Tsai

NONLINEAR RESPONSE MODELING OF LOW-RISE STRUCTURAL WALLS
by
Simon Karabulut
B.S., Civil Engineering, Boğaziçi University, 2012
Submitted to the Institute for Graduate Studies in

Science and Engineering in partial fulfillment of
the requirements for the degree of
Master of Science
Graduate Program in Civil Engineering

Boğaziçi University
2016
ii
NONLINEAR RESPONSE MODELING OF LOW-RISE STRUCTURAL WALLS
APPROVED BY:
Assoc. Prof. Kutay Orakçal ...................

(Thesis Supervisor)
Prof. Hilmi Lus ...................
Prof. Alper Ilki ...................
DATE OF APPROVAL: 11.01.2016

iii
To my mother,
iv
ACKNOWLEDGEMENTS
This thesis would not have been possible without the support of Assoc. Prof.
Kutay Orakçal. I would like to thank for his precious guidance and help during the
preparation of the thesis. I would like to acknowledge his patience and his positive
approach during the impasse moments of my studies.
I would like to express my special thanks to my colleagues M. Fethi Güllü and

Burak Horoz for their support during the implementation process of models.
I would also like to express my thanks to my friends Selahattin Akalp, Sarper

Saygi, and Mustafa Sekman for their help in the preparation of this thesis.
I would also like to thank Leonardo M. Massone from University of Chile and
Tevfik Terzioglu from Texas A & M University for providing the experimental data.
Most of all, I would like to express my special thanks to my family for their
invaluable support. I cannot compensate their patience, belief and love.
v
ABSTRACT
NONLINEAR RESPONSE MODELING OF LOW-RISE

STRUCTURAL WALLS
A finite element modeling approach was validated and improved in this study,
for simulating the behavior of reinforced concrete structural walls with aspect ratios
of 1.0 or less, whose behavior is governed by shear deformations as well as interaction
between shear and flexural responses. To validate the model, a total of fifteen squat
structural walls were calibrated and analyzed using Matlab. Mechanisms of shear
aggregate interlock in concrete and dowel action on reinforcing bars were explicitly
incorporated in model formulation, and cyclic degradation parameters associated with
these mechanisms were implemented in model for better representation of the in-plane
hysteretic lateral load behavior of squat structural walls. The effect of strain penetra-
tion within the wall foundation was considered in the model predictions, to capture
the initial stiffness and lateral load capacity of the walls in early drift levels more ac-
curately. Refined constitutive models were used in the model formulation to represent
the hysteretic stress-strain behavior of reinforcing steel and concrete. Features of the
experimental conditions, such as loading under double-curvature or single-curvature, or
presence of a weakened plane joint in the wall, were represented in the model calibration
and analysis conditions. Analytical predictions of the lateral load vs. displacement re-
sponses obtained using the finite element model were compared with the experimental
measurements. Comparisons revealed that the model reasonably predicts the overall
lateral load vs. displacement behavior, as well as important response attributes such as
the initial stiffness, lateral load capacity, cyclic strength and stiffness degradation, and
pinching characteristics of the walls investigated. While subject to further refinement,
the analytical model in its current improved form is shown to be a feasible candidate
for simulating the hysteretic lateral load behavior of squat structural walls.
vi
ÖZET
BODUR PERDE DUVARLARIN DOĞRUSAL OLMAYAN

DAVRANIŞLARININ MODELLENMESİ
Bu çalışma kapsamında boy/en oranı 1.0 ve daha az olan perde duvarların

davranışını irdelemek üzere bir sonlu eleman modelleme yöntemi kullanılmış, kul-
lanılan model deney sonuçları ile doğrulanmış ve geliştirilmiştir. Yatay yük davranışı
kayma ve eğilme davranışlarının etkileşiminin yanısıra çoğunlukla kayma deformasyon-
ları tarafından etkilenen bodur perde duvarların modellenmesi kapsamında bu çalışmada
on beş adet bodur duvar numunesi irdelenmiştir. Kullanılan model, incelenen deney nu-
muneleri için kalibre edilmiş ve Matlab ortamında analiz edilmiştir. Modelde betonda
agrega kenetlenmesi ve donatıda kama etkisi mekanizmaları iyileştirilmiş, bu mekaniz-
malar ile ilişkili çevrimsel dayanım azalması parametreleri modele işlenerek perdelerin
çevrimsel yanal yük davranışının daha doğru temsil edilmesi amaçlanmıştır. Model
sonuçları ile deney sonuçlarının karşılaştırılmasında perde temellerindeki gerinim pen-
etrasyonu etkileri gözetilmiş, böylelikle perdelerin başlangıç rijitliği ve düşük öteleme
oranlarında yanal yük kapasitesinin daha doğru temsil edilmesi hedeflenmiştir. Mod-
elde beton ve donatı çeliği için kapsamlı gerilme-şekil değiştirme bağıntıları kullanılmıştır.
Tekil ya da çift eğrilik altında yanal yükleme ve perde numunesindeki zayıflatılmış
kesit bölgeleri gibi deneysel koşullar model kalibrasyonunda ve analiz koşullarında
gözetilmiştir. Model sonuçları, perdelerin deneysel olarak elde edilmiş olan yanal yük-
öteleme davranışları ile karşılaştırılmıştır. Sunulan karşılaştırmalar ışığında analitik
modelin incelenen perde numunelerinin çevrimsel yanal yük-yerdeğiştirme davranışını,
başlangıç rijitliğini, yanal yük kapasitesini, döngüsel kapasite/rijitlik bozunması ve ori-
jin bölgesinde daralma gibi özellikleri makul doğruluk ile kestirebildiği gözlemlenmiş,
iyileştirilmiş hali ile kullanılan analitik modelleme yönteminin bodur perde duvarların
çevrimsel yanal yük davranışının kestirimi için uygun bir yaklaşım olduğu gösterilmiştir.
vii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
ÖZET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
LIST OF SYMBOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
LIST OF ACRONYMS/ABBREVIATIONS . . . . . . . . . . . . . . . . . . . . xx
1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2. Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3. Objectives and Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4. Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2. MODEL DESCRIPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2. Constitutive Panel Model . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3. Constitutive Material Models . . . . . . . . . . . . . . . . . . . . . . . 19
2.3.1. Constitutive Model for Reinforcement . . . . . . . . . . . . . . . 19
2.3.2. Constitutive Model for Concrete . . . . . . . . . . . . . . . . . . 20
2.3.3. Constitutive Modeling of the Shear Aggregate Interlock Effect in
Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.3.4. Constitutive Modeling of Dowel Action on Reinforcing Bars . . 24
2.3.5. Constitutive Modeling of Cyclic Degradation in Shear Aggregate
Interlock and Dowel Action . . . . . . . . . . . . . . . . . . . . . 27
2.3.6. Consideration of the Strain Penetration Effect . . . . . . . . . . 29
2.4. Finite Element Modeling Methodology . . . . . . . . . . . . . . . . . . 31
2.4.1. 4 Nodes, 8 Degree of Freedom Rectangular Element . . . . . . . 32
2.4.2. Finite Element Model Stiffness Assembly . . . . . . . . . . . . . 33
2.4.3. Internal Force Vector Assembly . . . . . . . . . . . . . . . . . . 34
2.4.4. Support Conditions and Constraints . . . . . . . . . . . . . . . 35
viii
2.4.5. Nonlinear Analysis Solution Strategy . . . . . . . . . . . . . . . 36

3. EXPERIMENTAL CALIBRATION OF THE MODEL . . . . . . . . . . . . 39
3.1. Description of the Wall Specimens . . . . . . . . . . . . . . . . . . . . . 39
3.1.1. Wall Specimens Tested at University of California Los Angeles . 39
3.1.2. Wall Specimens Tested at Boğaziçi University . . . . . . . . . . 43
3.2. Test Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.3. Material Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.4. Failure Modes of the Wall Specimens . . . . . . . . . . . . . . . . . . . 54
3.5. Calibration of the Model for the Wall Specimens . . . . . . . . . . . . . 55
3.5.1. Calibration of Model Element Dimensions and Reinforcement
Ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.5.2. Calibration of Constitutive Material Parameters . . . . . . . . . 56
3.5.2.1. Calibration of the Constitutive Model for Reinforcing
Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.5.2.2. Calibration of the Constitutive Model for Concrete . . 63
4. COMPARISON OF ANALYTICAL AND EXPERIMENTAL RESULTS . . 71
4.1. Wall Specimens Tested at University of California Los Angeles . . . . . 71
4.2. Wall Specimens Tested at Boğaziçi University . . . . . . . . . . . . . . 77
4.3. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . 87
5.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.2. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.3. Recommendations for Future Studies . . . . . . . . . . . . . . . . . . . 88
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
ix
LIST OF FIGURES
Figure 1.1. Coupled Model Element (Massone et al., 2006, 2009). . . . . . . . 5
Figure 1.2. Element Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Figure 1.3. Transfer Mechanism in Cracked RC Members (Ince et al., 2006). . 8
Figure 2.1. Uncracked Behavior of Concrete in the FSAM (Orakcal et al., 2012). 16
Figure 2.2. Behavior of Concrete after the First Crack Formation in the FSAM
(Orakcal et al., 2012). . . . . . . . . . . . . . . . . . . . . . . . . . 17
Figure 2.3. Behavior of Concrete after the Second Crack Formation in the
FSAM (Orakcal et al., 2012). . . . . . . . . . . . . . . . . . . . . . 18
Figure 2.4. Constitutive Model for Reinforcing Steel (Menegotto and Pinto,
1973). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Figure 2.5. Constitutive Model for Concrete (Chang and Mander, 1994). . . . 20
Figure 2.6. Compression Softening Effects Implemented in the Panel Model

(Ulugtekin, 2010). . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Figure 2.7. (a) Friction Based Constitutive Model for Shear Aggregate Inter-
lock (Orakcal et al., 2012), (b) Shear Friction Mechanism along a
Crack in ACI-318. . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Figure 2.8. Dowel Action of a Reinforcement Bar Crossing a Crack (He and
Kwan, 2001). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
x
Figure 2.9. Cracked Concrete Elements Containing a Reinforcement Bar Run-

ning (a) in the x-direction and (b) in the y-direction Crossing the
Cracks at Acute Angle (He and Kwan, 2001). . . . . . . . . . . . . 27
Figure 2.10. Response to Cyclic Loading, (a) Friction (b) Dowel (Kappos et al.,
2011). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Figure 2.11. Calculation and Update of the Degradation Factor and Envelopes
for Dowel Action (Kappos et al., 2014). . . . . . . . . . . . . . . . 28
Figure 2.12. Origin-Oriented Model for Dowel Action. . . . . . . . . . . . . . . 29
Figure 2.13. Strain Penetration Model (Massone et al., 2009). . . . . . . . . . . 30
Figure 2.14. Effect of Strain Penetration. . . . . . . . . . . . . . . . . . . . . . 31
Figure 2.15. Tangent Stiffness (Elasticity) Matrix Calculation Path. . . . . . . 33
Figure 2.16. Global and Local DOFs for a Wall Model. . . . . . . . . . . . . . 34
Figure 2.17. Stress Vector Calculation Path. . . . . . . . . . . . . . . . . . . . 34
Figure 2.18. Numeration of Internal Forces for a Mesh. . . . . . . . . . . . . . 35
Figure 2.19. Conversion of DOFs after the Assignment of Body Constraint (Gullu,
2013). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Figure 2.20. Representation of the Nonlinear Analysis Solution Strategy for a

Single Degree of Freedom System (Clarke and Hancock, 1990). . . 37
Figure 2.21. Iterative Strategy and Residual Displacements (Clarke and Han-
cock, 1990). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
xi
Figure 3.1. Specimen Name Description for Tests Conducted at UCLA. . . . . 40
Figure 3.2. Weakened Plate Joint on Wall Spandrels (Massone et al., 2009). . 41
Figure 3.3. Wall Spandrel Geometry and Reinforcement (Massone et al., 2009). 42
Figure 3.4. Geometry and Steel Reinforcement for Wall Piers (Massone et al.,
2009). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Figure 3.5. Specimen Name Description for Tests Conducted at Boğaziçi Uni-
versity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Figure 3.6. Geometry and Wall Reinforcement of Specimen T1-S1 (Terzioglu,

2011). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

2011). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

2011). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Figure 3.9. Test Setup (Massone, 2006). . . . . . . . . . . . . . . . . . . . . . 49
Figure 3.10. Displacement History (Massone, 2006). . . . . . . . . . . . . . . . 49
Figure 3.11. Test Setup (Terzioglu, 2011). . . . . . . . . . . . . . . . . . . . . . 50
Figure 3.12. Steel Tensile Stress-Strain Curves for φ 13 and φ 16 bars of Type
1 Wall Specimen (Massone, 2006). . . . . . . . . . . . . . . . . . . 51
Figure 3.13. Steel Stress-Strain Curves for Specimen Set 1 (Terzioglu, 2011). . 53
xii
Figure 3.14. Steel Stress-Strain Curves for Web Reinforcement of Specimen Set
2 and Set 3 (Terzioglu, 2011). . . . . . . . . . . . . . . . . . . . . 53
Figure 3.15. Steel Stress-Strain Curves for Boundary Reinforcement of Specimen

Set 2 and Set 3 (Terzioglu, 2011). . . . . . . . . . . . . . . . . . . 53
Figure 3.16. The Stress-Strain Behavior of Bare Rebar and Rebar Embedded in
Concrete. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Figure 3.17. Stress-Strain Relationship for a Bar Embedded in Concrete (Be-

larbi and Hsu, 1994). . . . . . . . . . . . . . . . . . . . . . . . . . 58
Figure 3.18. Compression and Tension Envelope Curves of the Model by Chang
and Mander (1994). . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Figure 3.19. Stress-Strain Diagrams for Concrete in Compression for WS-T3-S2. 66
Figure 3.20. Stress-Strain Relationship for Concrete in Tension (Belarbi and

Hsu, 1994). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Figure 3.21. Stress-Strain Diagrams for Concrete in Tension for SW-T2-S1-1. . 69
Figure 4.1. Lateral Load - Top Displacement Response of Specimen WS-T1-S1. 72
Figure 4.5. Lateral Load - Top Displacement Response of Specimen WP-T5-

N0-S1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
xiii

N5-S1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

N10-S2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Figure 4.8. Lateral Load - Top Displacement Response of Specimen SW-T1-S1-2. 78
Figure 4.9. Lateral Load - Top Displacement Response of Specimen SW-T1-

N5-S1-10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Figure 4.10. Lateral Load - Top Displacement Response of Specimen SW-T1-

N10-S1-11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Figure 4.14. Lateral Load - Top Displacement Response of SW-T6-S1-7. . . . . 82
Figure 4.15. Lateral Load - Top Displacement Response of SW-T3-S1-5 (Origi-

nal Model). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Figure 4.16. Lateral Load - Top Displacement Response of SW-T3-S1-5 (Con-

crete Friction Coefficient in Shear Aggregate Interlock Mechanism
Reduced in the Model). . . . . . . . . . . . . . . . . . . . . . . . . 84
Figure 4.17. Peak-Oriented Model for Contribution of Reinforcing Steel on Shear

Aggregate Interlock. . . . . . . . . . . . . . . . . . . . . . . . . . . 85
xiv
Figure 4.18. Lateral Load - Top Displacement Response of SW-T3-S1-5 (Peak-

Oriented). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
xv
LIST OF TABLES
Table 3.1. Test Specimen Details for Tests Conducted at UCLA. . . . . . . . 41
Table 3.2. Test Specimen Details for Tests Conducted at Boğaziçi University. 44
Table 3.3. Average Compressive Strength of Concrete (Massone, 2006). . . . . 52
Table 3.4. Average Compressive Strength of Concrete (Terzioglu, 2011). . . . 54
Table 3.5. Wall Model Geometry Details for Tests Conducted at UCLA. . . . 56
Table 3.6. Wall Model Geometry Details for Tests Conducted at Boğaziçi Uni-
versity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Table 3.7. Calibrated Values of the Constitutive Parameters of Reinforcing

Steel Bars for Specimens WS-T1-S1, WS-T1-S2, WS-T2-S1, and
WS-T2-S2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Steel Bars for Specimens WS-T3-S1, WS-T3-S2, WS-T4-S1, and
WS-T4-S2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Steel Bars for Specimens WP-T5-N0-S1, WP-T5-N0-S2, WP-T5-
N5-S1, WP-T5-N5-S2, WP-T5-N10-S1, and WP-T5-N10-S2. . . . . 61

Steel Bars for Specimens SW-T2-S1-1, SW-T1-S1-2, and SW-T2-
S2-3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
xvi

Steel Bars for Specimens SW-T2-S3-4, SW-T3-S1-5, SW-T4-S1-6,
and SW-T5-S1-7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

Steel Bars for Specimens SW-T6-S1-8, SW-T1-S2-9, SW-T1-N5-
S1-10, and SW-T1-N10-S1-11. . . . . . . . . . . . . . . . . . . . . 63
Table 3.13. Calibrated Values of the Constitutive Parameters for Concrete in

Compression for the UCLA Specimens. . . . . . . . . . . . . . . . 66

Compression for the Boğaziçi University Specimens. . . . . . . . . 67

Tension for the UCLA Specimens. . . . . . . . . . . . . . . . . . . 69

Tension for the Boğaziçi University Specimens. . . . . . . . . . . . 70
xvii
LIST OF SYMBOLS
Ag Gross concrete cross-section area

B Tension stiffening parameter for reinforcing steel
b Strain hardening ratio for compression
b0 Strain hardening ratio for tension
di Local nodal displacement
Di Global nodal displacement
Db Dowel diameter
{D} Global displacement vector
E0 Modulus of elasticity of steel
Ec Elastic modulus for concrete
Ep Plastic (strain hardening) modulus for bare steel bars
Ep∗ Plastic (strain hardening) modulus for steel bars embedded
in concrete
Es Elastic modulus for steel
[E] Elasticity matrix
fc0 Compressive strength of concrete
fcr Tensile cracking stress of concrete
FD Dowel force
FDeg Degradation in the dowel action
fy Yield stress of web reinforcement
fult Ultimate stress of web reinforcement
{f } Local internal force vector of one mesh
{F } Global internal force vector
[G] Strain matrix
[G]T Transpose of strain matrix
hw Height of wall
[k] Local stiffness matrix of one mesh
[K] Global stiffness matrix
ld Embedment length
xviii
lw Length of wall
h Height of mesh
m Mesh number in horizontal direction
n Mesh number in vertical direction
R Exponent controlling the transition between elastic and hard-
ening region
r Shape parameter of the compression envelope
s Slip
sf Slip due to the friction
sf,u Maximum slip due to the friction
t Thickness of wall
[T ] Transformation matrix for body constraint
[T ]T Transpose of transformation matrix for body constraint
xcrn Critical strain parameter of concrete in compression by Saat-
cioglu and Razvi Model
xcrp Critical strain parameter of concrete in tension by Belarbi and
Hsu Model
w Width of mesh
w/ With
w/o Without
ε0 Strain at the intersection point of two asymptotes

εy Yield strain of web reinforcement
εult Ultimate strain of web reinforcement
ε0c Strain of concrete at peak stress
εun Unloading strain in shear aggregate interlock model
εp Unloading strain in shear aggregate interlock model
εr Plastic strain in shear aggregate interlock model
εr Strain at the point of strain reversal
xix
θ Rotation
µ Interface shear friction coefficient
σ0 Stress at the intersection point of two asymptotes
σ Concrete stress
σn Normal clamping stress
σ∗n Normal stress at tension stiffening
σr Stress at the point of strain reversal
σs Axial stress of bars
σy Yield stress of reinforcing steel
ρ Reinforcement area ratio
ρb Boundary reinforcement ratio
ρl Longitudinal reinforcement ratio
ρt Transverse reinforcement ratio
τagr Shear resistance of the aggregate interlock mechanism
τD Shear stress resisted by dowel action in cracked reinforced
concrete
τdeg Degradation in the aggregate interlock mechanism
τf Friction resistance
τf,l Ultimate friction resistance at initial cycle
ξ Normalized local coordinates of a mesh in x direction
η Normalized local coordinates of a mesh in y direction
∆λ Initial load increment
∆δn Initial displacement increment
∆δ Displacement increment
∆δR Residual displacement increment
xx
LIST OF ACRONYMS/ABBREVIATIONS
ACI American Concrete Institute

