EVALUATION OF SHEAR-FLEXURAL INTERACTION MODEL FOR

REINFORCED CONCRETE STRUCTURAL WALLS

A Thesis

Presented to the

Faculty of

California State University, Fullerton

In Partial Fulfillment

of the Requirements for the Degree

Master of Science

in

Civil Engineering

By

Haider Kadhem Sakban

Thesis Committee Approval:

Kristijan Kolozvari, Committee Chair, Department of Civil Engineering

David Naish, Member, Department of Civil Engineering
Joel Lanning, Member, Department of Civil Engineering

Fall, 2016
 ABSTRACT

Reinforced concrete (RC) walls are the most commonly used structural elements

in buildings for resisting lateral forces induced by earthquakes. Therefore, reliable and

robust analytical models that can predict their behavior under earthquake actions are

essential for the design of new buildings and evaluation of existing buildings.

Performance-Base Seismic Design (PBSD) and evaluation currently of RC structural

walls typically relies on analytical models that do not capture the interaction between

shear and flexural responses. Previous experimental and analytical studies showed that

this interaction could be important for walls with shear span ratio ranges between 1.5 and

3.0 and that models which do not capture this interaction can overestimate wall strength

and stiffness by 30 to 50%. A novel modeling approach that captures shear-flexure

interaction (SFI) has been recently developed and implemented in computational

platform OpenSees. The model has been previously validated against five RC wall

specimens with aspect ratios of 1.5 and 2.0 that experienced significant SFI, and it

showed to be capable of reproducing successfully experimentally measured global and

local wall responses. However, the model has not been validated extensively over a wide

range of wall characteristics, aspect ratio, axial load, reinforcement configuration and

shape of the cross-section.

The objective of this study is to calibrate and validate the SFI modeling approach

recently implemented in computational platform OpenSees against a great number of

ii
specimens to assess its capability of successfully reproducing experimentally measured

wall responses over a range of wall characteristics. The model was validated against 12

RC wall specimens tested under cyclic loading conditions. Comparison of experimental

and analytical wall behavior focused on the overall load versus total top wall

displacement and load versus shear displacement (global response) over the plastic hinge

region (local response). Based on the results presented, it can be concluded that the model

is capable of reproducing successfully experimentally measured wall behavior over a

range of considered wall characteristics, including lateral load, stiffness, stiffness

degradation and shear displacement within plastic hinge. It has also been observed that

the model tends to overestimate the area of the hysteretic loops (i.e., underestimate

pinching) for specimens with zero axial load, due to currently implemented models that

represent shear transfer mechanisms along concrete cracks. Based on the results

presented, future model improvements are suggested.

iii
 TABLE OF CONTENTS

ABSTRACT ................................................................................................................... ii

LIST OF TABLES ......................................................................................................... vi

LIST OF FIGURES ....................................................................................................... vii

ACKNOWLEDGMENTS ............................................................................................. xii

Chapter
1. INTRODUCTION ................................................................................................ 1

1.1. Background .................................................................................................... 1

1.2. Scope and Objectives ..................................................................................... 3
1.3. Organization................................................................................................... 4

2. LITERATURE REVIEW ..................................................................................... 5

2.1. General Information ....................................................................................... 5

2.2. Models without Interaction of Axial, Shear, and Flexure Responses ............ 6
2.2.1. Three-Vertical-Line-Element Model and Its Modifications ................ 6
2.2.2. Multiple-Vertical-Line-Element Model ............................................... 12
2.2.3. PERFORM 3D - Shear Wall Element ................................................. 19
2.3. Models with Interaction of Axial, Shear and Flexure Responses .................. 20

3. DESCRIPTION OF SHEAR-FLEXURAL INTERACTION MODEL ............... 27

3.1. Shear-Flexural Interaction Model Assumptions ............................................ 27

3.2. Reinforced Concrete Panel and Fixed-Strut-Angle-Model ........................... 30
3.2.1. Concrete Behavior in the RC Panel Element ....................................... 31
3.2.2. Steel Reinforcement Behavior in the RC Panel Element .................... 34
3.3. Shear Resistance Models ............................................................................... 35
3.4. Uniaxial Concrete and Steel Constitutive Behavior ...................................... 37
3.5. Implementation in OpenSees ......................................................................... 42

iv
4. OVERVIEW OF EXPERIMENTAL RESULTS ................................................. 44

4.1. Oesterle et al. (1976, 1979) ............................................................................ 44

4.2. Thomsen and Wallace (1995) ........................................................................ 52
4.3. Dazio et al. (1999) ......................................................................................... 56

5. OPENSEES MODELING OF RC WALLS ......................................................... 59

5.1. Specimen Description .................................................................................... 59

5.2. Model Geometry Discretization..................................................................... 61
5.3. Calibration of Material Models...................................................................... 63
5.3.1. Steel Reinforcement ............................................................................. 63
5.3.2. Concrete ............................................................................................... 65
5.4. Shear Resistance Mechanism Parameters ...................................................... 69
5.5. Model generation ........................................................................................... 71

6. ANALYTICAL RESULT AND COMPARISON WITH EXPERIMENTAL

RESULTS ............................................................................................................. 74

6.1. Oesterle et al. (1976, 1979) ............................................................................ 74

6.1.1. Rectangular Cross-Section specimens ................................................. 75
6.1.2. Barbell Cross-Section Specimens ........................................................ 79
6.1.3. Flanged Cross-Section Specimens ....................................................... 91
6.2. Thomsen and Wallace (1995) ........................................................................ 95
6.3. Alessandro Dazio et al. (1999) ...................................................................... 97

7. SUMMARY AND CONCLUSIONS ................................................................... 100

7.1. Experimental Results and Calibration of the Materials ................................. 100

7.2. General Model Assessment and Model Capabilities ..................................... 102
7.3. Conclusion and Future Improvement ............................................................ 102

APPENDICES ............................................................................................................... 104

A. OPENSEES CODE FOR SPECIMEN B8..................................................... 104

B. CALIBRATION OF CONCRETE AND STEEL MATERIALS.................. 115

REFERENCES .............................................................................................................. 123

v
 LIST OF TABLES

Table Page

4.1. Geometry and Material Properties of Specimens Tested by Oesterle et al.

(1976, 1979) .......................................................................................................... 46

4.2. Steel Reinforcement Details in Specimens Tested by Oesterle et al. (1976,

1979) ..................................................................................................................... 47

4.3. Steel Yield Strength and Reinforcement Ratios for Specimens Tested by
Oesterle et al. (1976, 1979) ................................................................................... 50

5.1. Material Properties of B8 Specimen; Oesterle et al (1979) .................................. 60

5.2. Calculated Strain Hardening Ratios for Bars Used in Specimens B8 .................. 64

5.3. Input Parameters of the Saatcioglu and Razvi (1992) Model for B8 Specimen ... 67

6.1. Displacement, Strength and Stiffness Comparison Between Observed and

Simulated Specimen Results ................................................................................. 98

vi
 LIST OF FIGURES

Figure Page

2.1. Three-Vertical-Line-Element Model (TVLEM), Kabeyasawa et al. (1983)....... 7

2.2. Wall rocking and effect of neutral axis shift on vertical displacements ............. 7

2.3. Axial-Stiffness-Hysteresis Model (ASHM) (Kabeyasawa et al., 1983) ............. 8

2.4. Origin-Oriented-Hysteresis Model (OOHM) (Kabeyasawa et al., 1983) ........... 9

2.5. Axial-Element-in-Series Model (AESM) (Vulcano and Bertero, 1986) ............. 9

2.6. Axial force-deformation relationship of the AESM (Vulcano and Bertero,

1986) ................................................................................................................. 10

2.7. Modification of the TVLEM (Kabeyasawa et al., 1997) ................................ 11

2.8. Multiple-Vertical-Line-Element-Model (Vulcano et al, 1988) ....................... 13

2.9. Flexural deformations of the MVLEM element (Orakcal et al., 2004) ........... 13

2.10. Constitutive model for steel (Menegotto and Pinto, 1973) ............................. 14

2.11. Degradation of cyclic curvature ...................................................................... 15

2.12. Constitutive law adopted in the original MVLEM for uncracked concrete
(Colotti and Vulcano, 1987) ............................................................................ 15

2.13. Constitutive law adopted in the original MVLEM for cracked concrete
(Bolong et al., 1980) ........................................................................................ 16

2.14. Force-Deformation relationships adopted in the modified MVLEM

(Fischinger et al., 1990) -Vertical Springs ...................................................... 17

2.15. Force-Deformation relationships adopted in the modified MVLEM

(Fischinger et al., 1990) -Horizontal Spring.................................................... 18

vii
2.16. Uniaxial constitutive model for concrete (Chang and Mander1994;
Orakcal et al., 2004) ........................................................................................ 19

2.17. Fiber model used by Perform 3D .................................................................... 20

2.18. RC wall and Wall Panel Truss Model (Panagiotou et al., 2012)..................... 22

2.19. Coupled Compression-Tension Model for diagonal truss elements: (a)

coupled compression-tension truss elements, (b) concrete compressive
stress reduction due to normal tensile strain (Panagiotou et al., 2012) ........... 23

2.20. Coupled model element (Massone et al., 2006, 2009) .................................... 24

2.21. Constitutive material models. (a) Concrete, and (b) Steel .............................. 25

2.22. Lateral load – top displacement response of specimen RW2 (Massone,

2006) ................................................................................................................ 26

2.23. Lateral load – displacement responses at first story level (RW2) (Massone,
2006) ................................................................................................................ 26

3.1. Implementation of RC panel behavior into MVLEM: (a) original

MVLEM (b) RC panel element; and (c) SFI-MVLEM .................................. 28

3.2. Behavior of uncracked concrete in the FSAM: (a) strain field, (b) principal
strains, (c) principal stresses in concrete; and (d) concrete stresses
(Kolovari, 2013) .............................................................................................. 32

3.3. Concrete behavior after development of the first crack: (a) strain field (b)
concrete strut strains (c) concrete strut stresses; and (d) concrete stresses
(Kolozvari, 2013) ............................................................................................ 33

3.4. Concrete behavior after development of the second crack: (a) strain field (b)
concrete strut strains (c) concrete strut stresses, and (d) concrete stresses
(Kolozvari, 2013) ............................................................................................ 34

3.5. Steel reinforcement behavior in the FSAM (a) strain field (b) stress–strain
model (c) dowel action model; and (d) steel stresses (Kolozvari, 2013) ........ 35

3.6 Concrete and steel stresses superposition: (a) concrete stresses (b) steel
stresses; and (c) resultant stress field .............................................................. 35

3.7. Transformation of concrete shear stress across cracks in FSAM (shear

aggregate interlock model) (Kolozvari, 2013) ................................................ 36

viii
3.8. Dowel action on reinforcement (Kolozvari, 2013) ......................................... 37

3.9. Hysteretic constitutive model for concrete by Chang and Mander (1994) ..... 38

3.10. Compression and tension envelopes of the model by Chang and Mander
(1994) .............................................................................................................. 39

3.11. Steel constitutive model by Menegotto and Pinto (1973) and Filippou et al.
(1983) .............................................................................................................. 40

3.12. Cyclic curvature degradation........................................................................... 41

3.13. Comparison between two groups of cyclic curvature parameters (Orakcal,

2004) ................................................................................................................ 42

3.14. Implementation of SFI-MVLEM model into OpenSees platform .................. 43

4.1. Isometric view of three wall cross-sections tested by Oesterle et al. (1976,
1979) ................................................................................................................ 44

4.2. Boundary details of rectangular specimens tested by Oesterle et al. (1976) ... 48

4.3. Boundary details of barbell specimens tested by Oesterle et al. (1976, 1979) 48

4.4. Boundary details of flanged specimens tested by Oesterle et al. (1976,

1979) ................................................................................................................ 47

4.5. Concrete stress-strain relation for specimens tested by Oesterle et al. (1976,
1979) ................................................................................................................ 48

4.6. Steel stress-strain relation for specimens tested by Oesterle et al. (1976,
1979) ................................................................................................................ 48

4.7. Schematic of test setup (Thomsen and Wallace, 1995)................................... 53

4.8. Profile view of RW1 specimen with dimensions and reinforcement

distribution....................................................................................................... 53

4.9. Cross sections of specimen RW1 (Thomsen and Wallace, 1995) ................... 54

4.10. Concrete stress-strain relations for specimen RW1 ........................................ 55

4.11. Steel stress-strain relations for specimen RW1 ............................................... 55

4.12. Plan view for WSH6 Specimen, all dimensions in mm (1in = 25.4 mm) ....... 56
ix
4.13. Concrete stress-strain relations for WSH6 specimen ...................................... 57

4.14. Steel stress-strain relations for 6 mm rebar used for WSH6 specimen ........... 57

4.15. Steel stress-strain relations for 8 mm rebar used for WSH6 specimen ........... 58

4.16. Steel stress-strain relations for 12 mm rebar used for WSH6 specimen ......... 58

5.1. Dimensions and cross section of B8 shear wall specimen (Oesterle et al

(1979) .............................................................................................................. 61

5.2. Discretization and geometry of B8 specimen.................................................. 62

5.3. Concrete constitutive model and associated parameters ................................. 66

5.4. Calibration of concrete envelop curve in compression ................................... 66

5.5. Confined and unconfined concrete calibration with Saatcioglu and Razvi
(1992) .............................................................................................................. 68

5.6. Calibration of stress-strain relationship for concrete in tension; Kolozvari

(2013) .............................................................................................................. 69

5.7. Shear resisting mechanism and associated parameters ................................... 70

5.8. The effect of shear resistance parameters on predicted results ....................... 70

5.9. Choosing the best predicted results that match experimental results .............. 71

5.10. Defining constitutive materials of reinforced concrete panel ......................... 72

5.11. SFI-MVLEM elements construct RC wall ...................................................... 72

5.12. Displacement increments technique for applying lateral cyclic loading ......... 73

6.1. Top total lateral load-displacements plot for specimen R1 ............................. 75

6.2. Lateral load-shear distortion plot for specimen R1 ......................................... 76

6.4. Top total lateral load-displacements plot for specimen R2 ............................. 77

6.5. Lateral load-shear distortion plot for specimen R2 ......................................... 78

6.6. Top total lateral load-displacements plot for specimen B1 ............................. 79

x
6.7. Lateral load-shear distortion plot for specimen B1 ......................................... 80

6.8. Top total lateral load-displacements plot for specimen B2 ............................. 81

6.9. Lateral load-shear distortion plot for specimen B2 ......................................... 82

6.10. Top total lateral load-displacements plot for specimen B3 ............................. 83

6.11. Lateral load-shear distortion plot for specimen B3 ......................................... 84

6.12. Top total lateral load-displacements plot for specimen B6 ............................. 85

6.13. Lateral load-shear distortion plot for specimen B6 ......................................... 86

6.14. Top total lateral load-displacements plot for specimen B7 ............................. 87

6.15. Lateral load-shear distortion plot for specimen B7 ......................................... 88

6.16. Top total lateral load-displacements plot for specimen B8 ............................. 89

6.17. Lateral load-shear distortion plot for specimen B8 ......................................... 90

6.18. Top total lateral load-displacements plot for specimen F1 ............................. 91

6.19. Lateral load-shear distortion plot for specimen F1 ......................................... 92

6.20. Top total lateral load-displacements plot for specimen F2 ............................. 93

6.21. Lateral load-shear distortion plot for specimen F2 ......................................... 94

6.22. Top total lateral load-displacements plot for specimen RW1 ......................... 95

6.23. Lateral load-shear distortion plot for specimen RW1 ..................................... 96

6.24. Top total lateral load-displacements plot for specimen WSH6 ....................... 97

xi
 ACKNOWLEDGMENTS

I would first like to thank my thesis advisor Dr. Kristijan Kolozvari. The door to

Prof. Kolozvari office was always open for guiding and providing me with knowledge

throughout my MSc. studies at California State University, Fullerton. I would also like to

express my deepest thanks to the members of the thesis committee, Dr. David Naish and

Dr. Joel Lanning for lending their time and insightful comments. I must express my very

profound gratitude to my mother for encouraging and supporting me during my study in

the USA; this study is dedicated to her and the soul of my father. My endless gratitude for

the Higher Committee for the Education Development in Iraq for their financial support.

xii
 1

CHAPTER 1

INTRODUCTION

1.1 Background

Lateral resisting systems are very important for building to resist the lateral forces

caused by the action of wind and earthquake ground motions. Reinforced concrete (RC)

structural walls are considered one of the most effective lateral resisting systems used in

building structures. RC walls are used in order to employ their substantial strength for

resisting lateral loads, provide sufficient stiffness and the deformation capacities that

meet the requirements during the ground motions of the energetic earthquakes. Therefore,

it is important to study the nonlinear behavior of RC walls and propose an analytical

model and experimental investigations. Given the significant role played by the walls in

the performance of buildings during the seismic action, it is necessary to propose

effective analytical models with robust capabilities for capturing inelastic responses and

predicting of nonlinear behavior.

The behavior of walls is usually classified according to the height-to-length ratio

of the wall, which is called aspect ratio (hw / l w ) , or according to shear-span-to-depth ratio

( M / Vlw ) , as either shear-controlled walls or flexure-controlled walls. Shear-controlled

walls are associated with walls with an aspect ratio between or less than 1.0, while

flexural-controlled walls are associated with walls with aspect ratios between or greater

than 3.0. For walls with aspect ratios approximately between 1.0 and 3.0, which is stated
 2

as a moderate aspect ratio wall, the interaction between shear and flexural is expected.

