3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020
Paper N° C003956 (Abstract ID)
Registration Code: S-A01433

SIMPLIFIED ANALYTICAL MODELS FOR EVALUATING

RETROFITTED REINFORCED CONCRETE SHEAR WALLS

P. Okumus(1), S. Basereh(2)
(1) Associate Professor, Dept. of Civil, Structural and Environmental Engineering, Univ. at Buffalo, [email protected]
(2) Graduate Research Assistant, Dept. of Civil, Structural and Environmental Engineering, Univ. at Buffalo, [email protected]

Abstract
Reinforced Concrete (RC) rocking shear walls have been studied over the past 20 years as seismic-resistant
components in buildings for minimizing damage and shortening downtimes after an earthquake. Although
design procedures, seismic response and macro modeling of such systems have been evaluated for newly built
structures, uses of RC rocking walls for retrofit of existing non-ductile, cantilever RC shear walls have received
limited attention. Past earthquakes showed that non-ductile, cast-in-place shear walls may experience
undesirable failure modes, rendering them unusable. This study investigates rocking as a retrofit method for
existing non-ductile cantilever RC shear walls through analytical modeling and evaluates simplified analytical
modeling techniques for their suitability for efficient nonlinear response history analyses. Detailed 3D finite
element modeling, fiber modeling, and lumped plasticity based modeling techniques are compared in terms of
global response parameters. The retrofit method follows the concepts of weakening and self-centering.
Weakening involves saw-cutting wall base across concrete and/or across vertical reinforcing bars. Self-
centering is provided by added vertical post-tensioning. The overall goal of the retrofit is to replace the non-
ductile response of the original wall by the rocking response, leading to minimized damage to the building
after an earthquake. To implement the retrofit method, an RC shear wall was designed so that it did not meet
the requirements of ACI 318-14 on vertical spacing of confining bars in the boundary elements, extension of
the confining reinforcement into the foundation and spacing of transverse reinforcement in the web. The wall
was then retrofitted using the concepts of weakening and self-centering. The RC shear wall was analyzed under
cyclic displacements up to 2% drift ratio. Detailed and simplified finite element modeling of RC walls were
conducted using commercial and open-source (OpenSees) analysis software, respectively. Both pre- and post-
retrofit walls were simulated. Simplified models used a lumped plasticity approach with a monolithic beam
analogy and a distributed plasticity approach with fiber elements. Modeling approaches were compared in
terms of energy dissipation, lateral strength, residual displacement, and secant stiffness. Detailed 3D finite
element model was used as a baseline for comparison. Both lumped plasticity and fiber models were able to
capture initial stiffness and lateral strength with reasonable accuracy. On the other hand, lumped plasticity
model overestimated energy dissipation capacity of the wall. Residual displacement prediction of fiber model
was closer to that of detailed 3D finite element analysis.
Keywords: Non-ductile RC shear wall; Rocking; Weakening; Self-centering; Earthquake resiliency

1. Introduction
Use of rocking shear walls as a primary lateral load-resisting system for new buildings started with the Precast
Seismic Structural Systems (PRESSS) program [1]. A 5-story building having a self-centering precast frame
and a jointed rocking wall system in orthogonal directions was designed and tested. After the PRESSS
program, single or jointed precast rocking walls [2-4], and precast walls with end columns (PreWEC) [5,6]
equipped with different types of energy dissipating mechanisms have been introduced. Design methods, testing
procedures, and acceptance criteria for precast rocking shear walls have been published through American
Concrete Institute’s (ACI) Innovation Task Groups (ITG) [7,8]. However, use of rocking walls as a retrofit
solution for non-ductile, cast-in-place RC shear walls received very little attention. The study presented in this
paper aims to assess the feasibility of converting non-ductile RC shear walls to rocking walls as a retrofit