DOF Degree of freedom
FEM Finite element model
FSAM Fixed strut angle model
MVLEM Multiple-Vertical-Line-Element Model
RC Reinforced Concrete
SFI Shear Flexure interaction
UCLA University of California Los Angeles
1
1. INTRODUCTION
1.1. General
Improving the seismic performance of reinforced concrete (RC) building-type

structures with means to improve their lateral stiffness and lateral load capacity, pro-
motes the use of structural walls. Structural walls are designed and detailed to pro-
vide adequate stiffness and strength, as well as sufficient ductility to attain favorable
structural performance under moderate and severe earthquake demands. Codes and
recommendations for design of new buildings enforce the walls to exhibit ductile flex-
ural behavior, with sufficient shear capacity to prevent brittle failures. Experimental
studies have shown that well designed and detailed slender walls exhibit failure under
flexure, and shear failure is not expected. Squat walls, on the other hand, are suscepti-
ble to fail in shear, which results in early loss of lateral strength and stiffness. Hence, to
characterize and to analytically represent the behavior of medium-rise or squat struc-
tural walls under earthquake actions is a significant area of research, towards a more
reliable design and performance of buildings incorporating such walls.
The so-called aspect (height-to-width) ratio (hw /lw ) is commonly used to classify
structural walls. Shear behavior governs the response of structural walls with aspect
ratios less than 1.0-1.5 (squat structural walls), whereas for structural walls with as-
pect ratios exceeding 2.5-3.0 (slender structural walls), flexural actions predominantly
control the wall response. For structural walls with moderate aspect ratios (between
1.5 and 2.5), both flexural yielding and nonlinear shear deformations (which are usually
coupled) contribute to wall behavior. For such walls, nonlinear shear deformations can
constitute up to 30% to 50% of lateral wall displacements, as investigated experimen-
tally by Tran and Wallace (2012). Reliable behavioral modeling of such structural walls
with predominant shear-flexure interaction (SFI) behavior is of particular interest, es-
pecially because fiber-based modeling methodologies commonly used in practice for
performance-based design of buildings typically consider uncoupled shear and flexural
response components. However, analytical models with uncoupled axial, flexure, and
2
shear responses are shown by previous research to underestimate compressive strains

even in relatively slender RC walls controlled by flexure (Orakcal and Wallace, 2006),
and overestimate the lateral load capacity of RC walls with moderate aspect ratios
(Tran, 2012) and low aspect ratios (Massone et al., 2006). So, there is a need for rela-
tively simple modeling approaches that consider interaction (coupling) between axial,
flexural, and shear responses, and capture important hysteretic response features for a
wide range of wall geometries and reinforcing details.
The lateral load behavior of structural walls controlled by nonlinear flexural de-
formations has been investigated both experimentally and analytically by numerous
researchers. The lateral force versus deformation response of slender walls in flexure
can be captured reasonably well using simple (moment-curvature based) analytical
models and improved predictions can be obtained using more detailed (fiber-based)
models. However, research focusing on the nonlinear shear behavior and interaction
between nonlinear shear and flexural responses of walls is based on various assump-
tions and simplifications, and are subject to limitations. There is still a need for a
reliable and generalized, yet relatively simple modeling approach that can simulate
the coupled flexural and shear responses of walls with various geometries and aspect
ratios. Therefore, the primary objective of this study was to provide a robust yet prac-
tical modeling approach for representing the nonlinear shear behavior as well as the
interaction between nonlinear shear and flexural responses of walls.
Accordingly, the finite element modeling methodology developed by Gullu (2013)

was adopted and improved in this study for simulating the response of squat walls with
coupled shear and flexural responses. The behavior of the constitutive panel elements
in the finite element model formulation is described by a previously-developed constitu-
tive relationship named as the Fixed-Strut-Angle Model (Ulugtekin, 2010). Improved
constitutive models for shear aggregate interlock and dowel action mechanisms were
implemented in the constitutive model formulation for simulating the transfer of shear
stress across cracks. The finite element wall model with the improved constitutive
formulation was calibrated for wall specimens with low aspect ratios, where coupled
flexural yielding and nonlinear shear deformations were observed, resulting in reduced
3
lateral stiffness, strength and ductility. The model was calibrated for the squat wall
specimens tested by Massone (2006), Orakcal et al., (2009), and Terzioglu (2011), and
model predictions were compared with the experimentally-observed hysteretic lateral
load vs. displacement responses.
1.2. Literature Review
Numerous experimental and analytical studies are available in the literature that
investigate the lateral load behavior of slender structural walls, the response of which is
governed by nonlinear flexural deformations. On the contrary, relatively few analytical
studies focus on squat structural walls, whose behavior is controlled by nonlinear shear
deformations as well as interaction (coupling) between shear and flexural responses.
Squat walls are very common in low-rise construction and basement levels of tall
buildings, and are likely to experience multiple deformation cycles exceeding yield in
maximum earthquake action. Identification of the response and failure modes of squat
walls plays a significant role in the design of RC systems as well as in the modeling of
such walls. Experimental approaches showed that shear contribution to displacement
of squat walls exceed the estimations using elastic shear response (Massone, 2006).
In the past 60+ years, various experimental studies on squat RC walls have been
conducted. First monotonic loading tests were conducted at Massachusetts Institute
of Technology and Stanford University (Galletly 1952, Benjamin and Williams 1953),
which focused on the shear strength of walls with barbell cross-sections. Cyclic tests
were also conducted on squat walls in Japan (Hirosawa, 1975) and in the United States
(Barda, 1972). Although extensive experimental work was conducted throughout this
period, code provisions and building manuals do not correlate well with the experi-
mental data (Wood, 1990).
As shear deformations significantly contribute to the response of squat walls,

their failure modes are typically associated with shear. Paulay and Priestley (1992)
stated that there are three types of shear failure observed in squat walls. These failure
4
types are diagonal tension, diagonal compression, and sliding shear. The failure mode
is considered mixed if the flexural strength is achieved by the wall followed by a shear
failure. Mixed failure is more common in walls with aspect ratios between 1.0 and 2.0
(Gulec, 2009). The behavior and failure type of squat walls are also affected by design
parameters such as vertical and horizontal web reinforcement, wall geometry, and axial
force.
Diagonal tension failure is generally observed in squat walls with light horizon-
tal web reinforcement. One or more cracks form with respect to loading direction
and damage is concentrated in the inclined cracks rather than the overall wall web.
The orientation of failure plane is assumed to form at approximately 45o and the web
reinforcement yield as cracks widen and displacement demand increases on the rein-
forcement. If adequate horizontal web reinforcement is provided detailing the wall,
inclined cracks open and close with successive cyclic loading. This process deteriorates
the concrete compression struts in the web and crushing of these struts results in di-
agonal compression failure. Sliding shear failure, on the other hand, is a phenomenon
when a heavily reinforced wall is subjected to large number of displacement cycles. It is
a result of concrete crushing as in diagonal compression failure. Inclined shear cracking
forms in sliding shear failure and the inclined cracks intersect with displacement cycles
leading to the deterioration of concrete within these cracks. Since the wall is heavily
reinforced, damage is spread on the overall wall web and a weakened horizontal plane
is formed near the base. Sliding shear is associated with the sliding of the upper part
of the wall on this weakened plane.
An analytical model should consider contributions from both shear and flexure
into the wall response. Flexural response was simulated by numerous researchers for
RC walls using uniaxial stress-strain relationships for concrete and steel. This approach
does not consider shear-flexure interaction (SFI) which results in variations from ex-
perimental data not only for squat walls but also for slender walls which are mainly
governed by flexure. Interaction between shear and flexure was first experimentally
observed by Oesterle et al., (1976) for moderately slender walls. The experimental
data indicated that nonlinear deformations in shear and flexure occurred almost simul-
5
taneously which confirmed the coupling of shear and flexure.
Several analytical approaches were presented regarding the uncoupled flexure and
shear interaction. The majority of these models neglect or underestimate the influence
of the shear-flexure interaction, which influences the response even for slender walls. In
late 1990s, Petrangeli (1999) introduced a fiber-based (multi-spring) model approach
into RC panel behavior to incorporate the interaction among nonlinear axial, flexural,
and shear responses of RC walls. Massone et al., (2006, 2009) assigned shear springs
to each uniaxial element and modified original Multiple-Vertical-Line-Element Model
(MVLEM) proposed by Vulcano et al., (1988) such that each element is treated as a
RC panel element subjected to in-plane uniform normal and shear stresses (Figure 1.1).
A rotating-angle modeling approach, such as the Modified Compression-Field Theory
(MCFT; Vecchio and Collins, 1986) or the Rotating-Angle Softened-Truss-Model (RA-
STM; Hsu, 1993; Belarbi and Hsu, 1994 and 1995; Pang and Hsu, 1995), was used to
model the constitutive panel behavior. The assumptions that form the model are as
follows: Plane section remains plane, strain fields acting on concrete and reinforcing
steel are identical (perfect bond between reinforcement and concrete), shear strains
are uniformly distributed along the wall length, principal stress and strain directions
coincide, and dowel action of reinforcement is equal to zero.
Figure 1.1. Coupled Model Element (Massone et al., 2006, 2009).
Kolozvari et al., (2012) introduced a new cyclic shear-flexure interaction modeling

approach in 2012. Experimental data obtained by Tran and Wallace (2012) were used
6
to validate the analytical model. The experimental study included five one-third scale
cantilever RC wall specimens which were heavily instrumented to gain detailed response
information. The walls were subjected to reversed cyclic lateral loading applied at the
top of the walls with a constant axial load. The test program aimed to obtain the
influence of level of axial stress and average shear stress on the wall failure modes
and lateral deformation capacity as well as the magnitude and distribution of lateral
deformations related with shear and flexure. For walls with moderate aspect ratios, the
shear-interaction model predicted the response reasonably accurately. In the proposed
model, each uniaxial element from the original formulation of the MVLEM is replaced
with a RC panel element subjected to membrane actions (Figure 1.2c). Based on
the Fixed-Strut-Angle-Model (FSAM) developed by Ulugtekin (2010) and extended
by Orakcal et al., (2012), a 2D constitutive RC panel model formulation was adopted
to describe shear aggregate interlock effects for the RC panel elements under reversed
cyclic loading conditions. The coupling of axial and shear responses is thus enabled
at the panel (macrofiber) level, which further allows coupling of flexural and shear
responses at the model element level. The model differentiates from the one proposed
by Massone et al., (2006) in terms of its capability to simulate cyclic responses as well
as the monotonic ones. Detailed information about the approach can be found in the
papers proposed by Kolozvari et al., (2014).
Figure 1.2. Element Models (a) Original MVLEM element, (b) RC panel element, (c)
SFI-MVLEM element (Kolozvari et al., 2014).
One of the most recent finite element modeling approach to investigate the inelas-
tic response of RC walls under reversed-cyclic loading was proposed by Gullu (2013).
7
The fixed-crack-angle modeling approach was implemented due to its ability to be used
for reversed cyclic loading conditions. The apprehension of the unexpected shear yield-
ing and nonlinear deformation behavior in slender RC walls are targeted with this mod-
eling approach. The analytical model successfully captured the stiffness degradation,
shape of the load-displacement hysteresis loops, plastic displacements, and pinching be-
havior in slender walls. On the other hand, the model was validated for slender walls,
which do not exhibit significant shear-flexure interaction behavior. Shear-aggregate-
interlock effects in concrete were implemented in the model, which contributed to the
shear stress transfer mechanism across the cracks. Later, Orendil (2014) proposed a
new constitutive relationship for the dowel action on reinforcing steel bars and investi-
gated the lateral load responses of squat RC walls with predominantly shear-controlled
responses. The model reasonably predicted the experimentally-observed wall responses.
Horoz (2015) implemented an origin oriented dowel model for reinforcing bars and in-
troduced a revised constitutive model to represent shear-aggregate-interlock effects in
concrete. The accuracy of the model implemented was calibrated for medium-aspect-
ratio walls and global and local responses were captured fairly accurately, although the
model did not capture failure modes associated with bar buckling and sliding shear
failure.
As aforementioned, shear-flexure interaction behavior constitutes a significant re-

sponse characteristic of RC walls. Therefore, shear effects which lead to behavioral phe-
nomena such as multi-axial stress conditions, interlocking of cracks, and dowel action
have been important areas of research. Across a crack, shear forces are transmitted by
either through the interaction between rough surfaces of the crack (aggregate interlock
mechanism) and through reinforcement crossing the crack (dowel action) (Maekawa
and Qureshi, 1997). The relative contributions are affected by the geometry of the
member, the arrangement of the loads applied, the reinforcement pattern, and the
crack pattern. The contributions scatter through changes in internal stress distribu-
tions as concrete cracks and reinforcement bars yield (He and Kwan, 2001). Figure 1.3
shows a typical cracked member with shear transfer mechanisms illustrated.
8
Figure 1.3. Transfer Mechanism in Cracked RC Members (Ince et al., 2006).
Concrete-to-concrete friction (aggregate interlock) as a shear transfer mechanism

has been the object of numerous experimental investigations. On the contrary, an ana-
lytical model to express aggregate interlock is complex because cracks tend to “dilate”
as they slide due to aggregate particles sliding over each other (Sagaseta and Vollum,
2011). Hsu et al., (1987) divided the transfer mechanism into two categories which are
the transfer of shear across an uncracked plane and across an initially cracked plane.
Transfer of shear across an uncracked plane implies cracks to produce in a direction
inclined to the shear plane. Across an initially cracked plane, shear causes concrete
segments to slip relative to each other which results in widening of the crack through
roughness and irregularities that exist along the crack.
Mattock and Hawkins (1972) conducted monotonic loading tests on RC elements.

In the tests, cracks that exist prior to shear loading was investigated. These cracks,
which may form unrelated to shear such as due to tension forces caused by restrained
shrinkage or temperature deformations, were found to reduce the ultimate shear trans-
fer strength and increase slip at all loading stages. Additionally, for the initially cracked
specimens with adequate reinforcement and large compressive stresses applied to the
9
shear plane, the cracks in the shear plane were displayed to “lock up” and shear transfer
strength resembled the one with no initial cracks were present.
Walraven (1981) introduced analysis on aggregate interlock for initially cracked

sliding planes. The cracks were studied at both micro and macro-roughness levels and
the latter was shown to be the main shear transfer mode. Shear and normal stress
transferred across the crack were quantified statistically. The equations were derived
using slip and separation of the shear plane. Unlike Mattock and Hawkins (1972), these
equations were derived for shear transferred through rather than the shear capacity of
a member.
Bazant and Gambarova (1984) introduced the crack band model in which uniform
distribution of the relative displacements across the crack exists over a certain specified
width of the band.
Theodossius et al., (1987) investigated the aggregate interlock phenomenon through

experiments conducted for rough and smooth interfaces. For rough interfaces, it was
presented that formation of cracks introduce significant decrease in shear stiffness, and
as the surface deteriorates, the force-response deterioration is very high. The deterio-
ration was found to increase with cycling though decrease in the maximum dilatancy
was observed with increasing cycles.
Martin-Perez and Pantazopulou (1999) implemented a non-linear smeared-crack

/ smeared-reinforcement plain-stress model to investigate the dependence of shear resis-
tance of RC elements on deformation demand. In the formulation, concrete was treated
as and orthotropic material and the model was described using average stresses and
average strains. The analytical results showed that with increased imposed deforma-
tion, shear strength degradation occurs and the cause of this degradation is mainly the
susceptibility of concrete tensile and compressive strengths to increasing crack width.
Palieraki and Vintzileou (2009) did cyclic tests on RC specimens using two sepa-
rate blocks casted at different times for each specimen. It was concluded that cyclically
10
imposed slips lead to significant degradation of the shear resistance of interfaces. The
degradation was found to depend on the imposed cyclic slip and the anchorage length
of the reinforcing bars in between the blocks.
As aforementioned, Massone et al., (2009) implemented a rotating-angle model-

ing approach as panel behavior to represent the response of low-rise structural walls.
The interaction between shear and flexure were integrated through neglecting dowel
action on reinforcement and using only the uniaxial response of reinforcing bars while
a constitutive stress-strain model was used for concrete. In the modeling, a rotational
spring model was used to incorporate the effects of reinforcing bar extension within the
foundation which was proved to improve the results in terms of stiffness and overall
hysteretic behavior. Normal strains were also considered as the average horizontal nor-
mal strains measured during test rather than zero strain assumption. The comparisons
were made at wall mid-heights and shown that experimental and analytical results have
similarities. In a more recent attempt, Panagouli and Iordanidou (2013) applied a fi-
nite element analysis method to a shear wall which was constructed in Greece during
the 70s with the assumption of that cracks already formed. Fractal theory was used
to model cracks. The findings of the study included that cracked wall still has the
capacity to sustain monotonic horizontal loading. In addition, with higher resolution
of the interfaces, more accurate results can be obtained. As roughness of the contact
interface plays an important role in the response, with lower resolution values, interlock
between two faces of the crack is rather limited.
As stated above, another mechanism to transmit shear forces is through rein-

forcement crossing the crack. Until 80s, aggregate interlock was the only mechanism
considered to transmit shear. In addition, from 1985 to 1991, modeling of the dowel
action has never been mentioned in any of the papers due to the difficulties involved
(Darwin, 1993). Mainly, there are three difficulties in addressing the dowel action.
Firstly, the identification of the exact shear force transferred by dowel action within
other components of shear transfer mechanisms is challenging. Secondly, the compli-
cated nature of dowel action in a finite element analysis as it needs large number of
elements. Finally, limitations of testing conditions in which post-peak range is required
11
to investigate the effects of dowel action (He and Kwan, 2001). However, the increase
in the width of cracks under cyclic actions was shown to result in a substantial decrease
of concrete-toconcrete friction along the tensioned zone and dowel action becomes the
main contributor to the shear force transfer.
Vintzeleou and Tassios (1987) investigated the dowel action through experiments
on three concrete blocks and concluded that under fully reversed deformations, stiffness
degradation becomes significant and response degradation is important during the early
loading cycles. An analytical model was also presented using the experimental data.
In the model, the bar is assumed to behave like a horizontally loaded free-headed pile
embedded in cohesive soil, and dowel yields simultaneously as the concrete crushes.
The model uses the slip to calculate the dowel force.
Soroushian et al., (1988) conducted cyclic tests to simulate dowel action of beam
reinforcement at a cracked beam-column interface in reinforced concrete frames under
earthquake excitation. The results showed that repeated inelastic load cycles severely
decreases dowel bar stiffness and energy dissipation capacity. Additionally, the interface
crack width increases with the repetition of inelastic load cycles.
Vintzeleou and Tassios (1990) introduced an analytical model to predict the rein-
forcing bars placed eccentrically in concrete cross section and loaded against the core of
the section. Using two parameters, the top cover and the smaller side cover, the model
was compared with experimental work done by the researchers. The shear displace-
ment at failure was shown to increase with increase in the side cover of a reinforcing
bar.
A micro scale model was proposed by Maekawa and Qureshi (1996b) to predict
reinforcing bar response to axial pullout and transverse displacement. Axial pullout
and dowel action were combined in the formulation. The monotonic model was ex-
tended to path-dependent cyclic model by Soltani and Maekawa (2008) which solves
the nonlinear equilibrium and compatibility equations numerically to determine the
deformational behavior of deformed bars with respect to the loading path and history.
12
The shortcoming of this model is that it is time consuming as it considers the deformed
bar as three-dimensional member capable of developing coupled shears and moments
as well as axial force.
In the dowel action model by He and Kwan (2001), the dowel effects on the
reinforcing steel bars crossing cracks in concrete is represented in a smeared form in
order to be compatible with the smeared crack model and smeared reinforcement model
used in finite element analysis.
Ince et al., (2006) studied the size effect of shear plane depth through experiments
with push-off elements. The elements consisted of two concrete blocks connected by
two parallel dowel bars which crossed shear plane at an angle. The results of the
experimental work indicated that nominal strength at failure decreases as the specimen
size increases. Using fracture mechanics, the contribution of the dowel action to the
bearing capacity of cracked reinforced concrete elements was calculated. Additionally,
the shear capacity of the dowel specimen was reported to increase with an increase in
maximum aggregate size.
Husain et al., (2009) tested eight specimens with initially cracked form to inves-
tigate the dowel resistance. It was reported that as the area or the number of dowel
bars are increased, the slip tends to decrease due to the fact that the contribution of
the bar stiffness to the overall section stiffness increases.
In a more recent study, Moradi et al., (2012) proposed a macro scale model to
simulate the behavior of deformed bars across RC cracks. The model is computation-
ally efficient in attempt to be implemented in a smeared crack approach to analyze
large-scale structures. The test results showed that significant stiffness degradation
takes place during load reversals and it is more pronounced under fully reversed cyclic
loading than repeated loading. The plastic displacement was reported to depend on the
maximum applied shear and to have nonlinear relation under reversed cyclic loading.
13
1.3. Objectives and Scope
The lateral load behavior of structural walls has been an area of research both
experimentally and analytically by many researchers, and there are numerous model-
ing approaches available in the literature that use various techniques for simulating
their seismic response. However, most of the available modeling approaches focus on
characterization of the nonlinear flexural response of walls, or coupling of nonlinear
flexural deformations with moderate levels of shear deformation. Existing approaches
to model coupled shear and flexural responses typically incorporate ad-hoc modeling
assumptions, some of which have not been experimentally validated against experi-
mental data obtained for squat walls with shear-controlled responses. There is still a
need for a robust modeling approach to simulate the nonlinear hysteretic behavior of
such walls.
Accordingly, the objective of this study is to experimentally-validate and improve