Shear-Flexure-Interaction (SFI) behavior in RC walls was observed experimentally for

the first time by Oesterle, Fiorato, Johal, Carpenter, and Corley (1976) and Oesterle,

Aristizabal-Ochoa, Fiorato, Russel, and Corley (1979). More recent experimental data

have shown some evidence about the SFI. Massone and Wallace (2004) have concluded

that there is a coupling between the inelastic flexure deformation and inelastic shear

deformation, commonly called shear-flexure interaction (SFI), when they revealed that

the inelastic deformations for both shear and flexure occur near-simultaneously at the

same levels of lateral displacement even though the nominal shear strength of the wall is

approximately double the lateral force needed to develop flexural yielding. Shear-flexure

interaction has been detected experimentally not only in moderate RC walls but also in

slender RC walls that have aspect ratios equal to 3.0, such as walls tested by Thomsen

and Wallace (1995) with a contribution of shear deformation equal to approximately 30%

of the total lateral deformation at the first story level, and a wall with aspect ratio equal to

4.0, such as the wall tested by Sayre (2003) with a contribution of shear deformation

approximately equal to 10% of the total lateral deformation at the top of the wall

(Massone & Wallace, 2004). For RC walls with aspect ratios 1.5 and 2.0, Tran and

Wallace (2012) indicated that the contribution percent of shear deformation increases to

approximately 35% and 30% respectively, from the total deformation at the top of the

wall.
 3

1.2 Scope and Objectives

The analytical model proposed by Kolozvari, Orakcal, and Wallace (2015b) has

been shown to be a successful analytical tool for prediction of the nonlinear behavior of

RC walls under cyclic loading conditions. The modeling approach incorporates cyclic

constitutive RC panel model based on a fixed-crack-angle approach into the fiber-based

(multi-spring) model formulation to simulate the observed coupling between nonlinear

axial/flexural and shear responses in RC walls. The coupling of axial and shear responses

is accounted for at the RC panel (macro-fiber) level, which further allows coupling of

flexural and shear responses at the model element level. The model has been validated in

previous studies (Kolozvari et al., 2015b) against five RC wall specimens that

experienced significant shear-flexural interaction, and it showed to be an effective

analytical tool for prediction of global and local wall responses. However, model

validation against a great number of experimental results obtained from tests on walls

with a range of characteristics is needed to consider the modeling approach reliable for

further applications.

The objective of this study is to calibrate and validate the SFI modeling approach

recently implemented in the computational platform OpenSees against a great number of

specimens to assess its capability of successfully reproducing experimentally-measured

wall responses over a range of wall characteristics. In particular, the primary objectives

of the presented study are to:

1) Create a database of experimental results on structural walls with a range of

wall characteristics that will be used for model validation.

 4

2) Develop a unified procedure for calibration of material properties and

geometry of considered wall specimens, and employ the procedure to generate

computer models of the specimens in OpenSees.

3) Conduct assessment of model predictions by comparison experimentally

measured and analytically predicted wall responses for considered wall

specimens at global and local response levels.

4) Provide modeling recommendations and propose future model improvements

based on the assessment of analytical results.

1.3 Organization

This thesis is divided into seven chapters. Chapter 2 provides a summary of

previous research conducted on modeling of walls with and without SFI in RC structural

walls. A theoretical description of the SFI-MVLEM analytical model, including model

formulation, the use of RC panel, the uniaxial concrete and steel constitutive behavior,

and the implementation of the analytical models into the OpenSees platform are

described in Chapter 3. Chapter 4 provides an overview of experimental programs used to

validate the analytical model for RC walls with aspect ratios 2.4, 3, and 2.26. Chapter 5

provides an example for modeling RC walls into OpenSees and discusses detailed

information on calibration of the materials (concrete and steel) for experimental wall

specimens. Chapter 6 provides analytical and model result comparisons for walls at top

flexural deformation response and shear deformation response at various locations,

whereas Chapter 7 provides a summary and conclusions, along with general model

assessment and model capabilities, and recommendations for model improvements.

 5

CHAPTER 2

LITERATURE REVIEW

This chapter discusses an overview of prior studies done by some researchers on

modeling of the inelastic behavior of RC walls. This chapter was divided mainly into

general information and two main subjects including models without interaction of axial,

shear and flexure responses, and fiber models with the interaction of axial, shear, and

flexure responses.

2.1 General Information

Many analytical models have been developed to capture nonlinear behavior of

reinforced concrete structural walls. The purpose of these analytical models is to predict

the inelastic response of multi-story structural systems containing reinforced concrete

shear walls in simple and reasonably accurate manner and the contribution of each

element in this system. This can be achieved either by using macroscopic models that

depend on expected sensible accuracy and overall wall local behavior or by using

microscopic models that depend on a detailed simulation of local behavior and gain

results through the finite element methodology.

Microscopic models could present important aspects of the shear wall nonlinear

behavior and sophisticated details of local responses, but the efficiency, practicality, and

reliability of these models are related to the complexity of contributed factors in their

development and the experimental results interpretation. On the other hand, macroscopic

phenomenological models are practically useful and computationally efficient for

 6

capturing the nonlinear response of multi-story buildings that use RC structural walls

because the effort required to gain the output results is smaller than that required in

microscopic models. However, the trade-off between macro and micro models is still

controversial. Both macro and micro models are considered effective if they would be

interpreted simply with reasonable accuracy to obtain the results in a relatively short

time. Simulating the nonlinear hysteretic responses of RC shear walls under flexural

loading only without accounting for the interaction of other loadings such as shear or

axial loading is a kind of basic analysis and can be done with good accuracy, but for more

accuracy reasons, it is important to simulate their responses by incorporating the

interaction of shear, axial, and flexural loadings. The following sections represent an

overview of analytical models used to simulate the nonlinear behavior of RC structural

walls, including models with uncoupled and coupled axial/flexural and shear responses.

2.2 Models without Interaction of Axial, Shear, and Flexure Responses

2.2.1 Three-Vertical-Line-Element Model and Its Modifications

Kabeyasawa, Shiohara, Otani, & Aoyama (1983) developed a macroscopic model

named the Three-Vertical-Line-Element Model (TVLEM) after pseudo-dynamic

earthquake response tests of a full-scale seven-story reinforced concrete wall-frame

structure was conducted in Tsubaka, Japan. The shear walls were subdivided into three

vertical line elements with two rigid beams at the both top and bottom levels (see Figure

2.1); the central element consists of a single component model with three springs:

horizontal, vertical, and rotational located at the base, whereas the axial stiffness of the

boundary columns was represented by the two external truss elements. This model was

proposed to capture the important aspects of experimental behavior that could not be
 7

captured by the use of models consisting of one element such as the beam-column

element model. Some of these aspects are the fluctuation of neutral axis location along

the cross section of a wall, the wall rocking (see Figure.2.2), and the interaction with

other members connected to the wall.

Figure 2.1. Three-Vertical-Line-Element Model (TVLEM) Kabeyasawa et al. (1983).

 




(a) Beam-Column Element Model (b) Observed

Figure 2.2. Wall rocking and effect of neutral axis shift on vertical displacements.

The axial force-deformation relationship of the (TVLEM) model was defined by

the rules of the axial-stiffness-hysteresis model (ASHM) (Kabeyasawa et al., 1983)

(Figure 2.3), while an origin-oriented-hysteresis model (OOHM) Kabeyasawa et al.,

1983) was used for defining both the horizontal and rotational springs that concentrated

at the base of the central vertical element (Figure 2.4). The rotational spring stiffness
 8

characteristics were defined by stating that the wall area is limited to the area enclosed by

the inner faces of the two boundary columns (i.e., central element only); therefore, the

compatibility between the panel displacement and the boundary columns was not applied.

Regarding shear stiffness degradation, it was incorporated independently without any

interaction with axial force and bending moment. However, the model took into

consideration the neutral axis variation of the wall and the wall interaction with the other

surrounding frame elements connected to the wall (i.e., outrigging), and produced good

results compared to experimental results for each of the top displacement, base shear,

axial deformation at wall boundaries, and rotation at beam ends.

Figure 2.3. Axial-Stiffness-Hysteresis Model (ASHM) (Kabeyasawa et al., 1983).

 9

Yield

Cracking
Force

Displacement
Figure 2.4. Origin-Oriented-Hysteresis Model (OOHM) (Kabeyasawa et al., 1983).

Three-Vertical-Line-Element Model (TVLEM) has been modified by researchers

to predict flexural behavior and capture inelastic shear deformations by fixing the lack in

TVLEM. The two-axial-element-in-series model (AESM) shown in Figure 2.5 was used

instead of the axial-stiffness hysteresis model (ASHM) by Vulcano and Bertero (1986).

Figure 2.5. Axial-Element-in-Series Model (AESM) (Vulcano and Bertero, 1986).

AESM consist of two elements. Element 1 defines the linear elastic response and

characterizes the global axial stiffness of the column segments before the bond
 10

deterioration, while element 2 consists of two components to characterize the axial

stiffness of the rest of segments of steel and cracked concrete after the bond deterioration.

The relationship of axial force-deformation that was created by the two-axial-element-in-

series model is shown in Figure 2.6. Regarding the shear spring and the rotational spring

located at the centerline of wall, origin-oriented hysteresis model (OOHM) was reused

for capturing inelastic shear deformations.

Figure 2.6. Axial force-deformation relationship of the AESM (Vulcano and Bertero,
1986).

By assessing the accuracy of this model, some discrepancies were detected

between expected and experimental shear results of the wall specimens; however, a

global response such as base shear-top displacement was well-simulated compared to

measured behavior. Therefore, the authors found that the OOHM failed to predict shear

hysteretic behavior, whereas the flexural results predicted by AESM compared favorably

with experimental flexural behavior. Generally, the authors concluded that the
 11

incompatibility between the rotational spring and the boundary columns was not tackled

by the modified TVLEM; however, the modified TVLEM modeled the inelastic flexural

response of RC shear walls efficiently. It was also highlighted that the ad-hoc selection of

modeling parameters in this modified model, such as the bond deterioration

considerations, the yield strength of the horizontal spring, and the strain hardening ratio,

would affect the displacement components of this modified model sensitively.

Vulcano and Bertero (1987) also concluded that OOHM could not predict

inelastic shear deformations caused by cyclic loading for specimens exposed to high

values shear stress. Therefore, Kabeyasawa (1997) tried to solve this problem by

modifying the original Three-Vertical-Line-Element Model (TVLEM) to capture the

global behavior for both shear and flexural of RC walls under monotonic and cyclic

loading conditions. A new two-dimensional Panel-Wall Macro Element (PWME) was

used as an alternative to the horizontal, vertical, and rotational springs concentrated at the

base of the centerline of the RC wall (see Figure 2.7).

Figure 2.7. Modification of the TVLEM (Kabeyasawa et al., 1997).

It was found that the new alternative modification of TVLEM and PWME could

produce reasonable results for walls under monotonic, cyclic, and axial loading. But, both
 12

models were unable to simulate the behavior of walls with significant cyclic nonlinear

shear deformations and high axial loading conditions. The PWME model showed

inefficiency of simulating the shear response of RC walls as a function of axial load.

2.2.2 Multiple-Vertical-Line-Element Model

Fiber element models have become common approaches to simulate both flexure

and shear responses. One of the most common models is Multiple-Vertical-Line Element

Model (MVLEM) that was proposed by Vulcano, Bertero, and Colotti (1988) to obtain

the nonlinear response of the RC wall with the simplicity of Macro models and the

precision of micro models. A multi-uniaxial-element-in-parallel model was used to

simulate the flexural response of a wall member with two rigid beams at both the top and

the bottom floor levels. A structural wall was subdivided into an m number of MVLEM

elements (see Figure 2.8(a)). The two external elements characterized the axial stiffness

of the boundary columns (k1 and km), while the other two or more interior elements (k 2,

. . . kn-1) characterized the axial and flexural stiffness of the central wall panel (see Figure

2.8(b)). Inelastic shear response of the RC wall was represented by the horizontal spring,

with stiffness KH placed at a height (ch) where the rotation of the element will occur (see

Figure 2.8(b)), and hysteretic behavior described by the origin-oriented hysteresis model

(OOHM) (Kabeyasawa et al., 1983). For each element in the MVLEM, the center of

rotation of the wall member was to be located at a height (ch) assuming the wall

curvature distributed uniformly over the height of the wall elements in which curvature

would vary linearly between the top and bottom of the element (see Figure 2.9). The

selection of this parameter value was suggested by Vulcano et al. (1988) and confirmed
 13

by Orakcal and Wallace (2006) by comparing the model response with experimental

results.

(a) MVLEM Wall Model (b) Model Element

Figure2. 8. Multiple-Vertical-Line-Element-Model (Vulcano et al, 1988).

= (1-c)h

(1-c)h
h


ch Fiber Model

MOMENT CURVATURE

Figure 2.9. Flexural deformations of the MVLEM element (Orakcal et al., 2004).

The stress-strain relationship was used instead of force-deformation rules to

improve the hysteretic response of RC wall materials; hysteretic response of steel was

interpreted by the stress-strain relationship suggested by Menegotto and Pinto (1973) and
 14

developed by Filippou, Popov, and Bertero (1983) to incorporate the effects of isotropic

strain hardening (see Figure 2.10, Figure 2.11), while uncracked concrete was interpreted

by the stress-strain relationship recommended by Colotti and Vulcano (1987) (see Figure

2.12). Also, the cracked concrete was modeled by the stress-strain relationship

recommended by Bolong, Mingshun, and Kunlian (1980) in which the contact stresses

that are caused by the gradual opening and closing of cracks were accounted (see Figure

2.13). The stiffness of the tension effect was taken into consideration by incorporating the

non-dimensional parameter (c) such that the stiffness of tension of the uniaxial model

shown in Figure 2.9 and the actual stiffness of tension of the uniaxial RC member would

be equal.
 15

600
(r1,r1)
E1= bE0
(02,02) y 1

400
(MPa)

200
E0
1


0 y
Stress,

(1-b) *
*= b *+ 
1/ R
-200 (1 + * R )
- r
* = 
0 - r
-400
(01,01)  - r
( ,r ) *= 
0 - r
2 2
r
-600

-0.01 -0.005 0 0.005 0.01 0.015 0.02

Strain, 
Figure 2.10. Constitutive model for steel (Menegotto & Pinto, 1973).

1.5
2
/y

1 R0

0.5
Normalized Stress,

R(2)
0

R(1)
a1 
-0.5
R = R0 - a + 
2

-1 ( m - 0 )
 =
1 yy
-1.5

-5 -2.5 0 2.5 5 7.5

Normalized Strain, / y

Figure 2.11. Degradation of cyclic curvature (Menegotto & Pinto, 1973).

 16

Figure 2.12. Constitutive law adopted in the original MVLEM for uncracked concrete
(Colotti and Vulcano, 1987).

Figure 2.133. Constitutive law adopted in the Original MVLEM for cracked Concrete
(Bolong et al., 1980).

Although the MVLEM simulates essential response aspects such as neutral axis

fluctuation and axial force effect on stiffness and strength (including refined material

constitutive models such as confinement considerations, progressive opening and closing

of cracks effects in the analysis, and prediction of wall flexural responses in reason

accurate), the prediction of the shear behavior of RC structural walls was still unreliable
 17

and questionable and the simulation of slender RC wall behavior was unconfirmed.

Therefore, more investigations and research have been done to improve MVLEM.

Fischinger et al, (1990); Fajfar & Fischinger (1990); Fischinger, Vidic, & Fajfar (1991);

and Fischinger, Vidic, and Fajfar (1992) have modified MVLEM to validate, calibrate,

and capture the experimental behavior and simplify the hysteretic rules by the use of

vertical and horizontal springs; however, the results of modified model are based on

random force-deformation parameters, some of those parameters were easy to define, but

it was difficult to interpret the other parameters related to inelastic shear behavior or

those related to the fatness of the hysteresis loops (see Figure 2.14, Figure 2.15).

Figure 2.14. Force-deformation relationships adopted in the modified MVLEM

(Fischinger et al., 1990) - Vertical Springs.
 18

Figure 2.15. Force-deformation relationships adopted in the modified MVLEM

(Fischinger et al., 1990) - Horizontal Spring.

Orakcal, (2004); Orakcal, Conte, and Wallace (2004); and Orakcal and Wallace

(2006) have improved the MVLEM by incorporating the stress-strain relationship of

materials instead of force-deformation rules that were implemented by Fischinger et al.

(1990) and by updating and refining constitutive relationships for concrete and steel to

predict the nonlinear response of slender RC shear walls. Interpretation of concrete

uniaxial behavior was done by the constitutive model suggested by Chang and Mander

(1994), which addressed significant aspects of hysteretic behavior of concrete such as the

stiffness degradation of the unloading and reloading curves for increasing values of

strain, the confined and unconfined concrete hysteretic behavior, tension stiffening, and

progressive opening and closing of concrete cracks which had not been tackled

sufficiently in prior models (see Figure 2.16). Interpreting of the steel stress-strain

behavior was done by the Menegotto and Pinto (1973) model which is a well-known

model for the nonlinear relationship of steel reinforcement and extended by Filippou et

al. (1983) to include the effects of isotropic strain hardening (see Figure 2.11).
 19

Ascending Descending
Compression ( c' ,f c' )
Branch Branch

Compression or ( c' ,f c' ) or (t , ft )

Tension Envelope  (c0,0)
Tension (c0+ t,ft)


f
Stress,

r Linear
Branch
1
Ec
Esec 1

Ec
1
(pl,0) Epl (cr,0)
1

Strain, 
Figure 2.16. A uniaxial constitutive model for concrete (Chang & Mander, 1994; Orakcal
et al., 2004).

2.2.3 PERFORM 3D - Shear Wall Element

Computers and Structures Inc. (CSI) has produced software program to model

structural shear walls and other structural elements using a displacement-based design

based on the procedures that have been illustrated in ASCE 41: Seismic Rehabilitation of

Existing Buildings. This software model is able to model relatively slender walls and

squat walls by dividing the wall into some components and elements; each connects four

nodes with 24 degrees of freedom to capture the inelastic nonlinear analysis. The concept

of the fiber model used by Perform 3D is generally similar to the concept used in the

MVLEM model which was explained in Section 2.2.2 (see Figure 2.17). Perform 3D has

two options that can be used to model fiber elements of the structural wall: “General

Wall” element, which is kind of complex and recommended for squat walls and “Shear

Wall” element, which is simple and recommended for slender walls (Powel, 2007).
 20

Steel
Fibers

Concrete
Fibers

Reinforced
Concrete
Element
a) Fiber model of cross section b) RC Wall
Figure 2.17. Fiber model used by Perform 3D.

