Podium and Transfer Structure

Interference on Seismic Shear Demands

of Tower Walls in Buildings

Mehair Yacoubian

A thesis submitted in total fulfilment of the requirements for

the degree of Doctor of Philosophy

May 2018

Department of Infrastructure Engineering

The University of Melbourne
 ABSTRACT
The preference for the mixed-use of buildings has led to a paradigm shift in the way buildings
are designed, and their geometries are determined. A podium-tower building configuration that
caters for both commercial and residential functionalities has become a popular form of building
construction in many metropolitan cities around the world. The tower portion of a building with
this configuration may simply be extruded from part of the podium (i.e. a setback) or in some
cases is planted on top of a transfer plate. In the event of an earthquake, the podium (lower) storeys
of the building will function to redistribute the loads applied onto the tower. The interferences of
the podium structure on the lateral resistance of the building are central to ensuring that an
adequate load path is maintained up the height of the building. Nevertheless, the consequences of
these interferences on the seismic performance of the building are not well understood and often
require sufficiently detailed numerical simulations to model their effects. Even then, some aspects
of these interferences (by the podium) may be hidden from the user. The aim of this study has
been to investigate the different elements of the podium interferences on the response behaviour
of the cores and walls that support the tower structure above. The two forms of podium-tower
buildings (featuring a setback and a transfer plate) have been thoroughly investigated to develop
an understanding of the complex podium-tower interactions in a building under lateral loads. A
systematic investigation has been carried out by way of detailed numerical analyses on
representative building models to highlight the key contributors to these interferences and to
develop analytical tools to quantify their effects on the response behaviour of the building in
seismic conditions.

In a podium-tower building featuring a setback (in the floor area), the interferences by the
podium structure are believed to only occur in the lower (podium) storeys of the building. These
interferences are manifested by the shear force reversals in the tower walls and cores at the level
of the podium interface (commonly referred to as the backstay effect). This study has revealed
that significant redistributions of shear forces in the tower walls above the podium interface level
are attributed to the interferences by the podium structure when the building is subjected to lateral
loads. Furthermore, these interferences were shown to be most pronounced in a building where
the tower structure is not centred about the supporting podium. The direct result of these
redistributions was the high shear force concentration that has been revealed in some of the tower
walls above the interface level. As such, shear-critical conditions have been demonstrated to occur
in some tower walls which have reduced the capacity of the building to develop the required
lateral displacements and strengths when subjected to earthquake loads. More importantly, the
study emphasised that the described interferences (by the podium) would not have been
necessarily signalled by the engineer should some (common) modelling assumptions were made

i
in the early stages of the design. Analytical models have been developed to quantify the extent of
these interferences and the amount of the additional shear forces that are locally developed in
some tower walls above the level of the interface. These additional shear forces, which have
originated in the floor slabs connecting the tower walls (and cores), were proportional to the ratio
of the relative lateral stiffnesses of the tower and the podium. Accordingly, a wider podium at
only a few storeys in the lower storeys of the building pertains to a higher interference and a
higher shear force concentration in the tower walls.

In podium-tower building featuring a transfer plate, the interferences by the lower (podium)
structure mostly occur at the level of the transfer (which in this case is the podium interface level).
The redistributions of the lateral forces from the tower to the supporting podium are manifested
by the out-of-plane distortions (i.e. bending) of the transfer plate. Similar to a building with a
setback feature, these interferences by the podium (represented by the plate) also resulted in high
shear force concentrations in the tower storeys above the level of the transfer plate. An analytical
framework has been developed to estimate the extent of interferences by the transfer plate and to
model the anomalous increase in the shear force demand on the tower walls. In the developed
model, the in-plane strains (and forces) in the floor slabs have been found to be linearly
proportional to the difference in the transfer plate rotations evaluated at the base of the tower wall
(which is the primary parameter controlling the extent of the interference). The complexity of
estimating these deformations motivated the development of a simple displacement-based
solution to estimate the extent of the interferences by the transfer plate by the use of only a few
parameters that can be conveniently computed in a structural design office. As such, local
deformations of the plate have been formulated proportionally to the resulting rotations (drifts) of
the tower structure evaluated at its centre of mass. The study revealed that both the transfer plate
deformations and the consequent additional shear force demand on the tower walls displayed
displacement-controlled features which ultimately led to the development of a closed-form
analytical solution for quantifying these interferences. The applicability of the framework has
been tested on different case study buildings of different heights and plate thicknesses. In the
absence of detailed guidelines in modelling such effects, the introduced analytical technique
(which can be conveniently programmed on a single Excel spreadsheet) can potentially assist the
design engineer in accurately estimating the shear force demands on the tower walls which would
have been misrepresented in certain numerical simulations of the building.