a recently-developed finite element modeling methodology for simulating the in-plane
hysteretic lateral load behavior of squat walls (with aspect ratios smaller or equal to
1.0), the response of which are governed by nonlinear shear deformations. The behavior
of the constitutive panel (membrane) elements in the finite element model is described
by a previously-developed constitutive panel model (i.e., the Fixed-Strut-Angle Model
by Ulugtekin, 2010). Improvements were made on the constitutive panel model formu-
lation for better representation of the shear-aggregate-interlock effects in concrete and
dowel action in reinforcing bars, constituting the shear stress transfer mechanism across
the cracks of the constitutive panel elements. The model was extensively calibrated for
ten squat RC wall specimens tested by Massone (2006), four squat RC wall specimens
tested by Orakcal et al., (2009), and eleven wall specimens tested by Terzioglu (2011),
and model predictions were compared with the experimentally-observed responses for
fifteen squat wall specimens.
14
1.4. Thesis Outline
This thesis presents the modeling approach used, constitutive relationships incor-
porated in the model formulation, calibration of the model parameters, and comparison
of the experimental and analytical results for low-rise RC walls. In Chapter 1, an in-
troduction and a literature review, as well as the objectives and scope of this study
are provided. Descriptions of constitutive panel model, the finite element modeling
approach, and the shear transfer mechanisms across the cracks are presented in Chap-
ter 2. Chapter 3 summarizes the test programs used for experimental validation of the
model, as well as calibration of the model parameters. Chapter 4 focuses on comparison
of model predictions and test results for the lateral load-displacement response of the
squat wall specimens investigated. Finally, concluding remarks and recommendations
for model improvements are presented in Chapter 5.
15
2. MODEL DESCRIPTION
2.1. Overview
This chapter describes the constitutive panel (membrane) model formulation used
in this study, the constitutive material relationships implemented in the panel model,
the macroscopic finite element modeling methodology used, and the constitutive re-
lationships implemented to better represent transfer of shear stresses across cracks.
Finite element modeling of RC structural walls involves assembling of constitutive
panel elements for obtaining the overall model of a wall. To represent the behavior
of the wall model, the combination of hysteretic nonlinear material relationships along
crack directions of the constitutive panel, together with behavioral response charac-
teristics including compression softening, tension stiffening, hysteretic biaxial damage,
and shear stress transfer across cracks are used in combination.
The so-called Fixed Strut Angle Model (FSAM) proposed by Ulugtekin (2010)
was selected as the constitutive panel model in the finite element model assembly. Its
simple formulation and adequate accuracy makes it a feasible candidate for implemen-
tation. A summary of the FSAM formulation is provided in the next section, whereas
detailed information on the FSAM is available in M.Sc. thesis by Ulugtekin (2010).
The present study adopts the formulation of the original FSAM with modifica-
tions. The effects of shear aggregate interlock and dowel action across cracks were ne-
glected in the original FSAM. No dowel action assumption implies zero shear stresses
developing along crack surfaces, through which significant overestimation of sliding
shear strains along crack directions can take place. Formulations modified by Orakcal
et al., (2012), and later adopted by Gullu (2013), include a simple friction-based consti-
tutive model to be implemented into the panel model to simulate the shear aggregate
interlock effects in concrete. The shear aggregate interlock and nonlinear hysteretic
constitutive relationship were adopted to represent the contribution of reinforcing bars
(dowel action) to shear stresses developing along crack surfaces. Constitutive modeling
16
of shear aggregate interlock behavior in concrete and dowel effects on reinforcement in

the present model formulation are presented in subsequent sections in detail.
2.2. Constitutive Panel Model
The original FSAM model incorporates perfect bond assumption between con-
crete and reinforcing steel bars, which is reasonable under favorable anchorage con-
ditions. However, and importantly, it neglects shear aggregate interlock effects in
concrete along crack surfaces, and dowel action on reinforcing bars. As the perfect
bond assumption implies, the strain fields action on concrete component of a RC panel
are assumed to be equal to that acting on reinforcing steel component. To construct
the hysteretic constitutive modeling of reinforcing steel bars, uniaxial directions along
the rebar directions are used whereas for the concrete stress-strain behavior, biaxial re-
lations along the fixed strut (crack) directions are integrated. The biaxial stress-strain
behavior of uncracked concrete is simulated in a similar fashion to the Modified Com-
pression Field Theory (Vecchio and Collins, 1986) and the Rotating Angle Strut and
Tie Model (Pang and Hsu, 1995) through utilization of a rotating strut approach. For
the uncracked loading stages, a monotonic hysteretic behavior is assumed for concrete
for small strains prior to cracking. The principal strains directions acquired from the
applied strain field are assumed to coincide with principal stress directions in concrete.
Consequently, the principal strain directions are used to apply the constitutive material
model for concrete as seen from Figure 2.1.
Figure 2.1. Uncracked Behavior of Concrete in the FSAM (Orakcal et al., 2012).
When the cracking strain of concrete is exceeded by the principal tensile strain,
the first crack forms in the RC panel in perpendicular direction to the principal tensile
17
strain. For subsequent loading, this implies that the first “Fixed Strut” direction is
assigned, and it is parallel to the first crack. For subsequent loading stages, principle
strain directions continue to rotate with the applied strain field whereas principal stress
directions in concrete coincide with the first crack direction (parallel and perpendicu-
lar). As crack and principal stress directions coincide, zero shear stress develops along
the crack directions, implying zero shear aggregate interlock and zero dowel action,
which is the underlying assumption in the original FSAM formulation.
Following the formation of the first crack, the hysteretic stress-strain relationship
for concrete is used in parallel and perpendicular directions to the fixed strut. Calcula-
tion of the principal stresses in concrete is only achievable through the transformation
of the strain field into direction perpendicular and parallel to the fixed strut. The
uniaxial model for concrete adopted enables the calculation of concrete stress values
after the calculation of the strains that are perpendicular and parallel to the fixed
strut direction. Biaxial compression softening and biaxial damage parameters reduce
the calculated stresses. These parameters are described in the following section. The
calculation of the stresses in reinforcing steel is acquired by the application of the
hysteric uniaxial constitutive material model for reinforcing steel along the orthogonal
rebar directions. The stresses developing in the concrete and reinforcing steel are su-
perimposed to obtain the average stresses on the panel element. In Figure 2.2, concrete
behavior after the formation of the first crack in FSAM is illustrated.
Figure 2.2. Behavior of Concrete after the First Crack Formation in the FSAM
(Orakcal et al., 2012).
The zero shear stress assumption (no dowel action) along the crack directions
18
dictates the second crack to form perpendicular to the first crack. The second crack
forms as the tensile strain exceeds the cracking strain of concrete, upon unloading
from compression. The constitutive behavior proceeds in the form of a single fixed
strut mechanism until that stage. The perpendicularity of the first and second cracks
is a modeling approach adopted in other constitutive panel model formulations and is
referred as “orthogonal crack”. The second “fixed strut” forms as the first and seconds
cracks are perpendicular to each other. The loading direction dictates the fixed struts
to work under tension or compression. For subsequent loading stages, principle stress
directions are fixed along the fixed strut directions while principle strain values are
able to rotate freely. The fixed strut directions are used to calculate the principle
stresses in concrete rather than the principle strain directions. The applied strain field
is transformed into these fixed strut directions and the calculated strain values are
utilized to obtain the principle stresses in concrete through the uniaxial constitutive
model for concrete. The compression softening and biaxial damage parameters are
used to reduce the obtained stresses. The constitutive material model for reinforcing
steel is applied along the orthogonal rebar directions and superposition of stresses
developing in concrete and reinforcing steel gives the resultant average stresses on the
panel element. In Figure 2.3, concrete behavior after the formation of the second crack
in FSAM is illustrated.
Figure 2.3. Behavior of Concrete after the Second Crack Formation in the FSAM
(Orakcal et al., 2012).
19
2.3. Constitutive Material Models
2.3.1. Constitutive Model for Reinforcement
In the constitutive panel model (FSAM), the behavior of reinforcing steel is rep-
resented by the well-known nonlinear hysteretic relationship of Menegotto and Pinto
(1973), as extended by Filippou et al., (1983) to incorporate isotropic strain hardening
effects. Bauschinger’s effect is also incorporated into the model through cyclic degra-
dation of tangent stiffness in the unloading and reloading curves. A description of the
constitutive model is shown in Figure 2.4. In the cyclic case, σr and r are the stress
and strain at the point of strain reversal, which also forms the origin of the asymptote
with slope E0 , and parameters σ0 and 0 form the stress and strain at the intersection
point of two asymptotes.
In the model, the tension stiffening relationship proposed by Belarbi and Hsu
(1994) is implemented. The uniaxial stress-strain relationship of bare reinforcing bar
was replaced with an average stress-strain curve of defined for bars stiffened by concrete
between cracks.
Figure 2.4. Constitutive Model for Reinforcing Steel (Menegotto and Pinto, 1973).
20
2.3.2. Constitutive Model for Concrete
The uniaxial hysteretic constitutive model proposed by Chang and Mander (1994)
is implemented as stress-strain behavior for concrete (Figure 2.5). The Chang and
Mander model has been shown to provide accurate representation of the experimental
results presented by various researchers. The model is a rule-based, generalized, and
non-dimensional model that can simulate the hysteretic behavior of both confined and
unconfined concrete, under continuous compression and tension loading. Mander et al.,
(1988) calibrated the model for unconfined and confined concrete in cyclic compression.
This constitutive model implemented for concrete was modified by adding com-
pression softening (defined by Vechio and Collins, 1993), hysteretic biaxial damage
(defined by Mansour et al., 2002) and tension stiffening effects (defined by Belarbi and
Hsu, 1994) into the formulation. The aim of these modifications was to simulate the
behavioral features of concrete under biaxial loading conditions. The effects of these
parameters are described in the following sections.
Figure 2.5. Constitutive Model for Concrete (Chang and Mander, 1994).
The compression softening effect on concrete, stemming from the tensile strains
21
perpendicular to crack directions, is typically simulated by reduction in the compressive

stresses in concrete in the principal directions (Figure 2.6). The effects of the compres-
sion softening have been experimentally validated by many researchers (e.g., Vecchio
and Collins, 1986). The compression softening relationship proposed by Vecchio and
Collins (1993) was implemented into Fixed Strut Angle constitutive panel model for-
mulation. Although more complicated models are available to represent compression
softening effects, this relationship provides simplicity while achieving reasonably accu-
rate predictions of experimentally-observed behavior.
In addition to above factors, the biaxial loading which affects the cyclic damage on
concrete which is another considerable aspect in the modeling of an RC panel element.
A damage coefficient is used to represent the cyclic damage on concrete. This coefficient
differentiates from tension stiffening and compression softening parameters such that
it is a cyclic-strain-history dependent parameter. Depending on the maximum value
of the previously-applied compressive strain in the perpendicular direction, the biaxial
damage effect physically reflects as softening in the compressive stress-strain behavior
of concrete along one direction. Due to its dependence on previous loading steps, this
behavior is not observed under monotonic loading. This coefficient is implemented in
the model as a reducing multiplier to compressive stresses in the concrete as in the
case of compressive softening. There are two models for the damage coefficient in the
literature; Stevens (1987), Mansour and Hsu (2002). The biaxial damage formulation
proposed by Mansour and Hsu (2002) was used in the formulation of constitutive Fixed
Strut Angle Panel Model.
22
Figure 2.6. Compression Softening Effects Implemented in the Panel Model

(Ulugtekin, 2010).
2.3.3. Constitutive Modeling of the Shear Aggregate Interlock Effect in

Concrete
In the original formulation of the Fixed-Strut-Angle constitutive panel model

developed by Ulugtekin (2010), the shear aggregate interlock effect along crack surfaces
in concrete were neglected. To remedy this shortcoming of the original FSAM, the
cyclic shear aggregate interlock model proposed by Orakcal et al., (2012) starts with
linear loading/unloading behavior, relating the sliding shear strain along a crack to the
shear stress, via a simple linear elastic relationship between the sliding shear strain and
the resultant shear stress along the crack surface. When the concrete normal stress
perpendicular to the crack is tensile (crack open), the shear stress is restrained to zero
value and when the concrete normal stress is compressive (crack closed), is bounded
via the product of a friction coefficient and the concrete normal stress perpendicular
to the crack. The linear unloading/reloading slope of the shear stress vs. sliding strain
relationship was taken as a fraction of the concrete elastic modulus (a value 0.4Ec was
adopted, representing the elastic shear modulus of concrete), and a value of 1.0 was
assumed for the friction coefficient, which needs to be multiplied by the compressive
stress in concrete in perpendicular direction to the crack, in order to obtain the shear
stress capacity along the crack surface (Figure 2.7a). Under constant compressive stress
in concrete perpendicular to the crack, this model yields an elasto-plastic aggregate
23
interlock behavior under cyclic loading, similar to the cyclic stress-strain behavior of
reinforcing steel. Gullu and Orakcal (2014) also considered the clamping effect of
reinforcing steel bars straddling the crack surface to the sliding shear stress (shear
friction) capacity of the aggregate interlock model, similarly to the approach presented
in ACI-318-11 (Figure 2.7b).
As mentioned in the paper by Kappos et al., (2011), the mechanisms mobilized

along interfaces due to slip and their interaction is a rather complex mechanical issue,
especially under cyclic loading conditions where degradation should also be accounted
for. In addition, Kappos et al., (2011) recommends the shear aggregate interlock fric-
tion coefficient µ to be taken equal to a value of 0.44 regardless of whether the interlock
affects are induced by compressive stresses in concrete or clamping effect on reinforcing
bars. In this study, contribution of the clamping force created by the reinforcing steel
bars to the shear friction capacity is represented by a friction coefficient of µs = 0.35
and µs = 0.70 for specimens tested at Boğaziçi University and at UCLA, respectively,
based on preliminary analysis results obtained using the model. Contribution of the
compressive stress in concrete perpendicular the crack surface is considered using a
friction coefficient µ = 1, based on sensitivity studies conducted using the model on
previously-conducted tests on squat RC walls experiencing sliding shear failure (Orak-
cal et al., 2009). In addition, while sliding shear stresses on the crack surface resulting
from the compressive stress in concrete follows elasto-plastic hysteretic behavior (Fig-
ure 2.7a), those resulting from the clamping force of the reinforcing steel bars are
assumed to follow an origin-oriented hysteretic behavior (Figure 2.7b).
Figure 2.7. (a) Friction Based Constitutive Model for Shear Aggregate Interlock
(Orakcal et al., 2012), (b) Shear Friction Mechanism along a Crack in ACI-318.
24
2.3.4. Constitutive Modeling of Dowel Action on Reinforcing Bars
Accurate prediction of the nonlinear shear behavior of RC walls is a formidable

task, even with the use of finite element modeling approaches, due to complexity of
the shear transfer mechanism along cracks in concrete, and the lack of suitable models
for the various actions that contribute to the shear transfer. A constitutive model
representing dowel action on reinforcing bars was developed by He and Kwan (2001),
for incorporation in finite element analysis of reinforced concrete structures. This
constitutive dowel action model was implemented in the present study into the FSAM
formulation, due to its relative simplicity. An overview of the dowel action model by
He and Kwan (2001) is presented in this section.
In the dowel action model by He and Kwan (2001), the dowel effects on the
reinforcing steel bars crossing cracks in concrete is represented in a smeared form in
order to be compatible with the smeared crack model and smeared reinforcement model
used in finite element analysis. A representative cracked concrete element containing
a reinforcing steel bar crossing the crack perpendicularly is depicted in Figure 2.8.
When the concrete blocks at the two sides of the crack slide against each other,
the reinforcing bar embedded inside the concrete is subjected to dowel displacement,
which is defined as the relative transverse displacement of the two ends of the reinforcing
bar. Due to the dowel deformation so caused, transverse shear force and contraflexural
bending moment are developed in the reinforcing bar. However, only a certain length
of the bar in the vicinity of the crack is subjected to significant dowel deformation.
Denoting the length of the bar subjected to significant dowel deformation by l, the
dowel displacement ∆ may be calculated from the shear strain across the crack γ12
using Equation 2.1. The dowel force Vd developed in the bar varies with the dowel
displacement ∆ according to a certain dowel force displacement relationship. Using
the secant stiffness formulation, the dowel force and the dowel displacement can be
related by Equation 2.2 where Kd is the secant dowel stiffness of the reinforcement
bar. The dowel force Vd acts against the concrete and may be smeared over the crack
surface as a dowel stress acting across the crack. The dowel stress may be calculated
25
from the dowel force from Equation 2.3 in which ρs is the reinforcement ratio in the
crack direction and As is the sectional area of the reinforcement. Equation 2.4 is derived
by combining Equation 2.1 to Equation 2.3.
The case of a reinforcing bar crossing a crack at an acute angle is more complicated
because apart from the shear strain, the tensile strain across the crack can also cause
dowel displacement of the reinforcement bar. Consider the cracked concrete element
containing a reinforcement bar running in the x-direction in Figure 2.9a and the cracked
concrete element containing a reinforcement bar running in the y-direction in Figure
2.9b. The dowel displacements ∆x of the bar running in the x-direction and ∆y of
the bar running in the y-direction are given in Equation 2.5 and Equation 2.6. In
the present model formulation, dowel displacements ∆x and ∆y are defined as an
equal values that are calculated by using the half of the shear strain acting on the
panel element, γxy . The dowel forces are identified as Vdx and Vdy develop in the two
reinforcing bars running in the x and y directions, respectively. The corresponding
dowel stresses (τx , τy ) developing in the reinforcing bars are expressed as Equation 2.7
and Equation 2.8. In this study, only the contribution of reinforcing bars running in y
direction is considered and incorporated in the formulation of the FSAM.
∆ = ` × γ12 (2.1)
V d = Kd × ∆ (2.2)
d ρs
τ12 = × Vd (2.3)
As
d ρs
τ12 = × Kd × ` × γ12 (2.4)
As
26
∆x = `x (ε1 cosθ sin θ + γ12 cos2 θ) (2.5)
∆y = `y (ε1 cosθ sin θ − γ12 sin2 θ) (2.6)
ρx
τxd = × Vdx (2.7)
Asx
ρy
τyd = × Vdy (2.8)
Asy
Figure 2.8. Dowel Action of a Reinforcement Bar Crossing a Crack (He and Kwan,
2001).
In this dowel action model, the dowel force at ultimate limit state Vdu is expressed
as in Equation 2.1. In this equation, fc represents the compressive strength of the
concrete in MPa, db is the diameter of the bar in mm and fy is the yield strength
of reinforcing steel bar. This equation is based on experimental results conducted by
Dulacska (1972). A similar empirical equation has also been proposed by Vintzeleou
and Tassios (1987). Detailed information on derivation of this empirical equation is
presented in the paper proposed by He and Kwan (2001).
q
Vdu = 1.27 × d2b × |fc | × |fy | (2.9)
27
Figure 2.9. Cracked Concrete Elements Containing a Reinforcement Bar Running (a)
in the x-direction and (b) in the y-direction Crossing the Cracks at Acute Angle (He
and Kwan, 2001).
In the present study, the elasto-plastic dowel action model presented by He and
Kwan (2001) is implemented as the monotonic envelope of the dowel force vs. shear
strain constitutive relationship, and origin-oriented unloading and reloading paths pre-
viously adopted by Horoz (2015) are improved and used to represent hysteretic behav-
ior.
2.3.5. Constitutive Modeling of Cyclic Degradation in Shear Aggregate In-

terlock and Dowel Action
Under loading and unloading reversals, the shear aggregate interlock and dowel
mechanisms tend to degrade and this phenomenon is very common in squat structural
walls. Kappos et al., (2011) proposed Equation 2.10 and Equation 2.11 to incorporate
cyclic degradation in terms of number of cycles, concrete strength, steel yield strength,
and slip parameters.
 " ! !#1/3 
 fc |sf | 
τf,n = τf,1 × τdeg = τf,1 1 − 0.002 (n − 1) (2.10)
σN sf,u 
1√

