ITM Web of Conferences 15, 07012 (2017)

DOI: 10.1051/itmconf/20171507012
CMES’17

Selection of appropriate concrete model in numerical calculation

Krzysztof Śledziewski1,*
1
Lublin University of Technology, Faculty of Civil Engineering and Architecture, Department of Road and Bridges, Nadbystrzycka 40,
20-618 Lublin, Poland

Abstract. Modelling of composite and reinforced concrete structures requires very precise determination
of material parameters and constitutive relations between strain and stress. Erroneous selection of the
dependencies and incorrect modelling, in particular, of the performance of concrete in tension may generate
results in finite element method programs, that are far from the results obtained in an experiment. Using the
example of a concrete damage plasticity model, based on fracture mechanics, this paper describes the
physical interpretation and the method of the selection of parameters necessary for the appropriate
modelling of concrete in a complex stress state. The correctness of the assumed description of concrete was
verified on the basis of results of laboratory tests. A comparative analysis of the experimental and numerical
results showed that the application of the concrete damage plasticity model allowed correct determination of
the concrete element damage mechanisms for each level of strain.

1 Introduction media [6]. In the case of modelling a failure in concrete,

non-local models are also used [7], as well as the so-
The appropriate modelling of composite structures [1] called higher-order, or gradient, models [8]. The non-
and reinforced concrete structures [2] is a major local models introduce an averaging function which
challenge, in particular due to the material properties of converts local variables into non-local ones, according to
concrete. Unlike steel, concrete is a material that – in specific weights.
compression – behaves in a non-linear way from the In turn, the possibility of regularisation at the
very beginning. In addition, it undergoes degradation in numeric formulation level consists in the explicit
a much faster manner [3] when in tension. Hence, the introduction of the width of the plastic strain location
correct description of the concrete failure is the primary zone into a finite element [4]. In addition, an automatic
issue when considering the behaviour of concrete in refinement of finite element mesh is used, based on local
a structure with advanced strains. As a result of the errors caused by large gradients of internal variables, or
development of existing micro-cracks in concrete, the a displacement field discontinuity within the finite
material softens. On the grounds of mechanics of element is introduced [9]. This leads to the regularisation
continuous mediums, this phenomenon is defined as of the problem. All the above regularisation methods
damage during which the tangent stiffness matrix ceases bring some characteristic internal length scale, which
to be defined positively. In the case of concrete tension, provides information on the width of the strain location
due to the development of the damage, the tangent zone in the material model.
stiffness matrix is negatively defined, which leads to the
location of the strains. From the mathematical point of
view, a change occurs in a type of partial differential 2 Concrete damage plasticity model
equations, which govern the process, from elliptic to
hyperbolic. The problem is ill-conditioned, and 2.1 Mathematical description
consequently, the obtained solution depends on the
density of the applied discretisation [4]. The numerical analysis of the initial boundary value
The presented problem can be regularised via problem with the location of failure requires
different methods, among which one can distinguish two a comprehensive, constitutive modelling of the material.
main groups. The first group concerns the regularisation In this model, the scalar damage parameter is used for
at the level of a mathematical formulation, while the modelling of the failure of concrete separately in
other regularises the problem at the level of a numerical compression and tension. The scalar description of the
formulation. The first group includes, above all, material damage [10] was initially used to describe the
viscoplastic models in which the stress state depends on phenomenon of creep. The primary issue in the
the rate of the strain [5]. They are used mainly in numerical analysis of concrete structures is to define the
dynamics. This group also includes the Cosserat and correct mechanism of structure failure and its load
micro-polar models which are used for soils and granular carrying capacity. This requires the identification of
*
Corresponding author: [email protected]
© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution
License 4.0 (http://creativecommons.org/licenses/by/4.0/).
ITM Web of Conferences 15, 07012 (2017) DOI: 10.1051/itmconf/20171507012
CMES’17

