficient reinforced concrete members
for the following reasons: better bond characteristics; activation of controlled debonding that increases the ductility
of strengthened members; mechanical response to make the structure stiffer under serviceability loads; limited
damage by peeling-off failure due to flexural cracks; excellent fatigue resistance; reduced site work (due to not
necessary plaster removal, unevenness and irregularities of the concrete surface and removal of the weak laitance
layer on the concrete surface); easier anchorage to adjacent members to prevent debonding, where the maximum
moments typically occur at the ends of the member – flexural strengthening of columns); protection from
mechanical damage, accidental impact and vandalism; this aspect makes this technology particularly suitable for
the strengthening of negative moment regions, unchanged aesthetics [3].
The most commonly used CFRP strips have a thickness of 1,4 – 3,0 mm and a width of 15 – 30 mm, while FRP
bars of different diameters from 2 – 20 mm, which are bonded into the concrete grooves. As it was mentioned in
[45] two of the most efficient parameters guaranteeing the best bond behaviour are: the width of the strip and the
strips installation as deepest as possible into the slits providing the largest pull-out load and energy dissipation
(Fig. BD-J.12.1a). The research carried out by [46] indicated very promising conclusion that cutting of the bottom
horizontal arm of the steel stirrups in order to apply CFRP strips of an width larger than the concrete cover (Fig.
BD-J.12.1b) does not influence bending and shear capacity, because the bottom arms of the stirrups are safety
anchored. This solution in Fig. BD-J.12.1b) is not covered by Annex J. Another important issue mentioned in [45]
pointed an importance to reduce the number of NSM strips that make an increase in the distance between CFRP
strips and effects the interaction between steel reinforcement and strips by reducing the stiffness under the
longitudinal steel bars (Fig. BD-J.12.1c).

Figure BD-J.12.1: NSM flexural strengthening:

a) strips placed in the concrete cover;
b) strips placed after cutting of the stirrups arm (dimensions in mm);
c) Configurations for the relative position of strips and steel bar;
d) Configurations to assure higher pro-section against fire [45].

R. Kotynia, W. Finckh, A. Darby, C. Giaccio 21.12.2022 Page 31/34

For citations, copying and referencing see Note on Front page! page 689
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

[1] DAfStb-Guideline: Strengthening of concrete members with adhesively bonded reinforcement].

References
Deutscher Ausschuss für Stahlbeton, 2012, English version, Beuth, Berlin, 2014.
[2] UK Concrete Society, Design Guidance for Strengthening Concrete Structures Using Fibre Composite
Materials, Technical Report 55, 2015.
[3] fédération internationale du béton (fib): Externally applied FRP reinforcement for concrete structures,
fib bulletin 90. Lausanne, 2019.
[4] Zilch, K.: The German Design Guideline for Strengthening Concrete Structures with FRP. In: El-Hacha,
R.: Proceedings of the 7th International Conference on FRP Composites in Civil Engineering,
Vancouver, Canada 2014.
[5] Occhiuzzi, A.; Ceroni, F.; Franco, A.; Salzano, P.; Bonati, A.: Experimental results of a national
technical assessmentprocedure on commercial FRP for structural strengthening: wet-lay-up systems;
Materials and Structures (2020) 53:9.
[6] Song X; Gu X; Li Y; Chen T, (2013): “Mechanical Behavior of FRP-Strengthened Concrete Columns
Subjected to Concentric and Eccentric Compression Loading”, ASCE Journal of Composites in
Construction, 17(3), 336-346.
[7] Darby, A.P., Ibell, T.J., Evernden, M.C. and Coonan, R. (2011): “FRP Confined Square and
Rectangular Columns Under Concentric and Eccentric Loading”, Advanced Composites in
Construction, University of Warwick. pp. 264-275
[8] Hankers, C., (1996): Zum Verbundtragverhalten laschenverstärkter Betonbauteile unter nicht
vorwiegend ruhender Beanspruchung, (On the bond behaviour of steel-plate-reinforced concrete
members under cyclic loading) (in German), Ph.D. Thesis, Technische Universität Braunschweig.
[9] Bizindavyi, L., et al., (2003): Experimental Investigation of bonded fiber reinforced polymer-concrete
joints under cyclic, Journal of Composites in construction, pp. 127-133.
[10] Carloni, C., et al., (2012): Experimental determination of FRPconcrete cohesive interface properties
under fatigue loading, Composite Structures, Volume 94, Issue 4, Pages 1288-1296.
[11] Budelmann, H., et al., (2013): Praxisgerechte Bemessungsansätze für das wirtschaftliche Verstärken
von Betonbauteilen mit geklebter Bewehrung - Verbundtragfähigkeit unter dynamischer Belastung,
(Practical design approaches for the economic strengthening of concrete members with bonded
reinforcement – Bond behaviour under dynamic loading) DAfStb Heft 593, Beuth Verlag, Berlin.
[12] Leusmann, T. (2015): “Das Verbundtragverhalten geklebter Kohlefaserkunststoffe auf Beton unter
schwingender Beanspruchung. (Bond behaviour of cfrp glued on concrete under cyclic load) (in
German). Ph.D. thesis, Technische Universität Braunschweig.
[13] Dai, J.G.,et al., (2005): Static and fatigue bond characteristics of interfaces between CFRP sheets and
frost damage experienced concrete. Proc., 7th Int. Symp. On Fiber Reinforced Polymer Reinforcement
for Reinforced Concrete Structures, C.K. Shield, J.P. Busel, S.L. Walkup, and D.D. Gremel, eds.,
Kansas City, MO, American Concrete Institute (ACI), Farmington Hills, MI, 15151530.
[14] Ferrier, E., et al. (2005): Fatigue of CFRPs externally bonded to concrete, Materials and Structures 38,
39-46.
[15] Carloni, C.; Leusmann, T. and Budelmann, H.: Fatigue Behaviour, Chapter 6.2, in: Design Procedures
for the Use of Composites in Strengthening of Reinforced Concrete Structures – State of the Art Report
of the RILEM TC 234-DUC, (Eds: C. Pellegrino and J. Sena-Cruz), pp. 349-391, Springer, RILEM STAR
Book Series, 2016.
[16] Blaschko, M.: Zum Tragverhalten von Betonbauteilen mit in Schlitze eingeklebten CFK-Lamellen
[structural behaviour of concrete members with near-surface-mounted CFRP strips]. Dissertation,
Munich TU, Department of Concrete Structures, 2001.
[17] Blaschko M.: Bond Behaviour of CFRP Strips Glued into Slits, Proceedings of the sixth international
Symposium on FRP reinforcement for concrete structures (FRPRCS-6), pp.205-214, Singapore, 2003.
[18] Niedermeier, R.: Zugkraftdeckung bei klebearmierten Bauteilen [Tension reinforcement curtailment in
members with externally bonded reinforcement]. Dissertation, Munich TU, Department of Concrete
Structures, 2001.
[19] Niedermeier, R.: Abschlussbericht zum Forschungs- und Entwicklungsvorhaben Verstärkung von
Stahlbetonbauteilen durch Ankleben von Kohlenstoffaserfolien [Final report on the research and
development project - Strenghtening of reinforced concrete by externally bonded CFRP sheets]. No.
599526, dated 31.03.1999 (not published).
[20] Zilch, K., Niedermeier, R., Finckh, W.: Guideline for the Performance and interpretation of bond tests
on concrete with externally bonded CFRP strips. FRPRCS-9 Sydney, Australia, Monday 13 –
Wednesday 15 July 2009
[21] Deutscher Ausschuss für Stahlbeton: Erläuterungen und Beispiele zur DAfStb-Richtlinie “Verstärken
von Betonbauteilen mit geklebter Bewehrung” [Commentary on the DAfStb Guideline “Strengthening

CEN/TC250/SC2/WG1/TG1 21.12.2022
For citations, copying and referencing see Note on Front page! page 690
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

of concrete members with adhesively bonded reinforcement” with Examples], Schriftenreihe des
DAfStb No. 595, Beuth, Berlin, 2013.
[22] Zilch, K., Niedermeier, R., Finckh, W.: Praxisgerechte Bemessungsansätze für das wirtschaftliche
Verstärken von Betonbauteilen mit geklebter Bewehrung – Verbundtragfähigkeit unter statischer
Belastung [practical design methods for economic strengthening of concrete members with externally
bonded reinforcement – bond strength under static loading]. Schriftenreihe des DAfStb No. 592, Beuth,
Berlin, 2012.
[23] Zilch, K.; Finckh, W.; Niedermeier, R.: Strengthening of Concrete Structures with Adhesive Bonded
Reinforcement, Design and Dimensioning of CFRP Laminates and Steel Plates, Ernst und Sohn
Verlag, 2014, ISBN: 978-3-433-03086-8
[24] Finckh, W.; Zilch, K.: Strengthening and Rehabilitation of Reinforced Concrete Slabs with Carbon-Fiber
Reinforced Polymers Using a Refined Bond Model, Computer-Aided Civil and Infrastructure
Engineering 27 (5), pp. 333-346, 2012.
[25] Zehetmaier, G.: Zusammenwirken einbetonierter Bewehrung mit Klebearmierung bei verstärkten
Betonbauteilen [interaction of embedded reinforcement and externally bonded reinforcement in
strengthened concrete members]. Dissertation, Munich TU, Department of Concrete Structures, 2006
[26] Zehetmaier, G.; Zilch, K.: Interaction between internal bars and externally FRP Reinforcement in RC
Members, Proceedings of the sixth international Symposium on FRP reinforcement for concrete
structures (FRPRCS-6), pp.397-406, Singapore, 2003
[27] Husemann, U.: Erhöhung der Verbundtragfähigkeit von nachträglich aufgeklebten Lamellen durch
Bügelumschließungen [increasing the bond strength of retrofitted externally bonded strips by means of
shear wrapping]. Dissertation, Braunschweig TU. Institute of Building Materials, Concrete Construction
& Fire Protection, 2009
[28] Husemann, U., Budelmann, H: Increase of the Bond Capacitiy of Externally Bonded CFRP-Plates on
RC-Structures Due to Self-Induced Contact Pressure, In: Proceedings of the 9th International
Symposium on Fibre Reinforced Polymers in Reinforced Concrete Structures (FRPRCS 9), Sydney,
Australia, 2009
[29] Z-36.12-86: Bausatz StoCretec zum Verstärken von Stahl- und Spannbetonbauteilen durch schubfest
aufgeklebte CFK-Lamellen nach der DAfStb-Verstärkungs-Richtlinie, Deutsches Institut für
Bautechnik; 2021
[30] Jansze, W.: Strengthening of reinforced concrete members in bending by externally bonded steel
plates. Design for beam shear and plate anchorage. Dissertation, Delft University of Technology, 1997.
[31] DIN EN 1992-1-1 (2011): Eurocode 2: Design of concrete structures - Part 1-1: General rules and rules
for buildings.
[32] DIN EN 1992-1-1/NA: National Annex – Nationally determined parameters – Eurocode 2: Design of
concrete structures - Part 1-1: General rules and rules for buildings.
[33] Zilch, K., Niedermeier, R., Finckh, W.: Praxisgerechte Bemessungsansätze für das wirtschaftliche
Verstärken von Betonbauteilen mit geklebter Bewehrung – Querkrafttragfähigkeit [practical design
methods for economic strengthening of concrete members with externally bonded reinforcement –
shear strength]. Schriftenreihe des DAfStb No. 594, Beuth, Berlin, 2012.
[34] Finckh, W.: Einfluss bauteilspezifischer Effekte auf die Bemessung von mit CFK-Lamellen verstärkten
Stahlbetonbauteilen [influence of member-specific effects on the design of RC members strengthened
with CFRP strips]. Dissertation, Munich TU, Department of Concrete Structures, 2012.
[35] Finckh, W.; Zilch, K.: Influence of shear crack offsets on the bond behavior of EBR at the intermediate
crack element. In: The 6th International Conference on FRP Composites in Civil Engineering - CICE
2012, Rom, 2012
[36] DIN 1045: Structural use of concrete – Design and construction. Deutsches Institut für Normung. Beuth,
Berlin, 1988.
[37] Z-36.12-73: Verstärken von Stahlbetonbauteilen durch in Schlitze verklebte Kohlefaserlamellen
Carboplus nach DIN 1045-1:2008-08 [strengthening of RC members with near-surface-mounted
Carboplus carbon fibre strips to DIN 1045-1]. Deutsches Institut für Bautechnik, 2009.
[38] Borchert, K.: Verbundverhalten von Klebebewehrung unter Betriebsbedingungen [bond behaviour of
externally bonded reinforcement under operating conditions]. Schriftenreihe des DAfStb No. 575,
Beuth, Berlin, 2009.
[39] Zilch, K.; Borchert, K.: Bond behaviour of NSM FRP strips in service, structural concrete (9), pp.127-
142, 2008
[40] Finckh, W.; Niedermeier, R.: Bericht über Ausziehversuche an Probekörpern aus Normalbeton
niedriger Festigkeit mit in Schlitze verklebten CFK-Lamellen, 2012
[41] Zilch, K.; Finckh, W.: Gutachterliche Stellungnahme zur Eignung von in Schlitze verklebten S&P-
Lamellen zur Verstärkung von niedrigfesten Betonen, 2013