FD,n = FD,1 × Fdeg = FD,1 1− n−1 (2.11)
7
28
In Equation 2.10 and Equation 2.11, coefficient n is the number of cycles after the
capacity is reached, fc and σN are the concrete strength and clamping stress, respec-
tively. sf and sf,u represent slip at current step and the maximum slip due to friction,
respectively. τf,1 and FD,1 are the friction and dowel forces at each cycle, respectively,
calculated at initial step. Figure 2.10 illustrates degradation with increasing cycles for
both shear aggregate interlock and dowel mechanisms. Degradation of dowel force as
the cyclic load path is updated is displayed in Figure 2.11. Details of the degradation
parameters are presented in the paper proposed by Kappos et al., (2011).
Figure 2.10. Response to Cyclic Loading, (a) Friction (b) Dowel (Kappos et al., 2011).
Figure 2.11. Calculation and Update of the Degradation Factor and Envelopes for
Dowel Action (Kappos et al., 2014).
Preliminary analysis results showed that consideration of degradation parame-

ters significantly improved the model predictions for the squat walls investigated. The
29
degradation was implemented in the hysteresis loops beyond the elastic response re-
gion (i.e., after capacity is reached). For degradation of dowel action, the equation
proposed by Kappos et al., (2011) was implemented. For shear aggregate interlock,
extensive calibration was conducted for contribution of the clamping force created by
the reinforcing steel bars to the shear friction capacity. Through this calibration, the
degradation was introduced in the model in terms of number of hysteresis loops only
and no degradation was implemented for the contribution of concrete in the shear
aggregate interlock. The representation of the origin-oriented hysteretic dowel action
with cyclic strength degradation implemented in this study is illustrated in Figure 2.12.
Figure 2.12. Origin-Oriented Model for Dowel Action.
2.3.6. Consideration of the Strain Penetration Effect
When a wall experiences large bending moments at the base, as in the case
for early drift levels of the squat wall tests investigated in this study, large tensile
stresses (and therefore strains) develop in the longitudinal reinforcement at the base
of the wall. These strains gradually penetrate into foundation, which produces an
accumulated elongation (commonly referred to extension/slip) on the reinforcement
at the wall-foundation interface. The rotations stemming from the strain penetration
effect introduce additional top displacement on the wall, which may be a significant
portion of the total top displacement if the lateral stiffness of the wall is large (which
is the case for a squat wall). When strain penetration is neglected, analytical models
may overestimate the lateral stiffness of a RC member especially that of a squat wall
30
(Zhao and Sritharan, 2007).
To consider the effects of strain penetration on the response of the wall specimens
investigated in this study, an initial moment curvature analysis was performed for each
specimen. Yield moment and yield strain values were obtained from this analysis along
with neutral axis depth of each specimen. With the assumption of a crack along the
wall-foundation interface, the strains were integrated over the embedment length of
reinforcing bars, which was close to the development length, in the specimen pedestals
(foundations) to acquire the vertical elongation of the wall boundary reinforcement
at the interface, associated with strain penetration. A linear distribution of strains
within the pedestal was assumed in which strains are zero at a distance equal to the
development length of the bar from the interface (Figure 2.13). The rotations were
calculated as dividing the displacement by the neutral axis depth. The yield moment
was then divided by yield rotation to obtain an elastic rotational spring stiffness.
Using the rotational spring stiffness, any force accompanied by an incremental

displacement then was converted to moment at the top of the wall. Following the
rotations obtained from the moments, the additional top displacement resulting from
slip was calculated by multiplying the rotations with the height of the wall. Initial
cycles were affected by the contribution of slip more significantly which altered the
initial stiffness of the walls.
Figure 2.13. Strain Penetration Model (Massone et al., 2009).
In this study, the calculated displacements due to the strain penetration and
31
elongation of the reinforcement were discarded from the test results and drift histories
were implemented into model with lower displacement values, especially at early loading
stages. The results, in accordance with the calculations, give higher stiffness for both
analytical model and experimental outcomes due to this effect. As seen from Figure
2.14, for specimens tested under double curvature or single curvature, at early loading
levels, neglecting strain penetration and additional displacement associated with it
leads to the overestimation of initial stiffness. The discrepancy between two incidents
tends to decrease at high drift levels and both models, with or without considering
strain penetration effects, yield similar results.
Figure 2.14. Effect of Strain Penetration.
2.4. Finite Element Modeling Methodology
In this section, the two-dimensional finite element modeling methodology is de-

scribed. The methodology is used for assembling the constitutive panel elements de-
scribed in the previous section into a wall model, as developed by Gullu (2013) for
modeling of slender walls. In the following subsections, assembly of local stiffness ma-
trices and the internal force vector of mesh elements (local elements) are described.
Moreover, formulation of support conditions and constraints for modeling of walls are
summarized and the nonlinear analysis solution strategy is outlined. Further details on
the finite element model procedure are available in the M.S. Thesis by Gullu (2013).
32
2.4.1. 4 Nodes, 8 Degree of Freedom Rectangular Element
In the finite element modeling approach, rectangular two-dimensional membrane

elements are adopted. The rectangular element is formulated as a membrane element in
order to validate it for in-plane loading conditions. The rectangular membrane element
has four nodes with two degrees of freedom (DOFs) in each node, in vertical and
horizontal directions. No rotational DOFs are implemented in the model. Equilibrium
equations are utilized at the four nodes of the rectangular element to calculate the
known displacements along the DOFs. For the sake of simplicity, stress and strain
values are only calculated at the center of each rectangular element corresponding to
ξ = η = 0.
Each load step starts with the transformation of initial displacement values at
each node of the wall model into nodal displacement values defined for each model
element. Using a transformation matrix, the deformations of each element are calcu-
lated through known displacement values. This process is followed by the calculation
of average strain values in each model element after which the tangent stiffness matrix
is assembled for each element using the scheme depicted in Figure 2.15.
The next step is the generation of global tangent stiffness matrix for the structural
wall model by obtaining the local stiffness matrices for the model elements. Each local
stiffness matrix can be obtained using the tangent stiffness matrix and the geometric
properties of each element. To perform this operation, as described in the definition of
the FEM, the product of transpose of strain matrix [G]T , the elasticity matrix [E], and
the strain matrix [G] is integrated. The volume integration shown in Equation 2.12
x y
requires normalized coordinates ξ = w
,η = h
not to be considered constant as they
are dependent on x and y, respectively.
1
N1 = (1 + ξ) (1 + η) (2.12)
4
33
Figure 2.15. Tangent Stiffness (Elasticity) Matrix Calculation Path.
After the calculation of the local stiffness matrix [k] of each model element, the
assembly of these local stiffness matrices in modeling is constituted, which is described
in the following section.
2.4.2. Finite Element Model Stiffness Assembly
For the assembly operation, the rectangular wall model is divided into model
elements that are numbered as in advance. Next step is to superimpose the overlapping
terms of the 8x8 local stiffness matrices with each other to generate the global stiffness
matrix.
Local DOF numbers are displayed inside the model elements whereas global DOF
numbers outside the elements (Figure 2.16). Global stiffness matrix of this sample
model is calculated by superimposing the local stiffness terms at overlapping locations
within the global stiffness matrix. Equation 2.13 is an example for calculation of one
of the terms in the global stiffness matrix. Superscripts in Equation 2.13 indicate the
model element numbers while subscripts indicate the local stiffness term index for each
model element.
[K]99 = [k]111 + [k]233 + [k]322 + [k]444 (2.13)
Same procedure is applied to calculate each term in the global stiffness matrix of the
wall model as in Equation 2.13. As 18 global DOFs exist in this sample, the global
stiffness matrix will be an 18x18 matrix, with 324 terms.
34
Figure 2.16. Global and Local DOFs for a Wall Model.
2.4.3. Internal Force Vector Assembly
This section provides information about how the assembly of the global internal
force vector takes place using the local internal force vectors. Average strain values for
each model element are utilized to calculate the average stress values for each model
element using constitutive FSAM (Figure 2.17).
Figure 2.17. Stress Vector Calculation Path.
Stress values are multiplied with thickness and half of vertical or horizontal width
and superimposed to obtain the internal force vector for each element. Numbering
of the internal force vector components, which is similar to the local displacement
numbering scheme is illustrated in Figure 2.18.
35
Figure 2.18. Numeration of Internal Forces for a Mesh.
The order of the terms in the global force vector for the wall model is defined
by the global DOFs of the wall model. Local and global DOF numbers are displayed
inside and outside the model elements respectively. Equation 2.14 is an example for
calculation of one of the terms in the global internal force vector where superscripts
indicate the model element numbers while subscripts illustrate the local internal force
vector index for each model element.
{F}91 = {f}1 11 + {f}2 31 + {f}3 21 + {f}4 41 (2.14)
Same procedure is applied to calculate each term in the global internal force vector as
in Equation 2.14.
2.4.4. Support Conditions and Constraints
Displacement values are assigned as zero at the supports of a wall. This condition
depicts zero displacement along the DOFs of the model at the supports. The reaction
forces at the supports are calculated from the equilibrium equation, using the known
(zero) displacement values along the DOFs at the supports.
Moreover, to simulate the test conditions, constraint conditions along selected

DOFs have to be implemented to validate the wall model experimentally. The wall
specimens used for experimental validation of the model were cast monolithically with
36
stiff RC load-transfer-beams, which behaved like rigid bodies during testing. Therefore,
a body constraint was defined along the top DOFs of the FEM of the wall (Figure 2.19).
Figure 2.19. Conversion of DOFs after the Assignment of Body Constraint (Gullu,
2013).
The body constraint in the model represent the rigid load-transfer-beams located
at the top of RC wall specimens. A displacement transformation matrix [T] is used to
implement body constraint in the model formulation. The body constraint condition
enforces equal horizontal displacement in the horizontal direction while relating vertical
displacements to a global rotational displacement at the top of the wall through the
plane-sections-remain-plane condition. The transformation matrix, [T], is a matrix
that transforms an 18x18 stiffness matrix to a 15x15 stiffness matrix.
2.4.5. Nonlinear Analysis Solution Strategy
As described in the previous sections, through integration and stiffness assembly

of the model elements, the global stiffness matrix and the internal force vector for the
wall model were derived. Moreover, support conditions and constraints were imple-
mented in the model formulation. Then, a displacement-controlled nonlinear analysis
solution strategy (Clarke and Hancock, 1990; Simons and Powell, 1982) was adopted
for conducting displacement-controlled hysteretic lateral load analysis of RC walls so
37
that model results could be correlated with results of drift-controlled lateral load tests
conducted on wall specimens.
Two main stages constitute the applied strategy which are load incrementation
and equilibrium stages. The increment stage (denoted by “i”) includes the incrementa-
tion of the target displacements imposed on the model, whereas during the equilibrium
stage (denoted by “j”), iterations are performed on the model displacements, in order
to reach equilibrium. The load incrementation stage includes the incrementation of
the target displacements imposed on the model, whereas during the equilibrium stage,
iterations are performed on the model displacements, in order to reach equilibrium
between the internal and external forces, within a specified tolerance. The adopted
nonlinear analysis solution strategy, details of which are presented by Gullu (2013), is
graphically illustrated in Figure 2.20 and Figure 2.21.
Figure 2.20. Representation of the Nonlinear Analysis Solution Strategy for a Single
Degree of Freedom System (Clarke and Hancock, 1990).
38
Figure 2.21. Iterative Strategy and Residual Displacements (Clarke and Hancock,
1990).
39
3. EXPERIMENTAL CALIBRATION OF THE MODEL
The finite element model formulation, together with the constitutive material and
behavioral models described in the previous Chapter, were implemented in Matlab for
conducting nonlinear analysis to simulate the hysteretic lateral load behavior of squat
wall specimens. The calibrated and analytically compared tests include eleven squat
wall specimens tested by Terzioglu (2011) at Boğaziçi University Structural Engineer-
ing Laboratory, ten squat wall specimens tested by Massone (2006) at University of
California Los Angeles, and four squat wall specimens tested by Orakcal et al., (2009)
at University of California Los Angeles. This chapter provides a description of the ex-
perimental program and presents information on experimental calibration of the wall
model to represent the characteristics of wall specimens and the loading conditions.
3.1. Description of the Wall Specimens
In this study, fifteen specimens were selected to validate the proposed analytical
model. In the subsequent sections, description of the wall specimens tested at Univer-
sity of California Los Angeles and Boğaziçi University is illustrated. The selected wall
specimens all had relatively low aspect ratio and were expected to experience shear-
controlled responses. Detailed information about the tests conducted can be found
in Ph.D. Thesis by Massone (2006), the paper by Orakcal et al., (2009), and in M.S.
Thesis by Terzioglu (2011).
3.1.1. Wall Specimens Tested at University of California Los Angeles
Tests conducted at UCLA were spandrel and pier walls representative of wall seg-
ments in existing perforated perimeter wall buildings (from 1940 to 1970) in California.
The walls were 3/4-scale, spandrel walls included weakened plane joints (reduced con-
crete cross-section and cut longitudinal bars as crack initiator) while pier walls had
straight steel transverse (not hooked) reinforcement. Two types of spandrel walls were
tested by Massone (2006) and differentiated by boundary reinforcement ratio while
40
three types of pier walls tested by Massone (2006) were differentiated by the amount
of axial load. Two types of spandrel walls were tested by Orakcal et al., (2009) and
differentiated by concrete strength and loading history. For every type, two wall spec-
imens were tested. The specimens having the same aspect ratio, same horizontal and
vertical web reinforcement ratios, same boundary reinforcement ratio and location of
the weakened joint plate, if present, were considered as being of the same type. If any
of these properties would change, the type of the specimen would change. The naming
scheme used for the wall specimens is shown in Figure 3.1.
Figure 3.1. Specimen Name Description for Tests Conducted at UCLA.
A summary of the geometric and reinforcement properties of the wall specimens

are provided in Table 3.1. The steel reinforcement ratios were calculated as the total
steel area divided by the tributary area of concrete. For the boundary reinforcement,
the tributary area was defined as the full thickness of the wall multiplied by the length
of the boundary region, defined as the center-to-center distance between the boundary
reinforcing bars plus two clear covers (one per side). In Table 3.1, lw is length of
the wall, hw is height of the wall, ρt is transverse reinforcement ratio, ρl is transverse
reinforcement ratio, and ρb is boundary reinforcement ratio. The thickness for all
specimens was 15.2 cm.
41
Table 3.1. Test Specimen Details for Tests Conducted at UCLA.
Horizontal Vertical Boundary

Specimen Iw (cm) hw (cm)
Reinf. Bars ρt (%) Reinf. Bars ρl (%) Reinf. Bars ρb (%)
WS-T1-S1 152 152 φ13@33 cm 0.278 φ13@23 cm 0.428 4φ16 3.12
WS-T1-S2 152 152 φ13@33 cm 0.278 φ13@23 cm 0.428 4φ16 3.12
WS-T2-S1 152 152 φ13@33 cm 0.278 φ13@23 cm 0.4 1φ13 +1φ16 1.7
WS-T2-S2 152 152 φ13@33 cm 0.278 φ13@23 cm 0.4 1φ13 +1φ16 1.7
WS-T3-S1 152 152 φ13@28 cm 0.278 φ13@28 cm 0.256 2φ13 1.33
WS-T3-S2 152 152 φ13@28 cm 0.278 φ13@28 cm 0.256 2φ13 1.33
WS-T4-S1 152 152 φ13@28 cm 0.278 φ13@28 cm 0.256 2φ13 1.33
WS-T4-S2 152 152 φ13@28 cm 0.278 φ13@28 cm 0.256 2φ13 1.33
WP-T5-N0-S1 137 122 φ13@30.5 cm 0.278 φ13@33 cm 0.227 2φ13 1.33
WP-T5-N0-S2 137 122 φ13@30.5 cm 0.278 φ13@33 cm 0.227 2φ13 1.33
WP-T5-N5-S1 137 122 φ13@30.5 cm 0.278 φ13@33 cm 0.227 2φ13 1.33
WP-T5-N5-S2 137 122 φ13@30.5 cm 0.278 φ13@33 cm 0.227 2φ13 1.33
WP-T5-N10-S1 137 122 φ13@30.5 cm 0.278 φ13@33 cm 0.227 2φ13 1.33
WP-T5-N10-S2 137 122 φ13@30.5 cm 0.278 φ13@33 cm 0.227 2φ13 1.33
Both spandrel and pier specimens were secured to a loading frame and strong floor
by means of reinforced concrete pedestals at the top and bottom. The top and bottom
pedestals of the specimens were designed to resist shear forces and bending moments
generated by applied lateral loads. The weakened plane joints were attained by reducing
concrete cross section through introducing wood edges in the specimens form and by
reducing the effective longitudinal distributed steel reinforcement by cutting four out
of six bars (Figure 3.2). The wood wedges were placed along the whole length of the
specimen.
Figure 3.2. Weakened Plate Joint on Wall Spandrels (Massone et al., 2009).
Spandrel specimens were 152 cm in tall, 152 cm in long and 15.2 cm in thick. The
42
geometry and distribution of reinforcing steel of wall section are displayed in Figure
3.3. Distributed vertical and horizontal reinforcement were placed in one curtain only.
For Type 1 and 2 spandrels, the vertical distributed reinforcement was φ13@23 cm
while the horizontal distributed reinforcement was φ13@33 cm. For Type 3 and 4
spandrels, both the vertical and horizontal distributed reinforcement were φ13@28 cm.
As boundary reinforcement, 4φ16 were used for Type 1 walls, 1φ16 + 1φ13 were used
for Type 2 walls, and 2φ13 were used for Type 3 and Type 4 walls. At boundaries, the
horizontal bars were terminated with a 180o hook. The boundary bars were not cut at
the weakened plate joint.
Wall pier specimens were 122 cm in tall, 137 cm in long and 15.2 cm in thick.
The geometry and distribution of reinforcing steel of the wall specimen are shown in
Figure 3.4. Distributed vertical and horizontal reinforcement were placed in one curtain
only. The vertical distributed reinforcement was φ13@30.5 cm while the horizontal
distributed reinforcement was φ13@33 cm. As boundary reinforcement, 2φ13 were
used for all piers. The horizontal bars did not incorporate hooks at the ends. Some of
the wall pier specimens were also subjected to axial loads; the axial load levels applied
on the wall piers corresponded to 0%, 5%, and 10% of their axial compression capacity.
Figure 3.3. Wall Spandrel Geometry and Reinforcement (Massone et al., 2009).
43
Figure 3.4. Geometry and Steel Reinforcement for Wall Piers (Massone et al., 2009).
3.1.2. Wall Specimens Tested at Boğaziçi University
Wall tests conducted at Boğaziçi University included eleven squat wall specimens
with low aspect ratio. Three sets of specimens were tested. The first set of specimens
included three squat walls with aspect ratios of 0.5. The second set consisted of four
structural walls, one having 0.33 aspect ratio, one having 1.0 aspect ratio and two
having 0.5 aspect ratio. The third set included one 1.0 aspect ratio specimen and
three 0.5 aspect ratio specimens. Two of the specimens were tested under different
level constant axial loads. The specimens having the same aspect ratio, reinforcement
ratio, and testing sequence were considered as being of the same type. If any of these
properties changed, the type of the specimen would change. The naming scheme used
for the wall specimens is shown in Figure 3.5.
Figure 3.5. Specimen Name Description for Tests Conducted at Boğaziçi University.
44
A summary of the geometric and reinforcement characteristics of the walls is pro-

vided in Table 3.2. Reinforcement ratios were calculated as the total steel area divided
by the tributary area of concrete. For the boundary reinforcement, the tributary area
was defined as the full thickness of the wall multiplied by the length of the boundary
region, defined as the center-to-center distance between the boundary reinforcing bars
plus two clear covers (one per side). In Table 3.2, lw is length of the wall, hw is height
of the wall, ρt is transverse reinforcement ratio, ρl is transverse reinforcement ratio,
and ρb is boundary reinforcement ratio. The thickness for all specimens was 12 cm.
The specimen SW-T2-S1-1 and the specimen SW-T2-S2-3 differed only in terms of
their horizontal reinforcement anchorage conditions. Specimen SW-T2-S1-1 had U-cap
hooks at the ends of the horizontal reinforcement bars, for anchorage. On the other
hand specimen SW-T2-S2-3, as well as all other specimens, had 180o hooks on the
horizontal bars. Figure 3.6 shows the details of the hooks used. Another variation
existed in the Type1 specimens. The longitudinal boundary reinforcement of specimen
Type1-S2 was confined with ties, whereas no confinement was provided for specimen
Type1-S1, in order to observe the difference between these two specimens in terms of
strength and ductility.
Table 3.2. Test Specimen Details for Tests Conducted at Boğaziçi University.
Horizontal Vertical Boundary

