16.

D-A-CH Tagung Erdbebeningenieurwesen & Baudynamik (D-A-CH 2019)

26. und 27. September 2019
Universität Innsbruck

PREDICTING THE NONLINEAR RESPONSE AND

DAMAGE PATTERNS OF BARE AND INFILLED
REINFORCED CONCRETE FRAMES USING THREE-
DIMENSIONAL FINITE ELEMENT MODELS

Mohammad Reza Azadi Kakavand 1, Ertugrul Taciroglu 2 and Günter Hofstetter3

1
Unit of Strength of Materials and Structural Analysis, University of Innsbruck, Austria, E-Mail-Address:
[email protected]
2
Civil and Environmental Engineering Department, University of California, Los Angeles, USA, E-Mail-Address:
[email protected]
3 Unit of Strength of Materials and Structural Analysis, University of Innsbruck, Austria, E-Mail-Address:

[email protected]

ABSTRACT
In this study, the performance of three-dimensional finite element (FE) models to predict the behavior
of two bare and infilled reinforced concrete (RC) frames is evaluated. The two-story one-bay frames
were built at the scale of 1:3 and tested under reversed cyclic loads. To simulate the tested frames, FE
models are generated using a combination of continuum and truss elements for representing concrete
and reinforcing steel, respectively. The nonlinear responses of the constituents are described by a
damage-plasticity model for concrete and an elastoplastic model for reinforcing steel. Pushover analyses
of both frames were performed and the obtained results demonstrate the capability of the FE model to
capture the lateral capacities and ductilities of the tested frames, as well as damage evolution in concrete.
While both the bare frame’s ductile behavior and the infilled frame’s strength degradation are predicted
well by the FE models, the observed cracking pattern in the infill is not completely captured. The reason
for this lack of agreement between the FE model predictions and experimental observations can be
attributed to the fact that the employed concrete damage-plasticity model is only applicable for
simulating the concrete behavior under monotonic loading.

Keywords: reinforced concrete structure, constitutive model, damage-plasticity model, finite element
method, pushover analysis, earthquake engineering.

1. INTRODUCTION
Numerical models of reinforced concrete (RC) members/structures are typically developed based on
two approaches—namely, using either structural elements (beam elements), or continuum elements
combined with truss elements, which respectively represent concrete and reinforcing steel. The former
approach is computationally cheaper and is therefore suitable for global/structural scale analyses, while
the latter approach is usually only adopted to explore and optimize design details and characterize
behavior under complex loading scenarios as it can provide highly detailed information at sub-member-
scales such as concrete spalling, bar pull-out and buckling, etc. The application of computationally
expensive continuum FE models is therefore typically viewed as being a much less time-consuming and
relatively inexpensive solution for simulating the behavior of RC members in comparison to conducting
experimental studies.
The present study investigates the applicability of refined finite element (FE) models for simulating the
behavior of reinforced concrete (RC) structures under monotonic loading (i.e., pushover analyses). The
results obtained from numerical simulations are compared to measurements from various tests to explore
and demonstrate the capabilities of refined FE models, and to explore the potential shortcomings of
pushover analyses in earthquake engineering.
In the past decades, many numerical and experimental studies have been conducted on structural
components, particularly RC columns. In these tests, the effects of material and loading parameters on
the structural response during earthquakes were studied, such as the impacts of the magnitude of axial
loading [1-3], variable or constant axial loads [4-5], multiaxial versus uniaxial load cases [6-8], cyclic
loading patterns [9-10], and the transverse reinforcement ratios [10-11]. Several key numerical and
experimental studies demonstrated that the initial axial load and transverse reinforcement ratios
significantly affect the lateral displacement capacity of RC columns [12-17]. Several tests have been
conducted to investigate the behavior of bare and infilled frames under cyclic loading. Pinto et al. [18]
conducted pseudo-dynamic tests on two full-scale one-bay by three-bay by four story 3D frames, bare
and infilled with masonry. Altin et al. carried out cyclic tests on seven one-bay two-story 1:3-scale 2D
frames, bare and infilled with reinforced concrete so that the frames in both two studies were designed
with non-seismic detailing [19]. The latter test is chosen for the present study due to the availability of
detailed documentation of the test setup and the measurements made.
As mentioned above, the present study considers the continuum approach for modeling the mechanical
behavior of reinforced concrete. As the constitutive model for concrete is an essential ingredient for
nonlinear simulations of RC structures, many 3D constitutive models have been proposed up to now,
and a good overview and comparison of these models can be found in [20]. Herein, the damage-plasticity
model proposed in [21] is employed as the constitutive model for concrete due to its good performance
demonstrated in prior studies [20, 22]. In what follows, the employed damage-plasticity model is briefly
described first, and the bare and infilled frame tests presented in [19] are discussed. Next, details of the
FE models developed in the present study for pushover analyses are provided. Finally, comparisons of
computed and measured results are presented and the capabilities and shortcomings of pushover
analyses with refined FE models are discussed.