constitutive parameters of the model based on specific 2.2 Strength hypothesis and its parameters
laboratory experiments [11].
The constitutive equation [10] of the material with The failure of concrete can be divided into two types.
scalar damage is as follows: The first type refers to cracking in tension, which is
associated with the occurrence of a dominant crack and
(1  d ) D0 : (   ) 
el pl
D : (   )
el pl
(1) the decline in load-carrying capacity in the direction
perpendicular to it [12], while the other type is related to
where: σ – the Cauchy stress tensor, d – scalar damage crushing in compression, during which many small tight
parameter (representing stiffness degradation), ε – strain cracks develop and evolve as a result of the loss of most
tensor, 𝐷𝐷0el – initial tensor of elastic constitutive stiffness, of the concrete strength related to spalling.
𝐷𝐷el – degraded tensor of elastic constitutive stiffness. The strength of concrete in a complex stress state is
It is necessary to specify an effective stress tensor dependent on the complexity of the stress state present in
based on the equation (2): the material. Thus, it cannot be determined only based
on simple tests, such as uniaxial compression or tension.
 D0 : (   )
el pl
(2) The strength of concrete can be determined only by
considering the interaction of various stress state
where: ε pl – plastic strain tensor, and the evolution of the components.
damage parameter d as follows: Therefore, if one desires to describe the strength of
concrete with a triaxial state equation, one must present
d  d ( ,  pl ) (3) its plane in a three-dimensional stress space, where on
the surface there are stress states corresponding to the
This, variable d is assumed to be a function of the failure of the material, while inside there are safe
pl
effective stress 𝜎𝜎 and the equivalent plastic strain  . In performance states. In addition, inside the space there is
the concrete damage plasticity model, the stiffness of the the so-called plastic potential surface, after exceedance
material is initially isotropic, but during the process of of which two situations occur [13]:
stiffness degradation, it is determined by two variables: • an increase in strains without changing the stress (ideal
dc in the compression and dt in the tension. Ultimately, plasticity),
the Cauchy stress tensor σ is proportional to effective • material softening – destruction of the material.
stress tensor 𝜎𝜎, and the coefficient of proportionality is One of the most widely used hypotheses regarding
(1 – d): the strength of concrete is the Drucker-Prager hypothesis
[14], under which the failure depends on the shear strain
 (1  d ) (4) energy and the failure surface in the stress space takes
the shape of a cone (Fig. 1 [15]). An advantage of using
The damage of the material in compression and such a failure criterion is the surface smoothness and
tension is determined independently by two variables, c
pl thus no complications with numerical applications, but
pl a disadvantage is the incomplete compliance with the
and t , respectively, which specify the equivalent strains
real behaviour of concrete.
(in compression and tension):
(a) (b)
 cpl 
 pl   
 t 
pl
(5)
 tpl  h( ,  pl ) pl

Cracking (in tension) and crushing (in compression)

in concrete are governed by increasing hardening
(softening) variables. These quantities control the
evolution of the loading surface F and the material
stiffness degradation.
The loading surface determines the strain state as a Fig. 1. The Drucker-Prager failure surface in deviatoric cross-
function of stress and effective plastic strain. In the case section (a) and view (b) [15].
of non-viscous damage plasticity model, the state of
The CDP model used in Abaqus software is
stress and strain must satisfy the condition:
a modification of Drucker-Prager strength hypothesis. In
the recent years, subsequent modifications have been
F  ,  pl   0 (6)
made in it [16, 17] according to which the failure surface
in the deviatoric cross-section does not need to be
Plastic flow is determined by the plastic potential a circle and is regulated by parameter Kc. The physical
function G(σ) and a non-associated flow rule in the form: interpretation of parameter Kc is the ratio of the distances
from the hydrostatic axis to the compression and tension
G( )
 pl   (7) meridians in the deviatoric cross-section. This ratio is
 always larger than 0.5. If its value is 1, the deviatoric
cross-section of the failure surface becomes a circle, as is