Specimen Iw (cm) hw (cm)
Reinf. Bars ρt (%) Reinf. Bars ρl (%) Reinf. Bars ρb (%)
SW-T2-S1-1 150 75 φ8@12.5 cm 0.68 φ8@12.5 cm 0.68 4φ16 5.15
SW-T1-S1-2 150 75 φ8@12.5 cm 0.34 φ8@25 cm 0.34 4φ16 5.15
SW-T2-S2-3 150 75 φ8@12.5 cm 0.68 φ8@12.5 cm 0.68 4φ16 5.15
SW-T2-S3-4 150 75 φ8@12.5 cm 0.68 φ8@12.5 cm 0.68 4φ16 5.15
SW-T3-S1-5 150 75 φ8@12.5 cm 0.68 φ8@12.5 cm 0.68 2φ8 0.65
SW-T4-S1-6 150 50 φ8@12.5 cm 0.68 φ8@12.5 cm 0.68 4φ14 3.95
SW-T5-S1-7 150 150 φ8@12.5 cm 0.68 φ8@25 cm 0.34 4φ22 9.75
SW-T6-S1-8 150 150 φ8@25 cm 0.34 φ8@25 cm 0.68 4φ22 9.75
SW-T1-S2-9 150 75 φ8@25 cm 0.34 φ8@25 cm 0.34 4φ16 5.15
SW-T1-N5-S1-10 150 75 φ8@25 cm 0.34 φ8@25 cm 0.34 4φ16 5.15
SW-T1-N10-S1-11 150 75 φ8@25 cm 0.34 φ8@25 cm 0.34 4φ16 5.15
The specimens were connected to the strong floor by means of a bottom pedestal.
The top pedestal (load transfer beam) was connected to the horizontal actuator. The
beam and pedestal of the specimens were designed to resist shear forces and bending
moments generated by applied lateral loads. The bottom pedestal had dimensions
of 1.0 m width and 2.1 m length. The dimensions of the bottom pedestal were also
45
designed to create enough frictional area so that the specimen does not slide under
high lateral loads.
For the specimens, clear cover was 40 mm and the distance between boundary bars
along the length was 50 mm. Hence, the boundary tributary area width corresponded
to 130 mm. Specimens with three different aspect-ratio were constructed. There were
a total of eight 0.5 aspect ratio wall specimens, one 0.33 aspect ratio wall specimen,
and two 1.0 aspect ratio wall specimens.
The specimens with 0.5 aspect ratio had length of 1500 mm, height of 750 mm
and thickness of 120 mm. They differed by their web reinforcement ratio or boundary
reinforcement amount. Four of the Type-1 specimens had same transverse (horizontal)
and vertical (longitudinal) web reinforcement ratios of ρt = ρl = 0.34%. To differentiate,
two of these Type-1 specimens were tested under different axial load levels. These two
specimens were named as SW-T1-N5-S1-10 and SW-T1-N10-S1-11. The abbreviations
N5 and N10 represent the axial load level applied (e.g. N5 refers to 5% of wall axial load
capacity). Another Type-1 specimen was Type1-S2, which had confined boundaries
at the ends. This specimen was constructed in order to see the effect of confined
boundaries on the behavior. Specimen type T2-S1, which had horizontal (transverse)
and vertical longitudinal web reinforcement ratios of ρt = ρl = 0.68% and U-caps at
the end of the horizontal reinforcing bars. The parameters used for the construction
of the third Type 2 specimen (T2-S3) were the same as the T2-S2 specimens. All the
parameters were kept identical. The only parameter that differentiated these specimens
was the concrete compressive strength, which was higher for the T2-S3 specimen, due
to time of testing after concrete was poured. Details of specimen T1-S1 is illustrated
in Figure 3.6.
Uniformly distributed vertical and horizontal web reinforcement of φ8 reinforcing

steel bars were used in the wall specimens. Vertical reinforcement was placed uniformly
along the wall section. Since there was enough space for development length in top
and bottom pedestals, vertical reinforcing bars were not hooked at the ends. For all
specimens, the distance between the two curtains of vertical web reinforcement was 62
46
mm. The horizontal web reinforcement was distributed along the height of the wall
with equal spacing. The horizontal web bars were anchored with 180o hooks at the
ends. Only for the T2-S1 specimen, horizontal reinforcing bars were terminated at the
edges without hooks, and U-cap reinforcing bars were used for anchorage.
The longitudinal boundary bars were placed at both edges of the wall, inside the
hooks of the horizontal bars. However, the boundary regions were not confined with
ties, except for specimen T1-S2. Two different types of boundary reinforcement were
used for the 0.5 aspect ratio structural walls. Seven of the structural walls had 4φ16
reinforcing bars at their boundaries. T3-S1, which had similar web reinforcement with
Type 2 walls, had 2φ8 boundary reinforcement at both edges. The flexural yielding of
the boundary reinforcement was targeted in design of specimen T3-S1.
Figure 3.6. Geometry and Wall Reinforcement of Specimen T1-S1 (Terzioglu, 2011).
The specimen having a 0.33 aspect ratio was 1500 mm long, 500 mm tall and
120 mm thick. It had web reinforcement ratios of ρt = ρl = 0.68%. The geometry and
reinforcement details of specimen T4-S1 were displayed in Figure 3.7. The longitudinal
boundary bars were placed at both edges of the wall. They were placed inside the hooks
of the horizontal web bars. The boundary regions were not confined with ties. For this
specimen, 4φ14 boundary bars were provided at both ends of the wall. The amounts of
boundary bars provided targeted simultaneous reaching of flexural and shear capacities
47
of the wall.
Two structural walls having 1.0 aspect ratio were constructed, The Type 5 and
Type 6, with identical horizontal web reinforcement amounts. Both of the walls had
0.68% horizontal reinforcement ratios, whereas the Type 5 specimen had 0.34% vertical
web reinforcement ratio, and the Type 6 specimen has 0.68% vertical web reinforcement
ratio. Figure 3.8 shows the geometry and reinforcement details of specimen T5-S1.
The height of the wall specimens was 1500 mm, the length was 1500 mm, and the
thickness was 120 mm. For specimen T5-S1, φ8 vertical bars were placed with 250 mm
spacing, and for specimen T6-S1, they were placed with 125 mm spacing. The vertical
bars did not incorporate hooks at the ends, since there was enough embedment length
in the top and bottom pedestals for the bars to develop. φ8 horizontal web bars were
distributed along the height, with 125 mm spacing and terminated at the edges of the
wall with 180o hooks.
The boundary reinforcement was placed inside the horizontal bars at both edges.
Boundary regions were not confined with ties. 4φ22 bars were used at both edges as
boundary reinforcement. The boundary bars were anchored with 90o hooks inside the
top and bottom pedestals.
48
3.2. Test Setup
The UCLA specimens were tested in upright position. The general view of the
setup is shown in Figure 3.9. The specimens were designed to be tested under a zero
rotation at both ends (beam and pedestal), providing a constant axial load and a
variable cyclic lateral load applied at mid-height of the specimens, creating a double-
curvature loading condition, where the bending moment on each wall is zero at wall
midheight and maximum (with reverse signs) at the top and bottom cross-sections of
the wall, representing the actual loading conditions on wall piers and spandrels in a
perforated perimeter wall of a building. Two, vertically placed, 1780 kN hydraulic
actuators were used to apply the constant axial load and to prevent the top of the wall
from rotating. A horizontally placed 890 kN hydraulic actuator was used to apply the
lateral load. An out-of-plane frame was used to prevent twisting of each wall.
49
Figure 3.9. Test Setup (Massone, 2006).
Using the horizontal actuators, force-controlled cycles were performed followed

by displacement-controlled cycles. Two levels of force-controlled cycles were applied.
For two tests, the levels were 66.7 kN and 133.4 kN, whereas for the remainder, the
levels were 111.2 kN and 222.4 kN. Three full cycles (push and pull) for each force
level were applied. The forces were applied at rates from4.448 kN/s to 8.896 kN/s.
This procedure was followed by three complete cycles (push and pull) of the following
drift levels: 0.2%, 0.3%, 0.4%, 0.6%, 0.8%, 1.2%, 1.6%, 2.0%, 2.4%, and 3.2%. The
displacements were applied at rates from 0.0127 cm/s to 0.0762 cm/s. The applied
process is described in Figure 3.10.
Figure 3.10. Displacement History (Massone, 2006).

50
The first set of wall specimens tested at Boğaziçi University included specimens
T1-S1and T2-S1 and T2-S2. The second group consisted of specimens T2-S3, T3-
S1, T4-S1, and T5-S1. Remaining four specimens constituted the third group, three of
which were identical to Type1 specimen, and were tested under axial load. Lateral load
on the specimens was applied using an actuator with a capacity of 1000 kN. One end
of the actuator was fixed to a steel reaction wall. Figure 3.11 shows a photo of the test
setup. Two specimens were tested under 5% Agf c and 10% Agf c axial load levels. The
axial load was applied using four pre-stressed steel cables each having 250 kN axial
tension capacities. The axial load was applied through a hydraulic cylinder placed
between the steel beams and the top pedestal (load transfer beam) of each specimen.
An out-of-plane frame was designed to prevent twisting of the wall specimens.
Figure 3.11. Test Setup (Terzioglu, 2011).
Reversed cyclic lateral loading was applied at the top of each specimen, resulting
in a single curvature (cantilever) loading condition. Three full cycles were applied at
each drift level of 0.05%, 0.1%, 0.15%, 0.2%, 0.3%, 0.4%, 0.6%, 0.8%, 1.0%, 1.2%, 1.4%,
1.6%, 1.8%, 2.0%, 2.4%, and 3.2%. Depending on their residual lateral load capacity,
one or two more drift levels were imposed on selected specimens.
51
3.3. Material Properties
For the wall pier and spandrel specimens tested at UCLA, the reinforcing steel
was classified as Grade 280 (minimum yield strength of 280 MPa) by the manufacturer.
Tensile tests were performed for the reinforcing bars through tensile coupon samples
with a 203.2 mm (8 in) long reduced section (removed ribs). φ13 (6=4) and φ 16 (6=5)
bars were tested in tension. Test data revealed that for Type 1, Type 2, and Type
5 wall specimens, the yield stress capacity for φ 13 bars was 424 MPa (61.5 ksi) and
for 16 bars was 448.2 MPa (65 ksi). For Type 3 and Type 4 wall specimens, the yield
stress capacity for φ 13 bars was 351.6 MPa (51 ksi). Steel stress-strain curves for bars
used in Type 1 wall specimens are illustrated in Figure 3.12.
Figure 3.12. Steel Tensile Stress-Strain Curves for φ 13 and φ 16 bars of Type 1 Wall
Specimen (Massone, 2006).
For the UCLA specimens, the targeted compressive strength range for concrete
was 25 MPa - 30 MPa. Test-date average compressive strength values of the concrete
used in construction of the specimens, are listed in Table 3.3.
52
Table 3.3. Average Compressive Strength of Concrete (Massone, 2006).
Specimen f’c (Mpa)

WS-T1-S1 25.5
WS-T1-S2 43.7
WS-T2-S1 31.4
WS-T2-S2 31.0
WS-T3-S1 31.7
WS-T3-S2 33.6
WS-T4-S1 31.9
WS-T4-S2 33.0
WP-T5-N0-S1 29.9
WP-T5-N0-S2 31.0
WP-T5-N5-S1 31.9
WP-T5-N5-S2 32.0
WP-T5-N10-S1 28.3
WP-T5-N10-S2 31.4
For the wall specimens tested at Boğaziçi University, the targeted tensile strength
of reinforcing bars tested was 420 MPa. Tensile tests were performed for the reinforcing
bars through tensile coupon samples with 600 mm length. For the first set of specimens,
there were two different diameter types, φ8 and φ16. The tension test results for the
first set of specimens are provided in Figure 3.13. For the second and third set of
specimens, φ14 reinforcing bars were used for the boundary elements of 0.33 aspect
ratio specimen and φ22 reinforcing bars were used for the boundary bars of the 1.0
aspect ratio specimens. The tensile stress-strain properties of the bars used for the
second and third sets are provided in Figure 3.14 and Figure 3.15.
For the Boğaziçi University specimens, the targeted compressive strength of con-
crete was 25 MPa - 30 MPa. The concrete used in specimen construction exhibited
different test-date cylinder compressive strengths, ranging between 20 MPa - 35 MPa,
depending on the date of testing. For each wall specimen, 3 standard (150 x 300 mm)
cylinders were tested at the test date. Test-day average compressive strength values of
concrete used in the construction of the specimens are listed in Table 3.4.
53
Figure 3.13. Steel Stress-Strain Curves for Specimen Set 1 (Terzioglu, 2011).
Figure 3.14. Steel Stress-Strain Curves for Web Reinforcement of Specimen Set 2 and
Set 3 (Terzioglu, 2011).
Figure 3.15. Steel Stress-Strain Curves for Boundary Reinforcement of Specimen Set
2 and Set 3 (Terzioglu, 2011).
54
Table 3.4. Average Compressive Strength of Concrete (Terzioglu, 2011).
Specimen f’c (Mpa)

SW-T2-S1-1 19.3
SW-T1-S1-2 23.7
SW-T2-S2-3 25.8
SW-T2-S3-4 29.0
SW-T3-S1-5 32.1
SW-T4-S1-6 34.8
SW-T5-S1-7 35.0
SW-T6-S1-8 22.6
SW-T1-S2-9 24.0
SW-T1-N5-S1-10 26.3
SW-T1-N10-S1-11 27.0
3.4. Failure Modes of the Wall Specimens
Spandrel structural wall specimens tested at University of California Los Angeles

experienced diagonal tension failure except for Type 4 specimens for which the failure
mode was shear sliding. The important characteristic of the Type 4 specimens was the
location of weakened plane joint, as it was located at the wall-pedestal interface. The
wall pier specimens also experienced diagonal tension failure, due to their inadequate
web reinforcement ratios.
For the walls tested at Boğaziçi University, all three types of shear failure were
observed. Type 1 specimens failed under diagonal tension because of their relatively low
web reinforcement ratio. Type 2, 4, 5, and 6 walls experienced diagonal compression
failure. These walls incorporated relatively high reinforcing ratio in the web. For the
special case of the Type 3 specimen, which had relatively low longitudinal reinforcement
at wall boundaries, shear cracking at the wall-pedestal interface was observed, leading
to sliding failure at the base of the wall.
For all of the wall specimens investigated, shear was the predominant mode of
55
behavior, while shear flexure interaction also played an important role in the response
of the walls with an aspect ratio of 1.0. Degradation of lateral stiffness and lateral load
capacity at low drift levels were common in all walls, together with highly-pinched
lateral load-displacement behavior, indicating shear-controlled responses.
3.5. Calibration of the Model for the Wall Specimens
The finite element model configurations used for modeling and analysis of the
test specimens were calibrated for several aspects of the wall specimens, including
the geometric properties, reinforcement attributes, material characteristics, and load-
ing conditions. As-tested properties of the materials employed in the construction
of the specimens are used to calibrate the constitutive material parameters used in
the model formulation. In calibration of the parameters for concrete and reinforcing
steel in tension, tension stiffening effects were also considered. As well, the consti-
tutive material parameters were also calibrated in terms of the expected compressive
and the tensile stress-strain curves of concrete. Either as-tested material properties or
well-established empirical relationships presented in the literature are implemented for
calibrations. The material parameters were not adjusted to fit the model predictions
with the experimental results obtained for the wall specimens. Further information on
the calibration of the model is provided in this section.
3.5.1. Calibration of Model Element Dimensions and Reinforcement Ratios
Finite element models generated for the wall specimens were divided into an ap-
propriate number of constitutive panel elements (mesh sizes) in order to obtain almost
square-shaped constitutive elements. Corresponding dimensions and numbers of the
constitutive model elements (mesh dimension and mesh number) for tests conducted
at UCLA and Boğaziçi University are shown in Table 3.5 and Table 3.6, respectively.
The reinforcement ratios in the boundary and web regions of the walls were calibrated
according to the values presented in Table 3.1 and Table 3.2.
56
Table 3.5. Wall Model Geometry Details for Tests Conducted at UCLA.
Model Element Dimension Model Element Number

Horizontal Vertical Horizontal Vertical
Specimen
Boundary Region Web Region Boundary Region Web Region
WS-T1-S1 16.9 cm 16.9 cm 16.9 cm 2 7 9
WS-T1-S2 16.9 cm 16.9 cm 16.9 cm 2 7 9
WS-T2-S1 12.8 cm 12.6 cm 12.7 cm 2 10 12
WS-T2-S2 12.8 cm 12.6 cm 12.7 cm 2 10 12
WS-T3-S1 12.8 cm 12.6 cm 12.7 cm 2 10 12
WS-T3-S2 12.8 cm 12.6 cm 12.7 cm 2 10 12
WS-T4-S1 12.8 cm 12.6 cm 12.7 cm 2 10 12
WS-T4-S2 12.8 cm 12.6 cm 12.7 cm 2 10 12
WP-T5-N0-S1 12.8 cm 12.4 cm 12.2 cm 2 9 10
WP-T5-N0-S2 12.8 cm 12.4 cm 12.2 cm 2 9 10
WP-T5-N5-S1 12.8 cm 12.4 cm 12.2 cm 2 9 10
WP-T5-N5-S2 12.8 cm 12.4 cm 12.2 cm 2 9 10
WP-T5-N10-S1 12.8 cm 12.4 cm 12.2 cm 2 9 10
WP-T5-N10-S2 12.8 cm 12.4 cm 12.2 cm 2 9 10
Table 3.6. Wall Model Geometry Details for Tests Conducted at Boğaziçi University.
Model Element Dimension Model Element Number

Horizontal Vertical Horizontal Vertical
Specimen
Boundary Region Web Region Boundary Region Web Region
SW-T2-S1-1 13.0 cm 12.4 cm 12.5 cm 2 10 6
SW-T1-S1-2 13.0 cm 12.4 cm 12.5 cm 2 10 6
SW-T2-S2-3 13.0 cm 12.4 cm 12.5 cm 2 10 6
SW-T2-S3-4 13.0 cm 12.4 cm 12.5 cm 2 10 6
SW-T3-S1-5 11.4 cm 10.4 cm 10.7 cm 4 10 7
SW-T4-S1-6 13.0 cm 12.4 cm 12.5 cm 2 10 4
SW-T5-S1-7 13.0 cm 12.4 cm 12.5 cm 2 10 12
SW-T6-S1-8 13.0 cm 12.4 cm 12.5 cm 2 10 12
SW-T1-S2-9 9.7 cm 9.2 cm 9.4 cm 6 10 8
SW-T1-N5-S1-10 13.0 cm 12.4 cm 12.5 cm 2 10 6
SW-T1-N10-S1-11 13.0 cm 12.4 cm 12.5 cm 2 10 6
3.5.2. Calibration of Constitutive Material Parameters
Parameters of the constitutive material models implemented in the model for-

mulation were calibrated to represent the as-tested properties of the materials used
in the construction of the specimens, whenever available. Calibration studies were
also conducted on the concrete constitutive model for representation of the shape of
the stress-strain curve for concrete in compression. As well, the stress-strain behavior
of concrete in tension and reinforcing steel in tension were calibrated to incorporate
tension stiffening effects.
57
3.5.2.1. Calibration of the Constitutive Model for Reinforcing Steel. Belarbi and Hsu
(1994) presented the average stress-strain relationship of the reinforcing steel bars
covered by concrete and showed that the relationship differs from that of the bare
bars, especially when the yield stress is of concern. When embedded in concrete, the
yield stress of steel a bar due to the so-called tension stiffening effect. The stress-strain
behavior of a bare bar and a bar embedded in concrete is illustrated in Figure 3.16.
Figure 3.16. The Stress-Strain Behavior of Bare Rebar and Rebar Embedded in
Concrete.
Experimental results conducted by Belarbi and Hsu (1994) on uniaxially-loaded

column specimens form the basis for the empirical relationships between the behavior of
bare bars and bars embedded in concrete is derived based on. The parameters affecting
tension stiffening on the average yield stress of reinforcing steel are the cross-sectional
area ratio of the longitudinal steel (ρ), and the ratio of concrete tensile strength (fcr )
to the yield stress of the bare bar (σy ). Moreover, the strain hardening ratio of an
embedded bar in concrete and the strain hardening ratio of a bare bar are found to be
different.
The monotonic stress-strain envelope of bar embedded in concrete is represented