CEN/TC250/SC2/WG1/TG1 21.12.2022
For citations, copying and referencing see Note on Front page! page 691
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
Annex J Background to FprEN 1992-1-1:2023

[42] Alex, R.; Niedermeier, R.; Finckh, W.; Leusmann, T.: Requirements to Materials – CFRP reinforcement
for strengthening (Background document to Annex J.15), December 2022.
[43] Niedermeier, R.: Verstärkung von Stahlbetondruckgliedern durch Umschnürung. Habilitationschrift,
Technische Universität München, 2009.
[44] Niedermeier, R.: German Design Approach for Concrete Columns Strengthened with CFRP
Confinement. Proceedings of the 7th International Conference on FRP Composites in Civil Engineering
(CICE 2014), Paper 050, Vancouver, BC, Canada, 2014
[45] Barros, J.; Kotynia, R.: Possibilities and challenges of NSM for the flexural strengthening of RC
structures, Fourth International Conference on FRP Composites in Civil Engineering CICE2008, 2008,
Zurich, Switzerland.
[46] Kotynia, R.: Analysis of reinforced concrete beams strengthened with near surface mounted FRP
reinforcement. Arch. Civil Mech. Eng., 2006, LII(2), 305–17.
[47] prEN 1990:2020: Eurocode – Basis of structural and geotechnical design
[48] fédération internationale du béton (fib): fib Model Code for Concrete Structures 2010, Ernst & Sohn,
2013.
[49] Bilotta, A., Ceroni, F., Nigro, E. And Pecce, M. (2011): Design by testing procedure of debonding load
for RC elements strengthened with EBR FRP materials, in Proceedings of 10th FRPRCS International
Symposium, ACI SP-275 Fiber-Reinforced Polymer Reinforcement for Concrete Structures, R. Sen, R.
Seracino, C. Shield and W. Gold (eds), Tampa, Florida.
[50] Hassan, T. and Rizkalla, S. (2004), Bond mechanism of near-surface-mounted fiber-reinforced polymer
bars for flexural strengthening of concrete structures’, ACI Struct J, 101 (6), pp. 830-839.
[51] Rashid, R.; Ohelers, D.J. and Seracino, R. (2008), IC Debonding of FRP NSM and EB Retrofitted
Concrete: Plate and Cover Interaction Tests, J. Compos. Constr., 12(2), pp. 160-167.
[52] Youssef, M. N.; Mosallam, A. S. and Feng; M. Q. (2006): "Experimental Investigation on Large-Scale
FRP-Confined Axial Members," In: Proceedings of the International Conference on Civil Engineering
Infrastructure Systems (CEIS 2006) American University of Beirut(AUB), Beirut-Lebanon, June 12-14.
[53] Lam, L.; Teng, J. G. (2003a): Design-oriented stress-strain model for FRP-confined concrete.
Construction and Building Materials, 17, 471-489.
[54] Lam, L.; Teng, J. G. (2003b): Design-oriented stress-strain model for FRP-confined concrete in
rectangular columns. Journal of Reinforced Plastics and Composites, 22(13), 1149-1186.
[55] Binici, B. (2008): Design of FRPs in circular bridge column retrofits for ductility enhancenment.
Engineering Structures, Volume 30, Issue 3, pp. 766-776.

List of versions of this document

14.03.2021 1st draft
17.05.2021 2nd draft
21.05.2021 3rd draft
31.05.2021 4th draft
25.11.2022 5th draft
21.12.2022 6th draft

CEN/TC250/SC2/WG1/TG1 21.12.2022
For citations, copying and referencing see Note on Front page! page 692
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314

CEN/TC250/SC2/WG1/TG1 FprEN 1992-1-1:2023

Background document to subclause J.15

Requirements to Materials – CFRP
reinforcement for strengthening

Rolf Alex Deutsches Institut für Bautechnik (DIBt), DE

Roland Niedermeier Technische Universität München, DE
Wolfgang Finckh Ostbayerische Technische Hochschule Regensburg, DE
Thorsten Leusmann Technische Universität Braunschweig, DE
21.12.2022

For citations, copying and referencing see Note on Front page! page 693
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

List of content (only main clauses)

1 Introduction and scope
2 Product specific design and application parameters for strengthening kits
of externally bonded CFRP strips (EBR)
3 Product specific design and application parameters for strengthening kits
of externally bonded CFRP sheets (EBR)
4 Product specific design and application parameters for strengthening kits
of near surfaced mounted CFRP strips (NSM)
5 References
Annex BD-A Test program for externally bonded CFRP strips
Annex BD-B Test program for externally bonded CFRP sheets
Annex BD-C Test program for near surface mounted CFRP strips and bars

J.15 1 Introduction and scope

In Annex J.15 of FprEN 1992-1-1:2023 (E) are listed the design parameters for the strengthening of existing
concrete structures with Carbon Fibre Reinforced Polymers (CFRP) adhesively bonded (ABR) to the concrete
according to Annex J of FprEN 1992-1-1:2023 (E). The reinforcement can be externally bonded to the surface
(EBR) or near-surface mounted in the concrete (NSM).
The design rules of Annex J of FprEN 1992-1-1:2023 (E) are similar to the design rules, which since long time are
applied in Germany according to national approvals. Especially the formulae (J.2) and (J.38) to (J.50) for the
externally bonded reinforcement (EBR) as well as the formulae (J.2) and (J.56) to (J.61) for the near surface
mounted reinforcement (NSM) are well known in design practice in Germany.
In these formulae are some parameters which are product-specific and only may be determined with the help of
tests. But not all of these parameters can be derived directly from test results. The most important concern of this
document is to describe the procedures that are used for the uniform determination of these parameters.
According to the Construction Products Regulation (Regulation (EU) No 305/2011 (CPR)) on EU level test
procedures for the essential characteristics of construction products may be included in two types of technical
specifications for construction products:
 European harmonised product standards or
 European Assessment Documents (EAD).
At the time of this writing, neither technical specification for the above strengthening systems was available.
Because of this current situation, the following follows for this document:
 This document must not only contain methods for determining the design parameters for Annex J of
FprEN 1992-1-1:2023 (E) from test results, but also the descriptions of the test procedures themselves.
This is done in the following Annexes for the different types of kits:
 Annex BD-A Test program for externally bonded CFRP strips,
 Annex BD-B Test program for externally bonded CFRP sheets,
 Annex BD-C Test program for near surface mounted CFRP strips and bars.
 If at any time one of the technical specifications mentioned above is available for one or more types of
kits, this document must be amended as follows:
 The corresponding one or more appendices of this document shall be deleted.
 All references in the text of this document to the deleted Annexes shall then be replaced by
references to the European harmonised standards or European Assessment Documents.
In article 4 (1) of the CPR is written:
"When a construction product is covered by a harmonised standard or conforms to a European Technical
Assessment which has been issued for it, the manufacturer shall draw up a declaration of performance when such
a product is placed on the market."
As long as such technical specifications are not available, the manufacturers cannot explain the performance of
their kits via such a declaration of performance. These can currently only be stated in test reports for the tests
according to the Annexes BD-A to BD-C that are not binding with regard to EU Construction Products Regulation
(CPR).
But important for the application of the kits are not only the design parameters but also the application parameters
which are also derived from the tests according to Annex BD-A to BD-C.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 694
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

2 Product specific design and application parameters for strengthening kits of

externally bonded CFRP strips (EBR)
Systems (kits) of this type in general consist of the following components:
Component A unidirectional CFRP (Carbon Fibre Reinforced Polymers) strips
Component B adhesive (often also called structural bonding agent or structural bonding) for CFRP strips
according to EN 1504-4 [1],
Component C repair mortar with polymer binder,
Component D bonding agent for the repair mortar, repair concrete and shotcrete,
Component E cleaning agent for the CFRP strips.