2
ITM Web of Conferences 15, 07012 (2017) DOI: 10.1051/itmconf/20171507012
CMES’17

the case in the classic Drucker-Prager hypothesis The most reliable research in this area is the study
(Fig. 2 [15]). carried out by Kupfer in 1969. After its approximation
In the paper [13], in accordance with the results of with an equation of an ellipse, the uniform biaxial
the experimental studies, it is observed that this value is compressive strength fcc = 1.16248 [13] or by default
0.6 and slowly increases with the decrease in tension in 1.16 according to [15].
the case of normal average stress of equal to zero. In The last parameter describing the behaviour of
accordance with [15], the recommended value Kc is 2/3. concrete in a complex stress state is the dilation angle,
This shape is comparable to the strength criterion that is the angle of the inclination of the asymptote of the
formulated by William and Warnke [18], which failure surface in relation to the hydrostatic axis,
combined three mutually tangent ellipses. It is measured in the meridian plane (the angle of internal
a theoretical-experimental criterion built based on the friction in the concrete). Numerical analyses usually
results of tests in a triaxial stress state. assume the value of ψ = 36° or ψ = 40°.

2.3 The stress-strain curve in uniaxial compression

and tension
The most accurate method to describe the relationship
between stress and strain is to perform uniaxial
compression testing for the applied concrete and then
transform the variables from the obtained graph. In the
CDP model, one must determine the so-called inelastic
in
strain c by deducting the elastic portion that
Fig. 2. Deviatoric cross section of failure surface in concrete corresponds to the undestroyed material from the total
damage plasticity model and hyperbolic surface of plastic strain (recorded as part of a uniaxial compression test):
potential in meridional plane [15].

A similar change applies to the shape of the  cin  c   0cel (8)

meridians of the surface in the stress space, which are
curved lines. In the CDP concrete model, the plastic c
potential surface in the meridian plane takes the form of
 0cel  (9)
E0
a hyperbole (Fig. 2 [15]). The adjustment of the shape is
done via the eccentricity plastic potential parameter. This During the transformation of strains, one must
is the length of the segment measured along the determine the moment from which the material is
hydrostatic axis between the hyperbole vertex and the defined as non-linearly elastic (Fig. 4 [15]).
intersection of the asymptotes of this hyperbole (the
centre of the hyperbole). The eccentricity parameter can
be calculated as the ratio of the tensile strength to the
compressive strength [19]. For the CDP model, the
recommended value is є = 0.1. In a limit case where it is
equal to 0, the surface becomes a straight line in the
meridian plane – the classic Drucker-Prager hypothesis.

Fig. 4. The stress-strain curves of concrete in uniaxial compression

[15].

Laboratory tests – uniaxial tests – show that such

behaviour of material occurs almost from the beginning
of the process of compression, but it is negligible in the
initial stage for most numerical analyses. In the paper
[13], the author writes that the end of linear elasticity
Fig. 3. Strength of concrete under biaxial stress in CDP [15].
should grow with the increase in concrete strength and
Another parameter that describes the behaviour of should be more accepted from there, rather than
concrete is the point at which it fails in biaxial determined on the basis of tests. In addition, the author
compression (Fig. 3 [15]). The value σb0/σc0 (fb0/fc0) described it as a percentage of stress in relation to the
determines the ratio of the strength of concrete in strength of concrete, according to the relationship:
a biaxial state to the strength in the uniaxial state.

3
ITM Web of Conferences 15, 07012 (2017) DOI: 10.1051/itmconf/20171507012
CMES’17

f  such as splitting samples or a beam bending method. In

elim  1  exp  c  (10) the absence of laboratory tests, it is possible to derive
 80  tensile strength from the compressive strength in
accordance with [20]:
A simpler method for determining this ceiling is to
arbitrarily adopt it as 0.4·fcm. In accordance with [20],
f 0.30  f ck (2/3) (16)
the modulus of elasticity of concrete is defined as secant ctm