58
by two regions, the first is the elastic region with a slope equal to the elastic modulus
of reinforcing steel (Es ), and the second is the post-yield range with a slope E*p (Figure
3.17).
Figure 3.17. Stress-Strain Relationship for a Bar Embedded in Concrete (Belarbi and
Hsu, 1994).
The effective yield stress at the intersection of the two regions (σ∗n ) and the
plastic slope (E*p ) are given by the following expressions:
σn∗ = (0.93 − 2B)σy (3.1)
Ep∗ = (0.8b + 0.25B)Es (3.2)
where B is defined as:
!1.5
1 fcr
B= (3.3)
ρ σy
and b is the strain hardening ratio (Ep /Es ) defined for the bare bar.
59
Accordingly, the model of Belarbi and Hsu (1994) for modeling the average stress-
strain behavior of bars in concrete (Figure 3.17) is defined as:
If εs ≤ εn , σs = Es εs (3.4)
If εs > εn , σs = (0.93 − −2B) + (0.8b + 0.25B) Es (εs − εn ) (3.5)
where σs is the average stress, εs is the average strain and εn is the average strain
defined at the intersection of the two lines as defined in equation below:
σn∗ (0.93 − 2B) σy

εn = = (3.6)
Es Es
In the equations above, Es ,σy and b are the modulus of elasticity, yield stress, and
strain hardening ratio experimentally-measured for bare bars.
The calibrated parameters for the reinforcing steel bars in the wall specimens are
presented in the following tables, which include the reinforcement ratios (ρ), the strain
hardening slopes (E*p ), and the stress (σn ) and strain (εn ) values at the intersection of
the elastic and post-yield regions. Tension stiffening effects influence the constitutive
parameters for reinforcing bars in tension only, whereas the parameters for bars in
compression are identical to those obtained from bare bar tests.
Table 3.7, Table 3.8, and Table 3.9 represent the calibrated values for tests con-
ducted at UCLA. There are two different bar diameters, φ13 and φ16. φ13 bars have
different yield strength for Type 3 and Type 4 and remaining specimens.
60
Table 3.7. Calibrated Values of the Constitutive Parameters of Reinforcing Steel Bars
for Specimens WS-T1-S1, WS-T1-S2, WS-T2-S1, and WS-T2-S2.
Compression Tension
Specimen Mesh Location φ (mm) ρ
fy (Mpa) b σ∗n (Mpa) b’
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00660
Vertical φ16 3.12 448 0.00800 417 0.00642
Web Mesh
WS-T1-S1
Horizontal φ13 0.278 424 0.00800 394 0.00660
Vertical φ13 0.428 424 0.00800 394 0.00653
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 393 0.00670
Vertical φ16 3.12 448 0.00800 417 0.00642
Inner Mesh
WS-T1-S2
Horizontal φ13 0.278 424 0.00800 393 0.00670
Vertical φ13 0.428 424 0.00800 394 0.00660
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00664
Vertical φ13-φ16 1.7 436 0.00800 405 0.00644
Inner Mesh
WS-T2-S1
Horizontal φ13 0.278 424 0.00800 394 0.00664
Vertical φ13 0.4 424 0.00800 394 0.00656
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00663
Vertical φ13-φ16 1.7 436 0.00800 405 0.00644
Inner Mesh
WS-T2-S2
Horizontal φ13 0.278 424 0.00800 394 0.00663
Vertical φ13 0.4 424 0.00800 394 0.00656
for Specimens WS-T3-S1, WS-T3-S2, WS-T4-S1, and WS-T4-S2.
Compression Tension
Boundary Mesh
Horizontal φ13 0.278 351.6 0.00800 326 0.00671
Vertical φ13 1.33 351.6 0.00800 327 0.00647
Inner Mesh
WS-T3-S1
Horizontal φ13 0.278 351.6 0.00800 326 0.00671
Vertical φ13 0.256 351.6 0.00800 326 0.00674
Boundary Mesh
Horizontal φ13 0.278 351.6 0.00800 326 0.00673
Vertical φ13 1.33 351.6 0.00800 327 0.00647
Inner Mesh
WS-T3-S2
Horizontal φ13 0.278 351.6 0.00800 326 0.00673
Vertical φ13 0.256 351.6 0.00800 326 0.00676
Boundary Mesh
Horizontal φ13 0.278 351.6 0.00800 326 0.00672
Vertical φ13 1.33 351.6 0.00800 327 0.00647
Inner Mesh
WS-T4-S1
Horizontal φ13 0.278 351.6 0.00800 326 0.00672
Vertical φ13 0.256 351.6 0.00800 326 0.00674
Boundary Mesh
Horizontal φ13 0.278 351.6 0.00800 326 0.00672
Vertical φ13 1.33 351.6 0.00800 327 0.00647
Inner Mesh
WS-T4-S2
Horizontal φ13 0.278 351.6 0.00800 326 0.00672
Vertical φ13 0.256 351.6 0.00800 326 0.00675
61
for Specimens WP-T5-N0-S1, WP-T5-N0-S2, WP-T5-N5-S1, WP-T5-N5-S2,
WP-T5-N10-S1, and WP-T5-N10-S2.
Compression Tension
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00663
Vertical φ13 1.33 424 0.00800 394 0.00645
Inner Mesh
WP-T5-N0-S1
Horizontal φ13 0.278 424 0.00800 394 0.00663
Vertical φ13 0.227 424 0.00800 393 0.00668
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00663
Vertical φ13 1.33 424 0.00800 394 0.00645
Inner Mesh
WP-T5-N0-S2
Horizontal φ13 0.278 424 0.00800 394 0.00663
Vertical φ13 0.227 424 0.00800 393 0.00669
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00664
Vertical φ13 1.33 424 0.00800 394 0.00645
Inner Mesh
WP-T5-N5-S1
Horizontal φ13 0.278 424 0.00800 394 0.00664
Vertical φ13 0.227 424 0.00800 393 0.00669
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00664
Vertical φ13 1.33 424 0.00800 394 0.00645
Inner Mesh
WP-T5-N5-S2
Horizontal φ13 0.278 424 0.00800 394 0.00664
Vertical φ13 0.227 424 0.00800 393 0.00669
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00662
Vertical φ13 1.33 424 0.00800 394 0.00645
Inner Mesh
WP-T5-N10-S1
Horizontal φ13 0.278 424 0.00800 394 0.00662
Vertical φ13 0.227 424 0.00800 393 0.00667
Boundary Mesh
Horizontal φ13 0.278 424 0.00800 394 0.00664
Vertical φ13 1.33 424 0.00800 394 0.00645
Inner Mesh
WP-T5-N10-S2
Horizontal φ13 0.278 424 0.00800 394 0.00664
Vertical φ13 0.227 424 0.00800 393 0.00669
Tables Table 3.10, Table 3.11, and Table 3.12 represent the calibrated values for
tests conducted at Boğaziçi University. There are four different bar diameters. φ8,
φ14, φ16, and φ22 were used for boundary regions while φ8 was used to reinforce the
web regions in the specimens.
62
Table 3.10. Calibrated Values of the Constitutive Parameters of Reinforcing Steel

Bars for Specimens SW-T2-S1-1, SW-T1-S1-2, and SW-T2-S2-3.
Compression Tension
Boundary Mesh
Horizontal φ8 0.68 500 0.00323 465 0.00261
Vertical φ16 5.15 440 0.0043 409 0.00345
Web Mesh
SW-T2-S1-1
Horizontal φ8 0.68 500 0.00323 465 0.00261
Vertical φ8 0.68 500 0.00323 465 0.00261
Boundary Mesh
Horizontal φ8 0.34 500 0.00323 465 0.00266
Vertical φ16 5.15 440 0.0043 409 0.00345
Inner Mesh
SW-T1-S1-2
Horizontal φ8 0.34 500 0.00323 465 0.00266
Vertical φ8 0.34 500 0.00323 465 0.00266
Boundary Mesh
Horizontal φ8 0.68 500 0.00323 465 0.00262
Vertical φ16 5.15 440 0.0043 409 0.00345
Inner Mesh
SW-T2-S2-3
Horizontal φ8 0.68 500 0.00323 465 0.00262
Vertical φ8 0.68 500 0.00323 465 0.00262

Bars for Specimens SW-T2-S3-4, SW-T3-S1-5, SW-T4-S1-6, and SW-T5-S1-7.
Compression Tension
Boundary Mesh
Horizontal φ8 0.68 575 0.00815 535 0.00656
Vertical φ16 5.15 525 0.00282 488 0.00227
Inner Mesh
SW-T2-S3-4
Horizontal φ8 0.68 575 0.00815 535 0.00656
Vertical φ8 0.68 575 0.00815 535 0.00656
Boundary Mesh
Horizontal φ8 0.68 575 0.00815 535 0.00657
Vertical φ8 0.65 575 0.00815 535 0.00657
Inner Mesh
SW-T3-S1-5
Horizontal φ8 0.68 575 0.00815 535 0.00657
Vertical φ8 0.68 575 0.00815 535 0.00657
Boundary Mesh
Horizontal φ8 0.68 575 0.00815 534 0.00658
Vertical φ14 3.95 535 0.00395 498 0.00317
Inner Mesh
SW-T4-S1-6
Horizontal φ8 0.68 575 0.00815 534 0.00658
Vertical φ8 0.68 575 0.00815 534 0.00658
Boundary Mesh
Horizontal φ8 0.68 575 0.00815 534 0.00658
Vertical φ22 9.75 550 0.00647 511 0.00518
Inner Mesh
SW-T5-S1-7
Horizontal φ8 0.68 575 0.00815 534 0.00658
Vertical φ8 0.34 575 0.00815 534 0.00663
63

Bars for Specimens SW-T6-S1-8, SW-T1-S2-9, SW-T1-N5-S1-10, and
SW-T1-N10-S1-11.
Compression Tension
Boundary Mesh
Horizontal φ8 0.68 575 0.00815 535 0.00655
Vertical φ22 9.75 550 0.00647 511 0.00518
Inner Mesh
SW-T6-S1-8
Horizontal φ8 0.68 575 0.00815 535 0.00655
Vertical φ8 0.68 575 0.00815 535 0.00655
Boundary Mesh
Horizontal φ8 0.34 575 0.00815 534 0.00659
Vertical φ16 5.15 525 0.00282 488 0.00226
Inner Mesh
SW-T1-S2-9
Horizontal φ8 0.34 575 0.00815 534 0.00659
Vertical φ8 0.34 575 0.00815 534 0.00659
Boundary Mesh
Horizontal φ8 0.34 575 0.00815 534 0.00659
Vertical φ16 5.15 525 0.00282 488 0.00226
Inner Mesh
SW-T1-N5-S1-10
Horizontal φ8 0.34 575 0.00815 534 0.00659
Vertical φ8 0.34 575 0.00815 534 0.00659
Boundary Mesh
Horizontal φ8 0.34 575 0.00815 534 0.00660
Vertical φ16 5.15 525 0.00282 488 0.00226
Inner Mesh
SW-T1-N10-S1-11
Horizontal φ8 0.34 575 0.00815 534 0.00660
Vertical φ8 0.34 575 0.00815 534 0.00660
3.5.2.2. Calibration of the Constitutive Model for Concrete. The constitutive mate-
rial parameters were adjusted through calibration studies to represent the stress-strain
curve for concrete in compression and in tension. The compression envelope curve
of the model by Chang and Mander (1994) is defined by the initial slope Ec (elastic
modulus of concrete), the peak coordinate (ε0c , fc0 ), a parameter r from Tsai’s (1988)
equation defining the shape of the envelope curve, and the parameter x0cr > 1 to define
the normalized strain at which the stress-strain curve transforms into a straight line
(Figure 3.18).
64
Figure 3.18. Compression and Tension Envelope Curves of the Model by Chang and
Mander (1994).
The calibration of the material properties of concrete in compression and in ten-

sion was conducted using different criteria. For concrete in compression, the empirical
relationships recommended by Chang and Mander were utilized for the constitutive
parameters expect for the as-tested compressive strength. In order to calibrate the
critical strain parameter which defines the strain value at which the envelope curve
transitions into a straight line, the Saatcioglu-Razvi Model (1992) for concrete in com-
pression was also implemented. For concrete in tension, the constitutive parameters of
the model were calibrated to represent the empirical relationships suggested by Belarbi
and Hsu (1994), represent tension stiffening effects on the tensile stress-strain behavior
of concrete.
Concrete in Compression: The parameters constituting the compression envelope

curve are the compressive strength of concrete (fc0 ), the strain at compressive strength
(ε0c ), the initial Young’s Modulus (Ec ), the parameter r defining the shape of the enve-
lope and the normalized critical strain x̄cr (also named xcr ) where the curve transitions
into a straight line. Chang and Mander (1994) recommend empirical relationships for
65
the parameters Ec , ε0c and r as:
Initial modulus of elasticity : Ec = 8200(fc0 )(3/8) (3.7)
(fc0 )0.25
Strain at compressive strength : ε0c = (3.8)
28
fc0
Shape parameter : r= − 1.9 (3.9)
5.2
The unit of f’c in the empirical formulas above is MPa.
For the normalized critical strain parameter xcrn , the parameter was calibrated
graphically by comparing the shape of the stress-strain envelope of the Chang and
Mander (1994) model with the Saatcioglu and Razvi (1992) model for unconfined
concrete due to the absence of an empirical relationship. The calibration criterion
was to attain parallel straight lines in the descending region of the two stress-strain
relationships, as shown in Figure 3.19. In order to obtain the calibrated xcrn values,
numerous trials were conducted for each individual specimen. Calibrated constitutive
parameters for concrete in compression are presented in Table 3.13 and Table 3.14 for
all wall specimens.
66
Figure 3.19. Stress-Strain Diagrams for Concrete in Compression for WS-T3-S2.

Compression for the UCLA Specimens.
Specimen f’c (Mpa) Ec (Mpa) ε0c r Xcrn

WS-T1-S1 25.5 27778 0.00183 3.031 1.110
WS-T1-S2 43.7 33996 0.00236 6.551 1.054
WS-T2-S1 31.4 30033 0.00217 4.172 1.095
WS-T2-S2 31.0 29889 0.00204 4.095 1.086
WS-T3-S1 31.7 30140 0.00206 4.230 1.084
WS-T3-S2 33.6 30805 0.00209 4.598 1.075
WS-T4-S1 31.9 30211 0.00206 4.269 1.082
WS-T4-S2 33.0 30598 0.00208 4.482 1.078
WP-T5-N0-S1 29.9 29486 0.00248 3.882 1.145
WP-T5-N0-S2 31.0 29889 0.00266 4.095 1.155
WP-T5-N5-S1 31.9 30211 0.00231 4.269 1.105
WP-T5-N5-S2 32.0 30247 0.00246 4.288 1.118
WP-T5-N10-S1 28.3 28885 0.00210 3.573 1.116
WP-T5-N10-S2 31.4 30033 0.00254 4.172 1.130
67

Compression for the Boğaziçi University Specimens.
Specimen f’c (Mpa) Ec (Mpa) ε0c r Xcrn

SW-T2-S1-1 19.3 25022 0.00182 1.832 1.320
SW-T1-S1-2 23.7 27026 0.00191 2.683 1.155
SW-T2-S2-3 25.8 27900 0.00196 3.089 1.127
SW-T2-S3-4 29.0 29150 0.00201 3.708 1.098
SW-T3-S1-5 32.1 30282 0.00207 4.308 1.081
SW-T4-S1-6 34.8 31213 0.00211 4.830 1.070
SW-T5-S1-7 35.0 31280 0.00211 4.868 1.069
SW-T6-S1-8 22.6 26548 0.00189 2.470 1.178
SW-T1-S2-9 24.0 27153 0.00192 2.741 1.152
SW-T1-N5-S1-10 26.3 28101 0.00196 3.186 1.121
SW-T1-N10-S1-11 27.0 28380 0.00198 3.321 1.115
Concrete in Tension: The constitutive parameters to be calibrated for concrete in

tension consist of the initial Young’s modulus of the average stress-strain relationship
(Eg ), the tensile (cracking) strength of concrete (fcr ), and the strain at cracking (εcr ),
the parameter r representing the shape of the curve, and the normalized critical strain
at which the stress-strain curve becomes a straight line (xcrp ). Since material test
results were not available for these parameters, and the effects of tension stiffening
on concrete in tension needed to be incorporated, these parameters were calibrated so
that the tensile stress-strain relationship for concrete in tension is compatible with the
empirical relationships suggested by Belarbi and Hsu (1994) for concrete in tension.
The average stress-strain relationship proposed by Belarbi and Hsu (1994) for concrete
in tension (Figure 3.20) is defined by the following equations:
If εc > εcr then σc = Ec εc (3.10)
0.4
εcr

If εc > εcr then σc = fcr (3.11)
εc
68
where
q
Ec = 3875 fc0 (M P a) (3.12)
q
fcr = 0.31 fc0 (M P a) (3.13)
εcr = 0.00008 (3.14)
The stress-strain relationship of Belarbi and Hsu (1994) consists of two regions, the first
of which is the linear region, and the second of which is the exponentially-descending
region, as illustrated in Figure 3.20.
Figure 3.20. Stress-Strain Relationship for Concrete in Tension (Belarbi and Hsu,
1994).
The normalized critical strain value for concrete in tension, at which the stress-
strain curve becomes a straight line (xcrp ), was calibrated for a very large value (e.g.,
10000), so that the curve does not transform into a straight line and continues as a
69
curve, the shape of which resembles the descending region of the relationship by Belarbi
and Hsu (1994). The parameter rr (controlling the shape of the tension envelope) was
calibrated graphically by comparing the shape of the stress-strain diagram obtained
by the Chang and Mander (1994) model with the Belarbi and Hsu (1994) relationship,
so that similar shapes are obtained for the stress-strain curves, as displayed in Figure
3.21. The calibrated parameters of concrete in tension are presented in Table 3.15 and
Table 3.16 for all wall specimens.
Figure 3.21. Stress-Strain Diagrams for Concrete in Tension for SW-T2-S1-1.
Table 3.15. Calibrated Values of the Constitutive Parameters for Concrete in Tension
for the UCLA Specimens.
Specimen Ec (Mpa) fcr (Mpa) εcr rt Xcrp

WS-T1-S1 27778 1.5654 0.00008 1.2400 10000
WS-T1-S2 33996 2.0493 0.00008 1.2270 10000
WS-T2-S1 30033 1.7371 0.00008 1.2380 10000
WS-T2-S2 29889 1.7260 0.00008 1.2380 10000
WS-T3-S1 30140 1.7454 0.00008 1.2380 10000
WS-T3-S2 30805 1.7969 0.00008 1.2377 10000
WS-T4-S1 30211 1.7509 0.00008 1.2380 10000
WS-T4-S2 30598 1.7808 0.00008 1.2377 10000
WP-T5-N0-S1 29486 1.6951 0.00008 1.2380 10000
WP-T5-N0-S2 29889 1.7260 0.00008 1.2380 10000
WP-T5-N5-S1 30211 1.7509 0.00008 1.2380 10000
WP-T5-N5-S2 30247 1.7536 0.00008 1.2380 10000
WP-T5-N10-S1 28885 1.6491 0.00008 1.2390 10000
WP-T5-N10-S2 30033 1.7371 0.00008 1.2380 10000
70
Table 3.16. Calibrated Values of the Constitutive Parameters for Concrete in Tension
for the Boğaziçi University Specimens.
Specimen Ec (Mpa) fcr (Mpa) εcr rt Xcrp