It is not required for the shape of a shear wall element to be rectangular, but the

inclination of the shape should not be very high. Each element has three axes in three

different directions, a longitudinal direction, which is typically vertical, a transverse

direction, which is horizontal, and an out-of-plane direction, which is normal to the

element plane. Due to the fact that there is no rotational stiffness at any of the shear wall

element nodes, the connection of the wall with other structural elements such as beam

elements is pinned. A beam element must be embedded in the wall wherever connecting

between a beam and a wall is required for accounting the moment-resisting connection

purposes, and the bending stiffness of the embedded beam should be very high to make

the connection in rigid condition; however, this might reduce the accuracy of predicted

results. It is important to mention that the material models in Perform 3D are relatively

crude compared to sophisticated material models that are used in research-oriented

models.
 21

2.3 Models with Interaction of Axial, Shear and Flexure Responses

In recent research and studies, many models have been proposed to capture the

interaction axial, shear and flexure responses. One of these models was a strut-and-tie

(truss) model that was presented by Panagiotou, Restrepo, and Schoettler (2012), which

provided a new approach to capture interaction between shear and flexural responses in

RC walls by using inelastic elements (vertical, horizontal, and diagonal truss elements )

connected to each other by nodes. The vertical and horizontal truss elements symbolized

reinforcing bars and the concrete in contact with bars, while diagonal truss elements

symbolized the concrete between the bars (see Figure 2.18). The interaction of Shear and

flexure was interpreted by the compression stress reduction of diagonal concrete truss

elements due to transverse normal tensile strain by applying concrete uniaxial material

rules (see Figure 2.19).

Even though the model provided reasonable and viable simulation of nonlinear

responses and SFI capture, there was an overestimation of pre-cracking stiffness and wall

strength related to the overlapping areas of vertical, horizontal and diagonal concrete

elements. Also, the overall force-deformation response that was predicted by this model

had been found to be sensitive to the element size. As a result, the effect of element

length has to be considered not only in the concrete uniaxial stress-strain behavior but

also in the determination of the normal tensile strain and behavior of concrete

compressive stress-strain relationship.

 22

Figure 2.18. RC wall and wall panel truss model (Panagiotou et al., 2012).
 23

a) b)

Figure 2.19. Coupled compression-tension model for diagonal truss elements: (a) coupled
compression-tension truss elements, (b) concrete compressive stress reduction due to
normal tensile strain (Panagiotou et al., 2012).

Massone, Orakcal, and Wallace (2006, 2009) modified the MVLEM model by

adding a shear spring to all uniaxial elements that modeled the structural RC wall. Each

element was assumed to be an RC panel in which normal and shear stresses were

subjected (see Figure 2.20). The interaction of shear and flexure was incorporated in this

model by coupling the shear and axial stresses in each panel. The assumptions that

formulated this model were that the plane section before loading remains plane after

loading and there is no bond degradation between steel and concrete (i.e., the bond is

perfect) so the steel and concrete have the same strain fields. Further assumptions include

the distribution of shear strain is uniform along the wall length, principal stress and strain

directions coincide, and there is no reinforcement dowel action (i.e., equal to zero).
 24

Figure 2.20. Coupled model element (Massone et al., 2006, 2009).

Constitutive material that was used in this model to interpret the monotonic

behavior of steel was the same concept of Menegotto and Pinto model (1973) (see Figure

2.21(a)). The reduction in principal compressive stress that may happen due to the cracks

caused by tensile strain was taken in consideration by the concrete constitutive softened

which was calibrated and compared with many experimental results. Meanwhile, the

effect of tension stiffening was taken into consideration by the stress-strain relationship

suggested by Belarbi and Hsu (1994) (see Figure 2.21(b)).

 25

(a) (b)
Figure 2.21. Constitutive material models. (a) Concrete, and (b) Steel.

The shear flexural interaction was simulated with reasonable and accurate

analytical results for inelastic shear and flexural deformation compared to experimental

results, especially for moderately-slender walls (see Figure 2.22). One of the most

obvious deficiencies that can be noticed in Massone et al’s (2006, 2009) model was that

this model could not predict the cyclic responses because of the use of constitutive

monotonic material and a RC panel that was assumed at each element was modeled

according to the rotating-angle approach. Consequently, the model proposed by Massone

et al (2006, 2009) could not simulate cyclic degradation, which led to the overestimation

of the flexural deformations and the underestimation of the shear deformations (see

Figure 2.23).
 26

Figure 2.22. Lateral load – top displacement response of specimen RW2 (Massone et al.,
2006).

Figure 2.23. Lateral load – displacement responses at first story level (RW2) (Massone et
al., 2006).
 27

CHAPTER 3

DESCRIPTION OF SHEAR-FLEXURAL INTERACTION MODEL

In this chapter, the shear-flexure-interaction model (SFI-MVLEM from Kolozvari

[2013]) will be described, including the baseline model (MVLEM) descriptions and its

improvements to incorporate the shear-flexure interaction such as the constitutive RC

panels. Also, the uniaxial concrete and steel constitutive behavior will be explained.

3.1 Shear-Flexural Interaction Model Assumptions

The Multiple-Vertical-Line-Element-Model (MVLEM) (Vulcano et al., 1988) that

was implemented by Orakcal et al. (2004) has been selected as the standard model for the

new improvements to incorporate cyclic shear-flexural interaction (SFI) for many reasons

such as its simplicity in implementation, its full description of concrete and steel

behavior, stability in numerical analysis, efficiency, and its rationally precise predictions

of flexural responses (Orakcal & Wallace 2006). Reinforced concrete panel elements

have been incorporated in MVLEM instead of the uniaxial macro-fiber elements in order

to capture the shear-flexure interaction in RC shear walls (see Figure 3.1). Fixed-Strut-

Angle-Model (FSAM) that was proposed by Ulugtekin (2010) and Orakcal (et al. (2012)

and was used to describe the performance of RC panel elements but it was improved by

the incorporation of shear resistance of aggregate interlock behavior in concrete (Orakcal

et al., 2012) and shear resistance in reinforcement dowel action (Kolozvari, 2013) parallel

to inclined cracks of a reinforced concrete panels to improve the shear prediction of the

model.
 28

Rigid Beam

Rigid Beam
(a) (b) (c)

Figure 3.1. Implementation of RC panel behavior into MVLEM: (a) original MVLEM
(b) RC panel element; and (c) SFI-MVLEM.

The SFI-MVLEM model has been formulated based on three assumptions: (a)

plane sections before loading remain plane after loading (b) shear strains are uniformly

distributed across the wall cross section; and (c) the results of horizontal normal stresses

associated with steel and concrete (i.e., horizontal smeared stress in concrete [σ x ] ) along

the length of the wall are equal to zero. The first and second assumptions (a) and (b) were

taken from the baseline MVLEM model, whereas the third assumption (c) was involved

in the SFI-MVLEM model to provide equilibrium conditions in the horizontal axis (x-

axis) of the reinforced concrete panel elements in order to calculate the normal strain in

the horizontal axis (εx ) for each panel element. The second assumption does not

considerably affect the slender RC wall behavior prediction since the majority of lateral

displacements are due to flexural distortions. However, in a case of squat (short) walls

and short columns, this assumption may not be useful since there might be some

conditions that affect the prediction of RC walls behavior such as the probable existence

of different deformation modes, the mechanisms of transferring loads like the effect of
 29

shear deformations. Also, the third assumption related to the consequence that the

resultant normal stresses in the horizontal direction (x-axis) are equal to zero. This

assumption is more realistic with the slender RC wall behavior prediction than with strut

walls and short columns due to the aforesaid effects. According to Massone et al. (2006)

the three aforementioned assumptions that were used in the SFI-MVLEM model are more

accurate and reasonable with cantilever walls that have aspect ratios greater than 1.0.

Thus, it is expected that the SFI-MVLEM model might be appropriate for slender and

medium-rise RC columns and walls, with aspect ratios greater than 1.0. Generally, SFI-

MVLEM model is similar to the model proposed by Massone et al. (2006) in concept, but

there were some modifications and improvements that have been made to capture cyclic

shear-flexural interaction responses of structural walls. Regarding the degrees-of-

freedoms (DOFs), Each SFI-MVLEM element is characterized with six external DOFs.

The location of These DOFs is concentrated at the center of rigid beams (top and bottom)

of each model element (see Figure 3.1(c)).

The flexural response of the SFI-MVLEM elements can be captured by the action

of axial deformations along the y-axis or vertical direction in each RC panel elements

(see Figure 3.1(c)). The concept used in prediction of flexural response of the SFI-

MVLEM model was completely like the concept of prediction flexural response in

uniaxial fiber elements in the baseline model MVLEM (see Figure 3.1 (a)). For each RC

panel, the center of rotation of the wall member was to be located at a height (ch)

assuming the wall curvature distributed uniformly over the height of the wall elements in

which curvature would vary linearly between the top and bottom of the element (see

Figure 3.1(c)). Shear deformation of an SFI-MVLEM element, which occurs at a height

 30

equal to (ch), can be captured by shear deformations in each RC panel element by

assuming that shear strain distributed uniformly across the length of the wall (see Figure

3.1(b)). A value of (c = 0.4) was used in the SFI-MVLEM element similar to that value

that was suggested by Vulcano et al. (1988) and confirmed by Orakcal and Wallace

(2006) by comparing many experimental results with results that obtained from model

responses. More details can be found at Kolozvari (2013) and Kolozvari, Orakcal, and

Wallace(2015a).

3.2 Reinforced Concrete Panel and Fixed-Strut-Angle-Model

The behavior of RC panel in SFI-MVLEM element was based on the same

concept of the Fixed-Strut-Angle-Model (FSAM) which was suggested by Ulugtekin

(2010) and improved by Orakcal et al. (2012). By assuming a perfect bond between

concrete and reinforcement in RC panel, the concrete and reinforcement strain fields of

an RC panel element were assumed to be the same in value. In the original FSAM

(Ulugtekin, 2010), shear stresses in concrete along the cracks were assumed to be zero

because of the assumption stated that the principal stress directions of concrete coincide

with the cracks directions, but this was developed in SFI-MVLEM model by

incorporating shear aggregate interlock action in concrete according to the suggestions

that were proposed by Orakcal et al. (2012) and making some other developments in

FSAM by incorporating dowel action on reinforcement bars to improve the simulation of

shear-flexure interaction in shear walls and columns. These improvements were

implemented in OpenSees as will be shown in this chapter and these improvements were

incorporated into the SFI-MVLEM in order to gain reasonable and accurate simulation

results for cyclic responses in columns and structural walls.

 31

3.2.1 Concrete Behavior in the RC Panel Element

Concrete behavior in the RC panel element can be categorized into three

consecutive stages during the loading process: 1) Uncracked concrete, 2) Concrete

behavior after the development of the first crack, and 3) Concrete behavior after the

development of the second crack.

3.2.1.1 Uncracked Concrete. In this stage, the concrete stress–strain behavior is

characterized by a rotating strut approach (Vecchio & Collins, 1986) and the Rotating

Angle Strut and Tie Model (Pang & Hsu, 1995). Under a biaxial state of stress, it can be

confidently assumed that the behavior of concrete follows the monotonic stress–strain

relationship before the first cracking due to the use of concrete monotonic stress–strain

relationships in the same directions of principal strain. Principle strain and strain fields

which are transformed into the directions of principal strain directions are applied at first

on concrete (see Figure 3.2(a), Figure 3.2(b)) and then principal stresses are calculated

according to the concrete monotonic stress–strain relationships along the principal strain

directions (see Figure 3.2(c)). After the calculation of principle stresses, at the end of this

state, the stresses fields are converted into the original directions of principal strain (x-y

coordinate system) (see Figure 3.2(d)).

 32

(a) (b)

(a) (b)

(c) (d)
Figure 3.2. Behavior of uncracked concrete in the FSAM: (a) strain field, (b) principal
strains, (c) principal stresses in concrete; and (d) concrete stresses (Kolovari, 2013).

3.2.1.2 Concrete Behavior after Development of the First Crack. The first

crack in concrete occurs when the concrete principal tensile strain value exceeds the

concrete monotonic cracking strain value. In this stage, the direction of principal strain

that was corresponding to the first concrete cracking was considered as the first “fixed

strut” direction for the panel and it will continue fixed until the formation of the next

stage (second crack). Due to the fact that the first strut direction is fixed, a concrete

uniaxial hysteretic stress–strain relationship can be applied in this step in the same

direction of principal stress (parallel and perpendicular to the first cracking). As was

aforementioned, it can be implied from the original model of Ulugtekin (2010) that shear

resistance of aggregate interlock along a crack equal to zero; therefore, shear stresses in

concrete can be calculated depending on the shear strains parallel to the first crack by

using the model that was implemented to calculate shear aggregate interlock. The

concrete strain field was transformed into its components from the principal strain

directions perpendicular and parallel to the direction of the first fixed strut. At the end of
 33

this process, the stresses of concrete on the strut were transformed again into the original

x–y directions (see Figure 3.3).

(b)
(a)

(c) (d)
Figure 3.3. Concrete behavior after development of the first crack: (a) strain field (b)
concrete strut strains (c) concrete strut stresses; and (d) concrete stresses (Kolozvari,
2013).

3.2.1.3 Concrete Behavior after Development of the Second Crack. The

second crack develops when the value of strain parallel to the direction of the first strut

surpasses the value of cyclic cracking strain which is mainly based on the monotonic

cracking strain and the plastic strain. Since the direction of the first strut was considered a

principal stress direction, the direction of the second crack would be perpendicular to the

direction of the first crack. The second fixed strut would occur in the perpendicular

direction of the first strut direction and by increasing the loading, two perpendicular

independent fixed struts will be distinguished in concrete mechanism functioning as

switching tension and compression struts in the directions of the fixed struts depending

on the applied strain field. In this step, it is possible to use a concrete uniaxial hysteretic

stress–strain relationship due to the fact that the direction of both struts (perpendicular

and parallel) are fixed. For calculation purposes, concrete stresses for the two directions
 34

can be obtained by transforming the strain field in concrete into its components which

have the same directions of fixed strut directions (see Figure3.4).

(a) (b) (c) (d)

Figure 3.4. Concrete behavior after the development of the second crack: (a) strain field
(b) concrete strut strains (c) concrete strut stresses, and (d) concrete stresses (Kolozvari,
2013).

Also, for the calculation of concrete shear stresses, the technique that was used was the

same as the first crack development technique which was done by using Shear strains in

the both directions of fixed struts (cracks) by implementing the shear aggregate interlock

model. At the final step, the concrete stresses were retransformed into the original x–y

directions and superimposed.

3.2.2 Steel Reinforcement Behavior in the RC Panel Element

Regarding the steel reinforcement stresses in the FSAM, it can be obtained from

the horizontal and vertical steel reinforcement axial strains using the uniaxial stress–

strain relationship proposed by Menegotto and Pinto (1973) for steel reinforcement with

the implementation of the perfect bond assumption. In addition, the stress of shear acting

in horizontal and vertical directions (perpendicular to steel bars) can be calculated

 35

depending on a shear strain acting on the RC panel by using the model that proposed by

Kolozvari 2013 and implemented in the SFI-MVLEM model which is a linear elastic

dowel action model (see Figure 3.5).

(a) (b) (c) (d)

Figure 3.5. Steel reinforcement behavior in the FSAM (a) strain field (b) stress–strain
model (c) dowel action model; and (d) steel stresses (Kolozvari, 2013).

At the end of all processes, in order to obtain the smeared stresses acting on RC panel

element, the concrete and steel stress fields are superimposed by means of reinforcement

ratios in both vertical and horizontal directions (see Figure 3.6). More details of the RC

panel and the FSAM can be provided by Ulugtekin (2010) and Kolozvari (2013).

(a) (b) (c)

Figure 3.6. Concrete and steel stresses superposition: (a) concrete stresses (b) steel
stresses; and (c) resultant stress field.

3.3 Shear Resistance Models

As it was aforementioned, in order to capture shear response and improve the

FSAM described by Ulugtekin (2010), which assumed the zero-shear aggregate interlock
 36

along the cracks, two models have been involved to capture both the shear due to

aggregate interlock and the shear due to dowel action. For the shear aggregate interlock

model, its behavior begins first with a simple linear elastic loading/unloading relationship

attributable to the sliding shear strain and the shear stress acting parallel to the surface of

crack. When the crack is open, (i.e., the normal stress of concrete acting perpendicular to

the crack is tension), shear stresses of concrete will become zero. However, when the

concrete crack is closed (i.e., the normal stress of concrete is compressive), shear stresses

of concrete represent by the mean of a shear friction coefficient (η) (see Figure 3.7).

According to previous research, the value of the coefficient of shear friction is usually

equal to 0.45 - 1.4 from the concrete modulus of elasticity (Ec); however, Orakcal et al.

(2012) suggested a smaller value of a shear friction coefficient (0.2Ec) to provide more

reasonable and accurate results. Shear friction coefficient was applied in OpenSees under

a parameter called (nu) as it will be represented in chapter 5.

Figure 3.7. Transformation of concrete shear stress across cracks in FSAM (shear
aggregate interlock model) (Kolozvari, 2013).

Another model has been incorporated in the original formation of FSAM, which

is a simple linear elastic constitutive model for signifying the effect of longitudinal dowel

action in columns and structural wall (Kolozvari, 2013). When the crack is open, concrete
 37

shear resistance is assumed to be zero; however, shear resistance can be provided by dowel

action of longitudinal reinforcement in case of horizontal shear deformation of the column or

the wall. The shear strain in an RC panel element and the shear stress in longitudinal

reinforcement are related by using steel modulus of elasticity multiplied by dowel

stiffness parameter (α) (see Figure 3.8). The dowel stiffness parameter may affect the

shear deformations and pinching characteristics and provide more possibility for

convergence rate in the SFI-MVLEM (Kolozvari, 2013; Kolozvari et al., 2015a). The

dowel stiffness parameter was applied in OpenSees under a parameter called (alfadow) to

construct FSAM as it will be represented in chapter 5.

Figure 3. 8: Dowel action on reinforcement (Kolozvari, 2013).

3.4 Uniaxial Concrete and steel constitutive behavior

The Chang and Mander (1994) model was the hysteretic constitutive model that

has been used in SFI-MVLEM modelling to describe the constitutive concrete behavior.

The Chang and Mander (1994) model considers a developed, rule-based, generalized, and

non-dimensional model that provides a wide range of modification for calibrating the

constitutive material parameters and simulating the monotonic and cyclic responses of

unconfined and confined concrete, ordinary and high-strength concrete in both tension

and compression (see Figure 3.9). The original Chang and Mander model has been
 38

modified in order to clarify tension stiffening, compression softening, and hysteretic

biaxial damage.

The envelope curve of the Chang and Mander model was represented by the

monotonic curve for the same hysteretic behavior of stress-strain relationship (see Figure

3.10). Initial tangent slope (Ec) defines the pre-peak part of compression envelope curve

up to the peak point which is defined by (ε′c, f′c). The curve of the compression envelope

is defined by a parameter r which was represented by Tsai’s (1988) equation, and the end

point of the curve is represented by the normalized strain (ε -cr) with respect to (ε′c); after

this point, the envelope curve takes straight line shape until the spalling strain (ε sp) where

compressive stress is zero.

r
Stress, fc

Figure 3.9. Hysteretic constitutive model for concrete by Chang and Mander (1994).
 39

 sp
 cr
-
Compression
 c'
Straight
Line

f c'
Stress, fc

Ec
1
 0
1
O
f t Ec
Straight
Line
 t
Tension
 cr
+
Not to scale
 crk

Strain, c
Figure 3.10. Compression and tension envelopes of the model by Chang and Mander
(1994).

