Case Studies in Construction Materials 16 (2022) e00834

Contents lists available at ScienceDirect

Case Studies in Construction Materials

journal homepage: www.elsevier.com/locate/cscm

Case study

New approach for computing damage parameters evolution in

plastic damage model for concrete
R. Bakhti a, b, B. Benahmed b, *, A. Laib a, M.T. Alfach c
a
Department of Civil Engineering, Faculty of Science And Applied Science, University of Bouira, Algeria
b
Development Laboratory in Mechanics and Materials, University of Djelfa, Algeria
c
School of Mathematics, Computer Science and Engineering, Department of Civil Engineering, City, University of London, UK

A R T I C L E I N F O A B S T R A C T

Keywords: The Damage Plastic Model is one of the most widespread models used to simulate the real
Damage Plastic Model behavior of concrete. The high reliability of this model could be attributed to its ability to take
Damage parameters calculation into consideration the elastic stiffness degradation induced by plastic straining as well as the
Concrete behavior
stiffness recovery effects under cyclic loading. This model has been implemented in ABAQUS
Nonlinear finite element method
under the name of Concrete Damaged Plasticity Model (CDPM). Nevertheless, the use of CDPM in
Numerical simulation
Concrete structures ABAQUS is very sensitive due to the high complexity of the calibration process of comparing the
model parameters with experimental data, particularly, the stress-inelastic strain diagrams and
the damage parameters evolution. The main aim of this research is to develop a new methodology
to facilitate the use of this model without the need to calibrate the stress-inelastic strain diagrams
and the damage parameters evolution with experimental results. This methodology is based on
Lubliner formulas to identify the stress-inelastic strain diagrams, and on Alfarah formulas to
evaluate the damage parameters evolution. A new algorithm has been developed to determine the
dimensionless coefficients of Lubliner and Alfarah formulas according to the Model Code rec­
ommendations. This algorithm has been implemented in developed finite element code where the
results have been validated by comparing with experimental results and closed-form solutions.

1. Introduction

Concrete is one of the most used building materials in the world. In fact, the use of this material is twice time more than the use of all
other construction materials such as wood, steel, plastic, and aluminum. Consequently, the pervasive use of this material dictates a
thorough understanding of the actual concrete behavior. As the concrete stress-strain behavior under uniaxial compression or tension
loading is nonlinear, so simulating the concrete behavior by using linear models in the numerical modeling methods leads to inaccurate
results which makes the non-linear models is an inevitable solution to minimize the error margin. The outcomes of non-linear models
are deeply dependent on the adopted constitutive model, so it is very important to choose the suitable one to predict the reliable load-
displacement and/or stress-strain curves.
Recently, several constitutive models have been developed to simulate reliably the complex behavior of concrete. According to [1],
these models can be categorized as follows: empirical models [2–4], linear elastic models [5–7], nonlinear elastic models [8–11],
plasticity models [12–16], fracture models [17,18], endochronic models [19–21], damage models [22–24], microplane models

* Correspondence to: Civil Engineering Faculty, Development Laboratory in Mechanics and Materials, University of Djelfa, Algeria.
E-mail address: benahmed.tp@univ-djelfa.dz (B. Benahmed).

https://doi.org/10.1016/j.cscm.2021.e00834
Received 6 August 2021; Received in revised form 19 November 2021; Accepted 3 December 2021
Available online 5 December 2021
2214-5095/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

[25–27], and damage plastic models [28–34]. One of the most pioneering models for concrete behavior has been created by [28] under
the name of Damage Plastic Model (DPM), and improved by [30] to address the cyclic loading. This model has been implemented in a
standard finite element program for the first time by [35]. Later, [36] suggested a return-mapping algorithm to implement the recent
form of the DPM. The main advantage of this model is its ability to consider the elastic stiffness degradation caused by both tension and
compression plastic straining. The recent form of this model has been exploited from [37] to predict the crack propagation paths.
Furthermore, [38] have suggested a novel stress decomposition. Similarly, [39] have developed a new constitutive plastic-damage
model with evolutive dilatancy. The DPM has been implemented in the finite element code ABAQUS under the name of Concrete
Damaged Plasticity Model (CDPM). The package software ABAQUS has been widely used in the numerical modeling of concrete, for
instance [40] used CDPM to simulate the concrete damage. Likewise, [41] used CDPM in the numerical simulation of prestressed
precast concrete bridge deck panels. Furthermore, [42] used it to simulate ultra-high performance fibre reinforced concrete material
under impact loading rates at different damage stages.
The use of CDPM in ABAQUS requires the following parameters:

1. The stress-inelastic strain diagram for compression.

2. The stress-inelastic strain diagram for tension.
3. The ratio Kc of the second stress invariants on tensile and compressive meridians.
4. The eccentricity ε.
5. The ratio (fb0/fc0) of biaxial compressive yield stress to uniaxial compressive yield stress.
6. The dilation angle ψ in the p-q plane.

Moreover, optional parameters can be used, such as:

