COMPDYN 2019

7th ECCOMAS Thematic Conference on

Computational Methods in Structural Dynamics and Earthquake Engineering
M. Papadrakakis, M. Fragiadakis (eds.)
Crete, Greece, 24–26 June 2019

CYCLIC NONLINEAR MODELING OF SEVERELY DAMAGED AND

RETROFITTED REINFORCED CONCRETE STRUCTURES
George Markou 1, Christos Mourlas 2, Reyes Garcia 3, Kypros Pilakoutas4, Manolis Papadraka-
kis2
1
Department of Civil Engineering, University of Pretoria, South Africa
e-mail: george.markou@up.ac.za
2
Department of Civil Engineering, National Technical University of Athens, Greece
{mourlasch, mpapadra}@central.ntua.gr
3
School of Engineering, The University of Warwick, UK
reyes.garcia@warwick.ac.uk
4
Department of Civil and Structural Engineering, The University of Sheffield, UK
k.pilakoutas@sheffield.ac.uk

Abstract

Advanced numerical methods for seismic assessment of existing substandard reinforced con-
crete (RC) structures have significantly evolved in the last two decades. Nonetheless, existing
numerical tools have numerous limitations (e.g. numerical instabilities, applicable only in 1D
or 2D problems, computationally demanding, etc.) and lack of objectivity and accuracy in
providing robust, numerically accurate and objective solutions. As a result, new numerical
approaches are necessary to solve some of these limitations.
This paper presents a detailed 3D modelling approach that solves some of the current model-
ing limitations. The approach is used to predict the experimental results from i) a substand-
ard RC joint with inadequate detailing subjected to cyclic loading, and ii) the same RC joint
rehabilitated and retrofitted with carbon fiber reinforced polymer (CFRP) sheets and subject-
ed to cyclic loading. It is shown that the proposed numerical method reproduces the experi-
mental results of both substandard and CFRP retrofitted specimens in a robust and
computationally efficient manner. Current research is investigating the behavior of more RC
components and full-scale retrofitted structures. This study contributes towards providing en-
gineers and researchers with an advanced analytical tool to study the cyclic nonlinear behav-
iour of substandard RC structures.

Keywords: RC Joint, Nonlinear analysis, Finite Element Modelling, CFRP strengthening,

deficient structures
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

1 INTRODUCTION
Many numerical models have been proposed during the last decades for simulating the me-
chanical behavior of reinforced concrete (RC) structures. Nevertheless, a numerical approach
combined with a constitutive model for concrete behavior that produces accurate and compu-
tationally efficient results for any RC structural member is still an open research subject. An
accurate, objective and computationally efficient approach will constitute a powerful tool for
any professional Civil Engineer and scientist that study the seismic performance of existing
RC structures or perform the design of retrofitting interventions of severely damaged RC
structures.
Many models that are found in the international literature can only describe certain aspects
of the concrete material behavior and their implementation is limited to cases of small practi-
cal interest since they place emphasis on post-peak material characteristics, which are de-
scribed by the introduction of several material parameters that have no physical meaning. A
detailed literature review in regards to the modeling of RC structures can be found in [1, 2].
The numerical approach which was proposed in [1], is the latest research work that emphasiz-
es on the ability of realistically predicting the behavior of concrete structures in a wide range
of problems ranging from static cyclic to dynamic loading conditions, performed in a compu-
tationally efficient way.
The model that was proposed by Mourlas et al. [1], presents a numerical procedure which
is based on the brittle nature of the material of concrete. The constitutive model is based on
the experimental findings of Kotsovos and Newman [3] and a mathematical approach de-
scribed in Kotsovos and Pavlovic [4]. This model is also used for the analysis of large-scale
structures as described in Markou et al. [5], where a 4-storey RC building specimen was mod-
eled through the Hybrid Modeling (HYMOD) approach [6, 7]. The under study RC building
[5] was retrofitted with RC infill walls and carbon fiber reinforced polymer (CFRP) jacketing
at its base, where three different load histories were applied until the frame reached its ulti-
mate capacity. According to the HYMOD approach, the shear-dominated areas (RC joints,
shear walls) are simulated by a detailed model that foresees the use of 8-noded hexahedral
elements for discretizing the concrete domains and embedded beam elements for discretizing
the steel reinforcement. The rest of the structure (bending dominated structural members) are
modeled through the use of the beam-column finite element. It has to be noted here that in
their work [5], the building was also retrofitted with RC infill walls and CFRP jacketing.
Even though the proposed modeling method in [5] managed to capture the experimentally de-
rived curves with high accuracy, the pinching effect was not reproduced accurately given that
the concrete material deterioration (due to the multiple imposed cycles) and the potential rebar
slippage at the base of the building were physical phenomena that were not numerically ac-
counted for based on the adopted constitutive material models [1].
In this research article, a parametric investigation of a newly proposed material factor is
performed, where it takes into account the loss of bond between the steel reinforcement and
the surrounding damaged concrete, thus reduces the energy dissipation of the RC specimen’s
overall response based on the level of concrete damage. The methodology used in this re-
search work foresees the validation of the proposed modeling method through the use of bare
and retrofitted RC joints that were tested experimentally by Reyes et al. [8]. The numerically
obtained P-δ curves are compared with both the experimental and numerical data in order to
test the accuracy of the proposed method.
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