The concrete tension envelope curve shape of the Chang and Mander (1994)

model has the same shape characteristics of the compression envelope except that its

origin coordinates are moved to a new point due to the unloading strain condition from

the compression envelope. The tension envelope curve parameters such as the concrete

tensile strength (ft), the tensile strain (εt), the aforementioned parameter (r) which defines

the tension curve shape based on Tsai’s (1988) equation, and a parameter (ε +cr) which

defines normalized strain with respect to the tensile strain (ε t) in which the tension

envelope curve takes a straight line up to zero tensile stress at a cracked strain (ε crk), can

also be calibrated by empirical relations suggested by the Belarbi and Hsu (1994) model

or other similar models. Based on experimental results, the value of the parameter (ε +cr)

can be assumed to be very large (e.g., 10,000); therefore, concrete tension stiffening can

be expected not to be cracked completely. All aforementioned compression and tension

envelope parameters of concrete are executed in OpenSees for simulating the constitutive

material of concrete structural walls and columns as will be shown in Chapter 5.

 40

The model proposed by Menegotto and Pinto (1973) and extended by Filippou et

al. (1983) to incorporate isotropic strain hardening is the hysteretic model that was used

to describe the behavior of reinforcing steel in SFI-MVLEM model. The stress-strain

relationship consists of the straight line with a slope defined by the modulus of elasticity

(E0) connected by a transient curve to another straight line with a slope defined by yield

modulus (E1 = bE0) where the parameter b is the strain hardening ratio (see Figure 3.11).

The stress and strain at the point of strain reversal are represented by the Parameters (σ r)

and (εr), while the stress and strain that form at the intersection point between the two

straight lines are represented by the Parameters (σ0) and (ε0). The transition curve

between the two straight lines (asymptotes) is defined by the cyclic curvature parameter

(R).

600
(r1,r1)
E1= bE0
( , ) y
0
2
0
2 1

400 R0
(MPa)

200
E0
1


0 y
Stress,

(1-b) *
*= b *+ 
1/ R
-200 (1 + * R )
- r
* = 
0 - r
-400
(01,01) - r
( , ) *= 
0 - r
2 2
r r
-600

-0.01 -0.005 0 0.005 0.01 0.015 0.02

Strain, 
Figure 3.11. Steel constitutive model by Menegotto and Pinto (1973) and Filippou et al.
(1983).

Based on the investigational research, Menegotto and Pinto recommended the

following formula to define the parameter (R):

 41

R = R0 – (3-1)

Where (R0) is the initial value of the parameter R for monotonic loading, and (a 1

and a2) are the parameters obtained from experimental studies representing the curvature

degradation within subsequent cycles. The parameter ξ is representing the absolute strain

difference between the current asymptote intersection point and the previous maximum

or minimum strain reversal point (see Figure 3.12). Based on data obtained from steel

bars cyclic tests, two researchers suggested two groups of values for parameters R 0, a1,

and a2: Menegotto and Pinto (1973), recommended R 0 = 20, a1 = 18.5, and a2 = 0.15,

while Elmorsi et al. (1998) recommended R 0 = 20, a1 = 18.5, and a2 = 0.0015 (see Figure

3.13).

1.5
2
/y

1 R0

0.5
Normalized Stress,

R(2)
0

R(1)
a1 
-0.5
R = R0 - a + 
2

-1 ( m - 0 )
 =
1 yy
-1.5

-5 -2.5 0 2.5 5 7.5

Normalized Strain, / y

Figure 3.12. Cyclic curvature degradation.

 42

600

400
Stress,  (MPa)
200

-200

-400 R0=20 a1=18.5 a2=0.0015

R0=20 a1=18.5 a2=0.15
-600

-0.005 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035

Strain, 
Figure 3.13. Comparison between two groups of cyclic curvature parameters (Orakcal,
2004).

3.5 Implementation in OpenSees

As a consequence of the studies that have been done by several researchers, some

elements and models and their constitutive materials have been implemented into

OpenSees (Kolozvari et al., 2015a). These new element and material models are listed in

the following points (see Figure 3.14):

 Element MVLEM: macroscopic element with uncoupled flexural and shear

behavior

 Element SFI_MVLEM: macroscopic element with shear-flexure interaction

 nDMaterial FSAM: plane-stress RC panel constitutive behavior

 uniaxialMaterial SteelMPF: uniaxial material model for steel

 uniaxialMaterial ConcreteCM: uniaxial material model for concrete

Vulcano et al. (1988) and Orakcal et al. (2004) proposed and developed the

MVLEM model, while Kolozvari (2013), proposed SFI-MVLEM model. In addition,

 43

Ulugtekin (2010) and Orakcal et al. (2012) proposed implementing the FSAM model to

describe and simulate the behavior of RC panel elements. Menegotto and Pinto (1973),

and Filippou et al. (1983) proposed and extended their model for describing uniaxial steel

material, and Chang and Mander (1994) proposed a model to describe the uniaxial

concrete material. All the above models have been added to the computational analysis of

OpenSees platform for improving the nonlinear analysis and simulating the inelastic

responses of RC columns and RC structural walls. More information about the analytical

models and commands of the models are available in Kolozvari et al. (2015). Also,

detailed information about input files and executed examples can be found on the

OpenSees Wiki page.

Figure 3. 14 Implementation of SFI-MVLEM model into OpenSees platform.

 44

CHAPTER 4

OVERVIEW OF EXPERIMENTAL RESULTS

In this chapter, a brief overview of experimental results and material properties of

specimens used for validation of SFI-MVLEM model will be presented. Three different

cross-section shear walls have been used in this study in order to represent how these

different cross sections affect the model prediction of top deformation and shear

distortion of structural shear walls. Besides the cross-section shape, these specimens

differ from each other in some parameters such as axial load, confinement condition,

longitudinal and horizontal reinforcement ratio for web and boundary elements, and some

other parameters.

4.1 Oesterle et al. (1976, 1979)

Two groups of experimentally isolated shear wall specimens were conducted at

the Portland Cement Association (PCA) by Oesterle et al. (1976, 1979). The first group

consists of nine shear walls and the second group consists of seven specimens; the total

number of specimens are sixteen isolated shear walls. The research done by Oesterle et

al., (1976, 1979) focused on the determination of ductility and the energy absorption as

well as lateral load-deformation characteristics such as the flexure and shear responses to

develop new recommendations for designing structures in an area prone to earthquake

action. The cross-section shape of specimens consisted of three main shapes rectangular,

barbell, and flanged (see Figure 5.1). Both of the two investigations involved the setting

up and constructing of the sixteen cantilever shear walls and testing of the walls with

approximately 1/3 scale compared to the actual structural walls.

 45

Both of the two groups of specimens were designed based on ACI 318-71. The

monotonic load was applied on some specimens and cyclic lateral loads were applied to

the rest of the specimens to test the specimens and obtain the results of failure responses.

Ten specimens have been chosen out of sixteen to validate and assess the SFI-MVLEM

model in this study.

a) Rectangular b) Barbell c) Flanged

Figure 4.1. Isometric view of three wall cross-sections tested by Oesterle et al. (1976,
1979).

The dimensions of all walls were 180 in. height, 75 in. width, and 4 in. thickness.

The thickness of boundaries in walls with a barbell shape was 12 in, while the thickness

of boundaries in walls with a flanged shape was 36 in. along 4 in. All the specimens had

an aspect ratio equal to 2.4 due to the ratio of the height to the width. The range of the

maximum unconfined concrete strengths (f′c) varied between 3.17 ksi to 7.78 ksi and the

range of yield strengths of steel (fy) varied between 59.5 ksi to 74.2 ksi. Regarding the
 46

axial loads applied on the wall specimens, the values of axial loads varied in a range

between approximately zero to 0.134Agf’c. The names of chosen tested walls and

controlled parameters associated involved such as geometry dimensions of the walls with

their aspect ratios, walls cross-section shape, peak compressive strength of concrete, and

axial load can be illustrated in one table (see Table 4.1).

Table 4.1. Geometry and material properties of specimens tested by Oesterle et al. (1976,
1979)

ACI
Max. Load
Aspect P Confinement Design
ID Shape f'c (ksi)
Ratio Ag. f c Condition Flexure
Calc.(Kips) Obs.(Kip)
Shear
R1 Rectangular 2.4 6.48 0 Unconfined 0.22 29.1 26.6
R2 Rectangular 2.4 6.73 0 Confined 0.43 57.3 48.7
B1 Barbell 2.4 7.69 0 Unconfined 0.56 72.1 61
B2 Barbell 2.4 7.78 0 Unconfined 1.02 170.9 152.8
B3 Barbell 2.4 6.86 0 Confined 0.56 73.4 62
B6 Barbell 2.4 3.16 0.134 Confined 1.19 190.5 185.5
B7 Barbell 2.4 7.15 0.076 Confined 1.17 256.2 220.4
B8 Barbell 2.4 6.09 0.09 Confined 0.93 241.4 219.8
F1 Flanged 2.4 5.57 0 Unconfined 1.04 242.6 187.9
F2 Flanged 2.4 6.61 0.07 Confined 1.15 240.8 199.5

The details of steel reinforcement in both web and boundaries in the horizontal

and vertical directions can be shown in Table 4.2. The specimens have different features

in confinement, cross-section shape, and vary the range of reinforcement ratios for both

longitudinal and transverse (lateral) reinforcements in x-direction and y-direction (see

Figures 4.2- 4.4). The reason for these differentials is related to the investigation

purposes to develop new approaches for designing structural walls and understanding the

behavior of inelastic responses of walls.

 47

Table 4. 1: Steel reinforcement details in specimens tested by Oesterle et al. (1976, 1979)

Specimen
ID A B C D
R1 4#3 D3@4” 6mm@9” 6mm@8”
R2 6#4 [email protected]” 6mm@9” 6mm@8”
B1 8#4 D3@8” 6mm@9” 6mm@8”
B2 12#6 D3@8” 6mm@9” 6mm@4”
B3 8#4 [email protected]” 6mm@9” 6mm@8”
B6 12#6 [email protected]” 6mm@9” 6mm@4”
B7 12#6 [email protected]” 6mm@9” 6mm@4”
B8 12#6 [email protected]” 6mm@9” #3@4”
F1 28#4 [email protected]” 6mm@9” [email protected]”
20#4
F2 [email protected]” 6mm@9” 6mm@4”
12-6mm
A = Longitudinal reinforcement at boundaries in Y-direction shown in Figures 4.2-4.4.
B = Horizontal reinforcement at boundaries in X-direction shown in Figures 4.2-4.4.
C = Longitudinal reinforcement at the web in Y-direction shown in Figures 4.2-4.4.
D = Horizontal reinforcement at the web in X-direction shown in Figures 4.2-4.4.

The stress-strain relations of materials (concrete and steel) used in the specimens

tested by Oesterle et al. (1976, 1979) can be shown in Figures 4.5 and 4.6.

Figure 4.2. Boundary details of rectangular specimens tested by Oesterle et al. (1976).
 48

Figure 4.3. Boundary details of barbell specimens tested by Oesterle et al. (1976, 1979).

Figure 4.4. Boundary details of flanged specimens tested by Oesterle et al. (1976, 1979).
 49

Figure 4.5. Concrete stress-strain relation for specimens tested by Oesterle et al. (1976,
1979).

Figure 4.6. Steel stress-strain relation for specimens tested by Oesterle et al. (1976,
1979).
 50

The yield strength of steel reinforcement, as well as percent ratio of steel

reinforcements in web and boundaries in both horizontal and vertical directions, can be

illustrated in Table 4.3.

Table 4.2. Steel yield strength and reinforcement ratios for specimens tested by Oesterle
et al. (1976, 1979)
Yield Strength of Steel Reinf. Steel Reinforcement Ratio %
(1)
Specimen fy.Xw fy.Yw(1) fy.Xb(1) fy.Yb(1)
ρXw(2) ρXb(2) ρYw(2) ρYb(2)
ID ksi ksi ksi ksi
R1 75.7 66.0 75.7 74.2 0.31 0.31 0.25 1.47
R2 77.6 77.6 77.6 65.3 0.31 2.07 0.25 4.00
B1 75.5 68.7 75.5 65.2 0.31 0.31 0.29 1.11
B2 77.2 67.1 77.2 59.5 0.31 1.28 0.29 1.11
B3 69.4 69.4 69.4 63.5 0.63 0.63 0.29 3.67
B6 74.2 70.7 74.2 63.9 0.63 0.81 0.29 3.67
B7 71.0 71.0 71.0 66.4 0.63 1.35 0.29 3.67
B8 69.9 65.8 65.8 64.9 1.38 1.35 0.29 3.67
F1 76.2 76.2 76.2 63.5 0.71 0.3 0.3 3.89
F2 67.3 67.3 67.3 62.4 0.63 1.43 0.31 4.35
(1) fy.Xw, fy.Yw, fy.Xb, fy.Yb = Steel yield strength in horizontal and vertical directions
at the web of the wall and the boundary elements of the wall respectively.
(2) ρXw, ρYw, ρXb, ρYb = Steel reinforcement ratio in horizontal and vertical directions
at the web of the wall and the boundary elements of the wall respectively.

According to the confinement, the specimens were distributed into walls with

confinement in the boundaries with more hoops and ties along the first 72 in. from the

total height of wall specimen and walls without confinement in the boundaries, so that the

reinforcement of these walls included the ordinary hoops and ties along the total height of

the walls boundaries.

 51

Based on the nominal shear stress values, the behavior of the test walls can be

divided into two main groups: walls that were subjected to low nominal shear stresses

less than 3.1√f′c psi and walls that were subjected to high nominal shear stresses greater

than 7√f′c psi. Cracks and flexure failure mode were the most common features used to

distinguish between the walls subjected to high and low shear stresses. For walls with

low nominal shear stresses, the cracks occurred in the horizontal direction at the

boundaries where the transverse reinforcement levels located. The cracks continued in

inclined directions with bigger spaces. On the other hand, for the walls with high nominal

shear stresses, the cracks developed in boundaries had the same behavior as walls with

low nominal shear stresses, while the cracks in the web were steeper and with smaller

spaces than those in walls with low nominal shear stresses. The range of the average

shear stresses on the wall specimens varied between 1.4 √f′c psi to 13.8 √f′c psi. The

lateral deformation caused by the shear and the total lateral deformation are contributed

to each other.

Specimens R1, B1, and B3 were failed by inelastic buckling of reinforcement.

The longitudinal reinforcements were buckled at the first 36 in. of the boundaries. The

failure mode of the specimen R2 was the instability of the compression zone. Large

displacement in an out-of-plan direction at the first 36 in. of the wall led to instability of

the wall specimen and inability for carrying the loads. Web crushing was the failure

mode that occurred for specimens B2, B6, B7, B8, F1, and F2. Web crushing is

considered one of the most common failure modes that occurs in beams and walls,

especially those that have “large flanges and relatively thin web subjected to high shear

stress" (Oesterle, 1986). According to Oesterle (1986), the strength of the wall to resist
 52

the web crushing depends on lateral deformation, the strength of concrete, and axial and

lateral loads. Also, the effect of deformation history can reduce the compressive strength

of the web of RC wall and when the reduced compressive strength becomes less than the

stresses applied, the web crushing will occur suddenly.

4.2 Thomsen and Wallace (1995)

A study has been conducted by Thomsen and Wallace (1995) to show how the

rectangular and T-shaped cross-section shear walls are behaving. The selection of these

walls depends on a high-rise prototype building used as an office building located in an

earthquake area action. In this study, one rectangular specimen has been chosen to

validate the SFI-MVLEM model which is RW1 specimen. The scale of the specimen

equal to one-fourth the actual shear wall designed in the prototype building. The axial

load which is equal to 90 kips (0.10 Ag f’c) and lateral load were applied using special

apparatus for transforming the loads efficiently to the wall (see Figure 4.7). The

dimensions of the specimen wall are 144 in. in height, 48 in. in width, and 4 in. in

thickness (see Figure 4.8). The design procedure of the walls depended on the

requirements of the Uniform Building Code (1994); however, Thomsen and Wallace

depend on the displacement based design approach which was proposed by Wallace

(1994, 1995) to evaluate the flexural and shear strength of specimens, and reinforcement

details in the boundary elements. Confined reinforcement at boundary elements placed at

3 in. along 48 in. from the base, and longitudinal steel bars comprised of eight #3 bars,

while web reinforcement comprised of two curtains of #2 bars spaced 7.5 in. (see Figure

4.9).
 53

Hydraulic Jacks
Hydraulic Actuator
Pancake
Load Cells
B.T.B. Channels 305 mm
12 in.

Load Cell Clip Angles

High-Strength
Post-Tensioning
Cables

144
3.66 m in.

Rectangular
Wall Specimen
Reaction Wall
1.22 m

Pedestal

Cable Anchored Pedestal

Strong Floor Here Tie-Downs 27mm
686 in.

Figure 4.7. Schematic of test setup (Thomsen and Wallace, 1995)

48 in.
1.22 m
FOURTH FLOOR

8 @ 3.7 in.

THIRD FLOOR

B B

SECOND FLOOR
8 @ 7.5 in.
3.66 m
144 in.

4.5 in.
FIRST FLOOR

16 @ 3 in.
A A
GROUND LEVEL

10 to16
0.69 m in. 27 in.

76 in.
1.93 m

Figure 4.8. Profile view of RW1 specimen with dimensions and reinforcement
distribution.
 54

Figure 4.9. Cross sections of specimen RW1 (Thomsen and Wallace, 1995).

The test results of the actual material strength (the concrete compressive strength

and the steel yield strength) were shown in Figures 4.9 and 4.10; however, typical

material properties were used in design which are 4.00 ksi and 60.0 ksi for concrete and

steel respectively. According to the measured test results of standard cylinder specimens

with dimensions 6X12 inches, the peak compressive strength results of concrete in the

first, second, third, and fourth level floor were 4.580 ksi, 5.871 ksi, 5.632 ksi, and 8.462

ksi respectively at the testing date, so the average value, which is equal to 6.136 ksi, was

taken for analysis purposes. The strain value at the peak compressive stress was 0.0021.