1. The viscosity parameter.

2. The compressive damage parameter evolution.
3. The tensile damage parameter evolution.

The outcomes of ABAQUS strongly depend on the values of these parameters, thus, all parameters must be calibrated with
experimental tests. Furthermore, both the high complexity and sensitivity of the calibration process deeply diminishes the CDPM
efficiency. Few research has addressed this problem with the aim to reduce the number of parameters needed in the calibration process
and identify their typical values. Szczecina, Winnicki [43] have recommended the values of 0.0001 and 5 degrees as typical values for
the viscosity parameter and the dilation angle, respectively. In a related study, [44] have advised assessing the dilation angle and the
fracture energy in compliance with the results obtained from Strut-and-Tie method and laboratory tests. While [45] have proposed a
closed-form solution for evaluating the compressive damage parameter. As well as [46] have examined the role of the viscosity
parameter in the numerical simulation of RC deep beams and concluded that 0.0005 is the typical value for the viscosity parameter.
Likewise, [47] have employed a new algorithm to evaluate the dilation angle. In a different manner, experimental results have been
exploited from [48] to calibrate the dilation angle ψ, the eccentricity ε, the ratio fb0 /fc0 , the ratio Kc, and the viscosity parameter μ.
Important research efforts have been employed to avoid the problems due to the complexity and the sensitivity of the calibration
process of both the stress-inelastic strain and the damage parameters diagrams. Thus, [28] have proposed closed-form solutions for
both the stress-inelastic strain and the damage parameters. Similarly, a closed-form solution has been adopted from [33] to compute
the damage parameters and by consequence evaluating the stress-strain diagrams. Differently, [49] have developed an analytical
approach to evaluate the damage parameters evolution in terms of corresponding inelastic strains. Moreover, they elaborated an
analytical approach to estimate the stress-strain diagram under compressive loading. Likewise, [50] have exploited an analytical
approach to compute the stress-strain diagrams. Additionally, they have created closed-form solutions to calculate the damage pa­
rameters evolution. While [47] have used a new algorithm to evaluate the stress-strain diagrams.
In this study, a numerical approach has been developed to estimate the stress-strain diagrams and the damage parameters evolution
in accordance with the Model Code recommendations [51]. In this approach, the stress-inelastic strain has been evaluated according to
[28] formulas, and the damage parameters evolution has been estimated according to [33] formulas, with a view to eliminate the mesh
size effect on the stress-strain diagrams and the damage parameters evolution (as demonstrated in this work).
The main objective of this study is to develop an algorithm to compute Lubliner coefficients following the fact that the peak point of
the compressive stress-strain curve has the same compressive strength and the same strain (at compressive strength) delivered in the
Model Code recommendations [51]. Furthermore, the developed algorithm has been implemented in our academic finite elements
code “Concrete v2.0′′ . Additionally, the outcomes of the algorithm have been validated by comparing with both the experimental
results and closed-form solutions to demonstrate the efficiency of our methodology. Moreover, the outcomes of [33] algorithm have
been examined to analyze the mesh size influence on the stress-inelastic strain and damage parameters evolution.

2. Damage plastic model

Damage Plastic Model (DPM) has been developed by [28] to simulate the complex behavior of concrete. The main concept of this
model is to substitute the hardening variable in the overall form of classical plasticity by the plastic damage variable which has two
values (0 and 1), where the zero value represents the undamaged concrete and the value of one represents the totally damaged concrete
with full loss of cohesion.

2
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

The fundamental equations of this model are:

a. The Yield function

1 ( √̅̅̅ )
F= 3αp + 3 J + β〈σmax 〉 − γ〈 − σmax 〉 − c (1)
1− α

With α, β and γ are dimensionless parameters given by:

( )/( )
f fb0
α = b0 − 1 2 − 1 (2)
fc0 fc0

fc0
β = R(1 − α) − (1 + α) with R = (3)
ft0
( )/( max )
γ = 3 1 − rmax
oct 2roct − 1 (4)

Here:
J Deviatoric stress,
p Mean total stress,
σmax The maximum principal effective stress,
fb0
fc0 The ratio of biaxial and uniaxial compressive yield strengths. According to [52],fb0 /fc0 takes a value between 1.10 and 1.20.
In ABAQUS [53], the default value is 1.16,
ft0 The initial uniaxial tensile yield stress,
c Cohesion,
rmax
oct Constant takes a value of 0.65 according to [35],
<X> Macaulay bracket and takes the form:〈±X〉 = X ±|X|
2

a. The potential function:

A non-associated potential plastic flow has been suggested by [28], where the potential function G takes the same form of the
classical Mohr-Coulomb yield function with substituting the friction angleϕby the dilation angle ψ . It takes the form:
( )
sin θ sin ψ
G = p sin ψ + J cos θ − √̅̅̅ (5)
3

Where:
J Deviatoric stress,
p Mean total stress,
ψ The dilation angle,
θ Lode angle.
The main weakness of the yield function and the potential function suggested by [28] is their inability to handle the dynamic
loading. Hence, a second form of this model has been developed by [30] to address the dynamic loading. The following modifications
have been proposed in this model:

a. The Yield function

1. Substituting cohesion “c” by effective compressive cohesion stress σc

2. New formulas forβand γparameters, given as:
⎧ σc
⎪
⎨ β = σ t (1 − α) − (1 + α)
⎪
(6)
⎪
⎩ γ = 3(1 − Kc )
⎪
2Kc − 1

Where:
Kc The ratio of second stress invariants on tensile and compressive meridians. For Mohr-Coulomb yield surface,Kc takes a value
of 0.7 [33].
σc The effective compressive cohesion stress.
σt The effective tensile cohesion stress.
As a result, the second form of the yield function suggested by [30] is written as:

3
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Fig. 1. Implementation of DMP in “Concrete v2.0′′ .

1 ( √̅̅̅ )
F= 3α p + 3 J + β〈σ max 〉 − γ 〈 − σmax 〉 − σc (7)
1− α

a. The potential function:

By using the Drucker-Prager hyperbolic function, a new formula for the potential function has been developed by [30] in the form:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2
G= εσ t0 tan ψ + 3J 2 + p tan ψ (8)

Here,
σ t0 The uniaxial tensile stress at failure,
ε The flow potential eccentricity (0.1 in ABAQUS).
According to the Model Code recommendations [51], the value of σt0 is related to the concrete grade:

1. For concrete grade ≤ C50

σt0 = ftm = 0.3016 f 2/3

ck (9)

2. For concrete grade > C50

σt0 = ftm = 2.12 ln(1 + 0.1(fck + 8)) (10)

With fck is the characteristic value of concrete compressive strength,

4
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Table 1
Default values of DPM parameters.
Parameter Default value/Estimation methodology

σt0 Eqs. (9,10)

ε 0.1 According to[53]
ψ 130 According to[54] and 50 According to[43]
Kc 0.7 According to[33]
fb0 /fc0 1.16 According to[53]
σc See the following sections
σt See the following sections

Fig. 2. a: Response of concrete to uniaxial loading in tension, b: Response of concrete to uniaxial loading in compression.

In fact, the yield function “F” and the potential function “G” have a major role in the finite element implementation of any non-
linear constitutive model. For instance, Fig. 1 shows the necessary steps adopted in the implementation of DPM in "Concrete v2.0"
for static loading.
To implement DPM, the following steps have been carried out:

1. Identification of the yield function value to determine the material state which be elastic behavior if F < 0, plastic (or elastoplastic)
behavior if F = 0 and impossible situation if F > 0.