Figure 2.1: Principle sketches of reinforced concrete sections with subsequent strengthening by externally
bonded CFRP strips
The product-specific design parameters for kits of EBR CFRP included in Annex J of FprEN 1992-1-1:2023 (E)
are the following:

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 695
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

Table 2.1: Product-specific design parameters included in Annex J of FprEN 1992-1-1:2023 €

Symbol in FprEN Meaning used in FprEN 1992-1-1

1992-1-1 (according to FprEN 1992-1-1, clause 3.5) (only clauses of Annex J are listed,
which refer to externally bonded
CFRP strips)
Design parameters for the CFRP strips (component A)
reduction factor applied to the
𝜂 tensile strength of the EBR CFRP 3.5.4, J.5.2 (1), J.15
sheet or strip
characteristic short-term tensile strength of 3.5.2, J.5.3 (1), formula (J.2),
𝑓
the ABR CFRP Figure J.1
3.5.1, J.5.3 (1), formula (J.3)
Mean modulus of elasticity in longitudinal
𝐸 Figure J.1, formula (J.35), (J.37) and
direction of ABR CFRP
(J.39), (J.40)
characteristic ultimate strain for ABR
𝜀 3.5.4, J.5.2 (1), Figure J.1
CFRP
𝐴 Cross sectional area of a CFRP system 3.5.1, J.5.1 (3)
Design parameters for the adhesive (component B) and the repair mortar (component C)
In all design formulae for the EBR CFRP in FprEN 1992-1-1:2023, Annex J no material parameter for the
adhesive and repair mortar are included but only material parameters of the ABR CFRP (component A) and
the concrete substrate. Therefore, when applying the formulae (J.35) to (J.42), it is assumed that only kits
are used that always have concrete failure in substrate and never adhesive failure.
Design parameters for the bond to the concrete
𝑘 , product-specific system factor 3.5.2, formula (J.42), (J.48), J.15
𝑘 , product-specific system factor 3.5.2, formula (J.41), J.15
𝑘 , product-specific system factor 3.5.2, formula (J.39), J.15
𝑓 average value of the compressive strength 3.5.2, J.11.1.1.2, formula (J.35),
of the concrete substrate (J.37), J.11.1.1.3, (J.42)
𝑓 , average value of the tensile surface 3.5.2, J.11.1.1.1, formula (J.34),
strength of the concrete substrate J.11.1.1.2, formula (J.35), (J.37),
J.11.1.1.3, formula (J.42)
𝛼 reduction factor for fatigue 3.5.4, J.10.2, formula (J.27)
Design parameters for confinement of columns
𝑘 product-specific system factor; 3.5.2, J.8.1.2
confinement factor for columns
strengthened with CFRP

2.1 Boundary conditions of the intended use of the kits

Kits tested in accordance to Annex BD-A are used since long time in Germany under the following environmental
conditions for a design life of 25 years:
 The strengthening kit may be used for normal concrete substrates of in the range of compressive
strength classes between 𝐶 to 𝐶 according to EN 206-1.
 Without additional protective measures for the FRP and adhesive, strengthening with adhesively
bonded reinforcement shall only be applied in exposure classes X0, XC1 (dry) or XC3 in accordance
with Table 6.1 in FprEN 1992-1-1:2023 (E).
 In addition, EBR CFRP shall not be exposed to direct UV radiation (direct sun radiation or indirect sun
from snow or water reflection) or alternating or permanent penetration of moisture in the area of the
strengthening. If necessary, appropriate protective layers shall be applied.
 The temperature of the concrete substrate does not exceed 𝑇 and does not fall below 𝑇 during
the entire design life of the strengthened structural element.
The range of compressive strength classes and the range of temperature concrete substrate are not given in
FprEN 1992-1-1:2023 (E). Therefore, it is necessary to be declared by the manufacturer and is largely confirmed
by the tests in accordance with Annex BD-A. If there is no confirmation by tests, the manufacturer's declaration
shall suffice or Annex BD-A shall be supplemented by suitable tests.
A further difference to the boundary conditions of kits tested by Annex BD-A is that in FprEN 1992-1-1:2023,
Annex J, clause J.6.3 other exposure classes are also allowed. It shall be checked whether this is also possible
without additional tests compared to Annex BD-A. If not, the corresponding tests in Annex BD-A shall be added.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 696
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

2.2 Boundary conditions of the installation of the kits

The manufacturer's instructions and safety information shall be considered. In particular, it shall be ensured that
the surface temperature of the concrete component to be strengthened is always between the minimum and
maximum installation temperature of the component of the kit installed on it (see Table 2.2 "Application
parameters for the kit").

2.3 Design and application parameters and their determination

Table 2.2: Sources of design and application parameters from test procedures according to Annex BD-A for kits
of externally bonded CFRP strips

Symbol according to Meaning Sources

FprEN 1992-1-1 Annex BD-A Test Additionally
procedure verified by
Design parameters for the CFRP strips (component A)
reduction factor applied to the A.2.2.5 and 2.3.1
𝜂 tensile strength of the ABR CFRP A.2.2.61)
strip
characteristic material strength of the A.2.21)
𝑓 𝑓
CFRP strips
mean value of the modulus of elasticity of A.2.21)
𝐸 𝐸
the CFRP strips
ultimate strain of the adhesively bonded A.2.21)
𝜀 𝜀
CFRP strips
𝑡,𝑏 𝑡 ,𝑏 thickness and width of the CFRP strip A.2.2.31)
the area of the cross section of CFRP A.2.2.31)
𝐴 𝐴 𝑡 ⋅𝑏
strips
Design parameters for the bond to the concrete
𝑘 , product-specific system factor A.2.3.51) 2.3.2
𝑘 , product-specific system factor A.2.3.51) 2.3.2
1)
𝑘 , product-specific system factor A.2.3.5 2.3.2
average value of compression strength of A.2.32)
𝑓 𝐶 the highest concrete strength class of the
tests
𝑓
average value of compression strength of A.2.32)
𝑓 𝐶 the lowest concrete strength class of the
tests
average value of the tensile surface A.2.32)
𝑓 , 𝑓 , 𝐶 strength of the concrete substrate with the
highest concrete strength class of the tests
𝛼 reduction factor for fatigue A.2.3.61) 2.3.3
Design parameters for confinement of columns
product-specific system factor; 2.4
𝑘 confinement factor for columns
strengthened with CFRP
Application parameters for the kit
𝑇 , minimum temperature of installation for A.2.3.42)
the adhesive (component B)
3)
𝑇 , maximum temperature of installation for
the adhesive (component B)
𝑇 , minimum temperature of installation for A.2.3.42)
the repair mortar (component C)
3)
𝑇 , maximum temperature of installation for
the repair mortar (component C)
3)
𝑇 minimum temperature of the
strengthening kit at intended use
𝑇 maximum temperature of the A.2.3.32)
strengthening kit at intended use
𝐶 highest concrete strength class of the A.2.32)
tests
𝐶 lowest concrete strength class of the tests A.2.32)

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 697
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

1)
determined by test procedure
2)
declared by the manufacturer and confirmed by test procedure
3)
declared by the manufacturer

The number of tests given in Annex BD-A, clauses A.2.1 to A.2.3 is the minimum number which is necessary.
Each applicant can expand it according to their needs.

2.3.1 The creep reduction factor ηf of CFRP strips derived from the tests according to Annex BD-A,
A.2.2.5 and A.2.2.6
The factor 𝜂 is the minimum of the factors 𝑅 from the test A.2.2.5 and 𝑅 from the test A.2.2.6
(𝜂 Min 𝑅 , 𝑅 . It is given separately for all alkaline solutions used in A.2.2.5 and A.2.2.6 (pH 9 and/or 11
and/or 13,7).

2.3.2 Parameters of the bilinear bond approach ksys,b1, ksys,b2 and ksys,b3 derived from the tests
according to Annex BD-A, A.2.3.5
According to the studies by Kuntz [2] and Holzenkämpfer [3], the bilinear approach for the bond law according to
Figure 2.2 is well suited for the evaluation of the bond tests according to A.2.3.5. Therefore, only the determining
parameters of this approach have to be determined for each individual test by the evaluation.

Figure 2.2: Energetic consideration of the bilinear bond approach

These are the maximum bond stress 𝜏 , the associated elastic limit deformation 𝑠 and the limit value of the
relative displacement 𝑠 .
The following procedure is used for this:
The differential formula for the bond of externally bonded reinforcement given in the Figure 2.3 can be completely
solved for the two branches of the bilinear bond approach. With an adjustment to the boundary conditions of the
bond test specimen according to Figure A.15, the following formulae are obtained, which describe the relationship
between the relative displacement 𝑠 at the load-side beginning of the bond length 𝑙 and the tensile force 𝐹 . For
𝑠 𝑠 , only elastic bond stresses occur within the total bond length 𝑙 . After exceeding the elastic limit
displacement (𝑠 𝑠 ), additional plastically softened bond areas occur.
𝐹 𝐸 ∙ 𝑏 ∙ 𝑡 ∙ 𝑠 ∙ 𝜔 ∙ tanh 𝜔 ∙ 𝑙 for 𝑠 𝑠
∙ ∙ ∙
𝐹 𝐸 ∙𝑏 ∙𝑡 ∙𝜔∙𝜆∙ for 𝑠 𝑠
∙ ∙

With:

𝜏
𝜔
𝑠 ∙𝐸 ∙𝑡

𝑠
𝜆
𝑠 𝑠

The length 𝑎 of the elastically deformed bond region enters into the second formula and can be determined with
the following expression. Because of the simultaneous occurrence of hyperbolic and trigonometric functions, 𝑎
cannot be determined in complete form but only iteratively.
𝜆 𝑠 𝑠
tanh 𝜔 ∙ 𝑎 ∙ 𝑠 𝑠 ∙ cot 𝜔 ∙ 𝜆 ∙ 𝑙 𝑎
𝑠 sin 𝜔 ∙ 𝜆 ∙ 𝑙 𝑎
The parameters 𝑠 , 𝜏 , and 𝑠 of the bilinear composite approach (cf. Figure 2.2) are determined using the least
squares method on the experimentally determined 𝑠 -𝜀 curves (cf. Figure 2.3) for each of the two sides of a
specimen. In view of the scope of the calculations to be performed, a limited number of supporting points is
recommended for the computational determination of the parameters.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 698
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

Figure 2.3: Evaluation of the bond tests

Subsequently, system-specific coefficients 𝑘 , and 𝑘 , can be determined in regression analyses according

to EN 1990, Annex D.8.
∙
For this purpose, the correlations 𝜏 0 ⇔𝑓 𝑓 , 0 and 𝐺 0 ⇔𝑓 𝑓 , 0 can be
used in addition. For strengthening systems with externally bonded reinforcement, in which the adhesive
(component B) does not lead to a hardening of the concrete layers near the surface, the following the following
simple functions 𝑓 and 𝑓 of the term 𝑓 ∙ 𝑓 , can be assumed for this purpose,

𝜏 𝑓 𝑓 ∙𝑓 , 𝑘 , ∙ 0,37 ∙ 𝑓 ∙𝑓 ,

𝑘 , ∙ 0,37 ∙ 𝑓 ∙𝑓 , ∙𝑘 , ∙ 0,2
𝐺 𝑓 𝑓 ∙𝑓 ,
2
𝑠 𝑘 , ∙ 0,2
The region before the maximum bond length is described in formula (J.38) via a parabola. However, the exact
solution from the bilinear composite approach is not exactly parabolic.
After the test interpretation, a better agreement of the approximation from the parabolic approach with the test
values can be achieved with the factor 𝑘 , .
For many known reinforcement systems with CFRP strips and cold-curing, quartz sand-filled epoxy resin
adhesives, 𝑘 , , 𝑘 , and 𝑘 , are equal 1,0.

