Magazine of Concrete Research Magazine of Concrete Research

http://dx.doi.org/10.1680/jmacr.16.00471
Paper 1600471
Shear friction strength of smooth Received 01/11/2016; revised 28/03/2017; accepted 08/05/2017
Keywords: joints/shear/structural design
construction joints and monolithic
ICE Publishing: All rights reserved
interfaces
Yang, Hwang and Kwon

Shear friction strength of smooth

construction joints and monolithic interfaces
Keun-Hyeok Yang Seung-Jun Kwon
Department of Plant & Architectural Engineering, Kyonggi University, Department of Civil & Environmental Engineering, Hannam University,
Suwon, South Korea (corresponding author: [email protected]) Daejeon, South Korea
Yong-Ha Hwang
Department of Architectural Engineering, Kyonggi University, Suwon,
South Korea

This study examined the shear friction behavior of smooth construction joints and monolithic interfaces according to
the equivalent compressive stresses generated from clamping forces of the transverse reinforcement crossing the
joint and external forces applied normal to the joint. Twenty-one push-off specimens were tested to measure the
relationship between direct shear stress and relative slip along the interfacial failure plane. Mechanical models for
monolithic joints proposed by Hwang and Yang are extended to estimate the shear friction strength of the
construction joints. From the mechanical model and test results, the cohesion and frictional angle of concrete along
smooth construction joints are determined to be 0·11( fc′)0·65 and 32·8°, respectively, where fc′ is the concrete
compressive strength. The construction joints have considerably lower shear capacities than the companion
monolithic joints, by approximately 70% for shear cracking strength and 55–70% for shear friction strength,
indicating that the reduction ratios decrease with the increase of the equivalent compressive stress. The reliability of
the previous empirical equations for the shear friction strength is highly sensitive to the roughness of the interfacial
failure plane, whereas the proposed model provides improved accuracy in predicting the shear friction strength of
the monolithic and smooth construction joints.

Notation νc effectiveness factor for concrete

Ac section area of interfacial failure plane compressive strength
Avf area of transverse reinforcement crossing the νt effectiveness factor for concrete tensile strength
interfacial failure plane ρc unit weight of concrete
c0 reference aggregate size (= 25 mm) ρo reference unit weight of concrete (= 2300 kg/m3)
c′ cohesion of concrete along construction joints ρvf transverse reinforcement ratio
da maximum size of aggregate σeq equivalent normal compressive stress (= ρvf fy + σx)
fco reference concrete compressive strength (= 10 MPa) σx axial stress applied normally to the interfacial
fc′ compressive strength of concrete failure plane
fc * effective compressive strength of concrete τcr initial shear cracking strength
ft tensile strength of concrete τn shear friction strength
f*t effective tensile strength of concrete ϕ friction angle of concrete along monolithic interface
fy yield strength of transverse reinforcement ϕ′ friction angle of concrete along construction joint
S0 relative slip at τn
Vn direct shear force at the interfacial failure plane
XIC horizontal coordinate of instantaneous centre Introduction
α angle between relative displacement at chord The design perspective associated with shear connections
midpoint and failure plane in structural systems is commonly governed by the concrete
γ ratio of test result to prediction interface found in details between columns and corbels, squat
γm mean of γ value shear walls and foundations, in dapped end beams and
γs standard deviation of γ value shear keys in precast members (ACI–ASCE Committee 426,
γv coefficient of variation of γ value 1973). The shear connections must be designed and detailed
δ relative displacement implicitly so that the applied shear forces to be resisted by the
ηcr normalised shear cracking strength (= τcr/( f c′ )1/2) connection can be transferred to the element and further on to
ηn normalised shear friction strength (= τn/( fc′)1/2) the overall load-resisting system. Shear friction is generally
θs inclination of transverse reinforcement relative to known to be a major mechanism of load transfer along
the interfacial failure plane the concrete–to–concrete interface subjected to simultaneous
μ coefficient of friction shear and normal compression. Most investigations (Haskett