1. Identification of the potential function value needed to evaluate the plastic strain.

The parameters to be identified for the evaluation of both functions F and G have been summarized in Table 1.

3. Estimation of the effective tensile and the effective compressive cohesion stress

The value of the yield function depends on the values of the parametersα, β,andγ. αandγwhich can be estimated by using Eq. (2) and
Eq. (6) respectively. From Eq. (6), also, the estimation of β depends on the values of the effective tensile and the effective compressive
cohesion stresses which can be estimated by:
⎧ σt
⎪
⎨ σt = 1 − d
⎪
(11)
t
⎪
⎪ σc
⎩ σc =
1 − dc

Where dc is the compressive damage parameter, dt is the tensile damage parameter, σc is the compressive stress, andσ t is the tensile stress.
As well as the compressive and tensile stress can be predicted from the stress-strain curves as shown in Fig. 2.a and b.
To quantify the stress-strain curves, two main alternatives can be used:

5
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

• User data: where the user supplied either the stress-inelastic strain data or the stress-strain data.
• Auto-estimation: in which the stress-inelastic strain data or the stress-strain data are auto-estimated depending on the characteristic
value of concrete compressive strength.

3.1. User data

The stress-inelastic strain or the stress-strain curve must be supplied by the user as a set of points. The strain values can be evaluated
according to the next correlations:
⎧ σt
⎪ ck
⎨ εt = εt + E
⎪
(12)
0
⎪
⎪ in σc
⎩ εc = εc +
E0

Where E0 is the initial undamaged stiffness, εck in

t is the tensile inelastic strain (the cracking strain), εc is the compressive inelastic strain
(crushing strain).
In “Concrete v2.0′′ , the damage parameters have been estimated by:

1. in tension state [55]:

⎧
⎨ dt = 0 if εt ≤ εtm
σt (13)
⎩ dt = 1 − if εt > εtm
ftm

1. In compression state [55,56]:

⎧
⎨ dc = 0 if εc ≤ εcm
σc (14)
⎩ dc = 1 − if εc > εcm
fcm
Finally, the damage parameter can be evaluated as:
d = 1 − (1 − st dc )(1 − sc dt ) (15)

Where sc and st are the stress state, they take the form of:
{
sc = 1 − hc (1 − r∗ (σ11 ))
(16)
st = 1 − ht r∗ (σ11 )

Where

1. hc andht are weighting factors (with values vary between 0 and 1)

r∗ (σ 11 )is the unit step. For uniaxial loading:

{
1 if σ11 > 0
r∗ (σ11 ) = (17)
0 if σ11 < 0

For multiaxial loading conditions, the damage parameter expression is generalized by replacing the unit step function r∗ (σ 11 ) by a
multiaxial stress weight factorr(σ ):
∑3
〈σi 〉
r(σ) = ∑i=13
; 0 ≤ r(σ) ≤ 1 (18)
i=1 |σ i |

Another formula can be used to evaluate the damage parameters provided by [57] takes the following form:
√̅̅̅̅̅̅̅
σ
d =1− (19)
E0 ε

3.2. Auto-estimation

3.2.1. Lubliner methodology for calculating damage parameters evolution and stress-strain diagrams

a. Stress-inelastic strain diagrams

6
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Fig. 3. a: Parts of tension energy dissipated by damage, b: Parts of compressive energy dissipated by damage.

According to [28], the relations of stress- inelastic strain both in tension and in compression are given by:
{ [ in in
]
σc = fc0 (1 + ac )e− bc εc − ac e− 2bc εc
[ ck ck
] (20)
σt = ft0 (1 + at )e− bt εt − at e− 2bt εt

Wherefc0 andft0 are the compressive and tensile stresses that correspond respectively to zero crushing (εin ck
c = 0) and zero cracking (εt =
0).
ac , at , bc , andbt are dimensionless coefficients evaluated from the correlation of tensile/compressive energies per unit of volume
dissipated by damage along entire deterioration process (See Fig. 3.a-b), which are given by:
⎧ ∫∞
⎪ gc =
⎨ σc dεinc
0
∫∞ (21)
⎪
⎩g =
t σt dεck
t
0

To evaluateac , at , bc , andbt , the values of tension and compression stresses obtained from Eq. (20) has been used in Eq. (21) to
obtain:
⎧
⎪
⎪ fc0
⎪ gc =
⎨ (1 + 0.5 ac )
bc
(22)
⎪
⎪ ft0
⎪
⎩ gt = (1 + 0.5 at )
bt
By consideringgc = Gch /Leq andgt = GF /Leq , the coefficientsbc andbt can be estimated by:
fc0 Leq
bc = (1 + 0.5 ac ) (23)
Gch

ft0 Leq
bt = (1 + 0.5 at ) (24)
GF

Where:
Gch The crushing energy per unit area.
GF The fracture energy per unit area.
Leq The mesh size (finite element characteristic length). For a brick element, Leq is taken as the volume divided by the largest
face area.
According to [33], by zeroing derivatives of σc and σ t in Eq. (20) with respect to the compressive and tensile inelastic strains,
respectively, the maximum valuesfcm and ftm are obtained by:

fc0 (1 + ac )2 ft0 (1 + at )2
fcm = ; ftm = (25)
4ac 4at
Therefore, coefficients ac andat can be evaluated as:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2 ̅
fcm fcm fcm
ac = 2 − 2 − (26)
fc0 fc0 fc0

7
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Fig. 4. a: Mesh size effect on the tensile stress -inelastic strain, case: fcm = 25MPa, b: Mesh size effect on the Compressive stress - inelastic strain,
case: fcm= 25 MPa.