within the range of 0–0.4·fcm. It is good at this stage to The CDP model uses the notion of cracking strain
adopt such a ceiling of the beginning of the inelastic ck
phase so that the initial value of the Young's modulus t . This allows to consider the phenomenon called
and the secant value are consistent with each other. Most tension stiffening effect [21]. Concrete in tension is not
numerical analyses usually examine the phase of treated as an elastic-brittle body, but one takes into
reaching the strength of the material, rather than the account the effects of phenomena such as the aggregate
initial phase of its performance. Adoption of such a level interlock in a crack and the adhesion of concrete to steel
gives greater certainty regarding the convergence of the in the segment between the cracks. This assumption has
solution. a reason during the analysis with a fuzzy cracking image.
When defining the stress-strain variables, one must In this way, the decrease in the stress in concrete in the
specify the degradation variable dc, which is 0 for an tension zone does not occur suddenly but gradually. The
undestroyed material and 1 for a total loss of the ability post-cracking strains are defined as the difference
to carry stress: between the total strains and elastic strains of the
undestroyed material:
dc dc ( cpl , fi ) if 0  dc  1 (11)
 tck  t   0tel (17)
These values can also be read from the uniaxial
compression tests by calculating appropriately the ratio t
 0tel  (18)
of stress values of the declining branch of the graph to E0
the compressive strength of concrete. Hence, the CDP
model specifies strains as: pl
In turn, the plastic strains t are calculated in the
same way as for compression after defining the
dc  degradation parameter dt:
 cpl 
 cin  c (12)
(1  dc ) E0
d t d t ( tpl , fi ) if 0  d t  1 (19)
In Eq. (5), E0 denotes the initial modulus of elasticity
of the undestroyed material. Based on the plastic strain In order to build the stress-strain curve of concrete in
values, it is possible to determine the stress σc for uniaxial tension, one must determine the form of the
uniaxial compression and their effective values 𝜎𝜎c softening function (Fig. 5 [15]).
describing the size of the flow and failure surfaces:

 c (1  dc )  E0  ( c   cpl ) (13)

c
c   E0  ( c   cpl ) (14)
(1  dc )

The course of the material behaviour can also be

determined without uniaxial compression testing,
knowing only the average compressive strength of
concrete fcm. Then, the second value necessary to start
the analysis of the variability of the stress-strain curve is
the modulus of longitudinal elasticity of concrete Ecm. Its
value can be estimated based on dependencies available
in the code of practice [20]: Fig. 5. The stress-strain curves of concrete in uniaxial tension
[15].
Ecm 22  (0.1 f cm )0.3 (15)
In the absence of data, the stress can be reduced
Other values indicating the location of the linearly down to zero from the moment of reaching the
characteristic points on the graph are the strains while tensile strength in the case of total strains that are
reaching the strength (c1) and the limit strains at failure 10 times higher than at the time of reaching the value fctm
[15]. A detailed description of this function, however,
(cu1).
requires model calibration with experimental results in
In turn, the tensile strength of concrete in the uniaxial
a specific analysis case [22]:
tension state is rarely determined in a direct tension test
due to implementation difficulties and large result
variability. Therefore, one usually uses indirect methods,

4
ITM Web of Conferences 15, 07012 (2017) DOI: 10.1051/itmconf/20171507012
CMES’17

t 
Ec   t if  t   cr  cycles, the force was increased by approx. 200 kN until
0,4  the maximum assumed load of 700 kN was reached. The
   (20) maximum value was reached in each cycle by the
t 
f ctm   cr  if  t   cr 
 t   gradual increase of the force, introducing intermediate
load values. The load increase rate was on average
In Eq. (13), cr denotes the strain at concrete 10 kN/min.
cracking. As the effect of stiffening can significantly
affect the analysis results and additionally this
relationship requires calibration during a simulation, it is
recommended to use a modification of the formula by
Wang T. and Tsu for the softening function:
n
 