2.3.3 Reduction factor for fatigue derived from the tests according to Annex BD-A, A.2.3.6
The fatigue reduction factor 𝛼 used in formula (J.27) in the draft standard FprEN 1992-1-1:2023 (E), Annex J
contains information for an S-N curve, which is verified by means of fatigue tests. The tests are performed
according to the test procedure in Annex BD-A, Chapter A.2.3.6. The following load levels are checked:

Table 2.3: Load level for fatigue bond tests

Load level Lower load 𝐹bLi Upper load 𝐹bLi 1st test Upper load 𝐹bLi 2nd test
A 0,15 𝐹 0,45 𝐹 0,49 𝐹
B 0,3 𝐹 0,55 𝐹 0,58 𝐹
C 0,45 𝐹 0,64 𝐹 0,67 𝐹
D 0,6 𝐹 0,74 𝐹 0,76 𝐹
The load levels are characterised by the average value of the anchorage failure load 𝐹 determined for the same
configuration of specimen under A.2.3.5. During the test the number of load cycles is counted until a separated
length of 30 mm is reached, see Chapter A.2.3.6.
If at least 2 million load cycles are achieved at all load levels A to D in the first test and at least 90.000 load cycles
in the second test, the S-N-curve defined in FprEN1992-1-1:2023 (E), Annex J, formulae (J.29) and (J.30) can be
used.

2.4 Design parameter for confinement of columns and their determination

The factor kcc in equations (J.7) and (J.10) represents a system coefficient that depends on the strength and
stiffness of the CFRP used for confinement. The recommended values were derived from databases of tests
performed under specific boundary conditions with available systems. Coefficients for other systems should be
determined in addition to the mandatory material tests on the CFRP under comparable boundary conditions, which
have yet to be established.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 699
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

3 Product specific design and application parameters for strengthening kits of

externally bonded CFRP sheets (EBR)
The kits in general consist of the following components:
Component A unidirectional CF (Carbon Fibre) sheets
Component B1 primer for the concrete surface to improve the bond of the adhesive (not in all kits),
Component B2 adhesive (often also called structural bonding agent or structural bonding) for the concrete
surface,
Component B3 impregnation for the CF sheets (often it is the similar to the adhesive B2).
Component C repair mortar with polymer binder,
Component D bonding agent for the repair mortar, repair concrete and shotcrete.
In general there are two types of kits regarding their handling on the construction site:
Type 1 In these kits, the CF sheets are impregnated (component B3) before they are worked into a layer of
bonding agent (component B2) applied to the surface of the building. Afterwards, a further adhesive
layer (component B2) is then applied to the adhesive layer (component B2) with the partially
incorporated and impregnated CF sheet (component A), which in turn can be used to incorporate a
further layer of CF sheet (component A) or to close off the surface of the layer of bonding agent
(component B2).
Type 2 In these kits, the CF sheets are worked into a layer of bonding agent (component B2) applied to the
surface of the building in such a way that the CF-Sheets (component A) are only partially integrated into
the layer of bonding agent (component B2). Before conclusion with an additional layer of bonding agent
(component B2) the impregnation (component B3) is applied to the partially integrated CF-Sheet.
Afterwards, a further adhesive layer (component B2) is then applied to the adhesive layer (component
B2) with the partially incorporated and impregnated CF sheet (component A), which in turn can be used
to incorporate a further layer of CF sheet (component A) or to close off the surface of the layer of
bonding agent (component B2).

Figure 3.1: Principle sketches of reinforced concrete sections with subsequent strengthening by externally
bonded CFRP sheets

The product-specific design parameters for kits of EBR CFRP sheets included in Annex J of FprEN 1992-1-1:2023
(E) are the following:

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 700
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

Table 3.1: Product-specific design parameters included in of FprEN 1992-1-1:2023 (E)

Symbol in Meaning used in FprEN 1992-1-1

FprEN 1992-1-1 (according to FprEN 1992-1-1) (only sections of Annex J are listed,
which refer to externally bonded CFRP
sheets)
Design parameters for the CFRP sheets (component A)
𝜂 reduction factor applied to the tensile
3.5.4, J.5.2 (1), J.15
strength of the ABR CFRP strip
𝑓 characteristic material strength of the CFRP
3.5.2, J.5.3 (1), formula (J.2), Figure J.1
strips
𝐸 mean value of the modulus of elasticity of the 3.5.1, J.5.3 (1), formula (J.3)
CFRP strips Figure J.1, formula (J.35), (J.37) and
(J.39), (J.40)
𝜀 ultimate strain of the adhesively bonded CFRP
3.5.4, J.5.2 (1), Figure J.1
strips
𝐴 cross sectional area of the fibres per
3.5.1, J.5.1 (3)
meter of the ABR CFRP sheet
Design parameters for the adhesive (component B2) the and the repair mortar (component C)
In all design formulae for the EBR CFRP sheets in FprEN 1992-1-1:2023 (E), Annex J no material parameter
for the adhesive and repair mortar are included but only material parameters of the impregnated ABR CFRP
sheets (component A+B2) and the concrete substrate. Therefore, when applying the formulae (J.35) to
(J.42), it is assumed that only kits are used that always have concrete failure in substrate and never adhesive
failure.

Design parameters for the bond to the concrete

𝑘 , product-specific system factor 3.5.2, formula (J.42), (J.48), J.15
𝑘 , product-specific system factor 3.5.2, formula (J.41), J.15
𝑓 average value of the compressive strength of 3.5.2, formula (J.39), J.15
the concrete substrate
𝑘 , product-specific system factor 3.5.2, J.11.1.1.2, formula (J.35), (J.37),
J.11.1.1.3, (J.42)
𝑓 , average value of the tensile surface strength of 3.5.2, J.11.1.1.1, formula (J.34),
the concrete substrate J.11.1.1.2, formula (J.35), (J.37),
J.11.1.1.3, formula (J.42)
𝛼 reduction factor for fatigue 3.5.4, J.10.2, formula (J.27)
Design parameters for confinement of columns
𝑘 product-specific system factor; 3.5.2, J.8.1.2
confinement factor for columns
strengthened with CFRP

3.1 Boundary conditions of the intended use of the kits

Kits tested in accordance to Annex BD-B are used since long time in Germany under the following environmental
conditions for a design life of 25 years:
 The strengthening kit may be used for normal concrete substrates of in the range of compressive
strength classes between 𝐶 to 𝐶 according to EN 206-1.
 Without additional protective measures for the FRP and adhesive, strengthening with adhesively
bonded reinforcement shall only be applied in exposure classes X0, XC1 (dry) or XC3 in accordance
with FprEN 1992-1-1:2023 (E), Table 6.1.
 In addition, EBR CFRP sheets shall not be exposed to direct UV radiation (direct sun radiation or indirect
sun from snow or water reflection) or alternating or permanent penetration of moisture in the area of
the strengthening. If necessary, appropriate protective layers shall be applied.
 The temperature of the concrete substrate does not exceed 𝑇 and does not fall below 𝑇 during
the entire design life of the strengthened structural element.
The range of compressive strength classes and the range of temperature concrete substrate are not given in
FprEN 1992-1-1:2023 (E). Therefore, it is necessary to be declared by the manufacturer and is largely confirmed
by the tests in accordance with Annex BD-B. If there is no confirmation by tests, the manufacturer's declaration
shall suffice or Annex BD-B shall be supplemented by suitable tests
A further difference to the boundary conditions of kits tested by Annex BD-A is that in FprEN 1992-1-1:2023,
Annex J, clause J.6.3 other exposure classes are also allowed. It shall be checked whether this is also possible
without additional tests compared to Annex BD-A. If not, the corresponding tests in Annex BD-A shall be added.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 701
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

3.2 Boundary conditions of the installation of the kits

The manufacturer's instructions and safety information shall be considered. In particular, it shall be ensured that
the surface temperature of the concrete component to be strengthened is always between the minimum and
maximum installation temperature of the component of the kit installed on it (see Table 3.2 "Application
parameters for the kit").

Table 3.2: Sources of design and application parameters from test procedures according to Annex BD-B for
kits of externally bonded CFRP sheets

Symbol according to Meaning Sources

FprEN 1992-1-1 Annex BD-B Test Additionally
procedure verified by
Design parameters for the CFRP sheets (component A)
𝜂 reduction factor applied to the B.2.2.6 and 2.3.1
tensile strength of the ABR CFRP B.2.2.7
sheet
𝑓 𝑓 characteristic material strength of the B.2.21)
CFRP sheet
𝐸 𝐸 mean value of the modulus of elasticity of B.2.21)
the CFRP sheet
𝜀 𝜀 ultimate strain of the adhesively bonded B.2.21)
CFRP sheet
𝐴 𝐴 cross sectional area of the fibres B.2.2.41)
per meter of the CFRP sheet
Design parameters for the bond to the concrete
𝑘 , product-specific system factor B.2.3.51) 3.3.2
1)
𝑘 , product-specific system factor B.2.3.5 3.3.2
𝑘 , product-specific system factor B.2.3.51) 3.3.2
2)
𝑓 𝑓 𝐶 average value of compression strength of B.2.3
the highest concrete strength class of the
tests
𝑓 𝐶 average value of compression strength of B.2.32)
the lowest concrete strength class of the
tests
𝑓 , 𝑓 , 𝐶 average value of the tensile surface B.2.32)
strength of the concrete substrate with the
highest concrete strength class of the
tests
𝛼 reduction factor for fatigue B.2.3.6 3.3.3
Design parameters for confinement of columns
product-specific system factor; 3.4
𝑘 confinement factor for columns
strengthened with CFRP
Application parameters for the kit
𝑇 , minimum temperature of installation for B.2.3.42)
the primer (component B1)
3)
𝑇 , maximum temperature of installation for
the primer (component B1)
𝑇 , minimum temperature of installation for B.2.3.42)
the adhesive (component B2)
3)
𝑇 , maximum temperature of installation for
the adhesive (component B2)
𝑇 , minimum temperature of installation for B.2.3.42)
the impregnation (component B3)
3)
𝑇 , maximum temperature of installation for
the impregnation (component B3)
𝑇 , minimum temperature of installation for B.2.3.42)
the repair mortar (component C)
3)
𝑇 , maximum temperature of installation for
the repair mortar (component C)

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 702
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

3)
𝑇 minimum temperature of the
strengthening kit at intended use
𝑇 maximum temperature of the B.2.3.32)
strengthening kit at intended use
𝐶 highest concrete strength class of the B.2.32)
tests
𝐶 lowest concrete strength class of the tests B.2.32)
1)
determined by test procedure
2)
declared by the manufacturer and confirmed by test procedure
3)
declared by the manufacturer
The number of tests given in Annex BD-B, clauses B.2.1 to B.2.3 is the minimum number which is necessary.
Each applicant can expand it according to their needs.