Yield stresses for the three bars used in the construction of the wall were 63 ksi, 65 ksi,

and 63 ksi for steel bars #3, #2, and 3/16 in. respectively. The yield strain values were

0.0022, 0.003, and 0.003 for steel bars #3, #2, and 3/16 in. respectively.
 55

Figure 4.10. Concrete stress-strain relations for specimen RW1.

Figure 4.11. Steel stress-strain relations for specimen RW1.

4.3 Dazio et al. (1999)
 56

A study has been conducted by Alessandro Dazio, Wenk, and Bachmann (1999)

using six RC wall specimens exposed to static cyclic loading based on building a

prototype with one-half scale. The main purpose of this study was to investigate the

behavior of structural wall responses in the high seismicity regions, specifically in the

Europe countries. In this study, one specimen has been chosen to validate the SFI-

MVLEM model which is the WSH6 specimen. The dimensions of the specimen are 178

in. tall, 78.75 in. wide, and 6 in. thick. It should be recalled that the specimen was

subjected to an axial load equal to 0.11Ag f’c. At the boundary elements, longitudinal

reinforcement ratio was 1.54% by using 6-12mm bars and 2-8mm bars, while transverse

(lateral) reinforcement ratio was 0.754% by the using 6mm and 4.2 mm bars with 50mm

spacing from center to center of bars. At the web, the longitudinal reinforcement ratio

was 0.54% by the use of 18-8mm, whereas lateral reinforcement ratio was 0.25% by the

use of two curtains of 6mm bars with 150mm spacing (see Figure 4.11).

Figure 4.12. Plan view for WSH6 Specimen, all dimensions in mm (1 in = 25.4 mm).

The actual material test results (stress-strain relations) for specimen WSH6 can be

seen in Figures 4.12 to 4.15 for concrete and steel bars.

 57

Figure 4.13 Concrete stress-strain relations for WSH6 specimen.

Figure 4.14. Steel stress-strain relations for 6 mm rebar used for WSH6 specimen.
 58

Figure 4.15. Steel stress-strain relations for 8 mm rebar used for WSH6 specimen.

Figure 4.16. Steel stress-strain relations for 12 mm rebar used for WSH6 specimen.
 59

CHAPTER 5

OPENSEES MODELING OF RC WALLS

This chapter provides detailed information about the procedure used to perform

discretization of wall geometry, material calibration, calibration of shear resisting

mechanisms, and model generation. The calculations and other parameters will be

described step by step so that methodology can be clear and straightforward. A summary

about the experimental specimen is provided at the beginning to describe the properties

and take an overview about the experimental specimen. The OpenSees code of the shear

wall specimen B8 is provided in Appendix A.Specimen Description

In order to demonstrate the steps of applying input parameters of experimental

specimens into OpenSees, B8 specimen (Oesterle et al., 1979) was chosen to describe

how the shear walls have been modeled into OpenSees by the use of SFI-MVLEM

model. With Barbell shape, the height of the B8 shear wall is (180 in.) and its width is

(75 in.), while the thickness equals to (12 in.) at the boundaries and (4 in.) at the web (see

Figure 5.1). Vertical and horizontal reinforcing ratio for boundary elements and web were

illustrated in Table 4.1, and confinement reinforcement has been used in first (6 ft.) of

boundaries with spacing equals to (1.33 in) and (8 in) spacing at the rest of boundaries

height. During the time of the test, the specimen B8 was loaded axially with

approximately constant stress equal to 545 psi (i.e., 0.09 A g f’c). The properties of

materials used in the B8 specimen tested by Oesterle et al. (1979) can be illustrated in

Table 5.1.
 60

Table 5.1. Material Properties of B8 Specimen; Oesterle et al. (1979)

Concrete
Compressive strength Modulus of rupture Modulus of Elasticity
f'c at 77 days fr Ec
(ksi) (ksi) (ksi)
6.085 0.614 3900
Steel
Bar Size Properties
fy (ksi) 65.8
6mm fsu (ksi) 89.3
Es (103 ksi) 28.2
fy (ksi) 69.9
No. 3 fsu (ksi) 105.6
3
Es (10 ksi) 25
fy (ksi) 64.9
No. 6 fsu (ksi) 108.2
Es (103 ksi) 27.5

The reinforcement ratio of vertical bars for boundaries was calculated by dividing

the total area of 12#6 (i.e., As = 5.28 in2) over the total concrete area of boundary element

(i.e., Ag = 144 in2), while the vertical reinforcement ratio for web was calculated by

dividing the total area of vertical bars in web region 12 bars with 6mm size (i.e.,

As = 0.526 in2) over the total concrete area of the web (i.e., A g = 204 in2), and same

concept for other directions.

 61

8″

180″

(a)

24″

(b)
Figure 5.1. (a) cross-section and reinforcement distribution (b) Dimensions of B8 specimen;
Oesterle et al (1979).

5.2 Model Geometry Discretization

The B8 cantilever shear wall was divided horizontally along the width of the

shear wall into n = 9 stacks as was aforementioned in Chapter 3. Each stack represents

one element of the SFI-MVLEM model and each SFI-MVLEM element was divided

vertically into m = 5 of the RC panel fibers based on the boundaries, dimensions, and

steel reinforcement distribution of web zone along the shear wall width. SFI-MVLEM

elements were separated by nodes (No. of nodes =10). The number of SFI-MVLEM

elements at the bottom region of the wall (plastic hinge region) was taken intentionally

greater than those at the top region of the wall for capturing significant features of failure

response for the wall and obtaining more accurate results in plastic hinge region (usually

between 1/3 to 1/4 of the wall width from the fixed edge, Tran (2012) says various failure

modes (crushing of concrete, buckling failure in reinforcement of boundaries, and shear

failure) were predicted to occur (see Figure 5.2). Boundary conditions of structural shear
 62

wall specimen were taken into consideration. Due to the fact that the specimen is an

isolated cantilever shear wall, the B8 specimen was assigned to be fixed at the bottom of

the wall in contact with the concrete pedestal; therefore, the first node of the shear wall

was assigned to be fixed (see Appendix A). The No. of the degree of freedom (DOF) and

the No. of element dimensions for analysis purposes were assigned to be 3 and 2

respectively.

m=1 2 3 4 5

Boundary Web Boundary

Figure 5.2. Discretization and geometry of B8 specimen (a) Vertical (b) Horizontal.

5.3 Calibration of Material Models

As aforementioned in Chapter 4, SteelMPF and ConcreteCM are the uniaxial

material models for both steel and concrete that have been added to the OpenSees

platform recently as a part of the SFI-MVLEM analytical model. The input parameters

for both SteelMPF, which is based on the Menegotto and Pinto (1973) and Filippou et al.
 63

(1983) model, and ConcreteCM which is based on the Chang and Mander (1994) model

were calibrated and validated, so they can be used in the SFI-MVLEM model. Detailed

information about the calibration procedure for concrete and steel is presented in the

following sections.

5.3.1 Steel Reinforcement

Steel reinforcement parameters have been calibrated to represent the actual

properties of experimental reinforcement bars used in both horizontal and vertical

directions. Modulus of elasticity for all specimen was used as (Es = 29000 ksi) and

applied to all reinforcing bars in both directions. The steel bars with diameters; D6

(db = 6mm; 0.24 in), #3 (db = 9.53 mm; 0.375 in), and #6 (db = 19.05 mm; 0.75 in)

reinforcing bars have been used in B8 specimen in both web and boundary elements for

longitudinal and transverse (lateral) reinforcements. The general calibration method of all

the steel bars that have been used in all specimens was calibrated depending on the values

of the yield strength of steel (fy) and the strain hardening ratio (b) to fit the experimental

data obtained from the stress-strain relationship of tensile tests. Due to the lack of

information of strain-hardening parts in stress-strain relation for specimens tested by

Oesterle et al (1976, 1979) (see Figure 4.6), the strain-hardening ratio for specimen B8

was calculated based on the Ramberg-Osgood (R-O) equation for strain-hardening

suggested by Aktan, Karlsson, and Sozen (1973) as shown in equation (5.1):

= + for ε > 4.2447 εy (5.1)

ε

Where:

σ = stress (psi), ε = strain, E=29,000,000 psi

 64

σ0 = 0.7 of yield stress of the bar (psi),

σ
ε0 = , εy= yield strain of the bar.

m = 4.3 as recommended by Aktan et al. (1973).

More details about the strain hardening ratios can be illustrated in Table 5.2 which

presents the calculated values of strain hardening ratios for the steel bars used in

specimen B8 tested by Oesterle et al (1979).

Table 5.2. calculated strain hardening ratios for bars used in specimens B8

Steel Bars

Properties No.6 6mm No.3

Yield stress (σy) psi 64900 65800 69900

0.7*yeild stress (σ0) psi 45430 46060 48930

ε0 0.0016 0.0016 0.0017

Ultimate stress (σu) psi 108200 89300 105300

Strain at Ultimate stress (εu) 0.0691 0.0305 0.0492

Strain hardening ratio (b) 0.0155 0.0197 0.018

Cyclic degradation of reinforcing steel and the Bauschinger effect have been

taken into consideration by using the parameters R0, a1, and a2 in which R represents the

curvature coefficient that was explained in chapter 3. Values recommended by Menegotto

and Pinto (1973) for these parameters are used to define steel cyclic properties of

constitutive material (SteelMPF) in SFI-MVLEM model; these values are R 0, a1, and a2

equal to 20, 18.5, and 0.15 or 0.0015 respectively, and those parameters are assumed to

be constants for the steel material of all wall specimens modeled by SFI-MVLEM model

in this study (see Appendix A).

 65

5.3.2 Concrete

Stress-strain curve of concrete materials in compression and tension can be

controlled by some associated parameters such as concrete modulus of elasticity (E c),

concrete peak stress (f'c), compressive strain at peak stress (ε'c), and the parameter (r) (see

Figure 5.3). The compressive strain at peak stress for unconfined concrete was calibrated

to fit the actual value by matching the experimental test results of concrete with the

model results (see Figure 5.4), while the value of the compressive strain at peak stress for

confined concrete (εcc) was calibrated by matching the data obtained from the Saatcioglu

and Razvi (1992) model for confined and unconfined concrete (ε 1) with the data obtained

from the model (see Table 5.3; Figure 5.5). The shape of stress-strain curve before the

peak (pre-peak) for compression can be defined by the parameter (r) which is compared

with and matched the envelope curve of the Saatcioglu and Razvi (1992) model for both

confined and unconfined concrete. The slope of stress-strain curve after the peak (post-

peak) can be defined by the parameter (εcr) (see Figure 5.3).

Ascending Descending
Compression ( 'c ,f 'c)
Branch Branch

Compression or ( 'c ,f 'c) or (t , ft )

Tension Envelope  (c0,0)
Tension (c0+ t,ft)


f
Stress,

r Linear
Branch
1
Ec
Esec 1

Ec
1
(pl,0) Epl (cr,0)
1

Strain, 
Figure 5.3. Concrete constitutive model and associated parameters.
 66

Figure 5.4. Calibration of concrete envelope curve in compression.

 67

Table 5.3. Input parameters of the Saatcioglu and Razvi (1992) model for B8 specimen

Specimen: B8
Input data Output data
fc' (ksi) = 6.085 Flx= 0.81
fy(ksi) = 65.8 Fly= 0.81
bcx (center to center of ties) (in) = 10.75 k2x= 0.53
bcy (center to center of ties) (in) = 10.75 k2y= 0.53
diameter of steel bar (in) = 0.25 flex= 0.43
Abx of ties (in2) = 0.19625 fley= 0.43
Aby of ties (in2) = 0.19625 fle= 0.43
s (in) = 1.333 k1 = 5.58
slx (in) = 3.75 K= 0.39
sly (in) = 3.75 f'cc= 8.47
α= 90 ε085= 0.0038
ε0 = 0.0031 ρ= 0.0137
ε1 = 0.009168 ε85= 0.0364

The parameter (εcr) represents the strain where the envelope curve takes straight

line shape until the spalling strain (εsp) where compressive stress equals to zero. The

parameter (εcr) was calibrated by matching the compression post-peak slope of the model

proposed by Saatcioglu and Razvi (1992) for unconfined and confined concrete (see

Figure 5.5).

Regarding the concrete stress-strain curve in tension, the Chang and Mander

(1994) model was used for analytical simulation of the concrete behavior of specimens in

tension; this model was calibrated by the use of the Belarbi and Hsu (1994) model. As

recommended by Belarbi and Hsu (1994), the peak tension strength of concrete (f t) in

each specimen was determined depending on peak compressive strength of concrete as

shown in equation (5.2):

 68

ft = 3.733 f c Or ft = 0.31 f c (5.2)

The value of strain at peak tensile stress (εt) was assigned to be 0.00008, whereas

the parameter (r) for tension in concrete was assigned to be 1.2. Both of (ε t) and (r)

parameters define the tension envelope shape of concrete and their values were chosen

precisely to reasonably match the average post-crack stress-strain relation suggested by

Belarbi and Hsu (1994) (see Figure 5.6).

Figure 5.5. Confined and unconfined concrete calibration with Saatcioglu and Razvi
(1992).
 69

Figure 5.6. Calibration of stress-strain relationship for concrete in tension; Kolozvari

(2013).

5.4 Shear Resistance Mechanism Parameters

As was mentioned in Chapter 3, two parameters have been incorporated in the

SFI-MVLEM model in order to account for shear resistance due to aggregate interlock

and shear resistance due to the dowel action of reinforcing steel. Both the aggregate

interlock parameter, which is associated with friction coefficient (η) and the dowel action

parameter, which is associated with stiffness of dowel reinforcement (α), were

incorporated in OpenSees’s coding using the names nu and alfadow respectively (see

Figure 5.7). Based on the comparison between the results of the SFI-MVLEM model and

experimental results of shear walls, Kolozvari (2013) recommended a specific range for

both of the parameters for obtaining reasonably accurate results; therefore, a loop for

these interval range values has been made to collect all results and distinguish the best

result.
 70

Figure 5.7. Shear resisting mechanism and associated parameters.

The suggested interval values for (nu) and (alfadow) used in the loop of analytical

model give deferent results varying with the change of the values of the shear resistant

parameters (see Figure 5.8). Best results obtained from an analytical model that match the

experimental results have been chosen and presented to compare with the experimental

results of shear distortion-lateral cyclic load (see Figure 5.9).

Figure 5.8. The effect of shear resistance parameters on predicted results.

 71

Figure 5.9. Choosing the best-predicted results that match experimental results.

5.5 Model generation

After the uniaxial materials were calibrated and defined separately, the uniaxial

constitutive materials added together to define the RC panel (see Figure 5.10). As was

aforementioned in Chapter 3, the Fixed-Strut-Angle-Model (FSAM) was modified by

using the RC panel instead of the uniaxial macro-fiber elements in order to obtain more

accurate results with the shear response particularly and other failure responses. The step

that comes after defining the FSAM model is defining the SFI-MVLEM elements by

adding the RC panels of web and boundaries to each other (i.e., assigning FSAM to SFI-

MVLEM). The final step of generating the model is by linking the SFI-MVLEM

elements by nodes to construct the structural wall by macro elements (see Figure 5.11).
 72

Figure 5.10. Defining constitutive materials of reinforced concrete panel.

It is important to mention here that the cyclic lateral load was applied based on

displacement control analysis (see Figure 5.12), and the axial load was applied as a point

load with a constant value. The solution algorithm that was used in this analysis was a

Newton-Raphson algorithm (first Newton’s iteration with current stiffness, then

Newton’s iteration with initial stiffness), more details about that can be reviewed in the

OpenSees wiki.

ConcreteCM
( 'c , f 'c )

Compression
y
FSAM


xy
O (0, 0)
Tension
Not to scale
(0+ t , ft)


SteelMPF
y E1= bE0

(b)
E0
O y

(c) SFI-MVLEM

(a) 
Figure 5.11. SFI-MVLEM elements construct RC wall (a) constitutive materials, (b)
FSAM, (c) RC wall.
 73

Figure 5.12. Displacement increments technique for applying cyclic lateral loading.
 74

CHAPTER 6

ANALYTICAL RESULT AND COMPARISON

WITH EXPERIMENTAL RESULTS

This chapter discusses detailed information on a comparison between

experimental test results and analytical results predicted by the use of SFI-MVLEM

model. Three different programs for experimental results of shear walls were discussed in

Chapter 4, Oesterle et al. (1976, 1979), Thomsen and Wallace (1995), and Dazio et al.

(1999). The main purpose of presenting the comparison is to assess the SFI-MVLEM

model ability to predict different responses characteristics especially the total top lateral

displacement versus lateral load response and shear distortion caused by the lateral shear

deformation versus lateral load response. The discussion includes the wall lateral

stiffness, capacities under cyclic loading, the cyclic degradation for each specimen, and

the effect of shear parameters of SFI-MVLEM values on the shear prediction.

6.1 Oesterle et al. (1976, 1979)

An experimental program by Oesterle et al. (1976, 1979) included specimens with

rectangular, barbell, and flanged cross-sections. Rectangular specimens considered from

the experimental program are R1 and R2. Barbell specimens are B1, B2, B3, B6, B7, and

B8. Flanged specimens are F1 and F2. The following section presents analytical and

experimental result comparison for each of the cross-section types.

 75

6.1.1 Rectangular Cross-Section specimens

Figure 6.1 shows the comparison between the predicted and the experimental

results of lateral load versus top total displacement response for the rectangular shear

wall specimen R1. The SFI-MVLEM analytical model succeeded in predicting the

capacities reasonably well for all drift levels in the positive loading side, while the model

missed the capacities in the first four drift levels by approximately 10% to 20% less than

measured capacities, but successfully captures the capacity in the fifth drift level in the

negative loading side. In terms of cyclic characteristics, which are comprising lateral

stiffness degradation, hysteretic shape, and pinching behavior, the model represented

accurate predictions for lateral stiffness in the positive and negative loading sides except

for the zone of the pinching.

Figure 6.1. Top total lateral load-displacements plot for specimen R1.
 76

Figure 6.2 shows the comparison of predicted and measured results of shear

distortion caused by the lateral shear deformation at approximately 6 ft. from the base of

the R1 shear wall specimen versus lateral load response. It can be observed from the

Figure 6.2 that the capacities captured by the analytical model underestimated the actual

capacities by approximately 8% in the negative loading side and approximately captured

all the capacities in the positive loading side except the last capacity, which was

overestimated by approximately 7%. Regarding the shear stiffness, the SFI-MVLEM

model could not capture the measured shear stiffness successfully. It can be clearly noted

that the predicted results pinched sharply in the pinching zone. The shear displacements,

in positive loading side, were underestimated by the analytical model in general but were

overestimated in the negative loading side.