√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2 ̅
ftm ftm ftm
at = 2 − 2 − (27)
ft0 ft0 ft0

By computing the values ofac , at , bc , andbt from Eqs: (23), (24), (26), and (27), the results indicate that the values ofbc andbt strongly
depend on the mesh size valueLeq for the reason that the energies per unit volume gc andgt (Eq. (22)) are replaced byGch /Leq andGF /Leq ,
respectively, where the influence of the mesh size can be observed. So, based on the dimensions of each element, the value of the mesh
size will be calculated, in consequence, unique values for bc andbt will be obtained. Thus, the stress-inelastic strain curves evaluated
from Eq. (20) will depend on the mesh size value as shown in Fig. 4.a and b. This weakness has been overcome in this study as will
demonstrate in the following sections.

a. Damage parameters evolution

To compute the damage parameters evolution, an initial form has been given by [28] as:
⎧ ∫ εinc
⎪
⎪ dc = 1
⎪
⎪ σ c dεinc
⎨ gc 0
∫ ck (28)
⎪
⎪ 1 εt
⎪
⎪
⎩ dt = σ t dεck
t
gt 0

3.2.2. Alfarah methodology for calculating damage parameters evolution and stress-strain diagrams

a. Stress-strain diagrams

To evaluate the stress-strain diagram under uniaxial compression loading, [31,33] have divided the stress-strain diagram into three
segments (Fig. 2.b):

1. First (linear) segment (till fc0 ):

σc(1) = E0 εc (29)

2. Second (ascending) segment (betweenfc0 andfcm ):

8
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

( )2
Eci fεcmc − εεcmc
σc(2) = ( ) fcm (30)
1 + Eci εfcm cm
− 2 εc
εcm

3. Third (descending) segment, given by:

( )− 1
2 + γ c fcm εcm ε2 γ
σc(3) = − γ c εc + c c (31)
2fcm 2εcm
Eci is the modulus of deformation of concrete for zero stress.

π2 fcm εcm εplc

γc = [ ( )]2 ; b = in (32)
εc
2 GLeqch − 0.5fcm εcm (1 − b) + b fEcm0

Gch The crushing energy per unit area.

Leq The mesh size (characteristic length).
εplc The compressive plastic strain.
εin
c The compressive inelastic strain.
In tension, according to [33], the stress-strain curve can be divided into two segments (Fig. 2.a):

1. First (linear) segment (till ftm ):

σt(1) = E0 εt (33)

2. Second (descending) segment given by:

[[ ( )3 ] ]
w w( )
σt(2) = ftm 1 + c1 e− c2 wwc
− 3 − c2
1 + c1 e (34)
wc wc

With:
w: The crack opening, given by:
w = (εt − εtm )Leq (35)

wc : The crack opening at fracture, given by:

GF
wc = 5.14 (36)
ftm
GF : The fracture energy by unit area, estimated according to the Model code recommendations [51] as:

GF = 0.073 f 0.18
cm (37)

c1 andc2 are dimensionless coefficients whose values are 3 and 6.93 respectively [58].

a. Damage parameters evolution

A closed-form expression for damage parameters has been developed by [33]:
⎧
⎪
⎪ 1 [ − bc εin
c − a e− 2bc εc
in
]
⎨ dc = 1 − 2 + a 2(1 + ac )e
⎪ c
(38)
c
⎪
⎪
⎪ 1 [ ck ck
]
⎩ dt = 1 − 2(1 + at )e− bt εt − at e− 2bt εt
2 + at

According to [33], the coefficients ac andat are evaluated by considering fc0 = 0.4 fcm and ft0 = ftm (from the Model code
recommendations), so the values of ac andat are 7.873 and 1, respectively. The coefficients bc andbt are estimated according to
Eqs. (23,24).
b. Alfarah algorithm

In brief, the algorithm proposed by [33] to evaluate the stress-strain curves and damage parameters evolution can be summarized

9
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Fig. 5. a: Tensile stress vs tensile strain (fcm = 25MPa), b: Compressive stress vs compressive strain (fcm = 25MPa).

Fig. 6. a: Tensile damage parameter vs tensile strain for fcm = 25MPa, b: Compressive damage parameter vs Compressive strain for fcm = 25MPa.

in the following points (All stress values are in MPa):

a. The input data are the concrete compressive strengthfck , the mesh sizeLeq , and the ratio b which its initial value is b = 0.9.
2/3
b. Calculate the compressive/tensile stress strength usingfcm = fck +8 and ftm = 0.3016 f ck
c. Take the strain at compressive stress strength as εcm = 0.0022
d. Calculate the initial tangent modulus of concrete deformationEci = 10000 f 1/3
cm and the undamaged modulus of deformationE0 =
( )
fcm
Eci 0.8 + 0.2 88 .
( )2
e. Calculate the crushing/fracture energy (N/mm) as Gch = ffcmtm
GF andGF according to Eq. (37).

f. Calculate the critical crack opening from Eq. (36)

g. Setting up the first, second and third segments of the concrete uniaxial compressive law using Eqs. (29,30), and (31).

10
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Fig. 7. Mesh size effect on the Compressive stress vs compressive strain curve, case: fcm = 25MPa.

h. Setting up the first and second segments of the concrete uniaxial tensile law using Eqs. (33,34).
i. Calculate the coefficients bc and bt using Eqs. (23,24). Taking ac = 7.873 and at = 1.
j. Calculate the compressive/tensile damage parameters using Eq. (38).
k. Calculate the compressive and tensile plastic strains:
σ c dc
εplc = εinc − (39)
E0 (1 − dc )

σt dt
εplt = εck
t − (40)
E0 (1 − dt )

1. Calculate the average value of the ratio b using Eq. (32) and compare it with the assumption in step A. Repeat until reaching
convergence.

a. Implementation of Alfarah algorithm in “Concrete v2.0′′

In this study, for auto-estimating the stress-strain diagrams and the damage parameters evolutions, Alfarah approach has been
implemented in the FE code “Concrete v2.0′′ where the mesh size influence has been examined. Figs. 5a, b, 6a, and b have been
generated by “Concrete v2.0′′ to demonstrate the influence of the mesh size on:

1. The stress-strain diagrams in both cases compression and tension stresses.

2. The compressive and tensile damage parameters evolution in terms of the corresponding strains.

The generated curves are based on different mesh size values (50 mm, 100 mm, 200 mm, and 400 mm).
However, some observations can be made based on the figures presented above (5.a, 5.b, 6.a, and 6.b):

1. The second segment of the tensile stress-strain curve and the third segment of the compressive stress-strain curve strongly depend
on the mesh size value.
2. The compressive and tensile damage parameters are also depending on the mesh size value.
3. The damage parameters are not compatible with the stress-strain diagrams in both cases compression and tension stresses. This
conclusion is based on the fact that the formulas of the damage parameters used in Alfarah method and presented in Eq. (38) are
derived from the stress-strain formulas developed by Lubliner (Eq. 20) while the used stress-strain diagrams are based on other
formulas (Section 3.2.2.a).