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

solution. This assessment was done through detailed 3D finite element analysis (FEA) and simplified analytical
models.
The retrofit method combines concepts of selective weakening and self-centering to convert non-ductile
cast-in-place RC shear walls to more ductile hybrid rocking walls. Here, “hybrid” refers to rocking walls which
have both unbonded post-tensioned strands (for self-centering) and yielding mild steel reinforcement (for
energy dissipation). In the retrofit method, walls are weakened through cutting some of the vertical reinforcing
bars and concrete at wall base, enabling the wall to rock about the base. This intervention is expected to weaken
the wall. The walls are also provided with external post-tensioning to facilitate self-centering. Self-centering
ensures that the wall is back to the plumb after an earthquake and lowers damage due to residual drift. Energy
is dissipated by hysteretic yielding of uncut vertical reinforcing bars at wall base. These bars are chosen to be
located closer to wall mid-length.
This paper discusses modeling of pre- and post-retrofit RC shear walls using detailed 3D FEA and
simplified models using both lumped plasticity and fiber models. Pre- and post-retrofit walls were analyzed
under quasi-static cyclic displacements up to 2% drift. In the absence of test data, a detailed finite element
model was used as a basis to compare all models in terms of energy dissipation, lateral strength, residual drift,
and secant stiffness. A 2D frame including the wall was then modeled using simplified and computationally
effective approaches and nonlinear response history analyses (NRHA) were conducted. Results of NRHA were
compared in terms of peak inter-story drift, peak residual drift, peak floor acceleration, and base shear.

2. Description of pre- and post-retrofit walls

A slender RC shear wall was designed to reflect pre-1970 construction practices. The RC shear wall was a 1:3
scale model of the main lateral load resisting mechanism of a 3 story building located in Rowland Heights,
California, USA. It had a height-to-length ratio of 2.1. Height, length and thickness of the wall were 126, 60,
and 6 in. (3200, 1524, and 152 mm), respectively. Axial load on the wall is 87 kips (387 kN) (i.e., 0.06Agf’c,
where Ag is the gross cross-sectional area of the wall and f’c is the uniaxial compressive strength of concrete).
The wall met the requirements of ACI 318-14 [9] in terms of longitudinal reinforcement ratio, while it was not
compliant with ACI 318-14 in terms of vertical spacing of confining bars in the boundary element, extension
of the confining reinforcement into the foundation and transverse reinforcement ratio in the web. Figure 1a
shows the elevation and cross-section views of the pre-retrofit wall.
The non-code-compliant RC shear wall was retrofitted to obtain a more desirable failure mode governed
by rigid body rotation (i.e., rocking). Preliminary design of the retrofitted wall was done per recommendations
of ACI ITG 5.1-07. ACI ITG 5.1-07 specifies low residual drift and sufficient energy dissipation for rocking
walls with unbonded post tensioning strands. Low residual drift was obtained by limiting the ratio of moment
contribution from energy dissipating reinforcing bars to the total moment capacity from both energy
dissipating reinforcing bars and axial load (i.e., gravity load plus post-tensioning force). Equation 1 shows the
recommended value of moment contribution ratio. Energy dissipation requirement was met by providing a
relative energy-dissipation ratio of 12.5%. Relative energy-dissipation ratio is an index of damping capability
and is the ratio of wall energy dissipation in the load cycle of interest to the theoretical maximum possible
energy dissipation for the same load cycle.
𝑀𝐸𝐷
≤ 0.4 (1)
𝑀𝐸𝐷 + 𝑀𝐴𝑥𝑖𝑎𝑙
Design variables were the amount of initial post-tensioning force and uncut reinforcing bars. The amount
of external unbonded post-tensioning strands and initial post-tension force was determined so that 1) the
moment contribution ratio requirement was met, 2) the wall had sufficient lateral strength, and 3) strands
yielded at the maximum displacement capacity. The amount of uncut reinforcing bars was selected to satisfy
the relative energy dissipation ratio requirement. The number of post-tensioning strands was determined as
two (i.e. one strand on each side of the wall). The strands were placed concentrically at mid-length of the wall,
with the initial prestressing stress of 0.55Fpy, where Fpy denotes the yield strength of post-tensioning strands.
Area of each post-tensioning strand was 0.217 in2 (140 mm2). The resultant initial post-tensioning force was

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

58 kips (258 kN). The amount of uncut reinforcing bars was set to 25% of that of the pre-retrofit wall (i.e., 4
reinforcing bars nearest to wall mid-length were left uncut). Figure 1b shows the elevation view of post-retrofit
wall.