2. CONSTITUTIVE MODELS FOR CONCRETE AND REINFORCING STEEL

In this study, the behavior of concrete is described using a damage-plasticity model by Grassl and Jirásek
[21]. The model represents isotropic hardening and softening material behavior, inelastic strains and
stiffness degradation due to damage by means of the combined theories of plasticity and continuum
damage mechanics. The nominal stress-strain relation reads

𝝈 = (1 − 𝜔) ℂ ∶ (𝜺 − 𝜺p ), (1)

with 𝝈 denoting the nominal stress tensor, 𝜔 the scalar isotropic damage variable ranging from 0
(undamaged material) to 1 (fully damaged material), ℂ the fourth order elastic stiffness tensor, 𝜺 and 𝜺p
the total and plastic strain tensors, respectively. The plastic part of the model is formulated in the
̅ given as
effective stress space, with the effective stress tensor 𝝈

1
̅=
𝝈 𝝈. (2)
1−𝜔
It is worth noting here that more information about the employed concrete model can be found in [22].
The performance of the model for biaxial stress states is evaluated, as illustrated in Fig. 1(a). In this
regard, the results obtained by the numerical model are compared to the biaxial compression tests
conducted by Kupfer et al. [23]. The material parameters are Young’s modulus 𝐸𝑐 = 32 GPa, Poisson’s
ratio 𝜈𝑐 = 0.18, uniaxial compressive strength 𝑓𝑐′ = 32.8 MPa and uniaxial tensile strength 𝑓𝑡𝑢 = 3.3 MPa.
The lateral confining pressure varies between 50% and 100% of the uniaxial compressive strength. It
can be followed from Fig. 1(a) that the model can capture the increase of strength and ductility capacities
with increasing the lateral pressure very well. Furthermore, the capability of the model to predict the
uniaxial tension response is evaluated using the experimental test reported by Li et al. [24], as shown in
Fig. 1(b). The material parameters for the tested specimen are set as 𝐸𝑐 = 35 GPa, 𝜈𝑐 = 0.2, 𝑓𝑐′ = 52.2
MPa, 𝑓𝑡𝑢 = 4.16 MPa and the specific mode I fracture energy 𝐺𝐹𝐼 = 0.15 N/mm. It should be mentioned
that the Poisson’s ratio 𝜈𝑐 is assumed on the basis of similar concrete grades and the specific mode I
fracture energy 𝐺𝐹𝐼 was estimated using the fib model code for concrete structures [25]. Fig. 1(b)
demonstrates that the model can capture the softening response in the post-peak region very well. It is
worthwhile to note that the characteristic length to compute the softening region of stress-strain diagrams
is employed as 100 mm for both biaxial compression and uniaxial tension tests, which follows from the
specimen dimensions provided in the test reports. A 1D elastic-plastic constitutive model is employed
for describing the behavior of steel reinforcements.

(a) (b)
Figure 1: Comparison of observed and predicted concrete behavior in (a) biaxial compression tests by
Kupfer et al. (1969) and (b) uniaxial tension test by Li et al. (1998).

3. EXPERIMENTAL STUDIES
Two two-story and one-bay bare and infilled RC frames on the scale of 1:3 were tested at the Structural
Mechanics Laboratory of Gazi University in Turkey [19]. The frames were subjected to cyclic lateral
loading by means of hydraulic jacks in the absence of axial load. The geometry and reinforcement details
for the bare frame are illustrated in Fig. 2. It should be noted that the geometry and reinforcement details
of beams and columns for the infilled frame are the same as the bare frame, and the thickness of the
infill is 50 mm. The information on the reinforcement of the infill can be found in [19]. In this
experimental study, the bare and infilled frames are characterized as specimen 1 and 2, respectively.
 Figure 2: Geometry and reinforcement details for the bare frame tested by Altin et al. [19].
The mechanical properties of concrete material and reinforcing steel are presented in Tables 1 and 2,
respectively. In Table 2, Es denotes the Young’s modulus of the reinforcing steel, and fy and fu denote
the yield and ultimate stresses of bars, respectively.

Table 1. The mechanical properties of concrete material.

Type of Type of f'c Ec υc GFI
specimen element (MPa) (MPa) - (N/mm)
Bare frame Frame 14.8 24100 0.2 0.15
Infilled frame Frame 15.0 24200 0.2 0.15
Infill 29.8 29600 0.2 0.15

Table 2. The mechanical properties of reinforcing steel.