t 
f ctm   cr  when  t   cr (21)
 t 
Fig. 7. Test piece load scheme together with the slab damage
where n denotes a parameter that determines the measurement area.
softening rate.
All computational tasks were solved in
Abaqus/Standard software using the incremental
3 Numerical analysis iterative Newton-Raphson method [15]. The problems
3.1 Test specimens related to achieving the convergence of the solution,
caused by the non-linearity of the material model, were
The correctness of the adopted concrete model was solved by viscosity stabilisation μ [24]. The size of the
verified on the basis of the results of experimental tests load increment was also reduced (0.01 – 1E-12) and the
of continuous composite beams subjected to bending [23]. maximum number of the load steps (max. 12000) was
Test specimens were made from steel girders increased. The selection of the parameter μ was made
(European I-beam, IPN 360 type) connected with iteratively after analysing the size of its impact on the
a concrete slab along its entire length using two rows of results of the task. Finally, μ=0.0001 was adopted,
pin fasteners. The fasteners with the diameter of 16 mm allowing one to solve the task in more than 1.200 load
and height of 75 mm were welded to the top flange of increments formed in approx. 4.000 iterations. The
the beam with a 20 cm spacing. Such a connection was analysis of the behaviour of the test configuration shows
intended to ensure the rigidity of the joint in the entire that such a value of the viscosity parameter makes it
load range. The primary reinforcement of the slab possible to reach a compromise between the
comprised twelve 10 mm smooth reinforcement bars, computational magnitude of the task and the accuracy of
laid in two rows upside and downside, along the entire the obtained results.
length, in groups of 6 bars with the spacing of 8 cm (the
reinforcement degree was 2%). The transverse 3.2 Geometrical models
reinforcement consisted of stirrups made of 4.5 mm
bars, arranged with a 20 cm spacing (Fig. 6). Due to the nature of the structure being modelled, which
consists of two different materials with clearly different
geometries, the numerical model included different types
of finite elements that best described the individual
components of the beam.
Hence, the concrete slab was modelled with an
8-node brick element with reduced integration (C3D8R),
and the I-beam with 4-node shell elements with reduced
integration (S4R) and the primary reinforcement bars
and stirrups with 2-node linear beam elements (B31).
Joining of the concrete slab with the upper flange of
the steel beam varied depending on the distribution of
Fig. 6. Tested beam parameters.
bending moments along the length of the beam (Fig. 7).
The places of application of external loads were In particular, the joint was modelled in the tension zone
chosen in such a way that the test piece included both the as a discrete connection representing the occurrence of
zone of negative moments causing tension of the upper rigid connectors in the elements subjected to the
fibres of the concrete slab and the zones of positive laboratory tests. On the other hand, in the positive
moments causing their compression (Fig. 7). The beams moment zones where the slab was compressed, the full
were loaded gradually by applying incremental joining of the upper surface of the steel beam with the
concentrated forces in four full load-unload cycles. The lower surface of the reinforced concrete slab by means
first cycle was always carried out in the range from 0 kN of a "continuous" (tie) connection was considered.
until the appearance of the first crack. In the subsequent