2 MATERIAL MODELING

2.1 Concrete Material Modeling and Damage Factor

The constitutive modelling of concrete has to describe a realistic behavior of concrete un-
der generalized three dimensional states of stress as mathematically described in Kotsovos
and Pavlovic [4] . The constitutive relations take the following form:
 0   0( h)   0( d )  ( 0   id ) / (3Ks )
(1)
 0   0( d )   0 / (2Gs )
(2)
where Ks and Gs are the secant forms of bulk and shear moduli, respectively. The secant
forms of bulk, shear modulus and σid are expressed as functions of the current state of stress
which derived by regression analysis of the experimental data found in [3].
The model treats cracking with the smeared crack approach (see Fig. 1). In this way the
models simulate the geometrical discontinuity by the assumption of displacement continuity.
The crack opening strategy during each load increment foresees the use of the unified total
crack approach (UTCA) proposed by Lykidis and Spiliopoulos [9], which foresees that the
state of crack formation or closure is treated in a unified way within every Newton-Raphson
internal iteration. When a crack is opened, the material properties normal to the crack plane
are set to zero. The concrete material assumes that it loses all of its carrying capacity along
the vertical direction of the crack, where it behaves in a brittle manner. The expression of the
strength envelope of concrete is provided in Eq. 1 and it’s based on the Willam and Warkne
[10] formulae.
2 0c ( 02c   02e )cos    0c (2 0e   0c ) 4( 02c   02e )cos 2   5 02e  4 02c 02e
 0u  (3)
4( 02c   02e )cos2   (2 0e   0c )2

where the rotational variable θ defines the deviatoric stress orientation on the octahedral
plane. The τ0e (θ=0°) and τ0c (θ=60°) correspond to the state of σ1=σ2>σ3 (triaxial extension)
and σ1>σ2=σ3 (triaxial compression), respectively and are expressed analytically by experi-
mental data.

Figure 1: Local axes for the case of two cracks at a specific Gauss point. [5]

The material modeling procedure also uses a flexible crack closing criterion introduced in
Mourlas et al. [1]. This numerical criterion was found to play a significant role during the cy-
clic loading analysis for achieving a faster convergence ratio. According to the proposed crite-
rion in [1], the closure of cracks is expressed based on the following equation:
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

 n 1 
 i  a   cr   b  cr    cr (4)
 ntot 

where εi is the current strain in the i-direction that is normal to the crack plane and εcr is the
strain that causes cracking formation. Parameter b is the number of the imposed displacement
branch of the load history, while ncr and ntot are the numbers of increment that the crack is
formed at and the total number of increments that an imposed displacement branch is divided
into, respectively.
When the criterion of crack-closure is satisfied at a Gauss Point, which had prior to that on-
ly one crack formation, then a part of the stiffness is lost along the previously crack plane
(material deterioration) that was assumed to form in an orthogonal direction to the maximum
principle tensile stress. Therefore, the constitutive matrix takes the following form:
 an  (1  Dc )  (2Gt   ) an  (1  Dc )   an  (1  Dc )   0 0 0 
 
 an  (1  D c )   an  (1  D c )  (2 G t   ) an  (1  D c )   0 0 0 
 a  (1  D )   a  (1  D )   2 G   0 0 0 
Cl''   n c n c t