3.3.1 The creep reduction factor ηf of CFRP sheets derived from the tests according
to Annex BD-B, B.2.2.6 and B.2.2.7
The factor 𝜂 is the minimum of the factors 𝑅 from the test B.2.2.6 and 𝑅 from the test B.2.2.7
(𝜂 Min 𝑅 , 𝑅 . It is given separately for all alkaline solutions used in B.2.2.6 and B.2.2.7 (pH 9 and/or 11
and/or 13,7).

3.3.2 Parameters of the bilinear bond approach ksys,b1 and ksys,b2 and ksys,b3 derived from the tests
according to Annex BD-B, B.2.3.5
The differential formula for the bond of externally bonded reinforcement given in the Figure 2.3 can be completely
solved for the two branches of the bilinear bond approach. With an adjustment to the boundary conditions of the
bond test specimen according to Figure B.15, the following formulae are obtained, which describe the relationship
between the relative displacement sf at the load-side beginning of the bond length 𝑙 and the tensile force 𝐹 . For
s 𝑠 , only elastic bond stresses occur within the total bond length 𝑙 . After exceeding the elastic limit
displacement (𝑠 𝑠 ), additional plastically softened bond areas occur.
𝐹 𝐸 ∙ 𝑏 ∙ 𝑡 ∙ 𝑠 ∙ 𝜔 ∙ tanh 𝜔 ∙ 𝑙 for 𝑠 𝑠
∙ ∙ ∙
F 𝐸 ∙𝑏 ∙𝑡 ∙𝜔∙𝜆∙ for 𝑠 𝑠
∙ ∙

With:

𝜏
𝜔
𝑠 ∙𝐸 ∙𝑡

𝑠
𝜆
𝑠 𝑠

The length 𝑎 of the elastically deformed bond region enters into the second formula and can be determined with
the following expression. Because of the simultaneous occurrence of hyperbolic and trigonometric functions, 𝑎
cannot be determined in complete form but only iteratively.
𝜆 𝑠 𝑠
tanh 𝜔 ∙ 𝑎 ∙ s 𝑠 ∙ cot 𝜔 ∙ 𝜆 ∙ 𝑙 𝑎
𝑠 sin 𝜔 ∙ 𝜆 ∙ 𝑙 𝑎
The parameters 𝑠 , 𝜏 , and 𝑠 of the bilinear composite approach (cf. Figure 2.2) are determined using the least
squares method on the experimentally determined 𝑠 -𝜀 curves (cf. Figure 2.3) for each of the two sides of a
specimen. In view of the scope of the calculations to be performed, a limited number of supporting points is
recommended for the computational determination of the parameters.
Subsequently, system-specific coefficients 𝑘 , and 𝑘 , can be determined in regression analyses according
to EN 1990, Annex D.8.
∙
For this purpose, the correlations 𝜏 0 ⇔𝑓 𝑓 , 0 and 𝐺 0 ⇔𝑓 𝑓 , 0 can be
used in addition. For strengthening systems with externally bonded reinforcement, in which the adhesive
(component B2) does not lead to a hardening of the concrete layers near the surface, the following the following
simple functions 𝑓 and 𝑓 of the term 𝑓 ∙ 𝑓 , can be assumed for this purpose,

𝜏 𝑓 𝑓 ∙𝑓 , 𝑘 , ∙ 0,37 ∙ 𝑓 ∙𝑓 ,

𝑘 , ∙ 0,37 ∙ 𝑓 ∙𝑓 , ∙𝑘 , ∙ 0,2
𝐺 𝑓 𝑓 ∙𝑓 ,
2
𝑠 𝑘 , ∙ 0,2
For many known reinforcement systems with CFRP sheets and cold-curing epoxy resin adhesives, 𝑘 , and
𝑘 , are greater than 1,0.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 703
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

The region before the maximum bond length is described in formula (J.38) via a parabola. However, the exact
solution from the bilinear composite approach is not exactly parabolic.
After the test interpretation, a better agreement of the approximation from the parabolic approach with the test
values can be achieved with the factor 𝑘 , .
If, on the other hand, a low viscosity adhesive or primer is used that penetrates the concrete matrix and leads to
a hardening of the concrete layers close to the surface, a more attenuated relationship between the target values
and the concrete strengths is to be assumed, e.g.
𝜏 𝑓 𝑓 ∙𝑓 ,

𝐺 𝑓 𝑓 ∙𝑓 ,

with exponents 𝑢 , 𝑢 , 𝑣 and 𝑣 smaller than or equal to 0,5.

In the case of a multi-layer sheets application, the influence of the adhesion thickness, which increases with the
number of layers, may also have to be taken into account.

3.3.3 Reduction factor for fatigue derived from the tests according to Annex BD-B, B.2.3.6
The fatigue reduction factor 𝛼 used FprEN 1992-1-1:2023 (E), Annex J, formula (J.27) contains information for
an S-N-curve, which is verified by means of fatigue tests. The tests are performed according to the test procedure
in Annex BD-B, Clause B.2.3.6. The following load levels are checked:

Table 3.3: Load level for fatigue bond tests

Load level Lower load 𝐹bLi Upper load 𝐹bLi 1st test Upper load 𝐹bLi 2nd test
A 0,15 𝐹 0,45 𝐹 0,49 𝐹
B 0,3 𝐹 0,55 𝐹 0,58 𝐹
C 0,45 𝐹 0,64 𝐹 0,67 𝐹
D 0,6 𝐹 0,74 𝐹 0,76 𝐹

The load levels are characterised by the average value of the anchorage failure load 𝐹 determined for the same
configuration of specimen under B.2.3.5. During the test the number of load cycles is counted until a separated
length of 30 mm is reached, see Clause B.2.3.6.
If at least 2 million load cycles are achieved at all load levels A to D in the first test and at least 90.000 load cycles
in the second test, the S-N-curve defined in FprEN 1992-1-1:2023 (E), Annex J, formulae (J.29) and (J.30) can
be used.

3.4 Design parameter for confinement of columns and their determination

The factor kcc in equations (J.7) and (J.10) represents a system coefficient that depends on the strength and
stiffness of the CFRP used for confinement. The recommended values were derived from databases of tests
performed under specific boundary conditions with available systems. Coefficients for other systems should be
determined in addition to the mandatory material tests on the CFRP under comparable boundary conditions, which
have yet to be established.

4 Product specific design and application parameters for strengthening kits of near
surfaced mounted CFRP strips (NSM)

Figure 4.1: Principle sketches of reinforced concrete sections with subsequent strengthening by near surface
mounted CFRP strips

The product-specific design parameters for strengthening kits of NSM CFRP included in Annex J of FprEN 1992-
1-1:2023 (E) are the following:

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 704
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

Table 4.1: Product-specific design parameters included in of FprEN 1992-1-1:2023 (E)

Symbol in Meaning used in FprEN 1992-1-1

FprEN 1992-1-1 (according to FprEN 1992-1-1, clause 3.5) (only clauses of Annex J are listed,
which refer to near surface mounted
CFRP strips)
Design parameters for the CFRP strips (component A)
𝜂 reduction factor applied to the tensile
3.5.4, J.5.2 (1), J.15
strength of the EBR CFRP sheet or strip
𝑓 characteristic short-term tensile strength of the
3.5.2, J.5.3 (1), formula (J.2), Figure J.1
ABR CFRP
𝐸 Mean modulus of elasticity in longitudinal 3.5.1, J.5.3 (1), formula (J.3)
direction of ABR CFRP Figure J.1,
𝜀 characteristic ultimate strain for ABR CFRP 3.5.4, J.5.2 (1), Figure J.1
𝐴 Cross sectional area of a CFRP system 3.5.1, J.5.1 (3)
Design parameters for the adhesive (component B)
𝛼 product-specific system factor for long-term 3.5.4, J.11.1.3, formula (J.59)
behaviour of the adhesive
𝑓 characteristic compression strength of the 3.5.2, J.5.1 (1), J.11.1.3, formula (J.60)
adhesive
𝑓 characteristic tensile strength of the adhesive 3.5.2, J.5.1 (1), J.11.1.3, formula (J.60)

Symbol in Meaning used in FprEN 1992-1-1

FprEN 1992-1-1 (according to FprEN 1992-1-1, clause 3.5) (only clauses of Annex J are listed,
which refer to near surface mounted
CFRP strips)
Design parameters for the repair mortar (component C)
In all design formulae (J.56) to (J.61) for the NSM CFRP of FprEN 1992-1-1, Annex J no material parameter
for the repair mortar (component C) are included but only material parameters of the ABR CFRP (component
A), the concrete substrate and the adhesive (component B).
Design parameters for the bond to the concrete
𝛼 product-specific system factor for long-term 3.5.4, J.11.1.3, formula (J.59)
behaviour of concrete
𝑓 average value of the compressive strength of 3.5.2, J.11.1.3, formula (J.61)
the concrete substrate

4.1 Boundary conditions of the intended use of the kits

Kits tested in accordance to Annex BD-C are used since long time in Germany under the following environmental
conditions for a design life of 25 years:
 The strengthening kit may be used for normal concrete substrates of in the range of compressive
strength between 𝐶 to 𝐶 according to EN 206-1.
 Without additional protective measures for the FRP and adhesive, strengthening with adhesively
bonded reinforcement shall only be applied in exposure classes X0, XC1 (dry) or XC3 in accordance
with FprEN 1992-1-1:2023 (E), Table 6.1.
 In addition, NSM CFRP shall not be exposed to direct UV radiation (direct sun radiation or indirect sun
from snow or water reflection) or alternating or permanent penetration of moisture in the area of the
strengthening. If necessary, appropriate protective layers shall be applied.
 The temperature of the concrete substrate does not exceed 𝑇 and does not fall below 𝑇 during
the entire design life of the strengthened structural element.
The range of compressive strength classes and the range of temperature concrete substrate are not given in
FprEN 1992-1-1:2023 (E). Therefore, it is necessary to be declared by the manufacturer and is largely confirmed
by the tests in accordance with Annex BD-C. If there is no confirmation by tests, the manufacturer's declaration
shall suffice or Annex BD-C shall be supplemented by suitable tests
A further difference to the boundary conditions of kits tested by Annex BD-A is that in FprEN 1992-1-1:2023,
Annex J, clause J.6.3 other exposure classes are also allowed. It shall be checked whether this is also possible
without additional tests compared to Annex BD-A. If not, the corresponding tests in Annex BD-A shall be added.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 705
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

4.2 Boundary conditions of the installation of the kits

The manufacturer's instructions and safety information shall be considered. In particular, it shall be ensured that the
surface temperature of the concrete component to be strengthened is always between the minimum and maximum
installation temperature of the component of the kit installed on it (see Table 4.2 "Application parameters for the
kit").