So, the mesh size value will affect the outcomes (displacements, stress, and strain) for various logical reasons, which are:

11
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Table 2
Values of εc1 for different concrete strength (Model code [51]).
fck (MPa) 12 16 20 25 30 35 40 45 50 55 60 70 80 90
fcm (MPa) 20 24 28 33 38 43 48 53 58 63 68 78 88 98
εc1 (‰) 1.8 1.9 2.0 2.1 2.2 2.25 2.3 2.4 2.45 2.5 2.6 2.7 2.8 2.8

Fig. 8. a: Assumed uniaxial model of concrete behavior (tension), b: Assumed uniaxial model of concrete behavior (compression).

1. The value of the effective compressive cohesion stress σ c introduced in the yield function depends on the values of the compressive
stress and the compressive damage parameter.
2. The value of β parameter delivered in the yield function depends on the values of the effective compressive and tensile cohesion
stress, and so, it depends on both the values of the compressive and tensile stresses and the compressive and tensile damage
parameters.

According to Alfarah approach, the yield function value strongly depends on the mesh size value which could be attributed to the
influence of the mesh size on:

1. The compressive stress

2. The tensile stress
3. The compressive damage parameter
4. The tensile damage parameter

4. The proposed approach for calculating damage parameters and stress-strain diagrams

4.1. General description

As demonstrated in Section 3.2.2, the estimation of coefficients ac , at , bc ,andbt according to Eqs. (23,24,26), and (27) represents a
major weakness concerning the depending of the compressive and the tensile stresses on the mesh size (Fig. 4.a, b). To overcome this
weakness and to evaluate the stress-inelastic strain curves according to Lubliner formulas and the damage parameters evolution in
terms of inelastic strain according to Alfarah formulas, we observed that by computing coefficients ac , at , bc andbt according to Eqs.
(23,24,26), and (27) for the same concrete compressive strength and different mesh size values, the peak points of all compressive

12
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

stress-strain curves approach from the same compressive strength value while the strain value for each peak point is different (Fig. 7),
so we suggest computing these coefficients in such a way the peak point will have the same strain and stress as delivered in the model
code recommendations [51]. The main objective is to develop an algorithm computes the previous coefficients which gives a peak
point value equal to the peak point given in the model code recommendations. This algorithm computes the coefficientsac , at , bc and bt
according to the following steps:

1. Substitution of fc0 by 0.4 fcm in Eq. (26) and replacing ft0 byftm in Eq. (27) to obtain the values of ac andat , which are 7.873 and 1,
respectively
2. Computing the coefficientsbc andbt where the peak point has the same compressive strength and the same strain for compressive
strength delivered in the Model Code recommendations [51]

Table 2 presents the strain values at peak stress for different concrete strength values according to the Model Code recommen­
dations [51].

4.2. Stress-strain diagrams

In order to evaluate the stress-strain diagram under uniaxial compression loading, the diagram has been divided into two segments
as shown in Fig. 8.b where:

1. In the first (linear) segment (tillfc0 ), the compressive stress can be evaluated from Eq. (29)
2. In the second segment, the compressive stress can be estimated from Lubliner formula (Eq. (20)), and the strain from Eq. (12):
⎧ [ ]
⎪ − bc εin − 2bc εin
⎨ σ c = fc0 (1 + ac )e c − ac e c

σc (41)
⎪
⎩ εc = εinc +
E0
Likewise, to estimate the stress-strain diagram under uniaxial tensile loading, the diagram has been divided into two segments as
shown in Fig. 8.a:

1. In the first (linear) segment (tillftm ), the tensile stress can be computed from Eq. (33)
2. In the second segment, the tensile stress can be estimated from Lubliner formula (Eq. (20)), and the strain from Eq. (12):
⎧ [ ]
⎪ − bt εck − 2bt εck
⎨ σ t = ft0 (1 + at )e t − at e t

σt (42)
⎪
⎩ εt = εck
t +
E0
As mentioned formerly, the main problem about the estimation of the stress-inelastic strain diagrams from Lubliner formulas is the
deep dependency of the dimensionless coefficientsac , at , bc , andbt on the mesh size value. Therefore, to eliminate the influence of the
mesh size value on the stress-inelastic strain diagrams, a new algorithm has been developed to compute these coefficients in
compliance with the Model Code recommendations.

4.3. Damage parameters evolution

In order to evaluate the compressive damage parameter evolution in terms of corresponding strain, the diagram has been divided
into two parts, in which:

1. In the first part (tillεc0 ), the compressive damage parameter value equal to zero
2. In the second segment, the compressive damage parameter can be estimated according to Alfarah formula (Eq. (38)), and the strain
from Eq. (12), so:
⎧
⎪ 1 [ − bc εin − 2bc εin
]
⎪
⎨ dc = 1 − 2 + a 2(1 + ac )e c − ac e c

c
(43)
⎪
⎪ σc
⎩ εc = εinc +
E0
The same procedure can be used to compute the tensile damage parameter evolution in terms of the corresponding strain:

1. In the first part (tillεtm ), the tensile damage parameter value equal to zero
2. In the second segment, the tensile damage parameter has been determined according to Alfarah formula (Eq. (38)), and the strain
according to Eq. (12), so:

13
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Fig. 9. Proposed algorithm for evaluatingac ,at , bc , andbt .

⎧
⎪ 1 [ − bt εck 2bt εck
]
⎪
⎨ dt = 1 − 2 + a 2(1 + at )e t − at e− t

(44)
t

⎪ ε = εck + σt
⎪
⎩ t t
E0
According to the Model Code recommendations [51], it is allowable to replace fc0 by0.4 fcm in Eq. (26) and replace ft0 byftm in Eq. (27)
to obtain the values of ac andat , which are 7.873 and 1, respectively. Whilst the coefficients bc andbt can be estimated from the algo­
rithm described in the next section.