Bars
Es fy fu
diameter
(MPa) (MPa) (MPa)
(mm)
4 190000 326 708
8 190000 592 964
10 190000 475 689

4. FINITE ELEMENT MODELS OF THE TESTED FRAMES

The FE models are discretized using 3D isoparametric continuum elements with quadratic shape
functions and full integration for representing concrete material. 1D truss elements with quadratic shape
function are employed for longitudinal and transverse reinforcements. The reinforcing bars are
embedded in the continuum elements and a perfect bond between concrete and reinforcing steel is
assumed. In the numerical simulations, two mesh sizes of 25 mm (fine mesh) and 50 mm (coarse mesh)
are used to demonstrate mesh-independence of the computed results. The coarse FE-mesh for the bare
and infilled frames consists of 900 and 1680 3D continuum elements with altogether 17,526 and 35,154
degrees of freedom, respectively. The fine FE-mesh for the bare and infilled frames is characterized by
7200 and 13,440 3D continuum elements with altogether 111,840 and 218,766 degrees of freedom,
respectively. Fig. 3 shows the coarse finite element discretization of two bare and infilled frames. In
Fig. 3, for each frame the left half shows the discretization of the concrete, whereas the right half displays
the discretized steel reinforcement.
Consequently, the pushover analyses of two tested frames were performed by the finite element software
package ABAQUS [26]. In the numerical simulations, first, the gravity load is applied and then the
pushover analyses were performed by increasing the lateral displacement of the exterior joints of the
first and second-level in an incremental manner up to failure. The applied lateral displacement ratio
between the first and second story is 1:2.

(a) (b)
Figure 3: FE discretization of the tested frames (a) bare frame, (b) infilled frame.

5. COMPARISON OF RESULTS: FE SIMULATIONS VS EXPERIMENTS

The comparison of the predicted base shear by numerical simulations with the measured ones during
tests for both bare and infilled frames in terms of the lateral drift is shown in Fig. 4. It can be seen from
Fig. 4(a) and 4(b) that the numerical model can capture the initial stiffness, the ultimate base shear and
the lateral ductility for both tested frames very well and the infilled frame’s strength degradation quite
well.

(a) (b)
Figure 4: Comparison of the measured and computed based shear for (a) bare frame, (b) infilled frame
in terms of the lateral drift.
Fig. 5 compares the predicted damage by pushover analyses with the observed ones during the cyclic
tests for the bare and infilled frames. The first row of the plots in Fig. 5 reveals the predicted and
observed damage in the bare frame. It is observed that the beam-column joints and the bottom of the
first-story columns in the tested bare frame experienced damage during cyclic lateral loading, which is
captured very well by the FE models under monotonic loading. Albeit, the observed damage pattern in
the infilled frame was not completely predicted by numerical simulations since the latter were restricted
to pushover analyses as the employed concrete damage-plasticity model is solely applicable for
monotonic loading.

Figure 5: Comparison of the predicted damage in bare and infilled frames by numerical simulations with the
ones observed during tests [19].

6. SUMMARY AND CONCLUSION

In this study, the performance of refined finite element models to predict the behavior of two types of
RC frames was evaluated by means of pushover analyses. The aforementioned FE models are
characterized by discretizing the concrete and reinforcement using 3D continuum and 1D truss elements.
The nonlinear behavior of concrete is described by an advanced 3D constitutive model, and a 1D elastic-
plastic constitutive model is used for the steel material. The FE-models were developed for two bare
and infilled frames tested by Altin et al. (2008). The frames were subjected to cyclic loads without axial
load. The obtained results by numerical simulations were compared to the experimental results by
presenting the base shear in terms of the lateral drift and by plots illustrating the evolution of concrete
damage pattern. The comparison of predicted and observed results demonstrates the capability of the
FE models to capture the initial stiffness, ultimate lateral strength capacity, lateral ductility capacity and
the strength degradation of the RC infill, the location and evolution of concrete damage. However, the
damage pattern of infill is not totally predicted by the FE-model since only pushover analyses were
performed as the concrete model is only applicable for monotonic loading.