4.3 Design and application parameters and their determination

Table 4.2: Sources of design and application parameters from test procedures according to Annex BD-A
and Annex BD-C for kits of near surface mounted CFRP strips

Symbol according to Meaning Sources

FprEN 1992-1-1 Annex BD-A Test Additionally
and BD-C procedure verified
Design parameters for the CFRP strips (component A)
𝜂 reduction factor applied to the A.2.2.5 and 4.3.1
tensile stress of the NSM CFRP strip A.2.2.61)
𝑓 𝑓 characteristic material strength of the A.2.21)
CFRP strips
𝐸 𝐸 mean value of the modulus of elasticity of A.2.21)
the CFRP strips
𝜀 𝜀 ultimate strain of the adhesively bonded A.2.21)
CFRP strips
𝑡,𝑏 𝑡 ,𝑏 thickness and width of the CFRP strip A.2.2.31)
𝐴 𝐴 𝑑 ⋅𝑏 the area of the cross section of CFRP A.2.2.31)
strips
Symbol according to Meaning Sources
FprEN 1992-1-1 Annex BD-A Test Additionally
and BD-C procedure verified
Design parameters for the adhesive (component B)
𝛼 𝑅 , , product-specific system factor for long- C.2.3.5,
term behaviour of the adhesive C.2.3.61)
𝑓 𝑓Gck characteristic compression strength of the A.2.1.2,
adhesive A.2.1.31)
𝑓 𝑓Gflk characteristic tensile strength of the A.2.1.2,
adhesive A.2.1.31)
Design parameters for the bond to the concrete
𝛼 𝛼 product-specific system factor for long- 𝛼
term behaviour of concrete 𝜂 ⋅𝑘
,
⋅𝑘
2)
𝑓 𝑓 𝐶 average value of compression strength of C.2.3
the highest concrete strength class of the
tests
𝑓 𝐶 average value of compression strength of C.2.32)
the lowest concrete strength class of the
tests
Application parameters for the kit
𝑇 , minimum temperature of installation for C.2.3.42)
the adhesive (component B)
3)
𝑇 , maximum temperature of installation for
the adhesive (component B)
𝑇 , minimum temperature of installation for C.2.3.42)
the repair mortar (component C)
3)
𝑇 , maximum temperature of installation for
the repair mortar (component C)
3)
𝑇 minimum temperature of the
strengthening kit at intended use
𝑇 maximum temperature of the C.2.3.32)
strengthening kit at intended use
𝐶 highest concrete strength class of the C.2.32)
tests
𝐶 lowest concrete strength class of the tests C.2.32)

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 706
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

1)
determined by test procedure
2)
declared by the manufacturer and confirmed by test procedure
3)
declared by the manufacturer

The number of tests given in Annex BD-A, clauses A.2.1 and A.2.2 and in Annex BD-C, clause C.2.3 is the
minimum number which is necessary. Each applicant can expand it according to their needs.

4.3.1 The creep reduction factor ηf of CFRP strips derived from the tests according to
Annex BD-A, A.2.2.5 and A.2.2.6
The factor 𝜂 is the minimum of the factors 𝑅 from the test A.2.2.5 and 𝑅 from the test A.2.2.6
(𝜂 𝑀𝑖𝑛 𝑅 , 𝑅 . It is given separately for all alkaline solutions used in 2.2.5 and 2.2.6 (pH 9 and/or 11 and/or
13,7).

References 5 References
[1] EN 1504-4:2004: Products and systems for the protection and repair of concrete structures. Definitions,
requirements, quality control and evaluation of conformity. Part 4: Structural bonding.
[2] Kuntz, S.: Verbundverhalten von laschenverstärkten Stahlbetonzuggliedern: Diplomarbeit, Institut für
Baustoffe, Massivbau und Brandschutz, TU Braunschweig; 1993.
[3] Holzenkämpfer, P.: Ingenieurmodelle des Verbundes geklebter Bewehrung für Betonbauteile;
Dissertation, TU Braunschweig; 1994.
[4] Niedermeier, R.: Zugkraftdeckung von klebearmierten Bauteilen; Dissertation, TU München; 2001.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 707
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

Annex BD-A Test program for externally bonded CFRP strips

A.1 Specific terms used in this Annex

A.1.1 Significant cross sections of one type of CFRP strips

If for one type of CFRP strips (see A.1.3) more than one cross section exist the significant cross
sections are those with the smallest and the highest thickness those with the smallest and highest
width as well as those with the section of the highest ratio between circumference and cross-section
area.

A.1.2 Coil diameter of CFRP strips

CFRP strips are usually delivered on site in form of coils. Without further investigations the inner coil
diameter should not be less than the diameter used in the assessment tests.

A.1.3 Type of CFRP strips

A type of CFRP strips is characterised by the same production technology, the same fibres and size
of fibres (see A.1.7), the same fibre content and the same matrix material.

A.1.4 Standard environmental conditions

In the framework of this Annex the standard environmental conditions are a temperature of 21 2°C
and relative air humidity of (50 ± 10)%.

A.1.5 Minimum temperature of installation

The following terms are used in this document:
𝑇 , , minimum temperature of installation of the structural bonding agent (component B),
𝑇 , , minimum temperature of installation of the repair mortar (component C and D).
The default values for these temperatures are 8°C. The default values are only used if the
manufacturer's instructions and safety information do not contain any other information.

A.1.6 Minimum and maximum temperature at the intended use

The default values for the minimum and maximum temperature of the repair mortar (component C and
D) 𝑇 , and 𝑇 , as well as for the bonding agent (component B) 𝑇 , and 𝑇 , at the intended
use is -10°C and 40°C. The default values are only used if the manufacturer's instructions and safety
information do not contain any other information.
For the intended use of the kit as a whole the minimum temperature results from 𝑇
max 𝑇 , , 𝑇 , and the maximum temperature from 𝑇 min 𝑇 , , 𝑇 , .

A.1.7 Size of fibres

During the spinning process the size is raised on the carbon fibres. It is a protective coating for the fibres
of less than 100 nm of thickness and improves the bond of the fibres to the matrix materials of the CFRP
strips. Mostly it consists of polymers.

A.1.8 Minimum and maximum thickness of layers of bonding agent (component B) and repair
mortar (component C and D)
The default minimum and maximum thicknesses
 of the layers of bonding agent (component B) are 𝑑 , = 1 mm and 𝑑 , = 5 mm and
 of the repair mortar (component C and D) 𝑑 , = 5 mm and 𝑑 , = 30 mm.
The default values are only used if the manufacturer's instructions and safety information do not contain
any other information.

A.1.9 Minimum curing time of the bonding agent (component B) and repair mortar (component C
and D) at standard environmental conditions (A.1.4) and at minimum temperature of
installation (see A.1.5)
The default values for the minimum curing time of the bonding agent (component B) at standard
environmental conditions (A.1.4) 𝑡 are 7 d and at minimum temperature of installation (A.1.5) 𝑡
28 d. The default values are only used if the manufacturer's instructions and safety information do not
contain any other information.
The default values for the minimum curing time of the repair mortar (component C and D) at standard
environmental conditions (A.1.4) 𝑡 are 7d and at minimum temperature of installation (A.1.5) 𝑡
28 d. The default values are only used if the manufacturer's instructions and safety information do not
contain any other information.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 708
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

A.1.10 Minimum and maximum strength class of the concrete substrate

The default values for the minimum and maximum strength classes of the concrete substrate are C12/15
and C50/60. The default values are only used if the manufacturer's instructions and safety information
do not contain any other information.

A.2 Tests program for kits of externally bonded CFRP strips

A.2.1 Tests to determine the product-specific properties of the adhesive (structural bonding agent)
and repair mortar

A.2.1.1 Glass transition temperature of the structural bonding agent (component B) and the repair
mortar (component C)
Purpose of the assessment
To identify the maximum temperature of intended use (see A.1.6) of the kit is necessary to determine the glass
transition temperature of the structural bonding agent (component B) 𝑇 and the repair mortar (component C)
𝑇 .
Assessment method
Basis for the tests is EN 12614. The glass transition temperature shall be determined on specimens of cured
structural bonding agent (component B) and repair mortar (component C) from one batch of resin and hardener.
The glass transition temperature must only be measured during the first heating cycle; the heating speed must
be 10°C/min. The preparatory cycle of temperature according to EN 12614, clause 6.2 shall be omitted if it causes
a change of the glass transition temperature.
The glass transition temperature shall be determined on at least two specimens of bonding agent (component B)
which have cured at least for 𝑡 (see A.1.9) and on at least two specimens of repair mortar (component C) which
have cured at least for 𝑡 (see A.1.9) under standard environmental conditions (see A.1.4).
Expression of results
Glass transition temperatures of the structural bonding agent (component B) 𝑇 together with the curing time 𝑡
(see A.1.9) and the repair mortar (component C) 𝑇 together with the curing time 𝑡 (see A.1.9) shall be given
in the test report.