 0 0 0 as  (1  Dc )    Gt 0 0  (5)
 0 0 0 0 as  (1  Dc )    Gt 0 
 
 0 0 0 0 0 as  (1  Dc )    Gt 

where β is a shear retention factor, an and as are constants with recommended values of 0.25
and 0.125, respectively, based on the parametric investigation presented in [11]. The proposed
expressions of the constitutive matrix describes the anisotropic behavior of concrete at the lo-
cal coordinate system, therefore, it has to be transformed to the global system by using the
standard coordinate system transformation laws as follows:
C g = T T Cl T (6)
where T is the transformation matrix consisted by the direction cosines that define the relative
orientation of the local to global axis. The factor Dc is a damage factor proposed in [11], de-
scribing the accumulated energy loss due to the number of times a crack has opened and
closed. As it was suggested in [11], the damage factor has the following form:
      
11 cr    cr 
 
   max     max 
Dc  e (1a )/ fcc  e fcc
e f cc
(7)
where fcc is the number of times where a crack has closed and is updated in every iteration at
every Gauss Point. A schematic representation of Eq. 7 can be seen in Fig. 2.
During the nonlinear analysis, the constitutive matrix is calculate according to Eq. 5 when a
crack is closed at a Gauss point, which had previously one or two cracks. After the crack clo-
sure, the stresses are corrected through the use of the following expression:

σ i  σ i 1  C g  Δεi (8)
Finally, when all the cracks have been closed (of a previously cracked Gauss Point) and the
reduction factor a (Eq. 4) of one of the previous cracks is larger than 0.5, then the constitutive
matrix takes the following form:
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

C g΄  (1  Dc )  C g (9)

Figure 2: Schematic representation of the damage factor Dc (Eq. 7) as a function of the parameter a and fcc. [11]

2.2 Steel Material Modeling and Modified Damage Factor

According to any nonlinear behavior of cracked RC structural members, the level of dam-
age that occurs due to the opening of cracks affects the respective contribution of the steel re-
inforcement to the surrounding concrete areas. Consequently, the cracks directly affect the
internal force transfer mechanism from the rebars to the surrounding concrete domain, espe-
cially in the anchorage areas. Therefore, a modification of the steel stress-strain relation of
Menegotto-Pinto [12] is proposed herein that is related to the level of concrete cracking sur-
rounding the rebar element during the nonlinear cyclic analysis. The loss of bonding between
steel reinforcement with the surrounding cracked concrete can indirectly be taken into account
by reducing the stiffness contribution of steel reinforcement [11], thus simulate the pinching
effect. Hence, the introduction of a bond-slip model, which will require the definition of new
material parameters and an additional dof at each reinforcement bar, can be avoided.
Based on the proposed formulation presented in [11], the average of all parameters a (found
in Eqs. 4 and 10, and expressed by Eq. 11) at the 8 Gauss Points of any concrete hexahedral
element can be used to determine the level of damage of the concrete material as expressed
through Eq. 10.
Ds  1  aElement  (10)
where,
ncr

a i
aElement  1
, ncr is the number of cracked Gauss Points (11)
ncr
In the case of unloading, when the structure reaches its initial deformation, a material dete-
rioration of the steel reinforcement is computed based on the following proposed formulae:

Es '  (1  Ds ) Es (12)
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

The material deterioration is applied when the criterion σs εs < 0 is satisfied, which de-
scribes the situations when crack closures and re-openings occur, where the pinching phe-
nomena are excessive.

Figure 3: Menegotto-Pinto steel model that uses the modified formulation for parameters E΄(Eq. 12) and R΄(Eq.
14) with different values of the damage factor DR and DS.