5
ITM Web of Conferences 15, 07012 (2017) DOI: 10.1051/itmconf/20171507012
CMES’17

3.3 Identification of constitutive parameters of

the concrete model
While identifying the constitutive parameters, it was
assumed that concrete behaved in a linear-elastic manner
up to 0.4·fc in compression and up to ft in tension. For
compression, the elastic limit is the point described by
the stress value of 28.06 MPa and the strain value of
0.0006450. Concrete hardens in compression from the
stress value of 28.06 MPa to the value of 70.15 MPa. In
this range, any strain relief occurs along a straight line,
whose slope is Ecm. Upon reaching stress that is equal to
70.15 MPa, concrete in compression softens, which is
accompanied by the degradation of stiffness as well.
After reaching the critical stress for concrete, the
Fig. 9. Determined stress-strain curves of tensioned concrete.
stiffness is reduced.
The stress-strain dependence of the non-linear Similar deliberations were made for tension. In
behaviour of concrete under uniaxial compression accordance with the earlier assumptions, it was assumed
(Fig. 8) was assumed in accordance with the that concrete behaved elastically up to the stress value of
recommendations [20]: 3.82 MPa. According to Eqs. (17)-(18), cracking strains
ck
t were determined, on the basis of which plastic strains
k  
2
pl
 c  f cm  when 0   c   cu1 (22) t are calculated taking into account the degradation
1  (k  2)
parameter dt (the maximum concrete degradation equal
to 0.99 was adopted). Based on the model calibration,
k 1.05  Ecm  c1 f cm
 (23)
the form of the softening function was determined with
the assumption that stress can be reduced linearly from
   c  c1 (24) the moment of reaching the tensile strength (Fig. 9).
Table 1 shows the material parameters that determine
The strain values corresponding to the greatest stress
the behaviour of concrete in a complex stress state.
and the limit strain were determined from:
Table 1. Adopted CDP model parameters.
 c1 
0.7  f cm0.31 and  c1  2.8 (25)
The parameters of CDP model
Dilatation
 cu1  2.8  27   0.01   98  f cm 
4
(26) 36° fbo/fco 1.16
angle ψ
Eccentricity є 0.10 Kc 0.667
Then, the variables were transformed into inelastic
in Concrete compression Concrete compression
strains c , according to Eqs. (8)-(9) and degradation hardening damage
variables dc were calculated as the ratio of stress values Crushing Crushing
in the descending branch of the graph to the compressive Stress [MPa] DamageC [-]
strain [-] strain [-]
strength of concrete. In this way, based on Eq. (12), 28.06 0.000000 0.0 0.000000
pl
plastic strains c are determined during the numerical 44.38 0.000113 0.0 0.000113
analyses. 57.87 0.000292 0.0 0.000292
66.84 0.000576 0.0 0.000576
70.15 0.000990 0.0 0.000990
69.69 0.001164 0.007 0.001164
68.24 0.001339 0.027 0.001339
Concrete tension stiffening Concrete tension damage
Cracking Cracking
Stress [MPa] DamageT [-]
strain [-] strain [-]
3.82 0.0 0.0 0.0
0.38 0.00032 0.99 0.00032

4 Results and discussion

Identification of the slab cracking was done on the
basis of the analysis of the maps of damage defined by
the changes in the size of the DAMAGE parameter, i.e.
the degradation of stiffness, which illustrates the material
Fig. 8. Determined stress-strain curves of compressed concrete. damage.
It should be remembered that the Concrete Damage
Plasticity model of the material does not enable the

6
ITM Web of Conferences 15, 07012 (2017) DOI: 10.1051/itmconf/20171507012
CMES’17

formation of cracks in a discrete way taking into account displacements of the tested beams and the numerical
the material spalling (its losses). It results only in the model (Fig. 12 and Fig. 13). The deflections were
gradual exclusion of finite elements from the interaction. measured at two points where the external load was
In this way, their “bonding” occurs, as well as their applied. In addition, readings of vertical displacements
further involvement in the transfer of strain to the were taken in the longitudinal beam axis, above each
adjacent elements. This is one of the imperfections in the support.
model, which has no significant effect on the behaviour The analysis of the relationship between the load and
of the entire test configuration, though. the displacement confirms the good agreement of the
The maps of damage defined by parameter dt in the solutions. The mean percentage differences between the
tension zone and parameter dc in the compression zone, computer calculation results and the measured values
presented in the Fig. 10 and Fig. 11, can be identified oscillate between ~5% and ~8%.
with the spots where cracks appeared in the concrete The graphs also show that the beam stiffness in the
slabs of the tested composite beams. It can be observed experiment and the stiffness of the numerical model are
that the image of damage obtained in the numerical similar to each other. This confirms the correctness of
analysis corresponds qualitatively to the picture of the the selection of degradation variables d (parameter dt, in
distribution of cracks obtained during the experimental particular) in the concrete model.
tests.
The analysis of the maps of damage defined by
parameter dt also allows for tracing the process of the
formation and development of cracks with an increasing
load. In the case of the analysed beams, the first damage
to the slab concrete appeared at the support axis, with the
load causing tensile stress in concrete that was equal to
the tensile strength. In the initial phase, as the load
continued to be increased, more cracks began to appear
simultaneously on both sides of the support, spreading
towards the centre of the span. The subsequent increase
in the load caused the density of the damage zones
(cracks) to increase. A similar phenomenon was
observed during the empirical research [23].
a)