1
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

et al., 2010; Kahn and Mitchell, 2002; Loov and Patnaik, to consider the reduced cohesion and frictional angle of
1994; Shaikh, 1978; Walraven et al., 1987; Yang et al., 2012) concrete along construction joints, and then those values
estimating the load transfer by the shear friction have been were generalised using test data. The reliability and limitation
conducted for a monolithic concrete interface. Harries et al. of previous empirical equations – the American Association
(2012) pointed out that the mechanism of shear friction is of State Highway and Transportation Officials (Aashto)
complex and the coefficient of friction is conceptual in explain- provision (Aashto, 2014) and the equation proposed by
ing the shear friction behaviour of concrete. The concrete Mattock (2001) – were also examined according to the vari-
interface frequently also has a construction joint between the ation of roughness along the interfacial failure plane through
precast element and the cast-in-place concrete or at the surface comparisons of measured and predicted shear friction
formed by placing one layer of concrete on an existing layer of capacities.
hardened concrete. When shear is applied to construction
joints, no aggregate interlock or tensile resistance are expected, Research significance
which leads to a considerable relative slip of the layers This study revealed that the cohesion and frictional angle of
accompanied by separation of the surfaces. Mattock (2001) concrete along a smooth construction joint are empirically
also pointed out that, compared with the shear transfer devel- generalised to be 0·11( fc′ )0·65 and 32·8°, respectively, where fc′
oped along monolithic interfaces, shear transfer by concrete is the concrete compressive strength. It was also ascertained
cohesion is considerably reduced in smooth construction joints. that the rate of increase in the shear friction strength with
ACI 318-14 (ACI, 2014) specifies a coefficient of friction (μ) respect to the equivalent normal compressive stresses is lower
for smooth construction joints that is approximately 60% lower for smooth construction joints than for monolithic interfaces.
than that for monolithic concrete interfaces. However, the The previously proposed empirical equations were found to
shear friction response of construction joints is still insuffi- have a severe discrepancy between applications for smooth
ciently explained with regard to the reduced cohesion and fric- construction joints because they were formulated under incor-
tional coefficient of concrete, particularly for high-strength rect limit states along the joints.
concrete. Furthermore, evaluations of the safety of code
equations (Aashto, 2014; ACI, 2014) for shear friction resist-
Experimental programme
ance are scarce for construction joints because of very little, if
any (Dahl, 1994), available test data. Test specimens
For this experimental programme, 21 push-off specimens
Kwon et al. (2017) and Yang et al. (2012) derived numerical were prepared with different values of roughness along the
approaches to evaluate the shear friction parameters associated interface, concrete compressive strength ( f c′ ), amount (Avf )
with the shear strength of a monolithic interface using the and configuration inclination (θs) of the transverse reinforce-
upper-bound theorem of concrete plasticity. They showed that ment crossing the interfacial plane and additional compres-
the angle of friction at monolithic concrete interfaces decreases sive stresses (σx) applied normally to the interface. Details of
as the compressive strength of concrete increases and the unit the specimens are provided in Table 1. The 21 specimens were
weight of concrete decreases, which indicates that the angle of classified into two groups: group M for monolithic interfaces
friction is not constant but variable depending on the brittle- and group C for construction joints. For each group, the
ness of materials. The variability of the frictional angle also design compressive strength of concrete was selected to be
affects the shear transfer by normal compressive stresses gener- 24 MPa (L-series) or 60 MPa (H-series). The applied com-
ated from the transverse reinforcement crossing the interfaces pressive stress was varied as 0, 0·15f c′ (≈ 9·4 MPa) and 0·3f c′
and forces applied normal to the interfaces. On the other (≈ 18·8 MPa) for the H-series concrete specimens and 0 and
hand, Nielsen and Hoang (2011) proposed that construction 0·15fc′ (≈ 4·5 MPa) for the L-series. For transverse reinforce-
joints have a constant cohesion and frictional angle of con- ments crossing the concrete interface, four deformed bars of
crete. However, further experimental evidence is required to diameter 10 mm were arranged at a 90° inclination (V-type
explain the effect of transverse reinforcement and additional bar) and a 45° inclination (X-type) to the interface, as shown
applied axial forces on the shear friction strength and frictional in Figure 1. Compared with the V-type bars, X-type bars at
angle at construction joints. the same inclination would provide less compressive force on
the shear plane but a greater force component parallel to the
The objective of this study was to evaluate the variation of the shear plane.
shear friction response of concrete interfaces according to the
existence of smooth construction joints with no special treat- The specimen notation shown in Table 1 was designated using
ment. In total, 21 push-off specimens (12 interfaces with con- the selected test parameters. The first and second letters
struction joints and nine monolithic interfaces) were prepared identify the specimen group according to the type of interface
to measure the relationship between direct shear stress and (C or M) and concrete compressive strength (H or L).
relative slip along the interfacial failure plane. The shear fric- The third letter refers to the configuration type of transverse
tion model developed for monolithic interfaces was extended reinforcement – N for non-reinforcing bars, V for θs = 90° bars

2
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research
Table 1. Details of push-off test specimens and summary of test results
τn: MPa (τn)Exp/(τn)Pre
Transverse reinforcement
Prediction

Specimen f′c: MPa Typea Amount ρvf ρvf fy: MPa σx: MPa σeq: MPa τcr: MPa S0: mm Exp. (1) (2)c (3) (4) (1)/(2) (1)/(3) (1)/(4)

Smooth construction joint

CHN-0 62·5 N — 0·0000 0·00 0 0 1·62 — 1·62 0·52 0·00 1·62 3·12 — 1·00
CHV-0 62·5 V 4 110 0·0090 4·25 0 4·3 1·63 0·21 4·32 3·07 2·55 4·36 1·41 1·70 0·99
CHX-0 62·5 X 4 110 0·0064 3·01 0 3·0 1·63 0·29 4·65 2·32 1·80 5·11 2·00 2·58 0·91
CHN-0·15 62·5 N — 0·0000 0·00 9·4 (0·15f c′ ) 9·4 4·12 0·50 6·69 5·50 — 7·66 1·22 — 0·87
CHV-0·15 62·5 V 4 110 0·0090 4·25 9·4 (0·15f c′ ) 13·6 3·82 0·54 9·27 5·50 2·55 10·40 1·69 3·64 0·89
CHX-0·15 62·5 X 4 110 0·0064 3·01 9·4 (0·15f c′ ) 12·4 3·78 0·38 11·46 5·50 1·80 11·15 2·08 6·35 1·03
CHN-0·3 62·5 N — 0·0000 0·00 18·8 (0·3f c′ ) 18·8 6·49 0·51 13·49 5·50 — 13·70 2·45 — 0·98
CHV-0·3b 62·5 V 4 110 0·0090 4·25 18·8 (0·3f c′ ) 23·0 — — — 5·50 2·55 16·44 — — —
CHX-0·3 62·5 X 4 110 0·0064 3·01 18·8 (0·3f c′ ) 21·8 7·21 0·51 18·36 5·50 1·80 17·19 3·34 10·18 1·07
CLN-0·15 29·8 N — 0·0000 0·00 4·5 (0·15f c′ ) 4·5 3·71 0·33 4·94 3·20 — 3·88 1·54 — 1·27
CLV-0·15 29·8 V 4 110 0·0090 4·25 4·5 (0·15f c′ ) 8·8 2·94 0·39 7·16 5·50 2·55 6·62 1·30 2·81 1·08
CLX-0·15 29·8 X 4 110 0·0064 3·01 4·5 (0·15f c′ ) 7·5 3·80 0·43 8·00 5·01 1·80 7·37 1·60 4·44 1·09