4.4. Proposed algorithm

The main idea in the present algorithm is to determine the coefficientbc where the peak point of the compressive stress-strain curve
has the same compressive strength and has the same strain at compressive strength delivered in the Model Code recommendations
[51]. The values of the strain at the peak stress delivered in the Model Code are estimated by [59] as:

εc1 = 0.5 f 0.31

cm ≤ 2.8 x 10− 3
(45)

14
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Table 3
Values of coefficients ac , at , bc , andbt for different concrete strength –Part1.
fck (MPa) 12 16 20 25 30 35 40

ac 7.873 7.873 7.873 7.873 7.873 7.873 7.873

bc 637.077 636.468 638.065 641.894 646.876 652.439 658.218
at 1.00 1.00 1.00 1.00 1.00 1.00 1.00
bt 6122.778 6059.292 6107.316 6240.193 6412.655 6604.052 6803.804

Table 4
Values of coefficientsac , at , bc , andbt for different concrete strength –Part2.
fck (MPa) 45 50 55 60 70 80 90

ac 7.873 7.873 7.873 7.873 7.873 7.873 7.873

bc 663.972 669.533 674.783 679.639 687.945 698.146 794.836
at 1.00 1.00 1.00 1.00 1.00 1.00 1.00
bt 7005.913 7206.661 7403.612 7595.106 7957.263 8334.567 9769.134

Fig. 10. Comparison between experimental results and proposed approach (uniaxial compression).

Wherefcm is the compressive strength in MPa. According to the Model Code recommendations [51], the compressive strength can be
estimated by:
fcm = fck + 8 (46)
In order to estimate the coefficientbt , the mesh size Leq has been computed from the Eq. (23):
bc Gch
Leq = (47)
fc0 (1 + 0.5 ac )
By substituting Eq. (47) in Eq. (24), the obtained form ofbt is completely independent from the mesh size (Leq ), which is given by:
ft0 Gch 1 + 0.5 at
bt = bc (48)
fc0 GF 1 + 0.5 ac
The proposed algorithm is illustrated in Fig. 9 (All stress values are in MPa).

15
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Fig. 11. a: Comparison between closed-form solutions and proposed approach (uniaxial tension). Case: fck = 32MPa, b: Comparison between closed-
form solutions and proposed approach (uniaxial tension). Case: fck = 25MPa, c: Comparison between closed-form solutions and proposed approach
(uniaxial tension). Case: fck = 20MPa.

4.5. Implementation in “Concrete v2.0′′

The presented algorithm has been implemented in “Concrete v2.0′′ in order to evaluate the stress-strain diagrams and the damage
parameters evolution. The coefficientsac , at , bc , andbt evaluated for different concrete strengths are summarized in Tables 3 and 4.
To validate the proposed algorithm, the stress-strain curves generated by “Concrete V2.0′′ according to the present approach has
been compared with experimental results for the compression case and with closed-form solutions for the tension case, the results are
illustrated in Fig. 10, 11a, b, and c.
Fig. 10 shows the compressive stress-strain curves of experimental tests provided by [60] together with the outcomes of the
proposed approach. Four samples with different compressive strengths (fck = 16.7MPa,fck = 25.3MPa,fck = 27.7MPaandfck = 32MPa)
have been chosen to compare with the results of the developed numerical approach. From this figure, we observe that for every sample,
both curves are very close, which validates the presented approach.
Fig. 11.a, b and c show the tension stress-strain curves generated by “Concrete v2.0′′ (by using the present algorithm) together with
the outcomes of the analytical approaches of [33,61]. The tension stress-strain curves generated according to Alfarah formulas (Eqs.

16
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

Fig. 12. a: Compressive stress vs compressive strain, b: Compressive damage parameter evolution.

Fig. 13. a: Tensile stress vs tensile strain, b: Tensile damage parameter evolution.

(33), (34)) have been calculated for a cubic sample with the following dimensions: 100 mm long, 100 mm wide and 100 mm high.
The tension stress-strain curves generated according to the approach of Thorenfeldt are linear up to the uniaxial tensile strength,
then determined by:
( )(0,7+1000εt )
εtm
σt = ftm (49)
εt
Fig. 11.a, b, and c show that the proposed approach gives similar results to Alfarah approach at the beginning, then, they start to
move gradually away until they reach the Thorenfeldt outcomes at the end of the curve. Fig. 12.a, 13a illustrate the compressive and
tensile stress-strain diagrams respectively, which are generated according to the proposed approach.
Fig. 12.b, 13b illustrate respectively the compressive and tensile damage parameters evolution generated according to the present
approach. The curves have been generated for four concrete strengths; which are: fck = 20, 30, 40and50Mpa.
Fig. 12.a illustrates the effect of compressive strength on the compressive stress-strain curves. The maximum values increased with

17
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

the augmentation of the compressive strength. While the effect of compressive strength on compressive damage parameter evolution is
negligible for compressive strainεc bigger than 60 × 10− 4.
From Fig. 13.a, we observe that the curves can be divided into three parts. The curves are quite close to each other in the first part
where represents the linear behavior of concrete, until they reach the peak points, where the maximum differences values are obtained.
Then the curves begin to get closer to each other in the second part, until they assemble in the third part. Similarly, the tensile curves
presented in Fig. 13.b consist of three parts. In the first part, we observed that the curves begin converging and reach a point of
intersection. In the second part, we notice that the curves begin to converge and reach the maximum values, then started to get closer
to each other until reaching the third part where the curves assemble. Moreover, we observe that the influence of the concrete strength
on the tensile damage parameter evolution is very limited.
As pronounced previously, all the stress-strain and the damage parameters curves plotted in Fig. 12.a, b, 13a, and b are completely
independent of the mesh size (Leq ) value. This conclusion is based on the fact that the stresses and the damage parameters values are
mainly based on the coefficients ac ,at ,bc , and bt according to Eq. (20) and Eq. (38), respectively. Furthermore, these coefficients are
estimated independently of the mesh size value (As demonstrated in Fig. 9).