Fig. 12. Comparison of the vertical beam displacements

measured and calculated in support section.

b)

Fig. 10. The comparison of crack patterns for bending beam in

support section: a) beam with CDP numerical model, b) the
fracture path, which is observed in experiment.

b) Fig. 13. Comparison of the vertical beam displacements

measured and calculated in span section.

5 Summary and conclusions

Fig. 11. The comparison of crack patterns for bending beam in The comparative analyses of the experimental for
span section: a) beam with CDP numerical model, b) the a numerical study results show that the application of the
fracture path, which is observed in experiment. concrete damage plasticity model allows correct
determination of the concrete element damage
In addition, the verification of the numerical model mechanisms. Considering the full non-linearity of
was carried out based on the comparison of the obtained

7
ITM Web of Conferences 15, 07012 (2017) DOI: 10.1051/itmconf/20171507012
CMES’17

concrete along with its progressive degradation as the 5. W. Wang, L. Sluys, R. de Borst, International
stress values increase, however, can cause some Journal for Numerical Methods in Engineering 40,
problems related to achieving the convergence of the 3839 (1997) http://dx.doi.org/10.1002/(SICI)1097-
solution [24]. These problems are mainly related to the 0207(19971030)40:20<3839::AID-
tensile response. FEM techniques based on reducing the NME245>3.0.CO;2-6
load increment size or increasing the maximum number 6. J. Tejchman, G. Gudehus, International Journal for
of steps while solving the problem using the Newton- Numerical and Analytical Methods in
Raphson approach may prove insufficient. Hence, the Geomechanics 25, 1 (2001)
Concrete Damage Plasticity model includes the viscosity http://dx.doi.org/10.1002/1096-
parameter μ, which is used for constitutive equation 9853(200101)25:1<1::AID-NAG115>3.0.CO;2-8
regularisation (allows to slightly exceed the plastic
7. J. Bobinski, J. Tejchman, Archives of Civil
potential surface in some sufficiently small steps). The
Engineering 52(1), 7 (2006)
very idea of viscoplastic regulation consists in such a
https://www.scopus.com/record/display.uri?eid=2-
parameter selection so that the ratio of the time step of
s2.0-
the task to the value of μ would approach infinity. Such
33845642530&origin=resultslist&zone=contextBox
a regulation method forces the parameter μ to be selected
multiple times to check its impact on the results of the 8. J. Pamin, Gradient-Enhanced Continuum Models:
problem and thus, determine its minimum value. Formulation, Discretization and Applications
The results of the studies make it possible to (Cracow, 2004)
formulate the following conclusions: 9. T. Belytschko, T. Black, International Journal for
- The material damage plasticity model is one of many Numerical Methods in Engineering 45, 601 (1999)
possibilities for modelling concrete [25], which can be http://dx.doi.org/10.1002/(SICI)1097-
used for computer simulation of the behaviour in 0207(19990620)45:5<601::AID-
advanced load states of the entire structure and its NME598>3.0.CO;2-S
individual elements. However, it is intended primarily 10. L. Kachanow, Izv. Ak. Nauk CCCP, Otd. Techn.
for the analysis of reinforced concrete structures and Nauk, 8, 26 (1958)
composite structures of various types, such as steel-