A thorough assessment of the seismic performance has been conducted on the two types of
buildings examined in this study. The main objective has been to evaluate and quantify the
response modification factors (the force reduction, ductility and overstrength factors) used in the
design of the building. The results of a large number of incremental dynamic analyses (IDA) using
ground motions with different intensities have been employed to assess the damage limit states

ii
sustained by the building. Seismic fragility curves were constructed to identify the limiting ground
motion demands corresponding to the different performance levels. The results highlighted the
high propensity for the building to develop local concentrations of inelastic demands in the storeys
immediately above the podium level. These concentrations were attributed to the unfavourable
interferences by the podium which has ultimately restricted the building from developing the
required displacement capacities in seismic conditions. While the computed values of the force
reduction factors were (generally) in good agreement with the code recommended values, the
limited-ductile response of the tower walls above the interface resulted in lower values of the
ductility factors than those assumed in the design. Thus, the assumed trade-off between strength
and deformation may be unconservative for the types of podium-tower buildings examined in this
study. A design framework has been proposed to mitigate the risks of developing the potential
unfavourable mechanisms above the podium level by adjusting the design seismic force
distributions up the height of the building.

iii
 Declaration

This is to certify that:

i. The thesis comprises only my original work

ii. Due acknowledgment has been made in the text to all material used
iii. The thesis is less than 100,000 words in length exclusive of tables, figures,
references and appendices

Mehair Yacoubian
May 2018

iv
 ACKNOWLEDGMENTS

First and foremost, I would like to express my sincere gratitude to my principal supervisor
Prof Nelson Lam, for his continued support and motivation throughout the course of my study,
his understanding, care and knowledge has been invaluable. My deepest gratitude goes to my co-
supervisor Dr Elisa Lumantarna for her encouragement, support and guidance along the way. I
would also like to sincerely thank Prof John Wilson of Swinburne University of Technology for
his support.

I am also grateful to all of the professors and staff who have helped over the years, with special
thanks to Prof Tuan Ngo, the chair of the advisory committee. The financial support provided by
the University of Melbourne is gratefully acknowledged.

I would also like to thank Dr Daniel Looi of the University of Hong Kong for generously
sharing his experimental results when needed and for the valuable discussions on numerous topics
relating to transfer structures.

To my colleagues and friends at the Department of Infrastructure engineering whom I have

the pleasure of making their acquaintance. I thank you all for your friendship, pep talks, laughs
and the valuable discussions we had over the years.

To my family, I thank you for being patient, loving and understanding in every step of the
way. I am deeply indebted to my wife for her unconditional love, support and the countless
sacrifices she made on my behalf throughout my study.

v
 LIST OF PUBLICATIONS
The research presented in this thesis has been published in three journal articles (one of which
is in press), a book chapter and five conference papers. Due acknowledgements to the below
publications are made in various chapters of this thesis.

Journal articles

Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. (2018). A simplified design approach
for modelling shear force demand on tower walls supported on a transfer structure in regions of
lower seismicity Earthquake and Structures, 15(1), 97-111. doi:10.12989/eas.2018.15.1.097
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. (2017b). Analytical Modelling of
Podium Interference on Tower Walls in Buildings. Australian Journal of Structural Engineering.
doi:10.1080/13287982.2017.1396870
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. L. (2017a). Effects of podium
interference on shear force distributions in tower walls supporting tall buildings. Engineering
Structures, 148, 639-659. doi :https://doi.org/10.1016/j.engstruct.2017.06.075
Book Chapter

Yacoubian, M., Lam, N., & Lumantarna, E. (2018). Simplified Design Checks for Buildings
with a Transfer Structure in Regions of Low Seismicity, in: Design of Buildings and Structures
in Low to Moderate Seismicity Regions (Chapter 8, pp. 113-140): Chinese National Engineering
Research Centre for Steel Construction (Hong Kong Branch) at the Hong Kong Polytechnic
University.
Conference papers

Yacoubian, M., Lam, N., Lumantarna, E., Wilson, J. (2015). Performance of Limited Ductility
Reinforced Concrete walls in low to moderate seismicity regions. Paper presented at the 10th
Pacific Conference on Earthquake Engineering, Sydney, Australia.
Yacoubian, M., Lam, N., Lumantarna, E., Wilson, J.L. (2016). Seismic Performance of high-
rise buildings featuring a transfer plate taking into account displacement-controlled behaviour.
Paper to be presented at the Australian Earthquake Engineering Society 2016, Melbourne,
Victoria.
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. L. (2016). Effects of podium
interferences on shear forces distributions in RC walls supporting tall buildings. Paper presented
at the Australasian Structural Engineering Conference, Brisbane, Australia.
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. L. (2017c). Simplified design checks
for buildings featuring transfer structures in regions of lower seismicity. Paper presented at the
The 2017 World Congress on Advances in Structural Engineering and Mechanics (ASEM17),
Ilsan (Seoul), Korea.
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. L. (2017d). Seismic performance
assessment of structural walls supporting towers in tall buildings. Paper presented in the
Australian Earthquake Engineering Society’s conference, Canberra, Australia (24th- 26th of
November 2017).