5. Conclusion

In this study, a new approach for determining the stress-strain curves and the damage parameters evolution in tension and
compression states have been developed accordance with the Model Code recommendations.
The main advantage of this approach, the use of the Damage Plastic Model is no longer related to the complected calibration of the
stress-strain and the damage parameters evolution with experimental tests. In fact, the only parameter needed for evaluating the stress-
strain and the damage parameters using the developed approach is the concrete compressive strength value. Several advantages of the
current method can be summarized as follows:

1. Only, the concrete compressive strength value is needed to evaluate the stress-strain and the damage parameters curves.
2. Intriguingly, the stress-strain curves and the damage parameters evolution are independent of the mesh size effect (Leq ).
3. Achieving a high level of accuracy in the estimation of damage parameters evolution, both in tension and compression.
4. Reducing the number of parameters needed to be calibrated according to the experimental tests.
5. The use of the Model Code recommendations and expandability to support other recommendations.
6. The developed method is quite appropriate to incorporate into other numerical codes.

Declarations Funding

Current research is supported by the directorate general for scientific research and technological development (DGRSDT) at the
Ministry of high education and scientific research of Algeria.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Acknowledgements

Current research is supported by the directorate general for scientific research and technological development (DGRSDT) at the
Ministry of high education and scientific research of Algeria.

Availability of data and material

Not applicable.

Code availability

Not applicable.

References

[1] R.R. Babu, G.S. Benipal, A.K. Singh, Constitutive modeling of concrete: an overview, Asian J. Civ. Eng. 6 (2005) 211–246.
[2] L.P. Sanez, Discussion of ’Equation for the stress-strain curve of concrete’ by Desayi and Krishnan, Acids J. Proc. 61 (1964) 1229–1235.
[3] D.J. Chu Carreira, H. Kuang, Stress-strain relationship for reinforced concrete in tension, Acids J. 84 (1986) 21–28.
[4] J.B. Mander, M.J.N. Priestley, R. Park, Observed stress-strain model for confined concrete, J. Struct. Eng. 114 (1988) 1827–1849.
[5] S.H. Ahmad, S.P. Shah, Complete triaxial stress-strain curves for concrete, J. Struct. Div. 108 (1982) 728–742.
[6] M.Y.H. Bangash, Concrete and Concrete Structures, Elsevier Publications, 1989.
[7] W.F. Chen, Constitutive Equations for Engineering Materials Vol. 1: Elasticity and Modelling, Elsevier Publications, 1994.