Yang, Hwang and Kwon

interfaces
construction joints and monolithic
Shear friction strength of smooth
Monolithic interface
MHN-0 62·5 N — 0·0000 0·00 0 0 4·83 0·11 5·75 2·8 0·0 6·7 2·09 — 0·85
MHV-0 62·5 V 4 110 0·0090 4·25 0 4·2 5·26 0·50 10·81 8·7 8·9 13·3 1·24 1·21 0·82
MHX-0 62·5 X 4 110 0·0064 3·01 0 3·0 4·94 0·37 12·88 8·7 8·9 14·9 1·48 1·45 0·87
MHN-0·15 62·5 N — 0·0000 0·00 9·4 (0·15fc′ ) 9·4 12·14 0·62 16·40 10·3 13·0 16·9 1·59 1·26 0·97
MHV-0·15 62·5 V 4 110 0·0090 4·25 9·4 (0·15fc′ ) 13·6 12·30 0·94 21·10 10·3 16·4 18·3 2·04 1·29 1·15
MHX-0·15 62·5 X 4 110 0·0064 3·01 9·4 (0·15fc′ ) 12·4 12·55 0·86 23·01 10·3 16·4 21·0 2·23 1·40 1·10
MLN-0·15 29·8 N — 0·0000 0·00 4·5 (0·15fc′ ) 4·5 8·58 0·58 11·54 7·5 6·6 8·4 1·55 1·76 1·38
MLV-0·15 29·8 V 4 110 0·0090 4·25 4·5 (0·15fc′ ) 8·7 9·02 0·75 14·63 7·5 8·9 9·2 1·96 1·64 1·58
MLX-0·15 29·8 X 4 110 0·0064 3·01 4·5 (0·15fc′ ) 7·5 8·87 0·76 14·24 7·5 8·9 12·2 1·91 1·59 1·17

Notes. f c′ is the concrete compressive strength (for construction joints, measured from the second casting on the other half of the push-off specimens), ρvf(= Avfsinθs/Ac), Avf and fy are, respectively, the ratio, total area and
yield strength of transverse reinforcement crossing the interface, Ac is the section area of the interface, σxis the externally applied compressive stresses normal to the interface, σeq (= ρvf fy + σx) is the equivalent compressive
stresses at the interface, τcr is the initial shear cracking strength at the interface, τn is the ultimate shear friction strength and S0 is the relative slip amount corresponding to τn
a
V, θs = 90°; X, θs = 45° (θs = configuration inclination of transverse reinforcement to the interface)
b
No test data were saved for specimen CHV-0·3
c
Predictions calculated using the equations of Aashto (2014), Mattock (2001) and the present study are given in columns (2), (3) and (4), respectively
3

Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

0
15
Ø16

Ø6

275
None

30
4Ø10

210
820
Testing zone

30
Construction joint
V-type

275
4Ø10

300

Monolithic interface Interface with construction joint

X-type
(a) (b)

Figure 1. Push-off specimen details and arrangement of reinforcement: (a) specimen details; (b) transverse reinforcement at the
interface (dimensions in mm)

and X for θs = 45° bars. The value following the letters indi- Materials
cates the σx/fc′ ratio. For example, specimen CHV-0·15 indicates Ordinary Portland cement was used to produce the concrete.
a construction joint made with a design compressive strength Locally available natural sand with a maximum particle size of
of concrete of 60 MPa (H-series), reinforced using four 10 mm 5 mm and crushed granite with a maximum particle size of
diameter bars perpendicularly crossing the interface, and sub- 25 mm were used as fine and coarse aggregates, respectively.
jected to an additional axial compressive stress of 0·15fc′. To achieve the initial concrete slump of 180 mm, a polycarbox-
Specimen MHV-0·15 is a monolithic interface with the same ylate-based high-range water-reducing admixture was added
parameters as the CHV-0·15. during mixing of the concrete ingredients. Control cylinders
150 mm in diameter and 300 mm high were cast and cured
The push-off specimens were designed and cast in a manner simultaneously with the push-off specimens to determine fc′.
similar to those used by previous researchers (Ahmed and The average values of the compressive strength measured from
Ansell, 2010; Mattock, 1976) to simulate the concrete interface testing the control cylinders were 62·5 MPa and 29·8 MPa for
subjected to direct shear stresses without flexural effects. The the H-series and L-series concrete mixtures, respectively. The
geometrical dimensions and reinforcement details of push-off concrete compressive strengths for the first casting of one half
specimens are shown in Figure 1. The outer dimensions of of the push-off specimens with construction joints were
total length, width and overall depth were 820 mm, 300 mm 61·2 MPa and 27·5 MPa for H-series and L-series, respectively.
and 150 mm, respectively. The length of the concrete interface The yield and tensile strengths of 10 mm steel bars used as the
under direct shear was 210 mm, producing a section area (Ac) transverse reinforcement were 472 MPa and 574 MPa,
of 31 500 mm2 for the shear plane under shear frictional resist- respectively.
ance. Nine specimens with a monolithic interface were cast
integrally, whereas the others were cast separately along the
shear interfacial plane to produce a smooth construction joint Test procedure and instrumentation
with only a steel-brushing treatment. The second casting was The test procedure was planned with reference to the exper-
made 3 d after the first casting on the scale of the other half. imental setup investigated by Mattock (1976). All push-off
All specimens were cured at room temperature and demoulded specimens were placed in a vertical position in a steel test
at an age of 3 d. Four reinforcing bars were symmetrically frame and loaded concentrically in line with the interfacial
arranged at the concrete interface. To lead crack propagation shear plane using a 1000 kN capacity hydraulic jack, as shown
and failure along the concrete interface, the top and bottom in Figure 2, to simulate direct shear along the joint without
stubs were reinforced using deformed bars of diameter 16 mm flexure. The applied shear loads were recorded by a load cell
and then externally strengthened using carbon fibre sheets attached to the hydraulic jack. Rod-shaped steel hinges were
before testing, as shown in Figures 1 and 2. positioned at the top and bottom surfaces of the specimen.