REFERENCES
[1] Lam, S. S. E., Wu, B., Wong, Y. L., Wang, Z. Y., Liu, Z. Q. & Li, C. S. (2003) Drift capacity of
rectangular reinforced concrete columns with low lateral confinement and high-axial load. Journal
of Structural Engineering, 129(6), 733-742.
[2] Ousalem, H., Kabeyasawa, T. & Tasai, A. (2004) Evaluation of ultimate deformation capacity at
axial load collapse of reinforced concrete columns. In: Proceedings of 13th World Conference on
Earthquake Engineering.
[3] Matamoros, A. B., Matchulat, L. & Woods, C. (2008) Axial load failure of shear critical columns
subjected to high levels of axial load. In: Proceedings of the 14th World Conference on Earthquake
Engineering.
[4] Lejano, B. A., Shirai, N., Adachi, H., Ono, A. & Amitu, S. (1992) Deformation properties and shear
resistance mechanism of reinforced concrete column with high and fluctuating axial force. In: Tenth
World Conference on Earthquake Engineering, Balkema, Rotterdam. Taylor & Francis, London,
3007-3012.
[5] Esmaeily, A. & Xiao, Y. (2004) Behavior of reinforced concrete columns under variable axial
loads. ACI Structural Journal, 101(1), 124-132.
[6] Staaciouglu, M. (1984) Reinforced concrete columns subjected to uniaxial and biaxial load
reversals. In: Proceedings of the 8th World Conference on Earthquake Engineering.
[7] Yoshimura, M. & Tsumura, K. (2000) Shear-failing reinforced concrete columns subjected to
multi-axial loading. In: Proceedings of the 12th World Conference on Earthquake Engineering.
[8] Osorio, E., Bairan, J. M. & Marí, A. R. (2017) Analytical modeling of reinforced concrete columns
subjected to bidirectional shear. Engineering Structures, 138, 458-472.
[9] Lynn, A. C., Moehle, J. P., Mahin, S. A. & Holmes, W. T. (1996) Seismic evaluation of existing
reinforced concrete building columns. Earthquake Spectra, 12(4), 715-739.
[10] Henkhaus, K., Pujol, S. & Ramirez, J. (2013) Axial failure of reinforced concrete columns damaged
by shear reversals. Journal of Structural Engineering, 139(7), 1172-1180.
[11] Wibowo, A., Wilson, J. L., Lam, N. T. K. & Gad, E. F. (2014) Drift performance of lightly
reinforced concrete columns. Engineering Structures, 59, 522-535.
[12] Elwood, K. J. (2004) Shake table tests and analytical studies on the gravity load collapse of
reinforced concrete frames, PhD Dissertation, University of California, Berkeley.
[13] Zhu, L., Elwood, K. J. & Haukaas, T. (2007) Classification and seismic safety evaluation of existing
reinforced concrete columns. Journal of Structural Engineering, 133(9), 1316-1330.
[14] Kakavand, M. R. A. (2012) Limit state material manual. Available at www.opensees.berkeley.edu,
University of California, Berkeley.
[15] Yavari, S., Elwood, K. J., Wu, C. L., Lin, S. H., Hwang, S. J. & Moehle, J. P. (2013) Shaking table
tests on reinforced concrete frames without seismic detailing. ACI Structural Journal, 110(6),
1001.
[16] Farahmand, H., Kakavand, M. R. A., Tafreshi, S. T. & Hafiz, P. H. (2015) The effect of mechanical
and geometric parameters on the shear and axial failures of columns in reinforced concrete
frames. Ciência e Natura, 37, 247-259.
[17] Kakavand, M. R. A. & Allahvirdizadeh, R. (2018) Enhanced Empirical Models for Predicting the
Drift Capacity of Less Ductile RC Columns with Flexural, Shear or Axial Failure Modes. Frontiers
of Structural and Civil Engineering. (Accepted for publication.)
[18] Pinto, A., Varum, H. & Molina, J. (2002) Experimental assessment and retrofit of full-scale models
of existing RC frames. Twelfth European Conference on Earthquake Engineering.
[19] Altin, S., Anil, Ö., & Kara, M. E. (2008). Strengthening of RC nonductile frames with RC infills:
An experimental study. Cement and Concrete Composites, 30(7), 612-621.
[20] Valentini, B. & Hofstetter, G. (2013) Review and enhancement of 3D concrete models for large‐
scale numerical simulations of concrete structures. International Journal for Numerical and
Analytical Methods in Geomechanics, 37(3), 221-246.
[21] Grassl, P. & Jirásek, M. (2006) Damage-plastic model for concrete failure. International Journal
of Solids and Structures, 43 (22), 7166–7196.
[22] Kakavand, M. R. A., Neuner, M., Schreter, M. & Hofstetter, G. (2018) A 3D continuum FE-model
for predicting the nonlinear response and failure modes of RC frames in pushover analyses. Bulletin
of Earthquake Engineering, 16(10), 4893-4917.
[23] Kupfer, H. B. & Gerstle, K. H. (1973) Behavior of concrete under biaxial stresses. Journal of the
Engineering Mechanics Division, 99(4), 853-866.
[24] Li, Z., Li, F., Chang, T.P. & Mai, Y.W. (1998) Uniaxial tensile behavior of concrete reinforced
with randomly distributed short fibers. Materials Journal, 95(5), 564-574.
[25] Taerwe L, Matthys S, et al (2010) CEP-FIB model code for concrete structures. Ernst & Sohn,
Wiley.
[26] ABAQUS (2014) Abaqus v6.14 documentation, Dassault Systèmes, Simulia Corporation.