(a) (b)
Figure 1. Elevation view of the walls a) Pre-retrofit b) Post-retrofit

3. Finite element modeling and analysis

The shear wall was modeled using a general-purpose nonlinear finite element software, LS-DYNA. A smeared
crack material model called Winfrith concrete model (MAT 084 in LS-DYNA) was used to model concrete
under cyclic actions [10]. Uniaxial compressive strength of concrete was set to 4 ksi. To account for post-peak
compression softening of concrete, element erosion technique was used. Concrete elements reaching the
maximum principal strain of 0.04 were removed from the model. Modulus of elasticity of concrete was 3605
ksi (24856 MPa) (per ACI 318-14). Tensile strength of concrete and crack opening at zero tensile strength
were extracted from fib Model Code [11]. Post-peak tensile behavior was considered to be linear.
Piecewise linear plasticity model (MAT024 in LS-DYNA) can capture the actual stress-strain behavior
of steel and account for fracture of reinforcing bars. This model was used to model reinforcing bars explicitly.
Conventional Grade 60 ASTM A615 bars common in the pre-1970 construction were considered for modeling
purposes. Buckling of reinforcing bars was not considered. Perfect bond was assumed between reinforcing
bars and concrete elements by merging their shared nodes. The foundation was assumed to remain linear
elastic. Cyclic lateral displacements up to 2% drift were applied at a linear elastic concrete cap beam located
on top of the wall.
To simulate the post-retrofit shear wall, a wall base cut, reinforcement cut and post-tensioning strands
had to be added to the model of the pre-retrofit shear wall. To model the base cut, a mortar-based, surface-to-
surface contact was utilized between the wall base and the foundation. The friction coefficient for the contact
was set to 0.6. Post-tensioning strands were simulated using elastic cable/discrete beam material (MAT071 in
LS-DYNA). This material is used when elastic tensile behavior is expected. Modulus of elasticity of strands
was set to 28,500 ksi (196500 MPa). More details on modeling of RC shear walls using LS-DYNA can be
found elsewhere [12,13]. Figure 2 shows the hysteretic force-displacement relationship of pre- and post-retrofit
walls. The strength of post-retrofit wall was 41% lower that the pre-retrofit wall.

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

Figure 2. Force-displacement relationship of pre- and post-retrofit walls

4. Simplified macro models

Detailed finite element models are computationally expensive, particularly for NRH analyses. Two simplified
macro models based on lumped plasticity or distributed plasticity approaches were discussed for efficiency in
NRH analyses. Both macro models were developed in OpenSees [14].
4.1. Lumped plasticity approach