18
R. Bakhti et al. Case Studies in Construction Materials 16 (2022) e00834

[8] B. Bresler, K.S. Pister, Strength of concrete under combined stresses, J. Proc. 55 (1958) 321–345.
[9] M.D. Coon, R.J. Evans, Incremental constitutive laws and their associated failure criteria with application of plain concrete, Int. J. Solids Struct. 8 (1972)
1169–1183.
[10] Z.P. Bazant, S.S. Kim, Nonlinear creep of concrete adaptation and flow, J. Eng. Mech. Div. 105 (1979) 429–446.
[11] Z.Y. Zhao, L.Q. Ren, Failure criterion of concrete under triaxial stresses using neural networks, Comput. -Aided Civ. Infrastruct. Eng. 17 (2002) 68–73.
[12] P.A. Vermeer, Non-Associated Plasticity for Soils. Concrete and Rock, Delft University of Technology, 1984.
[13] P.H. Feenstra, R. de Borst, A composite plasticity model for concrete, Int. J. Solids Struct. 33 (1996) 707–730.
[14] H.D. Kang, K.J. Willam, Localization characteristics of triaxial concrete model, J. Eng. Mech. 125 (1999) 941–950.
[15] I. Imran, S.J. Pantazopoulou, Plasticity model for concrete under triaxial compression, J. Eng. Mech. 127 (2001) 281–290.
[16] P. Grassl, K. Lundgren, K. Gylltoft, Concrete in compression: a plasticity theory with novel hardening law, Int. J. Solids Struct. 39 (2002) 5205–5223.
[17] A. Hillerborg, M. Modéer, P.E. Petersson, Analysis of crack formation and crack growth in concrete by means of fracture mechanics and finite elements, Cem.
Concr. Res. 6 (1976) 773–781.
[18] Z.P. Bazant, Concrete fracture models: testing and practice, Eng. Fract. Mech. (2002) 165–205.
[19] Z.P. Bazant, P.D. Bhat, Endochronic theory of inelasticity and failure of concrete, J. Eng. Mech. 102 (1976) 701–722.
[20] Z.P. Bazant, Endochronic inelasticity and incremental plasticity, Int. J. Solids Struct. 14 (1978) 691–714.
[21] Z.P. Bazant, H. Shieh, Endochronic model for non-linear triaxial behaviour of concrete, Nucl. Eng. Des. 47 (1978) 305–315.
[22] W. Suaris, C. Ouyang, V.M. Fernando, Damage model for cyclic loading of concrete, J. Eng. Mech. 116 (1990) 1020–1035.
[23] G.Z. Voyiadjis, J.M. Abu Lebdeh, Damage model for concrete using the bounding surface concept, J. Eng. Mech. 119 (1993) 1865–1885.
[24] G.P. Cabot, J. Mazars, Damage Models for Concrete. Handbook of Materials Behavior Models, Elsevier,, Amsterdam, The Netherlands, 2001, pp. 500–512.
[25] P.G. Gambarova, C. Floris, Microplane model for concrete subject to plane stresses, Nucl. Eng. Des. 97 (1986) 31–48.
[26] F.C. Caner, Z.P. Bazant, Microplane model M4 for concrete: II. Algorithm and calibration, J. Eng. Mech. 126 (2000) 954–960.
[27] Z.P. Bazant, F.C. Caner, I. Carol, et al., Microplane model M4 for concrete: I. Formulation with work- conjugate deviatoric stress, J. Eng. Mech. 126 (2000)
944–953.
[28] J. Lubliner, J. Oliver, S. Oller, E. Onate, A plastic-damage model for concrete, Int J. Solids Struct. 25 (1989) 299–326.
[29] J. Lee, G.L. Fenves, A plastic-damage concrete model for earthquake analysis of dams, Earthq. Eng. Struct. Dyn. 27 (1998) 937–956.
[30] J. Lee, G.L. Fenves, Plastic-damage model for cyclic loading of concrete, Struct. Eng. Mech. 124 (1998) 892–900.
[31] W.B. Kratzig, R. Polling, An elasto-plastic damage model for reinforced concrete with minimum number of material parameters, Comput. Struct. 82 (2004)
1201–1215.
[32] B. Ayhan, P. Jehel, D. Brancherie, A. Ibrahimbegovic, Coupled damage–plasticity model for cyclic loading: Theoretical formulation and numerical
implementation, Eng. Struct. 50 (2013) 30–42.
[33] B. Alfarah, F. López-Almansa, S. Oller, New methodology for calculating damage variables evolution in Plastic Damage Model for RC structures, Eng. Struct. 132
(2017) 70–86.
[34] C. Farahmandpour, S. Dartois, M. Quiertant, et al., A concrete damage-plasticity model for FRP confined columns, Mater. Struct. 50 (2017).
[35] S. Oller, E. Onate, J. Oliver, J. Lubliner, Finite element nonlinear analysis of concrete structures using a “plastic-damage model, Eng. Fract. Mech. 35 (1990)
219–231.
[36] J. Lee, G.L. Fenves, A return-mapping algorithm for plastic-damage models: 3-D and plane stress formulation, Int J. Numer. Meth Engng 50 (2001) 487–506.
[37] M.R. Javanmardi, M.R. Maheri, Extended finite element method and anisotropic damage plasticity for modelling crack propagation in concrete, Finite Elem.
Anal. Des. 165 (2019) 1–20.
[38] B. Ahmed, G.Z. Voyiadjis, T. Park, Damage Plast. Model Concr. Using Scalar Damage Var. a Nov. Stress Decompos. 191–192 (2020) 56–75.
[39] M. Poliotti, J.M. Bairán, A new concrete plastic-damage model with an evolutive dilatancy parameter, Eng. Struct. 189 (2019) 541–549.
[40] M.A.L. Silva, J.C.P.H. Gamage, S. Fawzia, Performance of slab-column connections of flat slabs strengthened with carbon fiber reinforced polymers, Case Stud.
Constr. Mater. 11 (2019), e00275.
[41] W. Ren, L.H. Sneed, Y. Yang, R. He, Numerical simulation of prestressed precast concrete bridge deck panels using damage plasticity model, Int. J. Concr. Struct.
Mater. 9 (2015) 45–54.
[42] H. Othman, H. Marzouk, A study on damage mechanism modelling of shield tunnel under unloading based on damage–plasticity model of concrete, Int. J.
Impact Eng. 114 (2018) 20–31.
[43] M. Szczecina, A. Winnicki, Calibration of the CDP model parameters in Abaqus. The 2015 World Congress on Advances in Structural Engineering and Mechanics
(ASEM1, Incheon,, Korea, 2015, p. 5).
[44] M. Szczecina, A. Winnicki (2015b) Numerical simulations of corners in RC frames using Strut-and-Tie method and CDP model. In: XIII International Conference
on Computational Plasticity. Fundamentals and Applications (COMPLAS XIII), Barcelona, Spain.
[45] Y. Sümer, M. Aktaş, Defining parameters for concrete damage plasticity model, Chall. J. Struct. Mech. 1 (2015) 149–155.
[46] A. Demir, H. Ozturk, K. Edip, et al., Effect of viscosity parameter on the numerical simulation of reinforced concrete deep beam behavior, Online J. Sci. Technol.
8 (2018) 50–56.
[47] R. Bhartiya, D.R. Sahoo, A. Verma, Modified damaged plasticity and variable confinement modelling of rectangular CFT columns, J. Constr. Steel Res. 176
(2021), 106426.
[48] L.M. e Silva, A.L. Christoforo, R.C. Carvalho, Calibration of concrete damaged plasticity model parameters for shear walls, Rev. Matér. 26 (2021).
[49] H. Behnam, J.S. Kuang, B. Samali, Parametric finite element analysis of RC wide beam-column connections, Comput. Struct. 205 (2018) 28–44.
[50] X. Yangjian, Z. Jianting, L. Yanling, X. Runchuan, Concrete plastic-damage factor for finite element analysis: Concept, simulation, and experiment, Adv. Mech.
Eng. 9 (2017) 1–10.
[51] CEB-FIP, Model Code 2010, Thomas Telford, London, 2010.
[52] J.Y. Wu, J. Li, R. Faria, An energy release rate-based plastic-damage model for concrete, Int. J. Solids Struct. 43 (2006) 583–612.
[53] Hibbitt, Karlsson, Sorensen, Abaqus analysis user’s manual Volume III: Materials version 6.10, Dassault Systémes, 2010.
[54] P.A. Vermeer, R. de Borst, Non-associated plasticity for soils, concrete and rock, Heron 29 (1984) 3–64.
[55] M. Hafezolghorani, F. Hejazi, R. Vaghei, et al., Simplified damage plasticity model for concrete, Struct. Eng. Int. (2017) 68–78.
[56] T. Yu, J.G. Teng, Y.L. Wong, S.L. Dong, Finite element modeling of confined concrete-II: Plastic-damage model, Eng. Struct. 32 (2010) 680–691.
[57] W. Demin, H. Fukang (2017) Investigation for plastic damage constitutive models of the concrete material. In: 6th International Workshop on Performance,
Protection & Strengthening of Structures under Extreme Loading (PROTECT2017), Guangzhou (Canton), China.
[58] D.A Hordijk, Tensile and tensile fatigue behavior of concrete; experiments, modeling and analyses, Heron 37 (1992) 9–79.
[59] A. Tomaszewicz (1984) Betongens Arbeidsdiagram. SINTEF Report N◦ STF 65A84065.
[60] A. Mohamad Ali, Bj Farid, A.I.M. Al-Janabi, Stress-strain relationship for concrete in compression madel of local materials, JKAU: Eng. Sci. 2 (1990) 183–194.
[61] E. Thorenfeldt, A. Tomaszewicz, J.J. Jensen (1987) Mechanical properties of high-strength concrete and application in design. in Proc Symposium on Utilization
of High-Strength Concrete, Tapir, Trondheim, Norway 149–159.

19