Figure 6.2. Lateral load-shear distortion plot for specimen R1.

 77

Figure 6.3 shows the comparison between the predicted and the measured lateral

load versus top total displacement response for the rectangular shear wall specimen R2.

Although the SFI-MVLEM analytical model underestimated almost all the capacities of

the R2 specimen in the cyclic loading history by approximately 10% to 20% compared to

the measured results in the positive and negative loading sides, the analytical model

captured the overall hysteretic shape. Regarding the other cyclic characteristics including

lateral stiffness degradation, and pinching behavior, the model showed good ability to

predict the lateral stiffness of the wall in both the positive and the negative loading sides

except the location of the pinching zone because the model failed to capture the pinching

behavior in all specimens that were not applied to axial force during the test of the

specimens.

Figure 6.3. Top total lateral load-displacements plot for specimen R2.
 78

Figure 6.4 shows the comparison between predicted and measured results of shear

distortion caused by the lateral shear deformation at approximately 6 ft. from the base of

the R2 shear wall specimen versus lateral load response. By taking a look at Figure 6.4,

the underestimation of the analytical model for the capacities located in the first three

drift levels can be noted in the positive loading side, and the last capacity was

overestimated by approximately 10%. In the negative loading side, all capacities were

underestimated by the analytical model by approximately 10% to 25% compared to the

measured capacities. The shear stiffness was not captured by the model for all the cycles.

It can be clearly noted that the predicted results pinched sharply in the pinching zone. The

shear displacements, in both positive and negative loading sides were predicted by the

analytical model in general.

Figure 6.4. Lateral load-shear distortion plot for specimen R2.

 79

6.1.2 Barbell Cross-Section Specimens

Figure 6.5 shows the comparison between the predicted and the experimental

results of lateral load versus top total displacement response for the barbell shear wall

specimen B1. The analytical model succeeded in predicting the capacities reasonably

well for all drift levels in the positive loading side, while the model underestimated the

capacities for drift levels between 0% and 2.0% of drift ratios by approximately 7% to

10% of measured capacities, but successfully captured the capacity at 2.5 drift ratio in the

negative loading side. By taking a look at the cyclic characteristics, which comprise

lateral stiffness degradation, hysteretic shape, and pinching behavior, the model

represented accurate prediction for lateral stiffness in the positive and negative loading

sides except the pinching zone because the model missed the pinching behavior.

Figure 6.5. Top total lateral load-displacements plot for specimen B1.
 80

Figure 6.6 shows the comparison of predicted and measured results of shear

distortion caused by the lateral shear deformation at approximately 6 ft. from the base of

the B1 shear wall specimen versus lateral load response. It can be observed from the

Figure 6.6 that the capacities captured by the analytical model overestimated the actual

capacities by approximately 8% in the positive loading side and approximately captured

all the capacities in the negative loading side except the last capacity which was

overestimated by approximately 7%. Regarding the shear stiffness, the SFI-MVLEM

model could not capture the measured shear stiffness successfully. It can be clearly noted

that the predicted results pinched sharply in the pinching zone. The shear displacements,

in both positive and negative loading sides, generally were underestimated by the

analytical model.

Figure 6.6. Lateral load-shear distortion plot for specimen B1.

 81

Figure 6.7 shows the comparison between the predicted and the experimental

results of lateral load versus top total displacement response for the barbell shear wall

specimen B2. The analytical model succeeded in predicting the capacities reasonably

well for all drift levels in the positive loading side except the last capacity due to the

failure mechanism of the actual test results. On the other hand, the model underestimated

the capacities in the range between 1 to 2.5% of drift ratio by approximately 3% to 5%,

but successfully captured the other capacities in the other drift level in the negative

loading side. By comparing the cyclic characteristics, which comprise lateral stiffness

degradation, hysteretic shape, and pinching behavior, with the predicted results, the

model showed high ability to predict the lateral stiffness in the positive and negative

loading sides except the pinching zone because the model missed the pinching behavior.

Figure 6.7. Top total lateral load-displacements plot for specimen B2.
 82

Figure 6.8 shows the comparison of predicted and measured results of shear

distortion caused by the lateral shear deformation at approximately 6 ft. from the base of

the B2 shear wall specimen versus lateral load response. It can be observed from the

Figure 6.8 that the model predicted the capacities accurately excluding the last capacity in

the positive loading side and missed the first two capacities in the negative loading side

by approximately 5% to 10% compared to the measured capacities, but the model

captured the rest of capacities reasonably in the negative side of loading. Regarding the

shear stiffness, the model could not capture the measured shear stiffness successfully; it

can be noted that the predicted results pinched sharply in the pinching zone as well. The

shear displacements, in the negative loading side, were approximately captured, but the

model underestimated the shear displacements in the positive loading side.

Figure 6..8. Lateral load-shear distortion plot for specimen B2.

 83

Figure 6.9 shows the comparison between the analytically predicted and the

experimentally measured lateral load versus top total displacement response for the

barbell shear wall specimen B3. The SFI-MVLEM analytical model missed almost all the

capacities in the positive and the negative loading sides for all drift levels. Compared to

the experimental capacities, the model underestimated the capacities located in between

0% to 2% drift ratio of approximately 20% to 25% in both positive and negative loading

sides and approximately 6% to 8% for the other levels of drift. The analytical model

failed to capture the lateral stiffness of the shear wall accurately, but it captured the

overall hysteretic shape. In the same content, the model failed to capture the pinching

behavior.

Figure 6.9. Top total lateral load-displacements plot for specimen B3.
 84

Figure 6.10 shows the comparison of predicted and measured results of shear

distortion at approximately 6 ft. from the base of the B3 shear wall specimen versus

lateral load response. It is clear that the analytical model failed to capture the shear

displacements in both of the negative and positive loading sides, but by taking a look to

the capacities that were captured by the analytical model and comparing them with the

experimental capacities, it can be observed that the model captured the shear capacities

approximately. Regarding the shear stiffness, the SFI-MVLEM model could not capture

the measured shear stiffness successfully. Also, predicted results pinched sharply in the

pinching zone. The causes that led the analytical model to fail in predicting the

experimental results reasonably were related to the failure in convergence when using

small values for shear mechanism parameters (α and nu); see Section 3.3 in Chapter 3.

Figure 6.100. Lateral load-shear distortion plot for specimen B3.

 85

Figure 6.11 shows the comparison between the predicted and test results of lateral

load versus top total displacement response for the barbell shear wall specimen B6. The

SFI-MVLEM analytical model succeeded in predicting the capacities accurately for all

drift levels in the positive loading side as well as in the negative loading side, but the

model could not capture the capacity at the drift ratio 0.5% on both negative and positive

loading sides. By taking a look at the cyclic characteristics, which are comprising lateral

stiffness degradation, hysteretic shape, and pinching behavior, the model showed good

ability to predict lateral stiffness in the positive and negative loading sides except the

zone of the pinching because the model could not capture the pinching behavior for all

specimens that have not been applied for axial loads. In the same content, the model

failed in simulate the cyclic degradation of capacities.

Figure 6.11. Top total lateral load-displacements plot for specimen B6.
 86

Figure 6.12 shows the comparison of predicted and measured results of shear

distortion at approximately 6 ft. from the base of the B6 wall specimen versus lateral load

response. It can be observed from Figure 6.8 that the model overestimated most the shear

capacities by approximately 4% to 8% comparing to the measured capacities in the

positive and the negative sides of loading. Regarding the shear stiffness, the model could

not capture the measured shear stiffness successfully; however, in some drift levels, the

model succeeded to predict the lateral stiffness of the shear wall in the loading part of the

drift cycles and failed in the reloading part of the same drift cycles. It can be noted that

the predicted results pinched sharply in the pinching zone in contrast with the pinching

zone behavior of total displacement versus lateral loads plot. The shear displacements in

the negative and the positive loading sides were approximately captured.

Figure 6.12. Lateral load-shear distortion plot for specimen B6.

 87

Figure 6.13 shows the comparison between the predicted and the measured lateral

load versus top total displacement results for the barbell shear wall specimen B7. By

comparing the analytical and measured capacities, it can be realized that the model

underestimated the capacities by approximately 10% at the first two drift levels in

between 0% and 1.5% drift ratio of the positive loading side, while the model succeeded

in capturing the rest of the capacities in the other drift levels. The model shows nearly the

same behavior for the negative loading side. By taking a look at the cyclic characteristics,

which comprise lateral stiffness degradation, hysteretic shape, and pinching behavior, the

model failed to capture the lateral stiffness of the shear wall in the positive and negative

loading sides, but the model showed the overall hysteretic shape, and the model succeeds

in capturing the pinching behavior in all cycles except the last cycle.

Figure 6.13. Top total lateral load-displacements plot for specimen B7.
 88

Figure 6.14 shows the comparison of predicted and measured results of shear

distortion at approximately 6 ft. from the base of the B7 wall specimen versus lateral load

response. It can be clearly observed from Figure 6.14 that the model succeeded in

predicting all the capacities accurately excluding the last capacity in the positive loading

side in which the different in predicted capacities approximately 5% to 10% greater than

the measured capacities. The same behavior was noted in the negative loading side, but

the model underestimated the last capacity by approximately 10% to 15%. For the shear

stiffness, the model could not capture the measured shear stiffness successfully. In the

same content, it can be noted that the predicted results pinched sharply in the pinching

zone. The shear displacements, in the positive loading side, was approximately captured,

while the model underestimated the shear displacements in the negative loading side.

Figure 6.14. Lateral load-shear distortion plot for specimen B7.

 89

Figure 6.15 shows the comparison between the predicted and test results of lateral

load versus top total displacement response for the rectangular shear wall specimen B8.

By taking a look at the positive loading side, the SFI-MVLEM analytical model

underestimated the capacities located in between 0.5% and 1.5% drift ratio; however, the

model succeeded in capturing the rest of the capacities reasonably well. In contrast, the

model succeeded in capturing the capacities located in between 0% and 1% drift ratio but

underestimated the rest of capacities in the negative loading side by approximately 10%

to 13% comparing to measured capacities. Regarding the cyclic characteristics, which

comprise lateral stiffness degradation, hysteretic shape, and pinching behavior, the model

represented accurate prediction for lateral stiffness in the positive loading side and

showed the well hysteretic shape and pinching behavior.

Figure 6.15. Top total lateral load-displacements plot for specimen B8.
 90

Figure 6.16 shows the comparison of analytically predicted and experimentally

measured results of shear distortion caused by the lateral shear deformation at

approximately 6 ft. from the base of the B8 shear wall specimen versus lateral load

response. It can be observed from the Figure 6.16 that the model succeeded in predicting

the shear capacities accurately, excluding the last capacity in the negative loading side

and underestimated the capacities in the positive loading side by approximately 5% to

10% comparing to the measured capacities. Regarding the shear stiffness, the analytical

model failed to capture the measured shear stiffness successfully, in the same content, it

can be clearly noted that the predicted results pinched sharply in the pinching zone as

well. The shear displacements in the positive loading side were approximately captured

as well as in the negative loading side.

Figure 6.16. Lateral load-shear distortion plot for specimen B8.

 91

6.1.3 Flanged Cross-Section Specimens

Figure 6.17 shows the comparison between the predicted and the experimental

lateral load versus top total displacement results for the flanged shear wall specimen F1.

Compared to the measured results, the SFI-MVLEM analytical model succeeded in

predicting the overall capacities reasonably well for all drift levels in the positive loading

side, while the model missed the capacity in the last drift level of the negative loading

side due to the failure mechanism in which the model cannot predict. Concerning the

cyclic characteristics, which comprise lateral stiffness degradation, hysteretic shape, and

pinching behavior, the model represented accurate prediction for lateral stiffness in the

positive and negative loading sides except for the zone of the pinching because the model

missed the pinching behavior, but it captured the general hysteretic shape.

Figure 6.17. Top total lateral load-displacements plot for specimen F1.
 92

Figure 6.18 shows the comparison of predicted and measured results of shear

distortion caused by the lateral shear deformation at approximately 6 ft. from the base of

the F1 shear wall specimen versus lateral load response. It can be observed from Figure

6.8 that the model predicted the capacities accurately excluding the last capacity in the

positive loading side and missed the first two capacities in the negative loading side by

approximately 5% to 10% compared to the measured capacities, but the model captured

the rest of capacities reasonably in the negative side of loading. Regarding the shear

stiffness, the model could not capture the measured shear stiffness successfully; it can be

clearly noted that the predicted results pinched sharply in the pinching zone as well. The

shear displacements, in the negative loading side, was approximately captured, while the

model underestimated the shear displacements in the positive loading side.

Figure 6.18. Lateral load-shear distortion plot for specimen F1.

 93

Figure 6.19 shows the comparison between the analytically predicted and the

experimentally measured results of lateral load versus top total displacement response for

the flanged shear wall specimen F2. The SFI-MVLEM analytical model underestimated

all the capacities for all drift levels in the positive loading side by approximately 10% to

17% compared to the measured capacities, and the model underestimated the capacities

in the negative loading side by approximately 5% to 10%. Regarding the cyclic loading

properties, which comprise lateral stiffness degradation, hysteretic shape, and pinching

behavior, the model represented good simulation for the lateral stiffness in the positive

and negative loading sides generally. Even though the analytical model could not capture

the pinching behavior accurately, it succeeded in capturing the overall hysteretic shape.

Figure 6.19. Top total lateral load-displacements plot for specimen F2.
 94

Figure 6.20 shows the comparison between analytically predicted and

experimentally measured results of shear distortion caused by the lateral shear

deformation at approximately 6 ft. from the base of the F2 shear wall specimen versus

lateral load response. It can be observed from Figure 6.8 that the model successfully

predicted almost all the capacities accurately excluding the last capacity in the positive

loading side due to the failure mechanism issues that cannot be predicted by the model.

Regarding the shear stiffness, the SFI-MVLEM model could not capture the measured

shear stiffness successfully. It can be clearly noted that the predicted results pinched

sharply in the pinching zone as well. The shear displacements, in the negative loading

side, was approximately captured, while the model underestimated the shear

displacements in the positive loading side.

Figure 6.20. Lateral load-shear distortion plot for specimen F2.

 95

6.2 Thomsen and Wallace (1995)

Figure 6.21 shows the comparison between the analytically predicted and the

experimentally measured results of the lateral load versus the top total displacement

response for the rectangular shear wall specimen RW1. The SFI-MVLEM analytical

model succeeded in predicting the capacities reasonably well for all drift levels in the

positive loading side, and all the capacities in the other drift level in the negative loading

side. Comparing between the cyclic characteristics predicted by the analytical model and

measured cyclic characteristics, which comprise lateral stiffness degradation, hysteretic

shape, and pinching behavior. The model showed accurate prediction for lateral stiffness

especially in positive loading side, and captured the hysteretic shape with good

simulation for pinching behavior.

Figure 6.21. Top total lateral load-displacements plot for specimen RW1.
 96

Figure 6.22 shows the comparison of analytically predicted and experimentally

measured results of shear distortion caused by the lateral shear deformation at

approximately 6 ft. from the base of the RW1 shear wall specimen versus lateral load

response. It can be observed from the Figure 6.22 that the model underestimated the

capacities in all the drift levels in both the negative and the positive loading sides by

approximately 10% to 15% comparing to the measured capacities. Regarding the shear

stiffness, the SFI-MVLEM model could not capture the measured shear stiffness

successfully in all the cycles. Also, it can be clearly noted that the predicted results

pinched sharply in the pinching zone as same as the other specimens. Concerning the

shear displacements, in the negative and the positive loading sides, the analytical model

captures the shear displacements in general.

Figure 6.22. Lateral load-shear distortion plot for specimen RW1.

 97

6.3 Alessandro Dazio et al. (1999)

Figure 6.23 shows the comparison between the predicted and the test results

lateral load versus top total displacement response for the rectangular shear wall

specimen WSH6. The SFI-MVLEM analytical model underestimated the capacities

located in between 0.5% and 1.5% drift ratio by approximately 5% to 10% compared to

the actual capacities and succeeded in predicting the capacities accurately for all other

drift levels in the positive loading side except the last capacity due to the failure

mechanism reasons, and the model represented same behavior in the negative loading

side. Cyclic characteristics, such as lateral stiffness degradation, hysteretic shape, and

pinching behavior, were predicted accurately by the analytical model except for the

pinching behavior which was missed by the model.

Figure 6.23. Top total lateral load-displacements plot for specimen WSH6.
 98

Table 6.1 summarizes the comparison between experimentally observed and

analytically simulated results for displacement at yield, shear load at yield, the maximum

shear strength, initial stiffness of load-total top deformation response, and initial stiffness

of load-shear deformation response. It can be noted that the ratio between simulated and

experimental yield displacement ranged between 0.80 and 1.10. The range of ratios

between simulated and experimental shear capacity at yielding was enclosed by 0.90 and

1.09, which reflects a good ability for the model to predict the shear capacity at yielding

point, and we see the same thing for maximum shear capacity in which the ratios range

between 0.98 and 1.06. The ratio range of simulated and experimental for both the initial

stiffness of load-total top deformation response, and initial stiffness of load-shear

deformation response were 0.9-1.17 and 0.92-1.27 respectively.