Figure 4: Material model of the CFRP jacketing. [13]

Additionally, so as to capture the pinching effect at a material level, the reduction of pa-
rameter R (see Eq. 13) of the Menegotto-Pinto [12] model is proposed.
a1
R  R0  (13)
2  
It must be noted here that R0, a1 and a2 are experimentally determined parameters and as-
sumed to be 20, 18.5 and 0.15, respectively, in this study. By using the same concept de-
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

scribed above, the reduction factor DR is proposed in order to decrease the parameter R into R΄
which is calculated though the following expression:
R'  (1  DR ) R , where DR  Ds (14)
The material deterioration is once more applied when the criterion σs εs < 0 is satisfied.
The modified Menegotto-Pinto material model’s numerical response is illustrated in Fig. 3,
according to the proposed formulation. Before moving to the next section, it is noteworthy to
state at this point that the material model used to simulate the stress-strain relationship of the
CFRP jacketing foresaw a linear behavior until complete failure for both tension and com-
pression states, as shown in Fig. 4. As it can be seen, it was assumed that when the ultimate
stress level at any Gauss Point of the CFRP hexahedral elements was reached, the model fore-
sees for a complete loss of its capacity. According to the reported material properties [8], the
ultimate CFRP tensile strength was set to 4,140 MPa and the Young Modulus of elasticity
was set equal to 241 GPa. In addition to that, it was also assumed that the CFRP hexahedral
elements that were used to discretize in detail the jacketing, had full bonding with the con-
crete finite elements.

3 NUMERICAL IMPLEMENTATION
The numerical investigation foresaw the use of finite element meshes that will be described
in this section, followed by a discussion on the obtained numerical findings. All analyses were
performed by using a displacement control algorithm that adopted an energy convergence cri-
terion expressed in Eq. 15. The energy convergence tolerance was set equal to 10-5, whereas
the CPU used to perform all the analyses presented in this section, had a 3.7 GHz computing
power.
 usj Fst  t  Rst  t
eerr   tolerance (15)
 us1 Fst  t  Rst

Figure 5: Geometric and reinforcement details of the beam-column joints tested in Garcia et al. [8].
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

The numerical implementation foresaw the study of an exterior joint that was tested by
Garcia et al. [8] under extreme cyclic loading (see Fig. 5). The column had a cross-section of
260 x 260 mm and the longitudinal reinforcement were 16 mm in diameter. The beam was
designed to have a cross section of 260x400 mm and was also reinforced with 16 mm rebars,
as it can be depicted in Fig. 5.
Two types of bar anchorage detailing were examined for the top beam reinforcement as
shown in Fig. 5. The beam reinforcement of type A and B (Fig. 5c) were anchored into the
joint for a length of 220 mm (approximately 14db) without foreseeing for any hooks nor bends.
This short anchorage was designed to be insufficient in order to develop slippage, thus ex-
treme nonlinearities within the joint volume. Furthermore, the flexural strength of the RC col-
umn was designed to be the same as the beam’s, a design strategy that does not follow the
modern design philosophy that foresees a strong column-weak beam approach. The speci-
mens were designed to fail at the core, where there are no confining stirrups to provide the
joint with the ability to behave in a ductile manner. In order to avoid the shear failure outside
the core of the joint, 8 mm transverse stirrups were provided at both column and the beam
members spaced every 150 mm. The mean concrete compressive strength was reported to be
32 MPa and 31.3 MPa for specimens JA2 and JB2, respectively, where the respective tensile
splitting strength was 2.44 MPa and 2.41 MPa for the two specimens, respectively. The yield
and tensile strength of the steel reinforcement were reported to be fy = 612 MPa and fu = 726
MPa for the 8 mm rebars and fy = 551 MPa and fu = 683 MPa for the 16 mm rebars. Addition-
ally, the elastic modulus was found [8] to be equal to Es = 209 GPa.
The frame joints were subjected to different cyclic loading sets according to the experi-
ment’s setup. The initial loading history applied on the bare specimens consisted of three
push-pull cycles at drift ratios of ±0.25%, ±0.5%, ±0.75%, ±1.0%, ±1.5%, ±2.0%, ±3.0%,
±4.0% and ±5.0%. Furthermore, a second actuator applied a constant axial load equal to N =
150 kN on the column as illustrated in Fig. 6.

Figure 6: Test setup (units: mm) and instrumentation of RC joints. [8]

 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

According to the experimental results reported in [8], the bare specimens suffered from se-
vere damages, where excessive diagonal cracking and partial concrete spalling at the region of
the core had been observed, as shown in Fig. 7.

Figure 7: Crack patterns after the failure of JB2 bare RC specimen. [8]

Figure 8: RC joint JB2. (Left) Hexahedral and (Right) embedded rebar finite element meshes.