4
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

reinforced with X-type transverse reinforcement showed several

cracks along the reinforcing bars.
Oil jack

1000 kN Shear stress–relative slip relationship

load cell V
Figure 4 shows typical shear stress–relative slip curves
Hinge measured along the interfacial shear plane of the H-series con-
Strengthened by crete specimens. No data were obtained for specimen CHV-0·3
Specimen carbon sheet
Steel rod to apply
due to a malfunction of the data logger. The shear stresses
external compressive stresses were calculated by dividing the applied shearing force by the
Expected failure plane area of the interfacial shear plane of the test zone. With the
Torque occurrence of a crack, the non-reinforced specimens failed
immediately, regardless of the existence of a construction joint;
LVDT frame
σx σx the relative slip after cracking for the non-reinforced specimens
was thus not measured. The reinforced specimens began to
reveal slip displacement along the interfacial shear plane with
200 kN LVDT (25 mm)
load cell
the occurrence of cracks. The shape of the shear stress–relative
Strengthened by slip curves was significantly affected by the existence of the
carbon sheet
construction joint, especially at the descending branch of the
curve after the peak stress.
Hinge

The reinforced construction joint showed a shear stress–relative

V slip curve that included a typical stress flow phenomenon, indi-
cating that the relative slip after the peak stress rapidly
Figure 2. Test setup increased without a noticeable drop of the applied shear stress.
This may be attributed to the fact that the applied shear after
the peak stress was mostly resisted by the dowel action
of the transverse reinforcement. Overall, the relative slip after
Immediately before applying the direct shear, constant axial
the peak stress mostly depended on the amount of transverse
compressive forces normal to the joint were introduced by
reinforcement (fib, 2008; Xu et al., 2015). The slope at the
means of tensile force in six high-tension bolts due to tighten-
ascending branch of the shear stress–relative slip curves was
ing of the nuts, which could be controlled by the externally
insignificantly affected by θs. However, the slope tended to
applied torque. To control and measure the applied axial
increase as σx increased, resulting in a smaller amount of rela-
forces, 200 kN capacity load cells were installed between the
tive slip. After the peak stress, the increase in relative slip was
nuts and the loading plate. Relative slip movement across the
independent of θs and the magnitude of σx.
joint was measured by 25 mm capacity linear variable differen-
tial transducers (LVDTs) mounted at the centre of the joint on
The reinforced monolithic specimens had a greater increase in
both the front and rear of the specimen. The test was termi-
shear stress up to the peak after initial cracking, and a greater
nated when either severe failure was observed or the applied
stress drop after the peak stress compared with the companion
load dropped suddenly.
construction joint specimens. After the peak, the applied shear
stress dropped sharply up to a level approximately correspond-
ing to the initial shear cracking stress and then remained con-
Test results and discussion stant, producing a stress flow phenomenon. This observation
Typical crack propagation patterns and failure planes of the was independent of θs and σx.
high-strength concrete (H-series) specimens are presented in
Figure 3. Cracks were concentrated on the interfacial shear In general, shear friction resistance at non-adhesive interfaces
plane of the test zone. Thus, all test specimens failed by direct between two materials is calculated as the product of the value
shear along the interfacial shear plane, resulting in separation of μ at the interface and applied axial forces normal to the
into two blocks. The propagation of cracks was not signifi- interface. The elongation of reinforcing bars crossing the inter-
cantly affected by the magnitude of applied axial stresses facial shear plane provides compression forces on the interface
normal to the joint. At failure, greater crack widths along the (Ali and White, 1999). Assuming that the separation of the
interfacial shear plane were observed for non-reinforced speci- shear crack faces at the ultimate state is sufficient to allow the
mens than for reinforced specimens. Most specimens showed stress of the transverse reinforcement to reach its yield point, a
similar crack propagation at the test zone, with a major linear clamping stress provided by the transverse reinforcement can
crack along the interface, whereas the monolithic specimens be expressed as ρvf fy, where ρvf and fy are the ratio and yield

5
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

(a) (b)

(c) (d)

(e) (f)

Figure 3. Typical crack propagation and failure plane: (a) non-reinforced construction joint; (b) construction joint with V-type
reinforcement; (c) construction joint with X-type reinforcement; (d) non-reinforced monolithic interface; (e) monolithic interface with
V-type reinforcement; (f) monolithic interface with X-type reinforcement

6
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

30 1·0
Smooth construction joint
CHN-0 CHN-0·15 0·9
Monolithic interface
25 CHV-0 CHV-0·15 0·8
CHX-0 CHN-0·15 0·7
Best-fit line for
Shear stress: MPa

20 0·6 monolithic interface

S0: mm
0·5
15 0·4
0·3
Best-fit line for smooth
10 0·2 construction joint
0·1 fib equation for smooth
construction joint
5 0
0 0·5 1·0 1·5 2·0 2·5 3·0 3·5 4·0 4·5 5·0

0 (ρvffy + σx)1/2: MPa

0 1 2 3 4 5 6 7 8 9 10
Relative slip: mm Figure 5. Relative slip at peak shear stress
(a)

30
MHN-0 MHN-0·15 Shear cracking strength
25 MHV-0 MHV-0·15 Figure 6 shows the effect of σx on the normalised shear crack-
MHX-0 MHX-0·15 ing strength (ηcr = τcr/( fc′ )1/2). The shear cracking strength was
Shear stress: MPa