For the lumped plasticity model, inelastic response is concentrated at discrete regions (near the base of a wall)
and is modeled using multi-axial or rotational springs. Palermo et al. [15],Pampanin et al. [16] used two zero-
length rotational springs in parallel to capture nonlinearities at a shear wall base, together with an elastic beam-
column to model the global response of rocking systems. Nazari et al. [17] modeled single precast rocking
walls using a single zero-length rotational spring that had flag-shape hysteretic moment-rotation behavior.
The model used in the study presented in this paper consisted of a linear elastic beam-column element
along with two nonlinear zero-length rotational springs at the base. One spring captured the effect of axial
load, while the second one represented energy dissipation of vertical reinforcing bar yielding. For the pre-
retrofit wall, the springs were calibrated using results of a sectional analysis of the wall. For the post-retrofit
wall, monolithic beam analogy (MBA) was used to calibrate the springs. MBA assumes that the total
displacement of a hybrid rocking RC wall is equal to that of a monolithic RC wall with identical geometry and
reinforcement. Moment-rotation relationship was determined using an iterative process [16].
The iterative process included the following steps. An initial rotation was imposed and neutral axis depth
was guessed. Deformation compatibility (rather than strain compatibility) was satisfied using MBA. By
satisfying deformation compatibility, concrete strain was determined. Concrete compressive force was
calculated and the neutral axis depth was adjusted iteratively until equilibrium was achieved at the wall base.
The procedure was then repeated by increasing the initially imposed rotation till the entire moment-rotation
relationship of the wall component was constructed. Figure 3 shows the model configuration along with
moment-rotation relationships for pre- and post-retrofit RC walls obtained using the lumped plasticity
approach.

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

(a) (b) (c)

Figure 3. a) Schematic of the lumped plasticity model, b) moment-rotation relationship of pre-retrofit wall and c)
moment-rotation relationship of post-retrofit wall

4.2. Distributed plasticity approach

Distributed plasticity model (fiber model) uses a number of elements that consisted of fibers representing cross
section with concrete or reinforcing bar and their proper stress-strain relationships. Kurama et al. [3] developed
and used the fiber model to simulate the response of a single rocking wall under monotonic quasi-static and
dynamic loading. Tensile strength of concrete was neglected to represent gap opening. Although the fiber
model was able to predict the global response of the wall with reasonable accuracy, it failed to predict local
response near gap opening region since the fiber model assumes that plane sections remain plane. Smith [18]
further extended the fiber model to reproduce the cyclic and dynamic behavior of hybrid walls with and without
perforations.
The distributed plasticity (fiber) model used in the study presented in this paper had a number of
displacement-based fiber beam-column elements. A displacement-based element formulation was calibrated
by increasing the number of elements. For the wall under consideration, 5 elements were sufficient to make
the model results insensitive to number of elements. Both confined and unconfined concrete materials were
modeled using concrete material with linear tension softening (Concrete02 in OpenSees). Material properties
for confined concrete were determined per Mander et al. [19]. Reinforcing bars were modeled using Giuffre-
Menegotto-Pinto steel model with isotropic strain hardening (Steel02 in OpenSees) [20]. For the post-retrofit
wall, gap opening was modeled by neglecting the tensile strength of concrete in the first two elements from
the wall base. To model cut reinforcing bars properly, an elastic-no tension material with high initial stiffness
(i.e., 100 times the stiffness of steel) and a conventional steel material representative of the response of
reinforcing bars (i.e., Steel02 in OpenSees) were used in series. Strands were modeled using truss elements.
Initial post-tensioning force was applied through an initial strain. Figure 4 shows the configuration of the
distributed plasticity model schematically.

Figure 4. Schematic of the distributed plasticity model (fiber model) of the pre- and post-retrofit walls

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

5. Quasi-static cyclic analysis

This section presents and compares results of detailed FEA and simplified models for both pre- and post-
retrofit walls under quasi-static cyclic loading conditions. Figure 5 shows the force-displacement response of
pre- and post-retrofit walls using the three modeling approaches. In general, both models based on lumped
plasticity or distributed plasticity approaches captured lateral strength and secant stiffness with reasonable
accuracy.
Results of the detailed models and simplified finite element models for pre- and post-retrofit walls were
compared in terms of relative-energy dissipation ratio, lateral strength, residual displacement and secant
stiffness. Relative energy-dissipation ratio was the ratio of the area of force-displacement hysteretic response
for the cycle of interest to the area of the circumscribing parallelograms defined by the initial stiffness and
maximum displacement of the cycle of interest. Residual drift was the ratio of the displacement at zero lateral
force upon unloading at each cycle to the total height of the wall specimen. Secant stiffness was determined
as the ratio of lateral force at the maximum displacement imposed in each cycle to the maximum displacement
of that cycle. In the absence of test data, detailed finite element analysis was used as the benchmark for
comparison. Figure 6Figure 7 show the comparison of simplified and detailed models in terms of parameters
of interest for pre- and post-retrofit cases, respectively.