Figure 9: Retrofitted RC joint JB2RF. Concrete and CFRP hexahedral finite element meshes.
For the needs of this research work, specimen JB2 was studied through the use of the pro-
posed numerical model that uses the damage factors for concrete and steel reinforcement, as
presented in section 2.2. So as to study the mechanical behavior of specimen JB2, the con-
crete domain was discretized with 8-noded hexahedral finite elements and the steel rein-
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

forcement was discretized with Natural Beam-Column Flexibility-Based elements [14]. A to-
tal number of 174 concrete and 500 steel elements were used to discretize the RC joint, where
18 hexahedral steel elements (red color) were used at the areas of the boundary conditions in
order to prevent local failure (see Fig. 8). Furthermore, for the retrofitted RC joint (JB2RF),
156 hexahedral elements were used for the CFRP jacketing as shown in Fig. 9. The FE mesh
and material details for both RC joints are given in Tables 1 and 2, respectively.
Total Embedded
Hexahedral
a/a Model Number of Rebar
Elements
Nodes Elements
1 JB2 396 192 500
2 JB2RF 620 348 500
Table 1: FE mesh details of the two RC Joint models.

Young Hardening Yielding Stress / Compressive Shear re- Ultimate

Poisson
Material Modulus Modulus Tensile Strength* Strength maining Strain
Ratio
(GPa) (GPa) (MPa) (MPa) strength β ε
Concrete 20 - 2.41* 31.3 0.03 0.2 -
High Strength
30 - 3.91* 55.3 0.05 0.2 -
Concrete
Steel Rebars 190 1 551 (Ø16) / 612 (Ø8) - - 0.3 15%
CFRP 241 - 4,140* - - 0.3 1.7%
Table 2: Material details used in the FE model.

Figure 10: JB2 Joint. Comparison between the numerical and experimental curves.
The computed force-displacement curves for the two specimens (JB2 and JB2FR) are com-
pared with the corresponding experimental data in Figs. 10 and 11, respectively. As it can be
seen, the numerical results match very well with the experimental data, where the stiffness
and the resulted load-carrying capacity of the specimen were predicted in an accurate manner.
The numerical results show that the concrete damage factors managed to capture the stiffness
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

and load-capacity degradation, where the pinching characteristics for both specimens were
described with an acceptable accuracy. It is evident that the proposed modeling approach
manages to captured this highly complicated physical phenomenon through the use of the
modified Menegotto-Pinto model and by adopting the proposed damage factors DS and DR
given in section 2.2 by Eqs. 10 and 11, respectively.

Figure 11: JB2FR Joint. Comparison between the numerical and experimental curves.
According to the numerical findings, the FE model developed for simulating the retrofitted
specimen JB2FR not only managed to capture numerically the hysteretic behavior of the spec-
imen and its strength’s corresponding degradation, but it was also able to predict the stiffness
degradation of a structural joint that foresaw three different solid domains:
a. Old damaged concrete,
b. New high-strength concrete found in the joint and
c. CFRP jacketing.
It must be noted herein that the numerical analysis failed to converge during the final cycle
of imposed displacements due to excessive cracking of the concrete domain within the joint
domain. Finally, the proposed numerical model for the case of the retrofitted specimen was
able to predict the overall strength enhancement with a deviation that was less than 5%.

4 CONCLUSIONS
In this research work, a 3D detailed modeling approach is used for the simulation of bare
and CFRP retrofitted RC joints. The proposed numerical model has been integrated with a
newly proposed damage factor for the steel material that is directly connected to the number
of opening and closing of concrete cracks that occurs during the nonlinear analysis. This dam-
age factor, which is an extension of the damage factor proposed in [11], was developed to ac-
count the loss of bonding between steel reinforcement and the cracked surrounding concrete
within areas that are expected to undergo significant cracking.
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