SW-T2-S1-1 25022 1.362 0.00008 1.25 10000
SW-T1-S1-2 27026 1.509 0.00008 1.25 10000
SW-T2-S2-3 27900 1.575 0.00008 1.25 10000
SW-T2-S3-4 29150 1.669 0.00008 1.24 10000
SW-T3-S1-5 30282 1.756 0.00008 1.24 10000
SW-T4-S1-6 31213 1.829 0.00008 1.23 10000
SW-T5-S1-7 31280 1.834 0.00008 1.23 10000
SW-T6-S1-8 26548 1.474 0.00008 1.25 10000
SW-T1-S2-9 27153 1.519 0.00008 1.25 10000
SW-T1-N5-S1-10 28101 1.590 0.00008 1.24 10000
SW-T1-N10-S1-11 28380 1.611 0.00008 1.24 10000
71
4. COMPARISON OF ANALYTICAL AND

EXPERIMENTAL RESULTS
This chapter presents comparison of the experimental results obtained from the
aforementioned wall test programs with predictions of the finite element model cali-
brated as described in the previous chapter, in an effort to validate the applicability of
the model in representing the response of structural walls with aspect ratio less than
or equal to 1.0. Lateral load vs. top displacement response comparisons for a total of
15 wall specimens are presented for brevity, since some of the wall specimens tested
were identical to each other. Analysis results presented in the comparisons consider
the influence of the improved shear aggregate interlock and dowel action constitutive
models with the cyclic strength degradation parameters implemented in this study.
4.1. Wall Specimens Tested at University of California Los Angeles
The squat wall tests conducted at UCLA were simulated as per their boundary
and loading conditions, with a fixed end condition at the bottom and zero rotation at
the top (via the rigid body constraint enforced at the top), together with a double-
curvature moment diagram along wall height. The number of loading cycles to each
target drift level was the same in the analyses as applied during testing, since the num-
ber of cycles is the governing parameter associated with cyclic strength degradation in
the constitutive shear aggregate interlock and dowel action models implemented. The
test results were processed to remove the contribution of strain penetration effects on
top lateral displacements as described in Section 2.3.4, and the processed top displace-
ment histories were used in the analyses. As a common response characteristic of all
wall spandrel specimens tested, which had weakened plane joints at wall mid-height
or wall-pedestal interface, the lateral load - displacement response was moderately un-
symmetrical in positive and negative loading directions. As there were two specimens
with identical properties for each specimen type, the comparisons are presented for one
specimen of each type only, except for the Type 1 specimens for which the concrete
72
compressive strength differs significantly (25.5 MPa and 43.7 MPa for WS-T1-S1 and
WS-T1-S2, respectively). All wall spandrels had an aspect ratio of 1.00 while all wall
piers had an aspect ratio of 0.89, and shear span-to-depth ratios for all specimens was
equal to half of the aspect ratio, due to the double-curvature loading condition.
Figure 4.1 shows the lateral load vs. top displacement response of Type 1 wall
specimen WS-T1-S1. The model predicts the initial stiffness of the wall, as well as the
reduction in the stiffness (more than 50%) due to the diagonal cracking, reasonably
well. The model overestimates the wall lateral load capacity in the negative loading
direction, and underestimates ductility of the wall. The specimen experiences pro-
nounced pinching due to closure of diagonal cracks, which is also present, although less
pronounced, in the model results.
Figure 4.1. Lateral Load - Top Displacement Response of Specimen WS-T1-S1.
The lateral load vs. top displacement response of Type 1 wall specimen WS-
T1-S2 is presented in Figure 4.2. The concrete compressive strength of this specimen
was relatively higher compared to the other specimen of same type (WS-T1-S1) which
increased the lateral load capacity of the wall which reflected on both the test results
and the model predictions. The model slightly underestimates the initial and cracked
stiffness of the wall, but captures its ductility reasonably well. The degradation esti-
mated by the model tends to decrease the capacity more mildly, compared to the test
results. The lateral load capacity of the wall in the negative direction was predicted
73
accurately by the model; however, the model slightly underestimates the capacity in
the positive loading direction. Pinching characteristics of the response are captured by
the model fairly well.
The lateral load vs. top displacement response comparison for the Type 2 wall
specimen WS-T2-S2, shown in Figure 4.3, indicates similar initial stiffness predicted
by the model and measured for the wall specimen. For this wall, lateral load response
does not degrade immediately after the capacity is reached, which is also reflected in
the analytical results. The beginning of loss of lateral load capacity is significantly
affected by the difference in the amount of boundary reinforcement among Type 1
and Type 2 specimens, as the behavior is more brittle in Type 1 walls with larger
boundary reinforcement, whereas capacity is maintained for several cycles in the case
of Type 2 specimens, due to flexural yielding of the lesser amount of boundary rein-
forcement. The lateral load capacity of specimen WS-T2-S2 in the positive loading
direction is captured by the model, while the capacity in the negative direction is un-
derestimated. Ductility prediction of the model in the positive loading direction is also
reasonable. At 1.5% drift level, crushing of concrete caused drastic reduction in the
experimentally-measured lateral load capacity of the wall specimen. Degradation in
lateral load capacity estimated by the model is relatively more gradual.
Type 3 wall specimens have relatively lower longitudinal web reinforcement ratio
74
compared to the first two types. With the presence of a weakened plane joint at wall
midheight, where some of the longitudinal web bars are cut, the lateral load capacity
of these specimens decrease. Figure 4.4 shows the response comparison for specimen
WS-T3-S1. The initial and cracked stiffness of the wall is well-estimated by the model.
The model also predicts the lateral load capacity of the wall fairly accurately, especially
in the positive loading direction. The lateral load degradation behavior is predicted
reasonably by the model, although the sudden decrease in lateral load observed in
the test result at 1.5% drift in the negative direction causes the model to slightly
overestimate the lateral load during loading cycles to larger drift. Pinching behavior
of the specimen is also represented by the model fairly well, although the analytical
response is slightly less-pinched.
All pier walls (Type 5) specimens tested at University of California Los Angeles
had identical dimensions and reinforcement details. Concrete compressive strength of
all wall pier specimens were also very similar. As no weakened plane joint is present
in the walls, the control parameter investigated was the influence of axial load level on
the wall response.

75
As can be observed in Figure 4.5, for Type 5 (wall pier) specimen WP-T5-N0-S1
with no axial load, the initial and cracked stiffness is well-captured by the model, while
lateral load capacity is estimated fairly well in the positive loading direction and is
overestimated in the negative loading direction. Ductility prediction of the model is
also reasonable. The model estimates a more gradual degradation in the lateral load
after capacity is reached. In the test results, a more sudden decrease in lateral load
takes place due to the formation of a single large diagonal crack on the specimen,
which decreases accuracy of the model in predicting the lateral load at larger drift
levels. For this specimen, the model also fails to represent the pronounced pinching
behavior observed in the test result, because of the same reason.
As depicted in Figure 4.6 for Type 5 specimen WP-T5-N5-S1, subjected to an

axial load corresponding to 5% of its axial load capacity, the model exhibits good
accuracy in predicting the stiffness, lateral load capacity, and ductility characteristics
of the wall, in both positive and negative loading directions. The failure mode of this
specimen was formation of diagonal cracks followed by crushing of concrete at the wall
center, resulting in pronounced degradation in the lateral load. This degradation in
lateral load, although captured by the model, is more gradual in the analytical results.
76
Figure 4.5. Lateral Load - Top Displacement Response of Specimen WP-T5-N0-S1.
A similar correlation was obtained for Type 5 specimen WP-T5-N10-S2, subjected

to as axial load corresponding to 10% of its axial load capacity. The response was well-
predicted by the model in terms of stiffness, lateral load capacity, and ductility. The
failure mode of this specimen was also crushing of concrete at the wall center, resulting
in sudden degradation in lateral load, which was captured by the model, although the
model predicted relatively more gradual degradation.
77
It is a significant attribute that the overall response prediction of the model

for the axially-loaded wall specimens (WP-T5-N5-S1 and WP-T5-N10-S2) is accurate,
because the model is apparently successful in representing the influence the axial load
on the shear-controlled response characteristics of walls. Considering that design codes
and standards neglect the influence of axial load in calculation of the shear strength
and stiffness of RC walls, mainly due to lack of experimental data, availability of a
modeling approach considering interaction between axial load and shear capacity is
promising towards improvement of code provisions on wall shear strength.
4.2. Wall Specimens Tested at Boğaziçi University
The wall tests conducted at Boğaziçi University were also simulated as per their
loading conditions, top displacement history, and the number of loading cycles applied
at each target drift level. As companion specimens with identical properties exist
for most of the specimen types, the comparisons between analytical and experimental
results are presented for one specimen of each type only, unless there is a significant
difference in the concrete compressive strength of the identical specimens or the applied
axial load level.
Among Type 1 wall specimens which all have an aspect ratio of 0.5, SW-T1-
78
S1-2 and SW-T1-S2-9 have the same properties in terms of dimensions, reinforcement
ratios, and loading history; and there is no axial load applied on these specimens.
The measured and predicted lateral load vs. top displacement responses of specimen
SW-T1-S1-2 are compared in Figure 4.8. The failure mode of this specimen was diag-
onal tension, where crushing of concrete propagated along diagonal struts. The initial
stiffness is reasonably captured by the model, although the cracked stiffness is over-
estimated. The lateral load capacity of the specimen is slightly overestimated in the
negative loading direction, whereas it is well-predicted in the positive loading direction.
For this specimen, at 1.2% drift level, concrete started to crush at the center of the
wall, initiating strength degradation. The model overestimates the ductility for the
wall; however, degradation of the lateral load and pinching behavior are represented
fairly well.
The remaining Type 1 wall specimens were subjected to axial load during the
tests. Specimen SW-T1-N5-S1-10, the properties of which were identical to SW-T1-
S2-9, was tested under 5% axial load level. The model captured the increase observed
in the lateral load capacity of the specimen due to applied axial load, as depicted in
Figure 4.8. The model overestimated the stiffness of the wall and underestimated its
ductility; however, degradation in the lateral load and pinching characteristics were
reasonably predicted.
Figure 4.8. Lateral Load - Top Displacement Response of Specimen SW-T1-S1-2.

79
Figure 4.9. Lateral Load - Top Displacement Response of Specimen SW-T1-N5-S1-10.
Specimen SW-T1-N10-S1-11 was another Type 1 specimen, and was tested un-
der 10% axial load level. Similarly to the case of specimen SW-T1-N5-S1-10 the model
captures the lateral load capacity of the specimen (especially in the positive loading
direction), overestimates its stiffness, underestimates its ductility, and reasonably rep-
resents its lateral load degradation and pinching characteristics as seen in Figure 4.10.
This specimen experienced sudden strength degradation in the positive loading direc-
tion due to crushing of concrete, whereas degradation in the analysis results is more
gradual, as is typically the case for the model.
Figure 4.10. Lateral Load - Top Displacement Response of Specimen

SW-T1-N10-S1-11.
80
Type 2 wall specimens had identical properties, with specimen SW-T2-S1-1 hav-
ing lower concrete compressive strength compared to the remaining Type 2 specimens,
due to the age of concrete during testing. Figure 4.11 shows the lateral load vs. top
displacement response comparison for specimen SW-T2-S1-1. Although the initial stiff-
ness of the wall is well captured by the model, the cracked stiffness is overestimated
and its ductility underestimated. In both positive and negative loading directions,
the model successfully predicts the lateral load capacity of the wall. For all Type 2
specimens, the failure mode was diagonal compression failure, leading to crushing at
the base of the wall. Concrete crushing decreased the experimentally-measured lateral
load capacity significantly at 1.5% drift. Although the model captures the degradation
in lateral load capacity, the degradation in the model results initiates earlier and is
more gradual. Similarly, for specimen SW-T2-S3-4 (Figure 4.12), the model accurately
predicts the initial stiffness, and lateral load capacity of the wall, underestimates its
ductility, and predicts a more gradual degradation in lateral load.

81
The Type 4 wall specimen SW-T4-S1-6 is the one with the lowest aspect ratio of
0.33, among the structural walls investigated in this study (FFigure 4.13). The failure
mode of this specimen was similar to the Type 2 walls, as this specimen also failed
due to crushing of concrete under diagonal compression. The associated degradation in
lateral load capacity is reflected in the model results; however, the model predicts a less
ductile behavior, with earlier and more gradual degradation in lateral load. As well,
the lateral load capacity of the wall is notably overestimated by the model, especially
in the negative loading direction. Pinching characteristics of the response as well as
the stiffness of the wall are reasonably predicted by the model.

82
The Type 5 and Type 6 wall specimens were the most slender walls tested at
Boğaziçi University, with an aspect of 1.00. Another common feature of these specimens
was that they had high boundary reinforcement ratios. In the test results, after the
lateral load capacity of these walls was reached, the walls maintained their capacity for
several cycles. The response comparison for specimen SW-T5-S1-7 is shown in Figure
4.14. Lateral stiffness of the wall is captured by the model. Lateral load capacity in
both loading directions, ductility, and pinching behavior are also reasonably predicted.
However, the model tends to estimate more pronounced degradation in lateral load
with increasing drift levels.
Figure 4.14. Lateral Load - Top Displacement Response of SW-T6-S1-7.
Type 3 specimens had a differentiating failure mode compared to the rest of the
wall specimens tested at Boğaziçi University. Specimen SW-T3-S1-5 had low longitu-
dinal boundary reinforcing ratio and failed under shear sliding along a single horizontal
crack that spanned across the wall-pedestal interface. Boundary reinforcement of the
wall yielded at very early drift levels, the horizontal crack propagated across the entire
length of the wall-pedestal interface, and the response was governed by shear sliding
along the crack. To represent this phenomenon, formation of a horizontal crack at the
base of the wall was imposed on the model formulation. As observed in Figure 4.15, the
analytical model fails to accurately represent the overall lateral load vs. top displace-
ment response of the wall. Especially, unfavorable predictions were obtained for the
lateral load capacity of the wall and pinching characteristics of the response. However,
83
parameters of the constitutive aggregate interlock and dowel action models (incorpo-
rating a friction coefficient of 1.0 for the compressive stresses in concrete perpendicular
to the crack and a friction coefficient of 0.35 for clamping effect of reinforcing steel
bars) had provided reasonable response predictions for the remaining specimens tested
at Boğaziçi University.
A second analysis was conducted for this specimen, using a friction coefficient
of 0.35 for both the compressive stresses in concrete perpendicular to the crack and
for the clamping effect of reinforcing steel bars. The results of the new analysis are
shown in Figure 4.16. As can be observed, decreasing the concrete stress friction coef-
ficient to a value of 0.35 provided an improved prediction of the lateral load capacity
of the specimen. However, the model still failed to represent the cyclic stiffness degra-
dation (e.g., unloading stiffness) and pinching characteristics of the experimentally -
observed response. In order to improve these features in the model prediction, a new
hysteretic constitutive model formulation was developed and implemented, to represent
the clamping effect of reinforcement on the shear aggregate interlock mechanism. Hys-
teric rules following the so-called Peak-Oriented Model (Figure 4.17) were implemented
for the clamping effect on aggregate interlock (to replace the origin-oriented hysteretic
rules in the original model formulation), where the unloading from the envelope follows
the elastic stiffness of the shear (friction) stress vs. shear (sliding) strain envelope, zero
stress is maintained until sliding direction is reversed, and reloading to the envelope
in the opposite direction is peak-oriented. The cyclic strength degradation parameters
were retained in the new Peak-Oriented hysteretic model.
84
Figure 4.15. Lateral Load - Top Displacement Response of SW-T3-S1-5 (Original

Model).
Figure 4.16. Lateral Load - Top Displacement Response of SW-T3-S1-5 (Concrete

Friction Coefficient in Shear Aggregate Interlock Mechanism Reduced in the Model).
Using the new Peak-Oriented constitutive model for the clamping effect of rein-
forcement on the shear aggregate interlock mechanism across cracks, a third analysis
was conducted for specimen SW-T3-S1-5. The lateral load vs. top displacement re-
sponse comparison obtained using the updated model formulation is presented in Figure
4.18. As can be observed in the figure, the model now provides significantly improved
predictions for the cyclic stiffness degradation and pinching characteristics of the re-
sponse. The updated model reasonably predicts the lateral load capacity and ductility
85
of the wall, as well as hysteretic characteristics of the load-displacement response, in-

cluding cyclic stiffness and strength degradation and pinching.
Figure 4.17. Peak-Oriented Model for Contribution of Reinforcing Steel on Shear

Aggregate Interlock.
Figure 4.18. Lateral Load - Top Displacement Response of SW-T3-S1-5

(Peak-Oriented).
86
4.3. Overview
The aim of this chapter was to present the results of correlation studies conducted
between the analytically-predicted and experimentally-observed lateral load vs. top
displacement response of the wall specimens investigated in this study. The analytical
model was subjected to the same loading conditions applied during testing including
loading protocol and axial load conditions. For the walls experiencing shear sliding
failure at the base, a horizontal crack was enforced on the model at the same location.
Response comparisons with test results on the low-rise wall specimens revealed
that the model provides reasonable predictions of the lateral stiffness, lateral load ca-
pacity, degradation in stiffness and lateral load, hysteretic response shape, and pinching
characteristics of the wall specimens investigated. The lateral load capacity of the walls
was estimated by the analytical model fairly well, except for specimen SW-T4-S1-6 with
an aspect (or shear-span-to-depth) ratio of 0.33, which is the lowest among all speci-
mens. The model tends to underestimate the ductility of the wall specimens tested at
Boğaziçi University, whereas it provides reasonably predictions of the ductility of the
UCLA specimens. Cyclic strength and stiffness degradation, and pinching behavior in
the load-displacement response of the walls were also captured reasonably well by the
model.
87
5. SUMMARY AND CONCLUSIONS
5.1. Overview
The aim of this study was to validate and improve a finite element modeling
methodology to simulate the hysteretic lateral load vs. displacement response of low-
rise structural walls, whose behavior is governed by shear deformations. The constitu-
tive modeling approach used is based on a fixed-strut-angle panel model formulation,
and improvements were made on the constitutive panel model for better representa-
tion of the shear-aggregate-interlock effects in concrete and dowel action on reinforcing
bars, which constitute the the shear stress transfer mechanism across cracks and thus
significantly influence the lateral load behavior of low-rise walls with aspect ratios of
1.0 or less.
The model was calibrated for squat wall specimens tested by Massone (2006),
Orakcal et al., (2009), and Terzioglu (2011), and model predictions were compared
with the experimentally-observed cyclic lateral load vs. displacement responses of rep-
resentative specimens. Cyclic degradation parameters were implemented in the consti-
tutive model formulations representing shear aggregate interlock effects along cracks
and dowel action on reinforcement. In order to obtain better response prediction for
walls experiencing shear sliding failure, an alternative peak-oriented hysteretic model
was developed and implemented for representing the contribution of the clamping ef-
fect of reinforcement on the shear aggregate interlock mechanism, and formation of a
horizontal crack at the base of the wall was imposed on the analytical model.
5.2. Conclusions
The following conclusions can be drawn based on the results obtained using the
analytical model adopted in this study, upon comparison with the experimentally-
observed response of low-rise structural walls:
88
• The adopted modeling approach is capable of simulating the nonlinear hysteretic

response of low-rise structural walls under in-plane reversed-cyclic loading condi-
tions, with an acceptable level of accuracy. Reasonable predictions were obtained
for the experimentally-observed cyclic response characteristics of the walls inves-
tigated, including their lateral load capacity, stiffness, ductility, cyclic stiffness
and strength degradation, hysteretic shape, and pinching behavior.
• The model tends to overestimate the lateral stiffness and underestimate the duc-
tility of some of the wall specimens investigated. As well, in most cases, the
degradation in lateral load predicted by the model is more gradual compared to
experimental results.
• The model successfully represents the influence of axial load on the shear-controlled
lateral load behavior of walls, which may be important considering that design
codes and guidelines neglect the influence of axial load in calculation of the shear
strength and stiffness of walls.
• It was observed that the cyclic strength degradation parameters implemented in
this study on the constitutive models associated with shear aggregate interlock
and dowel action mechanisms improves the accuracy of the model in predicting
the loss of lateral load capacity and lateral stiffness at high drift levels. As well,
consideration of strain penetration effects on the response provides improved
model predictions for the initial stiffness of walls.
For walls experiencing shear sliding type of failure, an alternative hysteretic