 99

Table 6. 1: Displacement, strength and stiffness comparison between experimentally observed and simulated specimen results

EIeff,sim /
Spec. y_Exp. y_sim / Vy_Exp Vy_sim / V.Max_Exp Vmax_sim / EIeff_Exp EIeff_Exp GAeff GAeff,sim /
ID (in) y_Exp. kips Vy_Exp kips Vmax_Exp Total Total Exp. GAeff
R1 0.53 0.85 21.8 0.92 26.62 1.01 45.28 1.08 41.13 1.08

R2 0.85 0.86 41.8 0.93 48.72 0.99 44.71 1.13 49.18 1.09

B1 0.70 1.00 51.0 1.08 61.04 1.06 71.43 1.00 72.86 1.08

B2 1.00 1.10 128.0 1.09 152.87 1.02 123.00 0.90 128.00 0.99

B3 0.70 1.00 51.5 1.01 62.02 1.03 67.07 0.96 73.49 1.01

B6 1.31 0.84 173.0 0.90 185.58 1.02 129.67 1.14 131.96 1.07

B7 1.38 0.80 178.5 1.01 220.49 1.02 136.05 1.14 129.17 1.27

B8 1.23 0.94 189.0 0.95 219.89 0.98 158.75 0.96 153.87 1.02

F1 1.00 1.00 150.6 0.92 187.99 1.09 155.00 1.02 150.60 0.92

F2 1.13 0.93 180.3 1.04 199.00 1.06 157.53 1.00 159.57 1.12

RW1 0.85 0.82 33.3 0.90 34.00 1.06 34.12 1.17 39.18 1.09

WSH6 0.90 0.94 -- -- -- -- 140.00 0.92 -- --

99
 100

All in all, the SFI-MVLEM model succeeded in predicting the capacities for

almost all the specimens with relatively high accuracy for lateral-load-versus-total-top-

deformation responses and lateral-load-versus-shear-distortion responses. Even though

analytically predicted results overestimated lateral stiffness for all specimens at drift

levels approximately 0.5% and lower, the analytical model represented a good ability to

simulate the measured lateral stiffness for the specimens used in this study at other drift

levels; however, the model’s ability to simulate the measured shear stiffness needs to be

improved. Regarding the pinching behavior, for the lateral-load-versus-total-top-

deformation response, it is noted that the model simulated reasonably well the pinching

behavior for all specimens applied to axial loads, but the model could not simulate the

pinching behavior for specimens without axial loads applied. In the same content, for the

lateral-load-versus-shear-distortion response, all predicted specimens pinched sharply.

Generally, the hysteretic shape was captured for all specimens with different simulation

accuracy. Regarding the failure mode simulation, the SFI-MVLEM model could not

capture the failure mode for any of specimens because the formulation of SFI-MVLEM

relied on constitutive models does not have the capability to predict the failure

mechanism.
 101

CHAPTER 7

SUMMARY AND CONCLUSIONS

This chapter provides a brief summary of the experimental results and calibration

techniques used for materials, general model assessment and model capabilities and some

important notes that were revealed as a result of this study about the SFI-MVLEM model

as well as some future improvements that can be added to the model are listed.

7.1 Experimental results and calibration of the materials

Twelve specimens from three different programs which are Oesterle et al. (1976,

1979), Thomsen and Wallace (1995), and Dazio et al. (1999) have been chosen to

validate and assess the MVLEM model. These specimens differ from each other in some

parameters such as axial load, confinement condition, cross-section shape, longitudinal

and horizontal reinforcement ratio for web and boundary elements, and some other

parameters. The main purposes of choosing these specimens is the difference in the

characteristics of each specimen and the existence of lateral load versus total top

deformation response and lateral load versus shear deformation response, except for the

specimen tested by Dazio et al. (1999), which has lateral deformation versus total top

deformation response only.

Parameters of the steel and concrete for all specimens were calibrated before

incorporating in the SFI-MVLEM model. The calibration of all steel bars was based on

the values of the yield strength of steel and the strain hardening ratio by fitting the results
 102

obtained from OpenSees based on the Menegotto and Pinto model (1973) with the

experimental data obtained from the stress-strain relationship of tensile tests. The strain-

hardening ratio for specimens was calculated based on the Ramberg-Osgood (R-O)

equation for the strain-hardening suggested by Aktan et al. (1973). In the same content,

cyclic degradation of steel was taken into consideration by selecting the same values for

parameters that simulate the degradation behavior of steel suggested by Menegotto and

Pinto (1973). Regarding concrete calibration, the compressive strain at peak stress for

unconfined concrete was calibrated to fit the actual value by matching the experimental

test results of concrete with the analytical model results, whereas the value of the

compressive strain at peak stress for confined concrete was calibrated by matching the

data obtained from the Saatcioglu and Razvi (1992) model for confined and unconfined

concrete with the data obtained from the analytical model. Peak stress of concrete for

confined concrete was obtained from the Saatcioglu and Razvi (1992) model. Other

parameters related to pre- and post-peak and curvature of the stress-strain relationship

were obtained from the matching of the Saatcioglu and Razvi (1992) model with the

analytical model results that based on Chang and Mander (1994) model. The peak tension

strength of concrete was calibrated by the use of equation recommended by Belarbi and

Hsu (1994).

7.2 General Model Assessment and Model Capabilities

Based on the comparison between experimental results and predicted results by

analytical model for wall responses, it can be concluded that the SFI-MVLE model

provided precise simulation of load-versus-total-top-deformation and load-versus-shear-

deformation responses for all wall specimens in terms of stiffness of wall, capacities, and
 103

pinching behavior, but there were difficulties in predicting the pinching behavior for

specimens that have not been subjected to axial loads. In the same content, for the lateral-

load-versus-shear-distortion response, all predicted specimens pinched sharply. Even

though analytically predicted results overestimated lateral stiffness for all specimens at

drift levels approximately 0.5% and lower, the SFI-MVLEM model showed a good

ability to simulate the measured lateral stiffness for the specimens at other drift levels;

however, the model capability for capturing the measured shear stiffness needs to be

improved. Overall, the hysteretic shape was simulated for all specimens with different

simulation accuracy. Regarding the failure mode simulation, the SFI-MVLEM model

could not capture the failure mode for any of specimens because the constitutive models

for materials used in SFI-MVLEM formulation do not have the capability to predict the

failure mechanism.

7.3 Conclusion and Future Improvement

Based on this study, it can be concluded that the SFI-MVLEM model can be

improved to capture the shear resistance responses by developing and incorporating

adequate model to describe the mechanism of shear resistance in concrete and

reinforcement instead of using ad-hoc parameters. In addition, the use of constitutive

models for materials incorporated in SFI-MVLEM model with the capability to capture

failure modes such as reinforcement buckling could improve the ability of the model to

predict the failure mode of walls. Also, incorporating the pinching effect would be very

important for improving the predicted results of SFI-MVLEM model.

 104

APPENDIX A

OPENSEES CODE FOR SPECIMEN B8

# ------------------------------------------------------------------------------------------------------------------------------------------
# Example 1: Simulation of wall cyclic behavior using SFI_MVLEM
# Specimen: B8 by Oesterle et al. (1979)
# Created by: Haider Sakban & Kristijan Kolozvari
# Date: 11/2016
# ------------------------------------------------------------------------------------------------------------------------------------------

# --------------------------------------------------------------------
# Start of model generation (Units: kip, in, sec, ksi)
# --------------------------------------------------------------------

# Set Up Directories
set modelName "$Dincr _ $alfadow _ $nu";
set dataDir B8_$modelName; # Name of output folder
file mkdir $dataDir;

# Create ModelBuilder for 2D element (with two-dimensions and 3 DOF/node)

model BasicBuilder -ndm 2 -ndf 3

# --------------------------------------------------------------------
# Set geometry, nodes, boundary conditions
# --------------------------------------------------------------------

# Wall Geometry
set H 180; # Wall height
set t 4; # Wall thickness
set tb 12; # boundary thickness

# Create nodes
# node nodeId xCrd yCrd
node 1 0 0;
node 2 0 9;
node 3 0 18;
node 4 0 27;
node 5 0 36;
node 6 0 54;
node 7 0 74;
node 8 0 108;
node 9 0 144;
node 10 0 $H;
 105

# Boundary conditions
fix 1 1 1 1; # Fixed condition at node 1

# Set Control Node and DOF

set IDctrlNode 10;
set IDctrlDOF 1;

# ----------------------------------------------------------------------------------------------
# Define uniaxial materials for 2D RC Panel Constitutive Model (FSAM)
# ----------------------------------------------------------------------------------------------

# STEEL ...........................................................
# uniaxialMaterial SteelMPF $mattag $fyp $fyn $E0 $bp $bn $R0 $a1 $a2
# steel X web
set fyX 69.9082; # fy No.3
set bx 0.0179; # strain hardening (calculated)

# steel Y web
set fyYw 65.8036; # fy 6mm
set byw 0.0197; # strain hardening (calculated)

# steel Y boundary
set fyYb 64.9044; # fy No.6
set byb 0.0155; # strain hardening (calculated)

# steel X boundary
set fyxb 65.8036; # fy x-axis 6mm
set bxb 0.0197; # strain hardening x-axis (calculated)

# steel misc
set Esy 29000.0; # Young's modulus
set Esx $Esy; # Young's modulus
set R0 20.0; # initial value of curvature parameter
set A1 0.925; # curvature degradation parameter
set A2 0.15; # curvature degradation parameter

# Build steel materials

uniaxialMaterial SteelMPF 1 $fyX $fyX $Esx $bx $bx $R0 $A1 $A2; # steel X web
uniaxialMaterial SteelMPF 2 $fyYw $fyYw $Esy $byw $byw $R0 $A1 $A2; # steel Y web
uniaxialMaterial SteelMPF 3 $fyYb $fyYb $Esy $byb $byb $R0 $A1 $A2; # steel Y boundary
uniaxialMaterial SteelMPF 4 $fyxb $fyxb $Esx $bxb $bxb $R0 $A1 $A2; # steel x boundary

# CONCRETE ........................................................
# uniaxialMaterial ConcreteCM $mattag $fpcc $epcc $Ec $rc $xcrn $ft $et $rt $xcrp <-GapClose
$gap>

# unconfined
set fpc 6.085; # peak compressive stress
set ec0 -0.0031; # strain at peak compressive stress
set ft 0.291; # peak tensile stress
set et 0.00008; # strain at peak tensile stress
set Ec 3900; # Young's modulus
set ru 7.0; # (1) shape parameter - curve before e0
 106

set xcrnu 1.172; # (2) post-peak slope - sensitive

set rt 1.2; # shape parameter - tension
set xcrp 10000; # cracking strain - tension

# confined
set fpcc 8.4671; # peak compressive stress
set ec0c -0.009168; # strain at peak compressive stress
set Ecc $Ec; # Young's modulus
set xcrnc 1.028; # cracking strain - compression
set rc 7.5; # shape parameter - compression

# Build concrete materials

# unconfined concrete
uniaxialMaterial ConcreteCM 5 -$fpc $ec0 $Ec $ru $xcrnu $ft $et $rt $xcrp -GapClose 0;
# confined concrete
uniaxialMaterial ConcreteCM 6 -$fpcc $ec0c $Ecc $rc $xcrnc $ft $et $rt $xcrp -GapClose 0;

# --------------------------------------------------
# Define 2D RC Panel Material (FSAM)
# --------------------------------------------------

# Reinforcing ratios
set rouXw 0.0138; # X web
set rouXb 0.0135; # X boundary
set rouYw 0.0029; # Y web
set rouYb 0.0367; # Y boundary
set rouXb2 0.0045; # Y boundary2 above 6 ft.

# Shear resisting mechanism parameters

set nu 1.0; # friction coefficient
set alfadow 0.005; # dowel action stiffness parameter

# nDMaterial FSAM $mattag $rho $sX $sY $conc $rouX $rouY $nu $alfadow
nDMaterial FSAM 7 0.0 1 2 5 $rouXw $rouYw $nu $alfadow; # Web (unconfined concrete)
nDMaterial FSAM 8 0.0 4 3 6 $rouXb $rouYb $nu $alfadow; # Boundary (confined concrete)
nDMaterial FSAM 9 0.0 4 3 6 $rouXb2 $rouYb $nu $alfadow; # Boundary (confined concrete
above 6 ft.)

# -----------------------------------------
# Define SFI_MVLEM elements
# -----------------------------------------

# element SFI_MVLEM eleTag iNode jNode m c -thick fiberThick -width fiberWidth -mat matTags
element SFI_MVLEM 1 1 2 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 8 8 7 7 7 7 7 7 8 8
element SFI_MVLEM 2 2 3 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 8 8 7 7 7 7 7 7 8 8
element SFI_MVLEM 3 3 4 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 8 8 7 7 7 7 7 7 8 8
element SFI_MVLEM 4 4 5 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 8 8 7 7 7 7 7 7 8 8
 107

element SFI_MVLEM 5 5 6 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 8 8 7 7 7 7 7 7 8 8
element SFI_MVLEM 6 6 7 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 8 8 7 7 7 7 7 7 8 8
element SFI_MVLEM 7 7 8 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 9 9 7 7 7 7 7 7 9 9
element SFI_MVLEM 8 8 9 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 9 9 7 7 7 7 7 7 9 9
element SFI_MVLEM 9 9 10 10 0.4 -thick $tb $tb $t $t $t $t $t $t $tb $tb -width 6 6 8.5 8.5 8.5 8.5 8.5 8.5
6 6 -mat 9 9 7 7 7 7 7 7 9 9

# -----------------------------------
# End of model generation
# -----------------------------------
# Initialize
initialize

# ------------------------------
# Recorder generation
# ------------------------------

# Nodal recorders
recorder Node -file $dataDir/MVLEM_Dtop.out -time -node $IDctrlNode -dof 1 disp
recorder Node -file $dataDir/MVLEM_DOFs.out -time -node 1 2 3 4 5 6 7 8 9 10 -dof 1 2 3 disp

# Element recorders
recorder Element -file $dataDir/MVLEM_Dsh.out -time -ele 1 2 3 4 5 6 7 8 9 ShearDef

# ------------------------
# Define Axial Load
# ------------------------

set N [expr 268.14]; # kips

# ----------------------------------------------------------------------
# Set parameters for displacement controlled analysis
# ----------------------------------------------------------------------

set Ncycles 3; # specify the number of cycles at each peak

# vector of displacement-cycle peaks in terms of wall drift ratio

set iDmax "0.001388889 0.001944444 0.003666667 0.005277778 0.011388889 0.016666667
0.022222222 0.027777778 0.033055556";
set TotalDrift [expr
$Ncycles*4*(0.001388889+0.001944444+0.003666667+0.005277778+0.011388889+0.016666667+0.0222
22222+0.027777778+0.033055556)];

#set iDmax "0.033055556"; # ULTIMATE DRIFT

#set TotalDrift [expr 0.033055556];

#set Dincr 0.092; # displacement increment for displacement controlled analysis (in.).
 108

set CycleType Full; # type of cyclic analysis: Full / Push / Half

set Tol 1.0e-5;

set LunitTXT "inch";

APPLYING GRAVITY LOADS

# ------------------------------------------------------------------------------------------------------------------------------------------
# Gravity load analysis
# ------------------------------------------------------------------------------------------------------------------------------------------

# Generate the model and load input variables

source B8.tcl

# Create a Plain load pattern with a linear TimeSeries

pattern Plain 1 "Linear" {

# Create the nodal load - command: load nodeID xForce yForce

load $IDctrlNode 0.0 [expr -$N] 0.0
}

# ------------------------------
# Analysis generation
# ------------------------------

# Create the integration scheme, the LoadControl scheme using steps of 0.1
integrator LoadControl 0.1

# Create the system of equation, a sparse solver with partial pivoting

system BandGeneral

# Create the convergence test, the norm of the residual with a tolerance of
# 1e-5 and a max number of iterations of 100
test NormDispIncr $Tol 100 0

# Create the DOF numberer, the reverse Cuthill-McKee algorithm

numberer RCM

# Create the constraint handler, the transformation method

constraints Transformation

# Create the solution algorithm, a Newton-Raphson algorithm

algorithm Newton # -initialThenCurrent

# Create the analysis object

analysis Static

# Run analysis
analyze 10

APPLYING LATERAL LOADS (DISPLACEMENT-CONTROLLED ANALYSIS)

# ------------------------------------------------------------------------------------------------------------------------------------------
# Displacement Controlled Analysis
 109

# ------------------------------------------------------------------------------------------------------------------------------------------

source gravity.tcl

puts "Model generated and gravity load applied successfully";

# Set the gravity loads to be constant & reset the time in the domain
loadConst -time 0.0

set Fact $H; # scale drift ratio by storey height for displacement cycles

# Set lateral load pattern with a Linear TimeSeries

set Plateral 1.0; # Reference lateral load
pattern Plain 200 "Linear" {

load $IDctrlNode $Plateral 0.0 0.0

}

set load_step 1;

# set up analysis parameters

source LibAnalysisStaticParameters.tcl; # constraintsHandler,DOFnumberer,system-
ofequations,convergenceTest,solutionAlgorithm,integrator

# perform Static Cyclic Displacements Analysis

source LibGeneratePeaks.tcl

set fmt1 "%s Cyclic analysis: CtrlNode %.3i, dof %.1i, Disp=%.4f %s"; # format for screen/file output of
DONE/PROBLEM analysis

foreach Dmax $iDmax {

set iDstep [GeneratePeaks $Dmax $Dincr $CycleType $Fact]; # this proc is defined above

for {set i 1} {$i <= $Ncycles} {incr i 1} {

set zeroD 0
set D0 0.0
foreach Dstep $iDstep {
set D1 $Dstep
set Dincr [expr $D1 - $D0]
integrator DisplacementControl $IDctrlNode $IDctrlDOF $Dincr
analysis Static
# ----------------------------------------------first analyze command------------------------
set ok [analyze 1]
# ----------------------------------------------if convergence failure-------------------------
if {$ok != 0} {
# if analysis fails, we try some other stuff
# performance is slower inside this loop global
maxNumIterStatic;# max no. of iterations performed before "failure to converge" is ret'd
if {$ok != 0} {
puts "Trying Newton with Initial Tangent .."
test NormDispIncr $Tol 2000 0
 110

algorithm Newton -initial

set ok [analyze 1]
test $testTypeStatic $TolStatic $maxNumIterStatic 0
algorithm $algorithmTypeStatic 500 1
}
if {$ok != 0} {
puts "Trying Broyden .."
algorithm Broyden 500
set ok [analyze 1 ]
algorithm $algorithmTypeStatic
}
if {$ok != 0} {
set putout [format $fmt1 "PROBLEM" $IDctrlNode $IDctrlDOF
[nodeDisp $IDctrlNode $IDctrlDOF] $LunitTXT]
puts $putout
return -1
}; # end if
}; # end if
# -----------------------------------------------------------------------------------------------------
set D0 $D1; # move to next step

# print load step on the screen

puts "Load Step: [expr $load_step]"
set load_step [expr $load_step+1]

}; # end Dstep

}; # end i

}; # end of iDmaxCycl
# -----------------------------------------------------------------------------------------------------
if {$ok != 0 } {
puts [format $fmt1 "PROBLEM" $IDctrlNode $IDctrlDOF [nodeDisp $IDctrlNode $IDctrlDOF]
$LunitTXT]
} else {
puts [format $fmt1 "DONE" $IDctrlNode $IDctrlDOF [nodeDisp $IDctrlNode $IDctrlDOF]
$LunitTXT]
}