According to the numerical investigation’s findings, the proposed model managed to cap-
ture the hysteretic behavior of both the bare and CFRP retrofitted RC joints, which were sub-
jected to ultimate limit cyclic loading conditions. The modification of the steel reinforcement
model was able to reproduce the severe pinching effects that occurred during the cyclic load-
ing of the bare joint. Additionally, the corresponding extreme pinching effect of the retrofitted
RC joint was also numerically reproduced with an acceptable accuracy, where the numerical
and experimental curves were found to be in a good agreement. Furthermore, for the case of
the CFRP retrofitted specimen, the proposed modeling approach managed to reproduce the
overall strength enhancement of the joint with an accuracy that foresaw a deviation of less
than 5% compared to the experimental data.
Future research work foresees the implementation of the proposed algorithm in predicting
the mechanical behavior of additional specimens that are characterized by extreme pinching
phenomena, while use the developed technology in the parametric investigation of the me-
chanical response of full-scale structures that undergo extreme cyclic loading. Finally, the
proposed modeling approach will be used for the study of the effects that different interven-
tions derive in-terms of the overall mechanical behavior and strength enhancement of RC
structural framing systems.

REFERENCES
[1] Mourlas C., Papadrakakis M., Markou G. A computationally efficient model for the cyclic
behavior of reinforced concrete structural members. Engineering Structures, Vol. 141, pp. 97–
125, 2017.
[2] Lykidis G.C., Spiliopoulos K.V. An efficient numerical simulation of the cyclic loading
experiments on RC structures. Computers and Concrete, Vol. 13(3), pp. 343–59, 2014.
[3] Kotsovos M.D., Newman J.B.. Behavior of Concrete Under Multiaxial Stress. ACI Journal
Proceedings, Vol. 74(9), pp. 213–23, 1977.
[4] Kotsovos M.D., Pavlovic M. Structural concrete : finite-element analysis for limit-state design.
T. Telford; 1995.
[5] Markou G., Mourlas C., Bark H., Papadrakakis M. Simplified HYMOD non-linear simulations
of a full-scale multistory retrofitted RC structure that undergoes multiple cyclic excitations –
An infill RC wall retrofitting study. Engineering Structures, Vol. 176, pp. 892–916, 2018.
[6] Markou G., Papadrakakis M. A simplified and efficient hybrid finite element model (HYMOD)
for non-linear 3D simulation of RC structures. Engineering Computations, Vol. 32(5), pp.
1477–524, 2015.
[7] Markou G., Mourlas C., Papadrakakis M. Computationally Efficient and Robust Nonlinear 3D
Cyclic Modeling of RC Structures Through a Hybrid Finite Element Model (HYMOD).
International Journal of Computational Methods, 2018.
[8] Garcia R., Jemaa Y., Helal Y., Guadagnini M., Pilakoutas K. Seismic Strengthening of
Severely Damaged Beam-Column RC Joints Using CFRP. Journal of Composites for
Construction, Vol. 18(2), pp. 04013048, 2014.
[9] Spiliopoulos K.V., Lykidis G.C. An efficient three-dimensional solid finite element dynamic
analysis of reinforced concrete structures. Earthquake Engineering & Structural Dynamics, Vol.
35(2), pp. 137–57, 2006.
[10] Willam K., Warnke E.P. Constitutive model for the triaxial behavior of concrete. Seminar on
concrete structures subjected to triaxial stresses, Instituto Sperimentale Modeli e Strutture,
Bergamo, Paper III-1., 1974.
[11] Mourlas C., Markou G., Papadrakakis M. Accurate and computationally efficient nonlinear
 George Markou, Christos Mourlas, Reyes Garcia, Kypros Pilakoutas and Manolis Papadrakakis

static and dynamic analysis of reinforced concrete structures considering damage factors.
Engineering Structures, Vol. 178, pp. 258–85, 2019.
[12] Menegotto M., Pinto P.E. Method of analysis for cyclically loaded reinforced concrete plane
frames including changes in geometry and non-elastic behavior of elements under combined
normal force and bending. Proceedings, IABSE Symposium on Resistance and Ultimate
Deformability of Structures Acted on by Well Defined Repeated Loads, Lisbon, Portugal, 15–
22., 1973.
[13] Markou, G., AlHamaydeh, M. (2018), 3D Finite Element Modeling of GFRP-Reinforced
Concrete Deep Beams without Shear Reinforcement, International Journal of Computational
Methods, 15(1), pp. 1-35.
[14] Markou, G., Papadrakakis, M., “Accurate and Computationally Efficient 3D Finite Element
Modeling of RC Structures”, Computers & Concrete, 12 (4), pp. 443-498, 2013.