20 found to be independent of Avf and θs, as shown in Figure 4,

implying that the applied shear stresses before the occurrence
15 of a crack are mostly resisted by the cohesion and tensile
strength of the concrete along the interface. Thus, the shear
10
cracking stress at the interfaces was significantly affected by fc′
and σx. The shear cracking strength of construction joints was
5
as much as 70% lower than that of the companion monolithic
0 interfaces. The value of ηcr increased in proportion to the
0 1 2 3 4 5 6 7 8 9 10 increase in σx, indicating that the rate of increasing ηcr was
Relative slip: mm greater for monolithic interfaces than for construction joints.
(b) This is because the applied axial compressive stresses reduced
the principal tensile stresses along the shear plane, and the
Figure 4. Typical shear stress–relative slip relationship (h-series
concrete): (a) construction joints (b) monolithic interfaces tensile resistance of concrete along the construction joint was
extremely low. From regression analysis using the obtained test
data, τcr at the interface can be simply formulated as (Figure 6)
strength of the transverse reinforcement crossing the interface,
σx pffiffiffiffiffi
respectively. Thus, the compressive stresses at the interfaces 1: τ cr ¼ A1 pffiffiffiffiffi þ B1 f c0
developed from the transverse reinforcement and externally f c0
applied axial forces are equivalent to the summation of
ρvf fy + σx, which is then defined to be the equivalent normal
compressive stress (σeq in MPa). The International Federation
for Structural Concrete (fib) guide for structural connections
3·0
(fib, 2008) notes that the amount of relative slip (S0) at peak Smooth construction joint (H-series concrete)
shear stress is significantly affected by σeq. According to the 2·5 Smooth construction joint (L-series concrete)
fib guide, S0 (in mm) for a smooth construction joint can be Monolithic interface (H-series concrete)
estimated as 0·15(σeq)1/2. Figure 5 shows the relationship 2·0
Monolithic interface (L-series concrete)
τcr /(f 'c)1/2

between S0 and σeq. The value of S0 tends to increase linearly 1·5

with the increase in (σeq)1/2, showing a slightly greater rate of
Best-fit line for
increase of S0 for monolithic interfaces than for smooth con- 1·0 monolithic interface
struction joints. At the same value of (σeq)1/2, a smooth con-
0·5
struction joint has approximately 35–50% lower S0 than Best-fit line for smooth
monolithic interfaces because the existence of construction construction joint
0
0 0·05 0·10 0·15 0·20 0·25 0·30 0·35
joints significantly reduces the shear friction strength of inter-
σx /f c'
faces, as shown in Table 1. The best-fit line determined from
the smooth construction joints is close to the estimation of the
Figure 6. Shear cracking strength at the interface
fib guide (fib, 2008).

7
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

where the experimental constants A1 and B1 can be set to be to the existence of the construction joint. At a construction
2·2 and 0·22, respectively, for construction joints and 6·4 and joint, concrete cohesion significantly decreases and no aggre-
0·63, respectively, for monolithic interfaces. gate interlock is expected (Nielsen and Hoang, 2011). The
ACI 318-14 (ACI, 2014) provision specifies the coefficient of
friction (μ) as 1·4 for concrete placed monolithically and 0·6
Shear friction strength for concrete against hardened concrete not intentionally rough-
Figure 7 shows the effect of σx on the shear friction strength ened, indicating that the shear friction resistance decreases by
(τn) of H-series concrete interfaces. Note that τn measured from as much as 60% due to the existence of a smooth construction
non-reinforced specimens was approximately identical to their joint. The reduction ratios of τn due to the existence of smooth
τcr values, as they failed immediately with the occurrence of a construction joints measured in the reinforced specimens were
crack at the interface. The shear friction strength tended to close to the those calculated from the μ values specified in ACI
increase in proportion to σx. The slope of the increasing rate in 318-14 (ACI, 2014).
the τn against σx curve was insignificantly affected by the
amount of transverse reinforcement, whereas a higher slope The effect of σeq on the normalised shear friction strength
was achieved for monolithic interfaces than for construction (ηn = τn/( fc′ )1/2) is shown in Figure 8. On the same figure, pre-
joints. At the same amount (four 10 mm diameter bars) of dictions calculated from the Aashto provision (Aashto, 2014)
transverse reinforcement, the X-type bars (θs = 45°) were more for H-series concrete are also plotted for comparison. The
effective than the V-type bars (θs = 90°) in enhancing τn of the Aashto equations consider τn as a summation of concrete
interfaces. Compared with τn of the companion specimens cohesion and frictional resistance provided by σeq. Based on
reinforced with V-type bars subjected to σx = 0·15fc′, the speci- previous test data, the Aashto equations set the concrete cohe-
mens reinforced with X-type bars had a higher value, by sion and μ as 2·76 MPa and 1·4, respectively, for monolithic
1·24 times, for construction joints and 1·09 times for mono- interfaces and 0·52 MPa and 0·6, respectively, for smooth con-
lithic interfaces. The ACI 318-14 equation (ACI, 2014) ignores struction joints. Furthermore, the Aashto equations do not
concrete cohesion and assumes that the applied shear force is allow τn to exceed a certain limit, which depends on fc′. As a
entirely transferred by the friction of transverse reinforcement, result, the Aashto equations give a constant value of τn when
which varies according to the inclination of the transverse bars σeq exceeds 5·5 MPa for monolithic interfaces and 9·0 MPa for
to the interface. According to the code equation, the highest construction joints with H-series concrete. On the other hand,
shear transfer capacity of transverse reinforcement is calculated the shear friction strength measured from the test specimens
at θs = 45°. This result is similar to the trend observed in the increased in proportion to σeq, showing a higher increasing
present tests. rate for monolithic interfaces than for construction joints.
Unlike the Aashto equations, no definite limit points were
Compared with τn of the companion monolithic interfaces, the observed in the test data. However, the relationship between
non-reinforced construction joint had a lower value by 72% for σeq and ηn has some deviations from the fitting line because
σx = 0 and 60% for σx = 0·15fc′, and the values for the the relationship does not account for the effect of θs and con-
reinforced construction joint were lower by 60% and 55%, crete cohesion on the shear friction action along the interfacial
respectively, under such values of σx (see also Table 1). The shear plane.
externally applied compressive stresses and the arrangement of
transverse reinforcements alleviated the reduction of τn owing

5·0
25 Smooth construction joint (H-series concrete)
Smooth Smooth construction joint (L-series concrete)
construction joints 4·0 Monolithic interface (H-series concrete)
20 Monolithic
Monolithic interface (L-series concrete) Best-fit line for
interfaces monolithic
15 3·0 Best-fit line interface
τn: MPa