11. T. Jankowiak, T. Łodygowski, Foundations of Civil
concrete and even concrete-concrete.
and Environmental Engineering, 6, 53 (2005)
- Before the start of the analysis of a structure, one
http://www.fcee.put.poznan.pl/sites/issues/2005.06/f
should determine the parameters of individual materials
ull/fcee_2005-06_053-069.pdf
and constitutive relations between the strain and the
stress. 12. A. Chen, W. Chen, Journal of the Engineering
- Due to the nature of the damage of concrete, some Mechanics Division 101(4), 465 (1975)
quantities should be adopted rather than determined as https://www.scopus.com/record/display.uri?eid=2-
part of laboratory tests. Hence, the correctness of these s2.0-
assumptions should be checked by comparing other 0016543486&origin=resultslist&zone=contextBox
parameters, e.g. the deflection of the modelled structural 13. S. Majewski, The mechanics of the structural
element. Therefore, the results of the numerical analysis concrete in elastic-plastic perspective (Gliwice,
often require that the calibration of the model parameters 2003)
be performed several times. 14. D. Drucker, W. Prager, Quarterly of Applied
Mathematics 10(2), 157 (1952)
The author would like to thank the Gotowski Budownictwo http://www.jstor.org/stable/43633942
Komunikacyjne i Przemysłowe Sp. z o.o. company for
providing the testing materials free of charge. Laboratory 15. Abaqus User’s Manual (2012)
studies were co-funded under the statutory measures of the 16. J. Lubliner, J. Oliver, S. Oller, E. Oñate,
Ministry of Science and Higher Education No. S-50/B/2012. International Journal of Solids and Structures 25(3),
229 (1989) doi:10.1016/0020-7683(89)90050-4
References 17. J. Lee, G.L. Fenves, Journal of Engineering
Mechanics 124(8), 892 (1998)
1. H. Hernández, J. Bonilla, G. Rodríguez, Revista http://dx.doi.org/10.1061/(ASCE)0733-
Ingeniería de Construcción 29(1), 5 (2014) 9399(1998)124:8(892)
http://dx.doi.org/10.4067/S0718- 18. K.J. Willam, E.P. Warnke, Concrete Structures
50732014000100001 Subjected To Triaxial Stresses 19, 1 (1974)
2. H.M. Gomes, A.M. Awruch, Advances in http://doi.org/10.5169/seals-17526
Engineering Software 32(4), 257 (2001) 19. Y. Sümer, M. Aktaş, Challenge Journal of Structural
http://dx.doi.org/10.1016/S0965-9978(00)00093-4
Mechanics 1(3), 149 (2015)
3. K. Śledziewski, Maintenance and Reliability 4(18), http://dx.doi.org/10.20528/cjsmec.2015.07.023
578 (2016) http://dx.doi.org/10.17531/ein.2016.4.13 20. Eurocode 2, Part 1-1 (2004)
4. S. Pietruszczak, Z. Mróz, International Journal for
Numerical Methods in Engineering, 17, 327 (1981)
http://dx.doi.org/10.1002/nme.1620170303

8
ITM Web of Conferences 15, 07012 (2017) DOI: 10.1051/itmconf/20171507012
CMES’17

21. S.M. Allam, M.S. Shoukry, G.E. Rashad, A.S. 24. P. Kmiecik, M. Kamiński, Archives Of Civil And
Hassan, Alexandria Engineering Journal 52(2), 163 Mechanical Engineering XI(3), 623 (2011)
(2013) http://dx.doi.org/10.1016/j.aej.2012.12.005 http://dx.doi.org/10.1016/S1644-9665(12)60105-8
22. T. Wang, T.T.C. Tsu, Computers and Structures 79, 25. L. Jason, A. Huerta, G. Pijaudier-Cabot,
2781 (2001) http://dx.doi.org/10.1016/S0045- S. Ghavamian, Computer Methods in Applied
7949(01)00157-2 Mechanics and Engineering 195(52), 7077 (2006)
23. K. Śledziewski, ET Sciences 5, 106 (2017) http://dx.doi.org/10.1016/j.cma.2005.04.017