(a) (b)
Figure 5. Force-displacement relationship of a) Pre-retrofit b) Post-retrofit walls using finite element analysis, lumped
plasticity and distributed plasticity approaches

Table 1 presents the ratio of parameters of interest (i.e. relative energy-dissipation ratio, lateral strength,
residual drift, and secant stiffness) predicted by simplified models to the ones predicted by FEA. Average of
ratios over all loading cycles (cycles 1-12) and between loading cycles 4 and 10 were reported. Loading cycles
4 to 10 were selected since the first three cycles were prior to yielding of vertical reinforcing bars and of lower
importance. There was also a divergence between simplified and detailed models for cycles 11 and 12, which
is attributed to the failure criterion defined only in detailed FEA.
Table 1 shows that the distributed plasticity model (2% to 33% error) was better at predicting relative
energy-dissipation ratio than the lumped plasticity model (55% to 102% error). Predictions of both the
distributed plasticity (3% to 6% error) and lumped plasticity (6% to 22% error) were reasonably close to the
FEA prediction. In terms of residual drift, the distributed plasticity model predictions (4% to 54% error) were
much closer to FEA predictions than the lumped plasticity model predictions (110% to 257% error). Both the
distributed plasticity (1% to 8% error) and lumped plasticity models (3% to 24% error) matched the secant
stiffness prediction of FEA reasonably well. Overall the distributed plasticity model results predicted the
response better (closer to FEA results) than the lumped plasticity model did.

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

Figure 6. Comparison of detailed and simplified models in terms of relative energy-dissipation ratio, lateral strength,
residual displacement, and secant stiffness for the pre-retrofit wall

Figure 7. Comparison of detailed and simplified models in terms of relative energy-dissipation ratio, lateral strength,
residual displacement, and secant stiffness for the post-retrofit wall

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

Table 1. Prediction of relative energy-dissipation ratio, lateral strength, residual drift, and secant stiffness using
simplified models divided by the prediction by detailed FEA

Simplified model results / detailed FEA results for pre-retrofit wall

Relative energy- Lateral Residual Secant
Model
dissipation ratio strength drift stiffness
Average of Distributed plasticity 1.02 1.06 1.54 1.08
normalized values
for all cycles Lumped 1.55 1.22 2.10 1.24

Average of Distributed plasticity 0.92 0.97 1.04 0.95

normalized values
for cycles 4 to 10 Lumped 2.02 1.08 3.04 1.06

Simplified model results / detailed FEA results for post-retrofit wall

Average of Distributed plasticity 1.33 1.03 1.04 1.01
normalized values
for all cycles Lumped 1.76 1.09 2.18 1.07

Average of Distributed plasticity 1.00 0.97 0.50 0.95

normalized values
for cycles 4 to 10 Lumped 1.75 1.06 3.57 1.03

6. Nonlinear response history analysis (NRHA)

The RC wall analyzed was a 1:3 scale of a wall from a 3 story RC building located in Rowland Heights,
California, USA. Seismic loading for this building (in full scale) was calculated using the equivalent lateral
force (ELF) method of ASCE 7-16 [21]. The main lateral load resisting system was ordinary RC shear wall in
the frames, therefore, the response modification factor was set to 5. The total seismic weight of the building
was 4800 kips (21351 kN), which was distributed to the floors per ASCE 7-16. Only a 2D frame in the shorter
direction was considered for the nonlinear response history analyses (NRHA). Figure 8 shows the footprint of
the building together with the elevation view of the frame under consideration (frame 6-6).