ηn

for smooth
construction joint
10 2·0

Non-reinforced
5 1·0
Reinforced with V-type bar Aashto for monolithic interface (H-series)
Reinforced with X-type bar Aashto for smooth construction joint (H-series)
0
0 0·05 0·10 0·15 0·20 0·25 0·30 0·35 0
0 2·5 5·0 7·5 10·0 12·5 15·0 17·5 20·0 22·5 25·0
σx = 9·4 MPa σx = 18·8 MPa σeq
σx /f 'c
Figure 8. Relationship between σeq(= ρvf fy + σx) and
Figure 7. Shear friction strength of the interfaces ηn (= τn/( fc′ )1/2)

8
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

Prediction models based on the upper-bound is the reference value for the unit weight (ρc) of concrete and
theorem c0 (= 25 mm) is the reference value for the maximum aggregate
size (da).
Shear friction strength of monolithic interfaces
Kwon et al. (2017) and Yang et al. (2012) derived an integrated
model for the shear friction strength of monolithic concrete Shear friction strength of construction joints
interfaces on the basis of the upper-bound theorem of concrete The failure mechanism of interfaces governed by shear friction
plasticity. The model accounted for the effects of applied axial is independent of the existence of the construction joint.
stresses and transverse reinforcement on the shear friction Hence, the concrete interfaces under direct shear and axial
action at interfacial shear cracks. Concrete was assumed as a loads are usually separated into two rigid blocks at failure, as
rigid perfectly plastic material obeying a modified Coulomb shown in Figure 3. One rigid block has two translational and
failure criterion. The friction angle (ϕ) of concrete was formu- rotational displacement components relative to the other rigid
lated as a function of the ratios of effective tensile and com- block. Thus, one rigid block can be assumed to be rotating
pressive strengths of concrete. From comparisons with previous about an instantaneous centre (IC), as schematically idealised
empirical equations and 103 push-off test specimens, the pro- in Figure 9. Similar to the failure mechanism of monolithic
posed model was verified to have improved accuracy in predict- interfaces, the shear plane at failure along the construction
ing the shear friction strength and consistent trends with test joint can be regarded as a plane strain problem. On the other
results in evaluating the effect of various parameters on the hand, the cohesion (c′) and friction angle (ϕ′) of concrete
shear friction strength. The model for a monolithic interface along construction joints are lower than those of monolithic
can be summarised as interfaces. Nielsen and Hoang (2011) demonstrated that the
relative displacement (δ) about the IC has a constant angle (α)
Vn 1 1 to the construction joint, which can be assumed to be equal
τn ¼ ¼ fc* ðl  m sin αÞ to ϕ′. Furthermore, no tensile resistance of concrete at the con-
Ac 2 cos α
2: struction joints can commonly be expected, implying that the
½cosðθs  αÞ
þ ρvf fy þ σ x tan α shear friction of concrete along the construction joint is
cos α entirely resisted by the concrete cohesion. Hence, neglecting
the tensile strength ( ft) of concrete, Equation 7 is obtained
from Equation 2 to estimate the shear friction strength of con-
" struction joints.
 09  16 #
f c0 ρo
3: fc* ¼ νc f c0 ¼ 079exp 003 f c0
fco ρc 1
τ n ¼ fc*
1
ð1  sin ϕ0 Þ
2 cos ϕ0
7:
½cosðθs  ϕ0 Þ
þ ρvf fy þ σ x tan ϕ0
cos ϕ0
  
1 2ðρvf fy sin θs þ σ x Þ
4: α ¼ sin1 m
l fc*

τ ( = V/Ac)

 021
ft* Rigid
5: ϕ ¼ 2065 block I
fc* IC
(XIC, YIC)
Failure plane along X IC ω
construction joint α
r = cos
θs

"     #043 σx σx
ft* f c0 c0 17 ρc 01 Transverse α = φ'
6: ¼ 0064
fc* fco da ρo reinforcing bar
δ

where l ¼ 1  2ð ft*=fc*Þðsin ϕ=1  sin ϕÞ, m ¼ 1  2ð ft*=fc*Þð1=1  sin ϕÞ, Y Rigid

block II
α is the angle between the relative displacement (δ) at the mid-
X
point of the yield line and the failure plane, fc* (= νc fc′ ) is
τ ( = V/Ac)
the effective compressive strength of concrete, ft* (= νt ft) is the
effective tensile strength of concrete, fco (= 10 MPa) is the Figure 9. Idealised failure mechanism of a construction joint
reference value for the compressive strength, ρo (= 2300 kg/m3)

9
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

Construction joints under sliding failure may be characterised The present study conservatively set the value of c′ for smooth
either by c′ and ϕ′ or fc* and ϕ′. Based on test results, Nielsen construction joints, although this value could be adjusted
and Hoang (2011) showed that the following relationship through further tests under extensive ranges of fc′. To determine
could be generalised for construction joints. the value of ϕ′ using the push-off specimens tested, Equation 9
is rearranged as
cos ϕ0
8: fc* ¼ 2c0 !
1  sin ϕ0 τ n  011ð f c0 Þ065  ρvf fy cos θs
11: ϕ0 ¼ tan1
ρvf fy sin θs þ σ x