A.2.1.2 Flexural and compressive strength of the structural bonding agent (component B) and the
repair mortar (component C)
Purpose of the assessment
These tests are intended to determine the relevant mechanical properties of the structural bonding agent
(component B) and the repair mortar (component C) after curing under standard environmental conditions (see
A.1.4). This includes the flexural and compressive strength.
Assessment method
The flexural and compressive strength according to EN 196-1 or EN 1015-11 of at least 6 specimens 160 mm x
40 mm x 40 mm (length x height x width) of cured structural bonding agent and repair mortar from one batch of
resin and hardener shall be determined. The specimens shall cure at standard environmental conditions (see
A.1.4). At least 3 tests of the flexural strength according to EN 196-1, 9.1 or EN 1015-11, clauses 8 and 6 tests of
the compressive strength according to EN 196-1, clause 9.2 or EN 1015-11, clause 9 shall be carried out at
standard environmental conditions (see A.1.4).
At least tests after curing of the bonding agent (component B) for 𝑡 (see A.1.9) and of the repair mortar
(component C) for 𝑡 (see A.1.9) shall be carried out.
Expression of results
From the failure loads of the bending tests with the specimens of structural bonding agent 𝐹Gfl, and the specimens
of repair mortar 𝐹Mfl, , the bending tensile strengths 𝑓Gfli and 𝑓Mfli shall be determined according to EN 196-1, clause
9.1 or EN 1015-11, clause 8. From these values, the mean values 𝑓Gflm and 𝑓 as well as the 5% quantile value
according to EN 1990, Table D.1 (Vx unknown) 𝑓Gflk and 𝑓 shall be determined and given in the test report.
From the failure loads of the compressive tests with the specimens of structural bonding agent 𝐹Gc, and the
specimens of repair mortar, 𝐹Mc, the compressive strengths 𝑓 and 𝑓Mci shall be determined according to EN 196-
1, clause 9.2 or EN 1015-11, clause 9. From these values, the mean value 𝑓Gcm and 𝑓Mcm as well as the 5% quantile
value according to EN 1990, Table D.1 (Vx unknown) 𝑓Gck and 𝑓Mck shall be determined and given in the test report.

A.2.1.3 Flexural and compressive strength of the structural bonding agent (component B) and the
repair mortar (component C) cured at minimum temperature of installation in dependence of
curing time
Purpose of the assessment
These tests are intended to determine the flexural and compressive strength of the structural bonding agent
(component B) and the repair mortar (component C) at minimum temperature of installation (see A.1.5) in
dependence of curing time.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 709
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

Assessment method
For different curing conditions regarding to temperature and time the flexural and compression strength of cured
specimens of bonding agent and repair mortar shall be determined according to A.2.1.2. At least specimens of
structural bonding agent (component B) cured at minimum temperature of installation 𝑇 , °C (see A.1.5) and
specimens of repair mortar (component C) cured at minimum temperature of installation 𝑇 , °C (see A.1.5)
shall be tested.
For each material at least 3 tests shall be carried out. The bending and compression test shall be carried out
immediately after the curing time ending. For the bonding agent (component B) at least tests after curing for tGTmin
and 2 tGTmin (see A.1.9) and for the repair mortar (component C and D) at least tests after curing for tMTmin and 2
tMTmin (see A.1.9) at the minimum temperature of installation 𝑇 , and 𝑇 , (see A.1.6) shall be carried out.

Expression of results
From the failure loads of the bending tests "i" for the curing time "j" of structural bonding agent 𝐹Gfl,ij and the
specimens of repair mortar 𝐹Mfl,ij the bending tensile strengths 𝑓Gflij and 𝑓Mtflj shall be determined according to
EN 196-1, clause 9.1 or EN 1015-11, clause 8. From these values, the mean value 𝑓Gtmj and 𝑓Mtmj as well as the
5% quantile value according to EN 1990, Table D.1 (Vx unknown) 𝑓Gflkj and 𝑓Mflkj shall be given in the test report
in conjunction with the used curing temperature and time.
From the failure loads of the compression tests "i" for the curing time "j" of the structural bonding agent 𝐹 , and
the specimens of repair mortar 𝐹 , the bending compression strengths 𝑓Gcij and 𝑓Mcij shall be determined
according to EN 196-1, clause 9.1 or EN 1015-11, clause 9. From these values, the mean value 𝑓Gcmj and 𝑓Mcmj
as well as the 5% quantile value according to EN 1990, Table D.1 (Vx unknown) 𝑓Gckj and 𝑓Mckj shall be given in
the test report in conjunction with the used curing temperature and time.

A.2.2 Tests to determine the product-specific properties of the CFRP strips

A.2.2.1 Fibre content of the CFRP strips

Purpose of the assessment
Determination of the fibre content of the CFRP strips.
Assessment method
The fibre content shall be determined according to EN 2564 method A. Tests shall be carried out at least on 6
specimens of known cross sections (see A.2.2.3) of all types of CFRP strips.
Expression of results
For every type of CFRP strips (see A.1.3) the average value and the characteristic value of the fibre content shall
be determined and reported in the test report.

A.2.2.2 Weight per meter of the CFRP strips

Purpose of the assessment
Determination of the weight of the CFRP strips.
Assessment method
Weight of 1m long pieces of CFRP strips mLi shall be determined. Tests shall be carried out at least on 6
specimens for the significant cross sections (see A.1.1) of all types of CFRP strips (see A.1.3).
Expression of results
For every type of CFRP strips (see A.1.3) the average value mLm and the 5% quantile value according to EN 1990,
Table D.1 (Vx unknown) mLk of the determined weight of 1m long pieces mLi shall be determined and reported in
the test report.

A.2.2.3 Dimensions of the CFRP strips

Purpose of the assessment
Determination of the dimensions of the CFRP strips.
Assessment method
Thickness 𝑡L and width 𝑏 of 1m long pieces of CFRP strips shall be measured. The thickness shall be measured
on at least 9 points (see figure A.1) and the width at least on 3 places per specimen (see figure A.2). Tests shall
be carried out at least on 6 specimens for the significant cross sections (see A.1.1) of all types of CFRP strips
(see A.1.3).

Figure A.1: Measure points for the thickness 𝑡L of CFRP strips

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 710
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

Figure A.2: Measure lines for the width 𝑏 of CFRP strips

Expression of results
For every significant cross section of every type of CFRP strips (see A.1.3) the average values 𝑡 and 𝑏 as
well as the 5% quantile value according to EN 1990, Table D.1 (Vx unknown) 𝑡 and 𝑏 of the determined
thicknesses 𝑡 and widths 𝑏 shall be determined and reported in the test report.
For every significant cross section of every type of CFRP strips (see A.1.3) also the average value 𝐴 𝑡 ⋅
𝑏 as well as the 5% quantile value according to EN 1990, Table D.1 (Vx unknown)
𝐴 𝑡 ⋅ 𝑏 of the cross section shall be reported in the test report.

A.2.2.4 Modulus of elasticity, tensile strength and strain at failure of CFRP strips (component A)
Purpose of the test
These tests are intended to determine the relevant mechanical properties of the CFRP strips (component A). This
includes the modulus of elasticity and the tensile strength, both related to the section of the CFRP strip as well as
the ultimate strain of CFRP strips.
Assessment method
The tensile load at failure 𝐹 (called 𝑃 in EN 2561, 7.2.7), the strains at 0,1 𝐹 as well as at 0,5 𝐹 and the strain
at failure 𝜀 (called 𝜀 in EN 2561, clause 7.2.7) of the CFRP strips shall be determined according to EN 2561,
Annex A, with specimens type B. In deviation from EN 2561, Annex A, Table A.1 the thickness of the test
specimens may correspond to the thickness of the CFRP strips. Tests shall be carried out at least on 5 specimens
of the significant cross sections (see A.1.1) of every type of CFRP strip (see A.1.3).
Expression of results
According to EN 2561 it is only possible to measure failure loads 𝐹 , the strains 𝜀 )A at 0,1 𝐹 (called 𝜀 )A in
EN 2561, clauses 8.3 and 8.4), the strains 𝜀 )B at 0,5 𝐹 (called 𝜀 )B in EN 2561, clauses 8.3 and 8.4) and
the strains at failure 𝜀 (called (11)R in EN 2561, clause 8.6). To determine strengths and secant moduli of
elasticity from these test results there are two possibilities. The strengths related to the sections of the CFRP
strips (called  in EN 2561, clause 8.1) and the strengths related to the sections of all fibres of the sections of
CFRP strips (called  in EN 2561, clause 8.2) as well as the moduli of elasticity related to the sections of the
CFRP strips (called 𝐸 in EN 2561, clause 8.3) and the moduli of elasticity related to the sections of all fibres of
the sections of CFRP strips (called 𝐸 in EN 2561, clause 8.4). With CFRP strips, it is customary to refer to the
cross sections of the strips. Therefore, the tensile strength and the moduli of elasticity shall only be determined
related to the section of the CFRP strips according to EN 2561, clause 8.1 and EN 2561, clause 8.3.
From the single values 𝐸 , 𝑓 , 𝜀 according to EN 2561, 8.3, 8.1 and 7.2.7 he average values of the secant
modulus of elasticity 𝐸 , of the tensile strengths 𝑓 and of the ultimate strain 𝜀 and the 5% quantile value
according to EN 1990, Table D.1 (𝑉 unknown) of tensile strengths 𝑓 shall be given in the test report for all types
of CFRP strips (see A.1.3).

A.2.2.5 Resistance of CFRP strips (component A) after storage in alkaline environment at maximum
temperature at intended use (see A.1.6)
Purpose of the assessment
These tests are intended to determine the change of the mechanical properties of the CFRP strips (component
A) according to A.2.2.4 after long-time storage in alkaline environment at maximum temperature at intended use
(see A.1.6). This includes the change of the secant modulus of elasticity, tensile strength and ultimate strain of
the CFRP strips.
Assessment method
At least in one of the following alkaline solutions the CFRP strips shall be stored at maximum temperature at
intended use (see A.1.6) for at least 1800 h:
 alkaline solution pH 9,0,
 alkaline solution pH 11,0,
 alkaline solution pH 13,7.
After this storage the specimens are washed and dried for 24 h at standard environmental conditions (see A.1.4).
Afterwards the tensile load at failure 𝐹 (called 𝑃 in EN 2561, clause 7.2.7), the strains at 0,1 𝐹 as well as at
0,5 𝐹 and the strain at failure 𝜀 (called 𝜀 in EN 2561, clause 7.2.7) of the CFRP strips shall be determined
according to EN 2561, Annex A, with specimens type B. In deviation from EN 2561, Annex A, Table A.1 the
thickness of the test specimens may correspond to the thickness of the CFRP strips.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 711
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

Tests shall be carried out at least on 5 specimens of CFRP strips for at least one thickness per type of CFRP
strips (see A.1.3). Preferably it should be the section with the highest ratio between circumference and cross-
section area among all the sections of this type of CFRP strips (see A.1.3).
Expression of results
According to EN 2561 it is only possible to measure failure loads 𝐹 , the strains 𝜀 )A at 0,1 𝐹 (called 𝜀 )A in
EN 2561, clauses 8.3 and 8.4), the strains 𝜀 )B at 0,5 𝐹 (called 𝜀 )B in EN 2561, clauses 8.3 and 8.4) and
the strains at failure 𝜀 (called 𝜀 in EN 2561, clause 8.6). To determine strengths and secant moduli of
elasticity from these test results there are two possibilities according to EN 2561. In accordance to A.2.2.4 it is
only necessary to determine the strengths related to the sections of the CFRP strips 𝑓 (called  in EN 2561,
clause 8.1) as well as the moduli of elasticity related to the sections of the CFRP strips 𝐸 (called 𝐸 in EN
2561, clause 8.3).
From the single values 𝐸 , 𝑓 and 𝜀 according to EN 2561, 8.3, 8.1 and 7.2.7 the average values 𝐸 ,𝑓
and 𝜀 as well as the ratios 𝑅 , 𝑅 and 𝑅 shall be determined as follows:

𝑅 𝑅 𝑅

where:
𝐸 ,𝑓 ,𝜀 are the average Modulus of elasticity, tensile strength and ultimate strain after
storage in alkaline solution
𝐸 , 𝑓 , 𝜀, are the average Modulus of elasticity, tensile strength and ultimate strain
according to A.2.2.4
The values 𝑅 , 𝑅 and 𝑅 shall be given in the test report for all types (see A.1.3) of CFRP strips separately
for all applied alkaline solutions (pH 9 and/or 11 and/or 13.7) used for the tests.