(a) (b)
Figure 8. a) Footprint of the building b) Elevation view of the RC wall

As mentioned previously, the RC shear wall had design details representative of pre-1970 construction
practices. It was non-code-compliant in terms of vertical spacing of confining bars in the boundary element,
extension of the confining reinforcement into the foundation and transverse reinforcement ratio in the web.
The full-scale wall had length and thickness of 180 inches (4572 mm) and 18 inches (457 mm), respectively.

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

Table 2 presents reinforcement layout for the full-scale RC shear wall in the building and ACI 318-14 web
and boundary element amounts calculated for the wall analyzed
Table 2. Reinforcement configuration of the wall in the frame

Transverse Boundary element

Web reinforcement Boundary element
reinforcement reinforcement ratio
ratio reinforcement ratio
ratio (confinement)

Wall design 0.012 (20#11 @ 13”) 0.026 (4#11 @ 13”) 0.002 (#5 @ 18”) 0.003 (#5 @ 12”)

ACI 318-14
0.0025 0.028 (4#11 @ 12”) 0.0025 0.006 (#5 @ 6”)
recommendation

Per ASCE 7-16, a suit of 11 ground motions was selected and scaled for amplitude. The fundamental
period of the pre-retrofit 2D frame was 0.24 seconds (post-retrofit frame had similar fundamental period).
Figure 9 shows the scaled spectra for individual ground motions (denoted with RSN in the figure) along with
the average of scaled records, risk-based maximum considered earthquake (MCER) spectrum, and Uniform
Hazard Spectrum (UHS).

Figure 9. Scaled individual records, average of scaled records, MCE, and UHS spectra

The 2D frame was modeled in OpenSees. Floor slabs were modeled using elastic beam-column
elements, including effective beam-width model [22]. A kinematic constraint was placed between nodes of the
wall and slab to establish deformation compatibility at floor levels. Seismic columns were modeled using
“BeamWithHinges” model of OpenSees, which uses a force-based element formulation with distributed
plasticity near the expected plastic hinge locations. A leaning column was added to the model to capture
possible P-Delta effects. A stiffness-proportional damping of 3%, considering tangent stiffness, was used to
capture the inherent damping of the RC structure. The pre-retrofit RC wall was modeled using the distributed
plasticity modeling approach, which was shown to capture the response of pre-retrofit RC wall more accurately
than the lumped plasticity approach. The wall composed of displacement-based beam-column elements with
5 elements per story. The post-retrofit RC shear wall was modeled using both distributed and lumped plasticity
approaches. Figure 10 shows the bottom part (i.e. the first floor and a portion of second floor) of both models
schematically.
Figure 11 compares three parameters of interest (i.e., peak inter-story drift, peak residual inter-story
drift, and peak floor acceleration) for the pre- and the post-retrofit walls for the 11 ground motions denoted as
RSN on the figure. The figure also shows the mean response under all ground motions. Performance of the
pre-retrofit wall was assessed per recommendation of FEMA 356 [23]. IO, LS, and CP on the figure stand for
immediate occupancy, life safety, and collapse prevention, respectively. Results of the NRHA demonstrated
that the proposed retrofit methodology resulted in 35% (using the lumped plasticity approach) and 42% (using
the distributed plasticity approach) increase in peak inter-story drift, on average across all floors, which can
be attributed to the softening of the structure.

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

(a) (b)
Figure 10. Schematic simplified models a) Distributed plasticity b) Lumped plasticity of the 2D frame

Pre-retrofit wall

Post-retrofit wall (lumped plasticity model)

Post-retrofit wall (distributed plasticity model)

Figure 11. Comparison of pre- and post-retrofit walls in terms of peak inter-story drift, peak residual drift and peak
floor acceleration