Substituting Equation 8 into Equation 7, the shear friction

strength of construction joints can be expressed as
The empirically obtained ϕ′ values for the present specimens
cosðθs  ϕ0 Þ with smooth construction joints commonly ranged between
9: τ n ¼ c0 þ ρvf fy þ σ x tan ϕ0
cos ϕ0 28·4° and 36·5°. The average value of 32·8° (μ = 0·64) is very
close to the typical value (37°) suggested by Nielsen and
Hoang (2011). For ordinary construction joints, ϕ′ is com-
monly found to be of the order of 30·9° (μ = 0·6) and 38·6°
Equation 9 is an extended version of Equation 2 for mono-
(μ = 0·8).
lithic interfaces in terms of the following: the shear transfer
capacity of concrete is calculated using the value of c′ by
In summary, the shear friction strength of smooth construction
neglecting the tensile strength of concrete along the construc-
joints can be generalised as
tion joints; the variable angle (α) is replaced by the constant
frictional angle (ϕ′) of concrete. Note that the shear friction ½cosðθs  328°Þ
strength of non-reinforced construction joints without τ n ¼011ð f c0 Þ065 þ ρvf fy
12: cosð328°Þ
additional axial stresses is resisted only by the concrete cohe- þ σ x tanð328°Þ
sion along the interfacial failure plane. Aashto (2014) and
Mattock (2001) assumed concrete cohesion to be a constant
value, even for monolithic interfaces. Nielsen and Hoang
(2011) revealed that concrete cohesion depends on fc′ and the Comparisons with test results
roughness along the interfacial plane. From regression analysis Using the current experimental data, a direct comparison was
using four specimens, they proposed the value of c′ for smooth made between predictions obtained from the proposed mech-
construction joints as 0·55( fc′ )1/2. Experimental programmes to anism models and previous empirical equations proposed by
determine the value of c′ of non-reinforced construction joints Aashto (2014) and Mattock (2001). The results are listed in
are very scarce. Plotting the results of push-off specimens Table 1. Figure 11 shows a comparison of the measured and
obtained in the current study and those reported by Choi et al. predicted shear friction capacities of the push-off specimens.
(1994) to show c′ for smooth construction joints as a function Table 2 also provides the mean (γm), standard deviation (γs)
of fc′ (Figure 10), the approximate relationship obtained is and coefficient of variation (γv) of the ratios between exper-
iments and predictions (γ = (τn)Exp/(τn)Pre). Since the shear fric-
c0 ¼ 011ð f c0 Þ
065
10: tion strengths (τn) of monolithic interfaces are fully examined
with different prediction models in the companion paper
(Kwon et al., 2017), the comparison presented here is focused
on smooth construction joints.
Equation 10 gives a lower value of c′ than that calculated
using the equation proposed by Nielsen and Hoang (2011).
The Aashto equation (Aashto, 2014) gives an underestimation
2·0
of the results for monolithic interfaces, with a γm value of 1·78.
This underestimation tends to be greater for construction
1·5 joints, with a γm value of 1·98. Furthermore, greater deviations
were observed for construction joints than for monolithic inter-
c ' : MPa

1·0
faces, giving approximately twice the value of γs for construc-
0·5
Best-fit curve tion joints. When σeq exceeded 15·0 MPa, the underestimation
y = 0·11x 0·65 for construction joints increased considerably (Figure 11(a)).
R2 = 0·91
0 As pointed out by Harries et al. (2012), the Aashto equation
0 10 20 30 40 50 60 70
does not capture the mechanism of shear friction and specifies
f 'c: MPa
incorrect limit states. Furthermore, the underestimation of the
Figure 10. Relationship between f c′ and c′ shear transfer capacity by concrete cohesion is more notable
for smooth construction joints than for monolithic interfaces.

10
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

5 monolithic interfaces is independent of σeq (Figure 11(b)).

Smooth construction joint
However, the values of γm and γs increased considerably for
4 Monolithic interface
construction joints, producing large scatter. This may be attrib-
(τn)Exp/(τn)Pre

3 uted to the fact that the equation was derived by regression

analysis using limited test data without additional axial stresses
2
normal to the joint. Furthermore, Mattock’s equation for the
1 shear friction strength of construction joints focuses on con-
crete with an intentionally roughened surface.
0
0 5 10 15 20 25
(a) The empirical equations show severe discrepancies when
applied to construction joints. On the contrary, the predictions
5
Smooth construction joint from the model proposed in this study are in better agreement
4 Monolithic interface with the test results (Figure 11(c)), yielding values of γm and γs
(τn)Exp/(τn)Pre

of 1·10 and 0·26, respectively, for monolithic interfaces and

3
1·02 and 0·11, respectively, for construction joints. In addition,
2 the proposed model gives good accuracy for interfaces under
high σeq (exceeding 15·0 MPa). Overall, it can be said that the
1
friction angle and cohesion of concrete derived in this study
0 are practically acceptable for predicting the shear friction
0 5 10 15 20 25
(b) strength of monolithic interfaces and smooth construction
joints. The present approach needs to be extended for rough
5 construction joints.
Smooth construction joint
4 Monolithic interface
(τn)Exp/(τn)Pre

3 Conclusions
The effect of equivalent compressive stresses (σeq) on the shear
2
friction behaviour of monolithic interfaces and smooth con-
1 struction joints was examined by testing 21 push-off specimens.
0
Mechanical models based on the upper-bound theorem were
0 5 10 15 20 25 formulated to predict the shear friction strength of such joints.
(c) The test results and model analysis showed that the clamping
σeq
stress (ρvf fy) of transverse reinforcement can be considered to
be equivalent to the applied compressive stress (σx) in enhan-
Figure 11. Comparison of measured and predicted shear friction
strengths: (a) Aashto equation; (b) Mattock equation; (c) proposed cing the shear friction strength of the interfaces. From the
mechanical models experimental observations and establishment of prediction
models, the following conclusions may be drawn.

Table 2. Statistical comparisons of measured and predicted shear (a) Reinforced construction joints showed a shear
friction strengths stress–relative slip curve with a typical stress flow
phenomenon, indicating that the relative slip after the
Statistical Aashto Mattock This
Interface type value (2014) (2001) study peak stress rapidly increased without a noticeable drop in
the applied shear stress. Meanwhile, reinforced
Smooth construction γm 1·98 4·53 1·02 monolithic interfaces showed a sharp shear stress drop
joint γs 0·72 2·91 0·11
γv 0·36 0·64 0·11
after the peak stress up to the level corresponding to the
Monolithic interface γm 1·78 1·45 1·10 initial shear cracking stress, and then remained constant.
γs 0·33 0·20 0·26 These observations were independent of inclination (θs)
γv 0·18 0·14 0·24 of the transverse reinforcement crossing the interface and
Total γm 1·89 2·89 1·05 the magnitude of the externally applied compressive
γs 0·57 2·49 0·19
γv 0·30 0·86 0·18
stresses (σx).
(b) The amount of relative slip (S0) at the peak shear stress
tended to increase linearly with the increase in (σeq)1/2,
showing a slightly greater rate of increase in S0 for
The equation proposed by Mattock (2001) yields compara- monolithic interfaces than for smooth construction joints.
tively good accuracy for monolithic interfaces, with γm and γs (c) The shear cracking strength can be formulated as a
equal to 1·45 and 0·20, respectively. The γ value calculated for function of σx and concrete compressive strength ( fc′ ),