(peak-oriented) modeling approach was developed to represent the clamping contri-
bution of reinforcing steel bars on the shear aggregate interlock mechanism along the
cracks. Implementation of this new hysteretic model resulted in improved prediction
of the lateral load response of walls exhibiting shear sliding failure.
5.3. Recommendations for Future Studies
The following recommendations can be made for future studies on the subject:
• A wider range of experimental data can be used to further validate the model and
89
improve its formulation (e.g., implementation of more refined constitutive models

for aggregate interlock and dowel mechanisms, incorporation of reinforcing bar
buckling and fracture under low-cycle fatigue, incorporation of bond slip), towards
developing a more robust model formulation to simulate the response of walls with
various response characteristics.
• Size effects (strain localization effects) can be considered in improving the for-
mulation of the model, and the accuracy of model predictions for local responses
(e.g., strains in concrete and steel) can be investigated.
• The model, either in its current state, or with a modified (possibly simplified)
formulation, can be extended to conduct analytical studies on structural systems.
90
REFERENCES
Barda, F., 1972, “Shear Strength of Low-Rise Walls with Boundary Elements”, Ph.D.
Dissertation, Lehigh University, Bethlehem, Pennsylvania.
Bazant, Z. P. and P. G. Gambarova, 1984, “Crack Shear in Concrete: Crack Band

Microplane Model”, Jounal of Structural Engineering, Vol. 110, No. 9, pp. 2015-
2035.
Belarbi, A. and T. C. Hsu, 1994, “Constitutive Laws of Concrete in Tension and

Reinforcing Bars Stiffened By Concrete”, American Concrete Institute Structural
Journal, Vol. 91, No. 4, pp. 465-474.
Benjamin, J.R., and H.A. Williams, 1953, “Investigation of Shear Walls, Part 3 - Ex-
perimental and Mathematical Studies of Reinforced Concrete Walled Bents under
Static Shear Loading”, Report No. 1, Department of Civil Engineering, Stanford
University, Stanford, California.
Chang, G.A. and J.B. Mander, 1994, “Seismic Energy Based Fatigue Damage Analysis
of Bridge Columns: Part I-Evaluation of Seismic Capacity”, National Center for
Earthquake Engineering Research.
Clarke, M. J. and G. J. Hancock, 1990, “A Study of Incremental-Iterative Strategies

for Non-Linear Analyses”, International Journal for Numerical Methods in Engi-
neering, Vol. 29, pp. 1365-1391.
Darwin, D., 1993, “Finite Element Analysis of Reinforced Concrete Structures II”,
American Society of Civil Engineers Journal of Structural Engineering, Vol. 1,
pp. 203-232.
Filippou, F.C., E.G., Popov and V.V. Bertero, 1983, “Effects of Bond Deterioration on
Hysteretic Behavior of Reinforced Concrete Joints”, Energy and Environmental
91
Rerearch Center Report No. 83/19 Earthquake, Engineering Research Center,

University of California, Berkeley, California.
Galletly, G.D., 1952, “Behavior of Reinforced Concrete Walls under Static Load”,
Department of Civil and Sanitary Engineering, Massachusetts Institute of Tech-
nology.
Gulec, C.K., 2009, Performance-Based Assessment and Design of Squat Reinforced

Concrete Shear Walls, Ph.D. Thesis, State University of New York.
Gullu, M.F., 2013, Finite Element Modeling Of Reinforced Concrete Structural Walls,
M.S. Thesis, Boğaziçi University.
Gullu, M.F., K. Orakcal, 2014, “Nonlinear Finite Element Modeling of Reinforced

Concrete Structural Walls”, Proceedings, 2nd European Conference on Earthquake
Engineering and Seismology, Istanbul, Turkey, No. 1485.
He, X.G. and A.K.H. Kwan, 2001, “Modeling Dowel Action of Reinforcement Bars for
Finite Element Analysis of Structures”, Computers and Structures, Vol.79, No. 6,
pp. 595-604.
Hirosawa, M., 1975, “Past Experimental Results on Reinforced Concrete Shear Walls
and Analysis on Them”, Kenchiku Kenkyu Shiryo, No. 6, pp. 277-280.
Horoz, B., 2015, Modeling of Coupled Nonlinear Shear and Flexural Responses in
Medium-Rise RC Walls, M.S. Thesis, Boğaziçi University.
Husain, H.M., N.K. Oukaili and H.S. Muhammed, 2009, “Dowel Action Between Two
Concretes”, Journal of Engineering, Vol. 15, No. 2, pp. 3583-3605.
Ince, R., E. Yalcin and A. Arslan, 2006, “Size-Dependent Response of Dowel Action in
RC Members”, Engineering Structures, Vol. 29, pp. 955-961.
Kolozvari, K., T. Tran, J.W. Wallace and K. Orakcal, 2012, “Modeling of Cyclic Shear-
92
Flexure Interaction in Reinforced Concrete Structural Walls”, Proceedings, 15th

World Conference on Earthquake Engineering, Lisbon, Portugal, No. 2471.
Kolozvari, K., T. Tran, J.W. Wallace and K. Orakcal, 2014, “Modeling of Cyclic Shear-
Flexure Interaction in Reinforced Concrete Structural Walls. I: Theory”, American
Society of Civil Engineers Journal of Structural Engineering, Vol. 141, No. 5, pp.
541-550.
Maekawa, K. and J. Qureshi, 1996b, “Computational Model for Reinforcing Bar Em-
bedded in Concrete under Combined Axial Pullout and Transverse Displacement”,
Proceedings of Japan Society of Civil Engineers, Vol. 31, No. 538, pp. 227-239.
Maekawa, K. and J. Qureshi, 1997, “Stress Transfer across Interfaces in Reinforced

Concrete Due to Aggregate Interlock and Dowel Action”, Journal of Materials,
Concrete Structures, Pavements, Vol. 34, No. 557, pp. 159-172.
Mander, J.B., M.J.N. Priestley and R. Park, 1988, “Theoretical Stress-Strain Model
for Confined Concrete”, American Society of Civil Engineers Journal of Structural
Engineering, Vol. 114, No. 8, pp. 1804-1826.
Mansour, M.Y., T.C. Hsu and J.Y. Lee, 2002, “Pinching Effect in Hysteretic Loops
of R/C Shear Elements”, American Concrete Institute Special Publications, Vol.
205, pp. 293-321.
Mansour, M.Y. and T.C. Hsu, 2005, “Behavior of Reinforced Concrete Elements under
Cyclic Shear”, American Society of Civil Engineers Journal of Structural Engi-
neering, Vol. 131, No. 1, pp. 44-53.
Martin-Perez, B. and S.J. Pantazopoulou, 1999, “Effect of Bond, Aggregate Interlock

and Dowel Action on the Shear Strength Degradation of Reinforced Concrete”,
Engineering Structures, Vol. 23, pp. 214-227.
Massone, L.M., 2006, Analytical and Experimental with Shear - Flexure Interaction in
93
RC Walls, Ph.D. Thesis, University of California.
Mattock, H.A. and M.N. Hawkins, 1972, “Shear Transfer in Reinforced Concrete -
Recent Research”, Precast/ Prestressed Concrete Institue, Vol. 1, pp. 55-75.
Mau, S.T. and T.C. Hsu, 1986, “Shear Design and Analysis of Low-Rise Structural
Walls”, American Concrete Institute Structural Journal, Vol. 83, No. 2, pp. 306-
315.
Menegotto, M. and E. Pinto, 1973, “Method of Analysis for Cyclically Loaded Rein-
forced Concrete Plane Frames Including Changes in Geometry and Non-Elastic
Behavior of Elements under Combined Normal Force and Bending”, Proceedings,
International Association for Bridge and Structural Engineering Symposium on
Resistance and Ultimate Deformability of Structures Acted on by Well-Defined
Repeated Loads, Lisbon, Portugal.
Moradi, A.R., M. Soltani and A.A. Tasnimi, 2012, “A Simplified Constitutive Model
for Dowel Action across RC Cracks”, Journal of Advanced Concrete Technology,
Vol. 10, pp. 264-277.
Oesterle, R.G., Fiorato, A.E., Johal, L.S., Carpenter, J.E., Russell, H.G., and Corley,
W.G. 1976, “Earthquake resistant structural walls-Tests of isolated walls”, Re-
port to National Science Foundation, Principal Component Analysis Construction
Technology Laboratories, Skokie, IL, 317.
Orakcal, K., D. Ulugtekin and L.M. Massone, 2012, “Constitutive Modeling of Rein-
forced Concrete Panel Behavior under Cyclic Loading”, Proceedings, 15th World
Conference on Earthquake Engineering, Lisbon, Portugal, No. 3573.
Orakcal, K. and J.W. Wallace, 2006, “Flexural Modeling of Reinforced Concrete Walls
- Model Calibration”, American Concrete Institute Structural Journal, Vol. 103,
No. 2, pp. 196-206.
94
Orakcal, K., L.M. Massone, and J.W. Wallace, 2006, Analytical Modeling of Reinforced
Concrete Walls for Predicting Flexural and Coupled - Shear - Flexural Responses,
Pacific Earthquake Engineering Research Report No. EEC-9701568, Berkeley.
Orakcal, K., L.M. Massone, and J.W. Wallace, 2009, “Modeling of Squat Structural
Walls Controlled by Shear”, American Concrete Institute Structural Journal, Vol.
106, No. 5, pp. 646-655.
Orendil, A.E., 2014, Analytical Modeling of Squat Walls with Shear-Controlled Re-
sponses, M.S. Thesis, Boğaziçi University.
Palieraki, V. and E.N. Vintzileou, 2009, “Cyclic Behavior of Interfaces in Repaired/

Strengthened RC Elements”, Architecture Civil Engineering Environment, No. 1,
pp. 97-108.
Panagouli, O. and K. Iordanidou, 2013, “Study of the Residual Strength of an RC Shear

Wall with Fractal Crack Taking into Account Interlocking Interface Phenomena”,
Mathematical Problems in Engineering, Vol. 2013, Article ID. 101484.
Pang, D. and T.C. Hsu, 1995, “Behavior of Reinforced Concrete Membrane Elements
in Shear”, American Concrete Institute Structural Journal, Vol. 92, No. 6, pp.
665-677.
Paulay, T., M.J.N. Priestley and A.J. Synge, 1982, “Ductility in Earthquake Resisting
Squat Shearwalls”, American Concrete Institute Journal, Vol. 79, No. 4, pp.
257-269.
Petrangeli, M., P.E. Pinto and V. Ciampi, 1999, “Fiber Element for Cyclic Bending and
Shear of RC Structures. I: Theory”, Journal of Engineering Mechanics, American
Society of Civil Engineers, Vol. 125, No. 9, pp. 994-1001.
Powell, G. and J. Simons, 1982, “Improved Iteration Strategy for Nonlinear Struc-
tures”, International Journal for Numerical Methods in Engineering, Vol. 17, pp.
95
1455-1467.
Saatcioglu, M. and S. Razvi, 1992, “Strength and Ductility of Confined Concrete”,

American Society of Civil Engineers Journal of Structural Engineering, Vol. 118,
No. 6, pp. 1590-1607.
Sagaseta, J. and R. L. Vollum, 2011, “Influence of Aggregate Fracture on Shear Transfer

through Cracks in Reinforced Concrete”, Magazine of Concrete Research, Vol. 63,
No. 2, pp. 119-137.
Soltani, M. and K. Maekawa, 2008, “Path-Dependent Mechanical Model for Deformed

Reinforcing Bars at RC Interface under Coupled Cyclic Shear and Pullout Ten-
sion”, Engineering Structures, Vol. 30, No. 4, pp. 1079-1091.
Soroushian, P., K. Obaseki, M. I. Baiyasi, B. El-Sweidan and K. Choi, 1988, “Inelastic

Cyclic Behavior of Dowel Bars”, American Concrete Institute Structural Journal,
Vol. 85, No. 1, pp. 23-29.
Stevens, N.J., 1987, Analytical Modeling of Reinforced Concrete Subjected to Monotonic

and Reversed Loadings, Ph.D. Thesis, University of Toronto.
Terzioglu, T., 2011, Experimental Evaluation of The Lateral Load Behavior of Squat
Structural Walls, M.S. Thesis, Boğaziçi University.
Thermou G.E., V.K. Papanikolaou, A.J. Kappos, 2011, “Analytical Model for Predict-
ing the Response of Old-Type Columns Rehabilitated with Concrete Jacketing
under Reversed Cyclic Loading”, Conference on Computational Methods in Struc-
tural Dynamics & Earthquake Engineering, Greece.
Thermou G.E., V.K. Papanikolaou, A.J. Kappos, 2014, “Flexural Behaviour of Rein-
forced Concrete Jacketed Columns under Reversed Cyclic Loading”, Engineering
Structures, Vol. 76, pp. 270-282.
Theodossius, P.T. and E. N. Vintzileou, 1987, “Concrete-to-Concrete Friction”, Journal

96
of Structural Engineering, Vol. 113, No. 4, pp. 832-849.
Tran, T.A., 2012, Experimental and Analytical Studies of Moderate Aspect Ratio Re-
inforced Concrete Structural Walls, Ph.D. Thesis, University of California, Los
Angeles.
Tran, T.A. and J.W. Wallace, 2012, “Experimental Study of Nonlinear Flexural and
Shear Deformations of Reinforced Concrete Structural Walls”, Proceedings, 15th
World Conference on Earthquake Engineering, Lisbon, Portugal, No. 3913.
Tsai, W.T., 1988, “Uniaxial Compressional Stress-Strain Relation of Concrete”, Amer-

ican Society of Civil Engineers Journal of Structural Engineering, Vol. 114, No.
9, pp. 2133-2136.
Ulugtekin, D., 2010, Analytical Modeling of Reinforced Concrete Panel Elements under
Reversed Cyclic Loadings, M.S. Thesis, Boğaziçi University.
Vecchio, F.J. and M.P. Collins, 1986, “The Modified Compression-Field Theory for
Reinforced Concrete Elements Subjected to Shear”, American Concrete Institute
Structural Journal, Vol. 83, No. 2, pp. 219-231.
Vecchio, F.J. and M.P. Collins, 1993, “Compression Response of Cracked Reinforced
Concrete”, American Society of Civil Engineers Journal of Structural Engineering,
Vol. 83, No. 2, pp. 219-231.
Vintzeleou, E.N. and T.P. Tassios, 1987, “Behavior of Dowels under Cyclic Deforma-
tions”, American Concrete Institute Structural Journal, Vol. 84, No. 1, pp. 18-30.
Vintzeleou, E.N. and T.P. Tassios, 1990, “Concrete-to-Concrete Friction”, American

Society of Civil Engineering Journal of Structural Engineering, Vol. 113, No. 4.
Vulcano, A., V.V. Bertero and V. Colotti, 1988, “Analytical Modeling of RC Structural
Walls”, Proceedings, 9th World Conference on Earthquake Engineering, Vol. 6,
Tokyo-Kyoto, Japan, pp. 41-46.
97
Walraven, J.C. and H.W. Reinhardt, 1981, “Theory and Experiments on the Me-
chanical Behavior of Cracks in Plain and Reinforced Concrete Subjected to Shear
Loading”, International Association for Bridge and Structural Engineering, Vol.
26, No. 1A, pp. 1-68.
Wood, S.L., 1990, “Shear Strength of Low-Rise Reinforced Concrete Walls”, American
Concrete Institute Structural Journal, Vol. 87, No. 1, pp. 99-107.
Zhao, J. and S. Sritharan, 2007, “Modeling of Strain Penetration Effects in Fiber-

Based Analysis of Reinforced Concrete Structures”, American Concrete Institute
Structural Journal, Vol. 104, No. 2, pp. 133-141.

Tsai

Uploaded by

Copyright:

Available Formats

Tsai

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tsai

Uploaded by

Copyright:

Available Formats

NONLINEAR RESPONSE MODELING OF LOW-RISE STRUCTURAL WALLS

Submitted to the Institute for Graduate Studies in

Graduate Program in Civil Engineering

NONLINEAR RESPONSE MODELING OF LOW-RISE STRUCTURAL WALLS

Assoc. Prof. Kutay Orakçal ...................

Prof. Hilmi Lus ...................

Prof. Alper Ilki ...................

DATE OF APPROVAL: 11.01.2016

I would like to express my special thanks to my colleagues M. Fethi Güllü and

I would also like to express my thanks to my friends Selahattin Akalp, Sarper

NONLINEAR RESPONSE MODELING OF LOW-RISE

BODUR PERDE DUVARLARIN DOĞRUSAL OLMAYAN

Bu çalışma kapsamında boy/en oranı 1.0 ve daha az olan perde duvarların

2.4.5. Nonlinear Analysis Solution Strategy . . . . . . . . . . . . . . . 36

Figure 1.1. Coupled Model Element (Massone et al., 2006, 2009). . . . . . . . 5

Figure 1.2. Element Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Figure 1.3. Transfer Mechanism in Cracked RC Members (Ince et al., 2006). . 8

Figure 2.6. Compression Softening Effects Implemented in the Panel Model

Figure 2.9. Cracked Concrete Elements Containing a Reinforcement Bar Run-

Figure 2.12. Origin-Oriented Model for Dowel Action. . . . . . . . . . . . . . . 29

Figure 2.13. Strain Penetration Model (Massone et al., 2009). . . . . . . . . . . 30

Figure 2.14. Effect of Strain Penetration. . . . . . . . . . . . . . . . . . . . . . 31

Figure 2.15. Tangent Stiffness (Elasticity) Matrix Calculation Path. . . . . . . 33

Figure 2.16. Global and Local DOFs for a Wall Model. . . . . . . . . . . . . . 34

Figure 2.17. Stress Vector Calculation Path. . . . . . . . . . . . . . . . . . . . 34

Figure 2.18. Numeration of Internal Forces for a Mesh. . . . . . . . . . . . . . 35

Figure 2.20. Representation of the Nonlinear Analysis Solution Strategy for a

Figure 3.1. Specimen Name Description for Tests Conducted at UCLA. . . . . 40

Figure 3.6. Geometry and Wall Reinforcement of Specimen T1-S1 (Terzioglu,

Figure 3.7. Geometry and Wall Reinforcement of Specimen T3-S1 (Terzioglu,

Figure 3.8. Geometry and Wall Reinforcement of Specimen T5-S1 (Terzioglu,

Figure 3.9. Test Setup (Massone, 2006). . . . . . . . . . . . . . . . . . . . . . 49

Figure 3.10. Displacement History (Massone, 2006). . . . . . . . . . . . . . . . 49

Figure 3.11. Test Setup (Terzioglu, 2011). . . . . . . . . . . . . . . . . . . . . . 50

Figure 3.15. Steel Stress-Strain Curves for Boundary Reinforcement of Specimen

Figure 3.17. Stress-Strain Relationship for a Bar Embedded in Concrete (Be-

Figure 3.19. Stress-Strain Diagrams for Concrete in Compression for WS-T3-S2. 66

Figure 3.20. Stress-Strain Relationship for Concrete in Tension (Belarbi and

Figure 3.21. Stress-Strain Diagrams for Concrete in Tension for SW-T2-S1-1. . 69

Figure 4.1. Lateral Load - Top Displacement Response of Specimen WS-T1-S1. 72

Figure 4.2. Lateral Load - Top Displacement Response of Specimen WS-T1-S2. 73

Figure 4.3. Lateral Load - Top Displacement Response of Specimen WS-T2-S2. 74

Figure 4.4. Lateral Load - Top Displacement Response of Specimen WS-T3-S1. 75

Figure 4.5. Lateral Load - Top Displacement Response of Specimen WP-T5-

Figure 4.6. Lateral Load - Top Displacement Response of Specimen WP-T5-

Figure 4.7. Lateral Load - Top Displacement Response of Specimen WP-T5-

Figure 4.8. Lateral Load - Top Displacement Response of Specimen SW-T1-S1-2. 78

Figure 4.9. Lateral Load - Top Displacement Response of Specimen SW-T1-

Figure 4.10. Lateral Load - Top Displacement Response of Specimen SW-T1-

Figure 4.11. Lateral Load - Top Displacement Response of Specimen SW-T2-S1-1. 80

Figure 4.12. Lateral Load - Top Displacement Response of Specimen SW-T2-S3-4. 81

Figure 4.13. Lateral Load - Top Displacement Response of Specimen SW-T4-S1-6. 81

Figure 4.14. Lateral Load - Top Displacement Response of SW-T6-S1-7. . . . . 82

Figure 4.15. Lateral Load - Top Displacement Response of SW-T3-S1-5 (Origi-

Figure 4.16. Lateral Load - Top Displacement Response of SW-T3-S1-5 (Con-

Figure 4.17. Peak-Oriented Model for Contribution of Reinforcing Steel on Shear

Figure 4.18. Lateral Load - Top Displacement Response of SW-T3-S1-5 (Peak-

Table 3.1. Test Specimen Details for Tests Conducted at UCLA. . . . . . . . 41

Table 3.3. Average Compressive Strength of Concrete (Massone, 2006). . . . . 52