10

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

This softening led to the increase of the displacement demand, as expected. On the other hand, damage in the
form of peak residual inter-story drift decreased by 90% per the lumped plasticity approach, on average. The
distributed plasticity approach predicted the reduction of residual inter-story drift to be 29%. The difference
in predictions of residual inter-story drift of the two modeling approaches can be attributed to the over-
estimation of energy dissipation by using lumped plasticity approach. Softening the structure through the
proposed retrofit strategy did not lead to a significant decrease of the peak floor acceleration using the lumped
plasticity approach (i.e., a decrease by 6%, on average). Implementing distributed plasticity approach led to
higher predictions of peak floor accelerations than predicted by lumped plasticity approach due to acceleration
spikes observed during unloading near zero displacement. These spikes may be explained by abrupt stiffness
changes near origin (i.e., zero displacement) along with high velocity [24-26].
To investigate the effect of the proposed retrofit strategy on the global force demand on the building, values
of base shear prior to and after retrofit were compared. Figure 12 shows that the mean base shear decreased by
28% (using the lumped plasticity approach) and 20% (using the distributed plasticity approach) when the
retrofit method was implemented. The base shear reduction under dynamic excitations was reasonably close
to the lateral force reduction under cyclic loading (i.e. the lateral strength reduction was 41% in this case) (see
Figure 2 and Figure 5).

Figure 12. Base shear demand for pre-retrofit and post-retrofit walls (lines show mean values)

7. Summary and conclusions

This study investigated a retrofit method targeting none-code-compliant RC shear walls using different
modeling approaches. The retrofit strategy utilized concepts of weakening and self-centering to convert the
monolithic cast-in-place RC shear walls to rocking walls. The goal of the retrofit method was to increase
ductility and reduce residual displacements.
RC shear walls were modeled using detailed FEA and simplified approaches. The results showed that
both lumped and distributed plasticity models were able to capture the lateral strength and secant stiffness of
pre- and post-retrofit walls with reasonable accuracy. Distributed plasticity approach predicted residual drift
and relative energy-dissipation ratio more accurately than the lumped plasticity approach.
A 2D frame that hosts the RC shear wall was subjected to nonlinear response history analyses. The
results showed that simplified modeling approaches predicted 35% to 42% increase of peak inter-story drift
and 20% to 28% decrease of base shear. Peak inter-story residual drift decreased using both modeling
approaches. However, predictions of peak inter-story residual drift and peak floor acceleration were highly
dependent upon the modeling technique.

Acknowledgements
This material is based upon work supported by the National Science Foundation under grant No. 1663063.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors
and do not necessarily reflect the views of the National Science Foundation. The authors would also like to

11

© The 17th World Conference on Earthquake Engineering - 3b-0097 -

3b-0097 The 17th World Conference on Earthquake Engineering

17th World Conference on Earthquake Engineering, 17WCEE

Sendai, Japan - September 13th to 18th 2020

thank Dr. Qingzhi (Andy) Liu of Englekirk Structural Engineers for his help with design and conceptualization
of the building used in analyses.