# Print the state at control node

print node $IDctrlNode

DEFINITION OF STATIC ANALYSIS PARAMETERS

# --------------------------------------------------------------------------------------------------
# static analysis parameters
# I am setting all these variables as global variables (using variable rather than set command)
# so that these variables can be uploaded by a procedure
# Silvia Mazzoni & Frank McKenna, 2006
# --------------------------------------------------------------------------------------------------
 111

# CONSTRAINTS handler -- Determines how the constraint equations are enforced in the analysis
(http://opensees.berkeley.edu/OpenSees/manuals/usermanual/617.htm)
# Plain Constraints -- Removes constrained degrees of freedom from the system of equations (only
for homogeneous equations)
# Lagrange Multipliers -- Uses the method of Lagrange multipliers to enforce constraints
# Penalty Method -- Uses penalty numbers to enforce constraints --good for static analysis with non-
homogeneous eqns (rigidDiaphragm)
# Transformation Method -- Performs a condensation of constrained degrees of freedom
variable constraintsTypeStatic Plain; # default;
if { [info exists RigidDiaphragm] == 1} {
if {$RigidDiaphragm=="ON"} {
variable constraintsTypeStatic Lagrange; # for large model, try Transformation
}; # if rigid diaphragm is on
}; # if rigid diaphragm exists
constraints $constraintsTypeStatic

# DOF NUMBERER (number the degrees of freedom in the domain):

(http://opensees.berkeley.edu/OpenSees/manuals/usermanual/366.htm)
# determines the mapping between equation numbers and degrees-of-freedom
# Plain -- Uses the numbering provided by the user
# RCM -- Renumbers the DOF to minimize the matrix band-width using the Reverse Cuthill-McKee
algorithm
set numbererTypeStatic RCM
numberer $numbererTypeStatic

# SYSTEM (http://opensees.berkeley.edu/OpenSees/manuals/usermanual/371.htm)
# Linear Equation Solvers (how to store and solve the system of equations in the analysis)
# -- provide the solution of the linear system of equations Ku = P. Each solver is tailored to a specific
matrix topology.
# ProfileSPD -- Direct profile solver for symmetric positive definite matrices
# BandGeneral -- Direct solver for banded unsymmetric matrices
# BandSPD -- Direct solver for banded symmetric positive definite matrices
# SparseGeneral -- Direct solver for unsymmetric sparse matrices
# SparseSPD -- Direct solver for symmetric sparse matrices
# UmfPack -- Direct UmfPack solver for unsymmetric matrices
set systemTypeStatic BandGeneral; # try UmfPack for large model
system $systemTypeStatic

# TEST: # convergence test to

# Convergence TEST (http://opensees.berkeley.edu/OpenSees/manuals/usermanual/360.htm)
# -- Accept the current state of the domain as being on the converged solution path
# -- determine if convergence has been achieved at the end of an iteration step
# NormUnbalance -- Specifies a tolerance on the norm of the unbalanced load at the current
iteration
# NormDispIncr -- Specifies a tolerance on the norm of the displacement increments at the current
iteration
# EnergyIncr-- Specifies a tolerance on the inner product of the unbalanced load and displacement
increments at the current iteration
# RelativeNormUnbalance --
# RelativeNormDispIncr --
# RelativeEnergyIncr --
variable TolStatic 1.e-5; # Convergence Test: tolerance
 112

variable maxNumIterStatic 100; # Convergence Test: maximum number of iterations that will be
performed before "failure to converge" is returned
variable printFlagStatic 0; # Convergence Test: flag used to print information on convergence
(optional) # 1: print information on each step;
variable testTypeStatic NormDispIncr; # Convergence-test type
test $testTypeStatic $TolStatic $maxNumIterStatic $printFlagStatic;
# for improved-convergence procedure:
variable maxNumIterConvergeStatic 2000;
variable printFlagConvergeStatic 1;

# Solution ALGORITHM: -- Iterate from the last time step to the current
(http://opensees.berkeley.edu/OpenSees/manuals/usermanual/682.htm)
# Linear -- Uses the solution at the first iteration and continues
# Newton -- Uses the tangent at the current iteration to iterate to convergence
# ModifiedNewton -- Uses the tangent at the first iteration to iterate to convergence
# NewtonLineSearch --
# KrylovNewton --
# BFGS --
# Broyden --
variable algorithmTypeStatic Newton
algorithm $algorithmTypeStatic;

# Static INTEGRATOR: -- determine the next time step for an analysis

(http://opensees.berkeley.edu/OpenSees/manuals/usermanual/689.htm)
# LoadControl -- Specifies the incremental load factor to be applied to the loads in the domain
# DisplacementControl -- Specifies the incremental displacement at a specified DOF in the domain
# Minimum Unbalanced Displacement Norm -- Specifies the incremental load factor such that the
residual displacement norm in minimized
# Arc Length -- Specifies the incremental arc-length of the load-displacement path
# Transient INTEGRATOR: -- determine the next time step for an analysis including inertial effects
# Newmark -- The two parameter time-stepping method developed by Newmark
# HHT -- The three parameter Hilbert-Hughes-Taylor time-stepping method
# Central Difference -- Approximates velocity and acceleration by centered finite differences of
displacement
integrator DisplacementControl $IDctrlNode $IDctrlDOF $Dincr

# ANALYSIS -- defines what type of analysis is to be performed

(http://opensees.berkeley.edu/OpenSees/manuals/usermanual/324.htm)
# Static Analysis -- solves the KU=R problem, without the mass or damping matrices.
# Transient Analysis -- solves the time-dependent analysis. The time step in this type of analysis is
constant. The time step in the output is also constant.
# variableTransient Analysis -- performs the same analysis type as the Transient Analysis object. The
time step, however, is variable. This method is used when
# there are convergence problems with the Transient Analysis object at a peak or when the time
step is too small. The time step in the output is also variable.
set analysisTypeStatic Static
analysis $analysisTypeStatic

DISPLACEMENT HISTORY GENERATION:

proc GeneratePeaks {Dmax {DincrStatic 0.01} {CycleType "Full"} {Fact 1} } {; # generate incremental
disps for Dmax
 113

###########################################################################
## GeneratePeaks $Dmax $DincrStatic $CycleType $Fact
###########################################################################
# generate incremental disps for Dmax
# this proc creates a file which defines a vector then executes the file to return the vector of disp.
increments
# by Silvia Mazzoni, 2006
# input variables
# $Dmax : peak displacement (can be + or negative)
# $DincrStatic : displacement increment (optional, default=0.01, independently of
units)
# $CycleType : Full (0->+peak), Half (0->+peak->0), Full (0->+peak->0->-peak->0)
(optional, def=Full)
# $Fact : scaling factor (optional, default=1)
# $iDstepFileName : file name where displacement history is stored temporarily, until
next disp. peak
# output variable
# $iDstep : vector of displacement increments
file mkdir data
set outFileID [open data/tmpDsteps.tcl w]

set Disp 0.

puts $outFileID "set iDstep { ";puts $outFileID $Disp;puts $outFileID $Disp; # open vector
definition and some 0

set Dmax [expr $Dmax*$Fact]; # scale value

if {$Dmax<0} {; # avoid the divide by zero

set dx [expr -$DincrStatic]

} else {

set dx $DincrStatic;

set NstepsPeak [expr int(abs($Dmax)/$DincrStatic)]

for {set i 1} {$i <= $NstepsPeak} {incr i 1} {; # zero to one

set Disp [expr $Disp + $dx]
puts $outFileID $Disp; # write to file
}
if {$CycleType !="Push"} {
for {set i 1} {$i <= $NstepsPeak} {incr i 1} {; # one to zero
set Disp [expr $Disp - $dx]
puts $outFileID $Disp; # write to file
}
if {$CycleType !="Half"} {
for {set i 1} {$i <= $NstepsPeak} {incr i 1} {; # zero to minus one
set Disp [expr $Disp - $dx]
puts $outFileID $Disp; # write to file
 114

}
for {set i 1} {$i <= $NstepsPeak} {incr i 1} {; # minus one to zero
set Disp [expr $Disp + $dx]
puts $outFileID $Disp; # write to file
}
}
}
puts $outFileID " }"; # close vector definition
close $outFileID
source data/tmpDsteps.tcl; # source tcl file to define entire vector
return $iDstep
 115

APPENDIX B

CALIBRATION OF CONCRETE AND STEEL MATERIALS

This part of study represents the Figures of the calibration of materials (concrete

and steel) as it was discussed in Chapter 5 of this study for all specimens from the three

programs used to assess and validate the SFI-MVLEM model.

Figure B.1. Confined and unconfined concrete calibration for R1 specimen.

 116

Figure B.2. Confined and unconfined concrete calibration for R2 specimen.

Figure B.3. Confined and unconfined concrete calibration for B1 specimen.

 117

Figure B.4. Unconfined concrete calibration for B2 specimen.

Figure B.5. Confined and unconfined concrete calibration for B3 specimen.

 118

Figure B.6. Confined and unconfined concrete calibration for B6 specimen.

Figure B.7. Confined and unconfined concrete calibration for B7 specimen.

 119

Figure B.8. Confined and unconfined concrete calibration for B8 specimen.

Figure B.9 Confined and unconfined concrete calibration for F1 specimen.

 120

Figure B.10. Confined and unconfined concrete calibration for F2 specimen.

Figure B.11. Confined and unconfined concrete calibration for RW1 specimen.
 121

Figure B.12. Confined and unconfined concrete calibration for WSH6 specimen.

Figure B.133. Calibration of strain hardening ratio of 6mm bar for WSH6 specimen.
 122

Figure B.14. Calibration of strain hardening ratio of 8mm bar for WSH6 specimen.

Figure B.15. Calibration of strain hardening ratio of 12mm bar for WSH6 specimen.
 123

REFERENCES

Aktan, A. E., Karlsson, B. I., & Sozen, M. A. (1973). Stress-strain relationships of

reinforcing bars subjected to large strain reversals. University of Illinois
Engineering Experiment Station. University of Illinois at Urbana-Champaign.

Bazant, Z. P., Xiang, Y., & Prat, P. C. (1996). Microplane model for concrete I: Stress-
strain boundaries and finite strain. Journal of Engineering Mechanics, 122(3),
245-254.

Belarbi, A., & Hsu, T. T. (1994). Constitutive laws of concrete in tension and reinforcing
bars stiffened by concrete. Structural Journal, 91(4), 465-474.

Belarbi A. and Hsu T.C. (1995). Constitutive Laws of Softened Concrete in Biaxial
Tension Compression. ACI Structural Journal, 119(12), 3590-3610.

Beyer, K., Dazio, A., & Priestley, N. (2011). Shear deformations of slender reinforced
concrete walls under seismic loading. ACI Structural Journal, 108(EPFL-
ARTICLE-162084), 167-177.

Bolong, Z., Mingshun, W., & Kunlian, Z. (1980). A Study of Hysteretic Curve of
Reinforced Concrete Members under Cyclic Loading Taking Account of Local
Contact Effects of Cracked Section. Journal of Tongji University, 1, 18-24.

Chang, G. A., & Mander, J. B. (1994). Seismic energy based fatigue damage analysis of
bridge columns: Part I-Evaluation of seismic capacity. Buffalo, NY: National
Center for Earthquake Engineering Research.

Dazio, A., Beyer, K., & Bachmann, H. (2009). Quasi-static cyclic tests and plastic hinge
analysis of RC structural walls. Engineering Structures. 31(7), 1556-1571.

Elmorsi, M., Kianoush, M. R., & Tso, W. K. (1998). Nonlinear analysis of cyclically
loaded reinforced concrete structures. Structural Journal, 95(6), 725-739.

Filippou F. C.; Popov E. G.; & Bertero V. V. (1983). Effects of Bond Deterioration on
Hysteretic Behavior of Reinforced Concrete Joints. EERC Report No.
UCB/EERC–83/19. Earthquake Engineering Research Center, University of
California, Berkeley, California.

Fischinger, M., Vidic, T., Selih, J., Fajfar, P., Zhang, H. Y., & Damjanic, F. B. (1990).
Validation of a macroscopic model for cyclic response prediction of RC walls.
Computer Aided Analysis and Design of Concrete Structures, 2, 1131-1142.
 124

Fischinger, M., Vidic, T., & Fajfar, P. (1991). Evaluation of the inelastic response of a
RC building with a structural wall designed according to Eurocode 8. Bı, 1(2),
487-498.

Fischinger, M. A. T. E. J., Vidic, T. O. M. A. Ž. & Fajfar, P. (1992). Nonlinear seismic

analysis of structural walls using the multiple-vertical-line-element model.
Elsevier Science Publishers Ltd., London and New York.

Kabeyasawa, T., Shiohara, H., Otani, S., & Aoyama, H. (1983). Analysis of the full-scale
seven-story reinforced concrete test structure. Journal of the Faculty of
Engineering, 37(2), 431-478.

Kabeyasawa, T. (1997). Design of RC shear walls in hybrid wall system. Proceedings,

Fourth Joint Technical Coordinating Committee, US-Japan Cooperative Seismic
Research on Composite and Hybrid Structures, Monterey.

Kolozvari K., Orakcal K., and Wallace J. W. (2015a). Shear-Flexure Interaction

Modeling of reinforced Concrete Structural Walls and Columns under Reversed
Cyclic Loading. Pacific Earthquake Engineering Research Center, University of
California, Berkley.

Kolozvari K., Orakcal K., Wallace J.W. (2015b). Modeling of cyclic shear-flexure
interaction in reinforced concrete structural walls. I: Theory. ASCE, Journal of
Structural Engineering, 141(5), 04014135.

Kolozvari K., Tran T., Orakcal K., Wallace J.W. (2015c). Modeling of cyclic shear-
flexure interaction in reinforced concrete structural walls. II: Experimental
Validation, ASCE, Journal of Structural Engineering, 141(5), 04014136.

Mander, J. B., Priestley, M. J., & Park, R. (1988). Theoretical stress-strain model for
confined concrete. Journal of Structural Engineering, 114(8), 1804-1826.

Massone L.M., and J. W. Wallace (2004). Load – deformation responses of slender

reinforced concrete walls. ACI Structural Journal, 101(1), 103-113.

Massone, L. M., Orakcal, K., & Wallace, J. W. (2006). Shear-flexure interaction for
structural walls. Special Publication, 236, 127-150.

Massone, L. M., Orakcal, K., & Wallace, J. W. (2009). Modeling of squat structural walls
controlled by shear. ACI Structural Journal, 106(5), 646.

McKenna, F., Fenves, G. L., & Scott, M. H. (2000). Open system for earthquake
engineering simulation. University of California, Berkeley, CA.

Oesterle, R. G., Fiorato, A. E., Johal, L. S., Carpenter, J. E., Russell, H. G., & Corley, W.
G. (1976). Earthquake resistant structural walls-tests of isolated walls. Report to
the National Science Foundation.
 125

Oesterle, R. G., Aristizabal-Ochoa, J. D., Fiorato, A. E., Russell, H. G., & Corley, W. G.
(1979). Earthquake resistant structural walls-tests of isolated walls-phase II.
Report to National Science Foundation.

Orakcal, K., & Wallace, J. W. (2004). Nonlinear modeling and analysis of slender
reinforced concrete walls. ACI Structural Journal, 5, 688-698.

Orakcal, K., & Wallace, J. W. (2006). Flexural modeling of reinforced concrete walls-
experimental verification. ACI Structural Journal, 103(2), 196-206.

Orakcal, K., Ulugtekin, D., & Massone, L. M. (2012). Constitutive modeling of

reinforced concrete panel behavior under cyclic loading. In Proceedings, 15th
World Conference on Earthquake Engineering (Vol. 3573).

Panagiotou, M., Restrepo, J. I., Schoettler, M., & Kim, G. (2012). Nonlinear cyclic truss
model for reinforced concrete walls. ACI Structural Journal, 109(2), 205-214.

Pang, X. B. D., & Hsu, T. T. (1995). Behavior of reinforced concrete membrane elements
in shear. Structural Journal, 92(6), 665-679.

Powell, G. H. (2007). Detailed Example of a Tall Shear Wall Building Using CSI's
Perform 3D Nonlinear Dynamic Analysis. Computers & Structures Educational
Services.

Saatcioglu, M., & Razvi, S. R. (1992). Strength and ductility of confined concrete.
Journal of Structural engineering, 118(6), 1590-1607.

Sayre B. (2003). Performance evaluation of steel reinforced shear walls. MS Thesis,

University of California, Los Angeles.

Thomsen, J. H., & Wallace, J. W. (1995). Displacement based design of reinforced

concrete structural walls: an experimental investigation of walls with rectangular
and t-shaped cross-sections: a dissertation. Doctoral dissertation, Clarkson
University.

Thomsen IV, J. H., & Wallace, J. W. (2004). Displacement-based design of slender

reinforced concrete structural walls-experimental verification. Journal of
Structural Engineering, 130(4), 618-630.

Tran T. A. (2012). Experimental and Analytical Studies of Moderate Aspect Ratio

Reinforced Concrete Structural Walls. Doctoral dissertation, University of
California, Los Angeles.

Tran, T. A., & Wallace, J. W. (2012). Experimental study of nonlinear flexural and shear
deformations of reinforced concrete structural walls. Proceedings in 15th World
Conference on Earthquake Engineering.
 126

Tsai, W. T. (1988). Uniaxial compressional stress-strain relation of concrete. Journal of

Structural Engineering, 114(9), 2133-2136.

Ulugtekin, D. (2010). Analytical modeling of reinforced concrete panel elements under

reversed cyclic loadings. MS Thesis, Bogazici University.

Vecchio, F. J., & Collins, M. P. (1986). The modified compression-field theory for
reinforced concrete elements subjected to shear. ACI Structural J, 83(2), 219-231.

Vulcano, A. and Bertero, V.V. (1986). Nonlinear Analysis of RC Structural Walls. In

Proceedings, 8th European Conference on EQ Engineering, 3, Lisbon, Portugal,
6.5/1-6.5/8.

Vulcano, A. and Bertero, V.V. (1987). Analytical Models for Predicting the Lateral
Response of RC Shear Walls: Evaluation of Their Reliability. EERC Report No.
UCB/EERC-87/19. Earthquake Engineering Research Center, University of
California, Berkeley.

Vulcano, A., Bertero, V. V., & Colotti, V. (1988, August). Analytical modeling of R/C
structural walls. In Proceedings of the 9th World Conference on Earthquake
Engineering (Vol. 6, pp. 41-46).