11
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research Shear friction strength of smooth
construction joints and monolithic
interfaces
Yang, Hwang and Kwon

regardless of the roughness of the interfacial failure Ahmed L and Ansell A (2010) Direct shear strength of high-strength
plane. The shear cracking strength of construction joints fibre concrete. Magazine of Concrete Research 62(5): 379–390,
http://dx.doi.org/10.1680/macr.2010.62.5.379.
was as much as 70% lower than that of the companion
Ali MA and White RN (1999) Enhanced contact model for shear friction
monolithic interfaces. of normal and high-strength concrete. ACI Structural Journal
(d) Compared with the shear friction strength of monolithic 96(3): 348–360.
interfaces, the non-reinforced construction joints had a Choi OC, Cho SS, Hong GS, Chung IY and Shin YS (1994) Interfacial shear
value lower by 72% for σx = 0 and 60% for σx = 0·15fc′, transfer characteristics of concrete joints. Journal of the
Architectural Institute of Korea 10(8): 89–96.
and the values of reinforced construction joints were
Dahl KKB (1994) Construction Joints in Normal and High
lower by 60% and 55%, respectively, under the same Strength Concrete. Department of Structural Engineering,
values of σx. With an increase in σeq, the reduction of Technical University of Denmark, Lyngby, Denmark,
shear friction strength due to the existence of the Report R-314.
construction joint decreased. fib (International Federation for Structural Concrete) (2008) Structural
Connections for Precast Concrete Buildings. Guide to Good
(e) The cohesion and friction angle of concrete at smooth
Practice. fib, Lausanne, Switzerland, Bulletin 43.
construction joints were proposed to be 0·11( fc′ )0·65 and Harries KA, Zeno G and Shahrooz B (2012) Toward an improved
32·8°, respectively. understanding of shear-friction behaviour. ACI Structural Journal
(f ) The reliability of previous empirical equations for the 109(6): 835–844.
shear friction strength was found to be highly sensitive to Haskett M, Oehlers DJ, Mohamed Ali MS and Sharmat SK (2010)
The shear friction aggregate interlock resistance across sliding
the roughness of interfaces. The equation proposed by
planes in concrete. Magazine of Concrete Research 62(12):
Mattock (2001) showed considerably larger deviations for 907–924, http://dx.doi.org/10.1680/macr.2010.62.12.907.
smooth construction joints than monolithic interfaces. Kahn LF and Mitchell AD (2002) Shear friction tests with high-strength
(g) The model proposed in this study was found to provide concrete. ACI Structural Journal 99(3): 98–103.
superior accuracy in predicting the shear friction strength Kwon SJ, Yang KH, Hwang YH and Ashour AF (2017) Shear
friction strength of monolithic concrete interfaces. Magazine
of monolithic interfaces and smooth construction joints,
of Concrete Research 69(5): 230–244, http://dx.doi.org/
with the mean and standard deviation of ratios between 10.1680/jmacr.16.00190.
experimental results and predictions of 1·10 and 0·26, Loov RE and Patnaik AK (1994) Horizontal shear strength of
respectively, for monolithic interfaces and 1·02 and 0·11, composite concrete beams with a rough interface. PCI Journal
respectively, for smooth construction joints. 39(1): 48–69.
Mattock AH (1976) Shear Transfer Under Monotonic Loading, Across
an Interface between Concrete Cast at Different Times. University
Acknowledgement of Washington, Seattle, WA, USA, Report SM76–3.
Mattock AH (2001) Shear friction and high-strength concrete. ACI
This work was supported by a National Research Foundation Structural Journal 98(1): 50–59.
of Korea (NRF) grant funded by the Korea Government Nielsen MP and Hoang LC (2011) Limit Analysis and Concrete Plasticity,
(MSIP) (no. NRF-2014R1A2A2A09054557). 3rd edn. CRC Press, Boca Raton, FL, USA,
Shaikh AF (1978) Proposed revisions to shear-friction provisions. PCI
Journal 23(2): 12–21.
Walraven JC, Frenay J and Pruijssers A (1987) Influence of concrete
REFERENCES strength and load history on the shear friction capacity of concrete
Aashto (American Association of State Highway and Transportation members. PCI Journal 32(1): 66–84.
Officials) (2014) Aashto LRFD Bridge Design Specifications, Xu J, Wu C, Li ZX and Ng CT (2015) Numerical analysis of shear transfer
7th edn. Aashto, Washington, DC, USA. across an initially uncrack reinforced concrete member.
ACI (American Concrete Institute) (2014) ACI 318-14: Building code Engineering Structures 102: 296–309, https://doi.org/10.1016/
requirements for structural concrete and commentary. ACI, j.engstruct.2015.08.022.
Farmington Hills, MI, USA. Yang KH, Sim JI, Kang JH and Ashour AF (2012) Shear capacity of
ACI–ASCE Committee 426 (1973) The shear strength of reinforced monolithic concrete joints without transverse reinforcement.
concrete members. Journal of the Structural Division ASCE 99(6): Magazine of Concrete Research 64(9): 767–779, http://
1091–1187. dx.doi.org/10.1680/macr.11.00107.

How can you contribute?

To discuss this paper, please submit up to 500 words to the
editor at [email protected]. Your contribution will be
forwarded to the author(s) for a reply and, if considered
appropriate by the editorial board, it will be published as a
discussion in a future issue of the journal.

12
Downloaded by [ University Of Newcastle] on [26/06/17]. Copyright © ICE Publishing, all rights reserved.