References
[1] Priestley, M.N., S. Sritharan, J.R. Conley, and S. Pampanin (1999): Preliminary results and conclusions from
the PRESSS five-story precast concrete test building. PCI Journal. 44(6), 42-67.
[2] Kurama, Y. (2002): Hybrid post-tensioned precast concrete walls for use in seismic regions. PCI Journal. 47
36-59.
[3] Kurama, Y., R. Sause, S. Pessiki, and L.-W. Lu (1999): Lateral load behavior and seismic design of unbonded
post-tensioned precast concrete walls. Structural Journal. 96(4), 622-632.
[4] Aaleti, S. and S. Sritharan (2009): A simplified analysis method for characterizing unbonded post-tensioned
precast wall systems. Engineering Structures. 31(12), 2966-2975.
[5] Aaleti, S. and S. Sritharan, Performance verification of the PreWEC concept and development of seismic design
guidelines. 2011, Iowa State University.
[6] Twigden, K., S. Sritharan, and R. Henry (2017): Cyclic testing of unbonded post-tensioned concrete wall systems
with and without supplemental damping. Engineering Structures. 140 406-420.
[7] ACI ITG (American Concrete Institute) (2009): Requirements for design of a special unbonded post-tensioned
precast shear wall satisfying ACI ITG-5.1 and commentary. Farmington Hills, MI.
[8] ACI ITG (American Concrete Institute) (2007): Acceptance Criteria for Special Unbonded Post-Tensioned
Precast Structural Walls Based on Validation Testing. Farmington Hills, MI.
[9] ACI (American Concrete Institute) (2014): Building code requirements for structural concrete (ACI 318-14):
commentary on building code requirements for structural concrete (ACI 318R-14). Farmington Hills, MI.
[10] LSTC, LS-DYNA keyword user's manual. 2017: Livermore, CA, Livermore Software Technology Corporation.
[11] fib (federation internationale du beton) (2013): fib model code for concrete structures 2010.
[12] Basereh, S., P. Okumus, and S. Aaleti. Reinforced concrete shear walls retrofitted using weakening and self-
centering: numerical modeling. Journal of Structural Engineering (Accepted).
[13] Basereh, S., P. Okumus, and S. Aaleti (2020): Seismic retrofit of reinforced concrete shear walls to ensure
reparability. Structures Congress 2020. St. Louis, MO.
[14] McKenna, F., G. Fenves, M. Scott, and B. Jeremic (2000): Open system for earthquake engineering simulation
(OpenSees). University of California, Berkeley, CA.
[15] Palermo, A., S. Pampanin, and D. Marriott (2007): Design, modeling, and experimental response of seismic
resistant bridge piers with posttensioned dissipating connections. Journal of Structural Engineering. 133(11),
1648-1661.
[16] Pampanin, S., M.N. Priestley, and S. Sritharan (2001): Analytical modelling of the seismic behaviour of precast
concrete frames designed with ductile connections. Journal of Earthquake Engineering. 5(03), 329-367.
[17] Nazari, M., S. Sritharan, and S. Aaleti (2017): Single precast concrete rocking walls as earthquake force‐
resisting elements. Earthquake Engineering & Structural Dynamics. 46(5), 753-769.
[18] Smith, B.J., Design, analysis, and experimental evaluation of hybrid precast concrete shear walls for seismic
regions. 2012, University of Notre Dame: Notre Dame, Indiana.
[19] Mander, J.B., M.J. Priestley, and R. Park (1988): Theoretical stress-strain model for confined concrete. Journal
of Structural Engineering. 114(8), 1804-1826.
[20] Filippou, F.C., V.V. Bertero, and E.P. Popov (1983): Effects of bond deterioration on hysteretic behavior of
reinforced concrete joints.
[21] ASCE (American Society of Civil Engineers) (2016): Minimum design loads and associated criteria for
buildings and other structures. Reston, VA.
[22] Elwood, K.J., A.B. Matamoros, J.W. Wallace, D.E. Lehman, J.A. Heintz, A.D. Mitchell, M.A. Moore, M.T.
Valley, L.N. Lowes, and C.D. Comartin (2007): Update to ASCE/SEI 41 concrete provisions. Earthquake
Spectra. 23(3), 493-523.
[23] FEMA (Federal Emergency Management Agency) (2000): FEMA 356: Prestandard and commentary for the
seismic rehabilitation of buildings. Washington, DC.
[24] Wiebe, L. and C. Christopoulos (2010): Characterizing acceleration spikes due to stiffness changes in nonlinear
systems. Earthquake engineering & structural dynamics. 39(14), 1653-1670.
[25] Belleri, A., M.J. Schoettler, J. Restrepo, and R.B. Fleishman (2014): Dynamic behavior of rocking and hybrid
cantilever walls in a precast concrete building: American Concrete Institute.
[26] Belleri, A., M. Torquati, and P. Riva (2013): Finite element modeling of “Rocking Walls”. 4th ECCOMAS
Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering.

12

© The 17th World Conference on Earthquake Engineering - 3b-0097 -