A.2.2.6 Resistance of CFRP strips (component A) in alkaline environment, under long-term load at
maximum temperature at maximum temperature at intended use (see A.1.6)
Purpose of the test
These tests are intended to determine the change of the mechanical properties of the CFRP strips (component
A) according to A.2.2.4 after long-time loading in alkaline environment at maximum temperature at intended use
(see A.1.6). This includes the change of the secant modulus of elasticity, of the tensile strength and of the ultimate
strain of the CFRP strips. In comparison to A.2.2.4 the secant modulus of elasticity and the tensile strength only
needs to be related to the cross section of the CFRP strips and not to the cross section of all fibres of the CFRP
strips.
Assessment method
At least in one of the following alkaline solutions the CFRP strips shall be stored at maximum temperature at
intended use (see A.1.6) and stressed with 50% of the characteristic value of tensile strength according to A.2.2.4
for at least 1800 h:
 alkaline solution pH 9,0,
 alkaline solution pH 11,0,
 alkaline solution pH 13,7.
After this storage the specimens are washed and dried for 24 h at standard environmental conditions (see A.1.4).
Afterwards the tensile load at failure 𝐹 (called 𝑃 in EN 2561, 7.2.7), the strains at 0,1 𝐹 as well as at 0,5 𝐹
and the strain at failure 𝜀 (called 𝜀 in EN 2561, 7.2.7) of the CFRP strips shall be determined according to
EN 2561, Annex A, with specimens type B. In deviation from EN 2561, Annex A, Table A.1 the thickness of the
test specimens may correspond to the thickness of the CFRP strips.
Tests shall be carried out at least on 5 specimens of CFRP strips for at least one thickness per type of CFRP
strips (see A.1.3). Preferably it should be the section with the highest ratio between circumference and cross-
section area among all the sections of this type of CFRP strips (see A.1.3).
Expression of results
According to EN 2561 it is only possible to measure failure loads 𝐹 , the strains 𝜀 at 0,1 𝐹 (called 𝜀
in EN 2561, clauses 8.3 and 8.4), the strains 𝜀 at 0,5 𝐹 (called 𝜀 in EN 2561, clauses 8.3 and 8.4) and
the strains at failure 𝜀 (called 𝜀 in EN 2561, clause 8.6). To determine strengths and secant moduli of
elasticity from these test results there are two possibilities according to EN 2561. In accordance to 2.2.4 it is only
necessary to determine the strengths related to the cross section of the CFRP strips 𝑓 (called  in EN 2561,
clause 8.1) as well as the moduli of elasticity related to the sections of the CFRP strips 𝐸 (called 𝐸 in EN
2561, clause 8.3) and not to the cross section of all fibres of the CFRP strips.
From the single values 𝐸 , 𝑓 and 𝜀 according to EN 2561, 8.3, 8.1 and 7.2.7 the average values 𝐸 ,𝑓
and 𝜀 as well as the ratios 𝑅 , 𝑅 and 𝑅 shall be determined as follows:

𝑅 𝑅 𝑅

where:
𝐸 ,𝑓 ,𝜀 are the average Modulus of elasticity, tensile strength and ultimate strain
after storage in alkaline solution

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 712
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

𝐸 ,𝑓 ,𝜀 are the average Modulus of elasticity, tensile strength and ultimate strain
according to A.2.2.4
The values 𝑅 , 𝑅 and 𝑅 shall be reported in the test report for all types (see A.1.3) of CFRP strips
separately for all applied alkaline solutions (pH 9 and/or 11 and/or 13.7) used for the tests.

A.2.3 Tests to verify the product-specific properties TGmin,a, TMmin,a, Tmax and Cmin, Cmax as well as to
determine the product-specific parameters of the bilinear bond approach ksys,b1 and ksys,b2 and
the reduction factor for fatigue loading fat2

A.2.3.1 Bond strength of specimens cured at standard environmental conditions (see A.1.4)
Purpose of the assessment
The tests are performed to determine the bond strength of the following specimens:
 specimens with layers of concrete, repair mortar (component C and D), bonding agent (component B)
and CFRP strips (component A),
 specimens with layers of concrete, bonding agent (component B) and CFRP strips (component A).
The bond strength is determined by pull-off tests according to EN 1542 after curing under standard conditions
(see A.1.4).
Assessment method
The test is based on EN 1542. The main differences to EN 1542 are the following:
 only steel dollies according to EN 1542, 4.7 are used, not dollies of aluminium,
 the bonding agent (Component B) is used also for the bonding of the steel dollies,
 the thickness of the layers of bonding agent (also for the dollies) is 𝑑 , (see A.1.8),
The preparation of the specimens with repair mortar shall be done as follows:
 On at least one specimen according to EN 1542, clause 4.12, with the maximum strength class of the
concrete substrate (see A.1.10) a layer of repair mortar (component C and D) with the maximum
thickness 𝑑 , 1 (see A.1.8) shall be installed according to the manufacturer’s instructions and safety
information (see Figure A.3a).
 The preparation of the surface according to EN 1542, clause 5.1, shall be done for the use in vertical
position.
 On the repair mortar (component C and D) a layer of CFRP strips bonded by a layer of bonding agent
(component B) with the minimum thickness 𝑑 , 2 (see A.1.8) shall be installed according to the
manufacturer’s instructions and safety information (see Figure A.3b). Preferably the width of the CFRP
strips should be at least 50 mm, to allow the drilling of the annular grooves without joints between CFRP
strips.
 After curing of the repair mortar (component C and D) and the bonding agent (component B) according
to the manufacturer's instructions and safety information at standard environmental conditions (see A.1.4)
on each specimen 5 annular grooves of 50 mm diameter shall be drilled through the multilayer specimen
until a depth of 15 ± 5 mm in the reference concrete is reached (see Figure A.3). The layers of CFRP
strips inside the annular grooves preferably should not contain joints between CFRP strips.

1 attached formwork
2 reference body of concrete according to EN 1542
3 layer of repair mortar (component C and D) with minimum thickness 𝑑 ,
4 layer of bonding agent (component B) with minimum thickness 𝑑 ,
5 CFRP strip (component A) externally bonded to the repair mortar (component C and D)
Figure A.3: Cross section through the test specimen with repair mortar (component C and D)
a) Specimen with repair mortar (component C and D)
b) Specimen with repair mortar (component C and D), bonding agent (component B) and externally
bonded CFRP strips (component A)

1
At least one test series with the maximum layer thickness of repair mortar is necessary.
2
At least one test series with the minimum layer thickness of bonding agent is necessary.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 713
 CEN/TC 250/SC 2 N2087
CEN/TC 250/SC 2/WG 1 N 1314
J.15 Background to FprEN 1992-1-1:2023

1 groove made by core drilling around the test surface

2 steel dolly with diameter of 50 mm
Figure A.4: Top view on the test specimen with annular grooves showing the areas of the steel dollies
The preparation of the specimens without repair mortar shall be done as follows:
 The preparation of the surface according to EN 1542, clause 5.1, shall be done for the use in vertical
position.
 On at least one specimen according to EN 1542, clause 4.12 (see Figure A.5a), with the maximum
strength class of the concrete substrate (see A.1.10) a layer of CFRP strips (component A) bonded by a
2
layer of bonding agent (component B) with the minimum thickness 𝑑 , (see A.1.8) shall be installed
according to the manufacturer’s instructions and safety information (see Figure A.5b). Preferably the
width of the CFRP strips should be at least 50 mm, to allow the drilling of the annular grooves without
joints between CFRP strips.

 After curing of the bonding agent (component B) according to the manufacturer's instructions and safety
information at standard environmental conditions (see A.1.4) on each specimen 5 annular grooves of
50 mm diameter shall be drilled through the multilayer specimen until a depth of 15 ± 5 mm in the
reference concrete is reached. The layers of CFRP strips inside the annular grooves preferably should
not contain joints between CFRP strips.

1 attached formwork
2 reference body of concrete according to EN 1542
4 layer of bonding agent (component B) with minimum thickness 𝑑 , (see A.1.9)
5 CFRP strip (component A) externally bonded to the concrete
Figure A.5: Cross section through the test specimen without repair mortar
a) Specimen without repair mortar
b) Specimen with bonding agent (component B) and externally bonded CFRP strips (component A)
The further treatment is the same for the test specimens with and without repair mortar:
 On each annular groove a dolly according to EN 1542, clause 7.2, shall be glued on the circular surfaces
with the structural bonding agent (component B). The minimum thickness of the layer 𝑑 , (see A.1.8)
of bonding agent shall be used.
 The structural bonding agent (component B) of the dollies shall cure according to the manufacturer's
instructions and safety information at standard environmental conditions (see A.1.4) before testing.
 Afterwards the bond strength 𝑓cti,surf shall be determined on each annular groove of the specimens with
and without repair mortar according to EN 1542, 7.4 at standard environmental conditions (see A.1.4).
The bond strength shall be determined on at least 5 annular grooves with and without repair mortar (component
C and D) for every type of CFRP strip.

R. Alex, R. Niedermeier, W. Finckh, T. Leusmann 21.12.2022

For citations, copying and referencing see Note on Front page! page 714