vi
 TABLE OF CONTENTS

Abstract ..................................................................................................................................... i

Declaration .............................................................................................................................. iv

Acknowledgments .................................................................................................................... v

List of Publications ................................................................................................................. vi

Table of Contents ................................................................................................................... vii

List of Figures ........................................................................................................................ xii

List of Tables ...................................................................................................................... xxiv

Chapter 1. Introduction ....................................................................................................... 1

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

1.2. Problem statement ..................................................................................................... 3

1.3. Research objectives ................................................................................................... 6

1.4. Thesis organisation .................................................................................................... 8

Chapter 2. Literature Review ............................................................................................ 11

2.1. Seismic response behaviour of buildings featuring podiums .................................. 11

2.1.1. A review of code classifications of vertical irregularities in buildings ........... 11

2.1.2. Buildings featuring a setback in elevation ...................................................... 15

2.1.3. Buildings featuring a transfer structure ........................................................... 20

2.2. Diaphragm actions in seismic conditions ................................................................ 24

2.2.1. Diaphragm forces ............................................................................................ 24

2.2.2. Compatibility force considerations in the structural design practice .............. 28

2.2.3. Diaphragm modelling in numerical simulations ............................................. 30

2.2.4. Nonlinear response of floor slabs .................................................................... 34

2.2.5. Assumptions for diaphragm behaviour ........................................................... 37

2.3. Critical review of modelling the shear response behaviour of walls ....................... 40

2.3.1. Shear force-deformation relationship for structural walls ............................... 42

vii
 2.3.2. A post-processing approach to the shear assessment of walls ........................ 49

2.3.3. The Modified Compression Field Theory (MCFT)......................................... 50

2.4. Seismic load criteria for Australia ........................................................................... 51

2.4.1. Seismicity of Australia .................................................................................... 51

2.4.2. Design response spectrum for Australian seismic conditions ......................... 53

2.4.3. Displacement controlled behaviour of buildings ............................................. 55

2.5. Seismic design and performance assessment of buildings ...................................... 56

2.5.1. Response modification factors for seismic design of buildings ...................... 56

2.5.2. Seismic performance assessment of buildings ................................................ 59

2.5.3. Seismic Fragility assessment of buildings....................................................... 60

2.6. Chapter Summary .................................................................................................... 63

Chapter 3. Podium interferences in setback buildings ...................................................... 65

3.1. Introduction ............................................................................................................. 65

3.2. Analyses on 2D planar podium-tower sub assemblages ......................................... 66

3.3. Effects of diaphragm modelling on shear force demands on the tower walls ......... 75

3.4. Verification studies on 3D FE model of buildings .................................................. 76

3.4.1. Linear time history analysis ............................................................................ 80

3.5. Inelastic shear behaviour of coupled tower walls.................................................... 83

3.5.1. Inelastic model development ........................................................................... 84

3.5.1. Shear demands in nonlinear time history analyses .......................................... 88

3.6. Discrete modelling of shear failure progression in podium-tower buildings .......... 90

3.6.1. Calibration and validation of the numerical model ......................................... 90

3.6.2. 2D planar model of the sub-assemblage.......................................................... 92

3.7. Chapter Summary .................................................................................................. 100

Chapter 4. Analytical modelling of the effects of podium interferences in setback buildings

103

4.1. Introduction ........................................................................................................... 103

4.2. Explicit Stiffness-based approach ......................................................................... 103

viii
 4.3. A simplified approach for rapid estimation of the strutting force distributions .... 109

4.3.1. Estimate of the backstay podium forces ........................................................ 109

4.3.2. Strutting (compatibility) force distribution ................................................... 116

4.4. Verifications using a case study building .............................................................. 120

4.5. Effects of in-plane yielding of the backstay diaphragm ........................................ 124

4.5.1. Numerical model definition........................................................................... 125

4.5.2. Inelastic features of the backstay slab ........................................................... 127

4.5.3. Quasi-static pushover analyses on podium-tower sub-assemblage models .. 129

4.5.4. Nonlinear time history analyses .................................................................... 131

4.6. Effects of expansion/settlement joints at the podium-tower interface .................. 139

4.7. Chapter summary .................................................................................................. 146

Chapter 5. Analytical modelling of the effects of transfer plate flexibility in buildings 147

5.1. Introduction ........................................................................................................... 147

5.2. Shear force anomalies in tower walls above the TFL ........................................... 147

5.3. In-plane strains between tower walls above TFL .................................................. 150

5.4. Analytical modelling of the Flexibility Index ....................................................... 153

5.4.1. FI for connected tower walls with different span lengths ............................. 155

5.4.2. Validation studies on 2D model of a building with multiple (connected) tower

walls 157

5.5. Analytical modelling of Peak Rotational Demand on the building ....................... 158

5.5.1. Deformation modes of buildings featuring a transfer plate ........................... 162

5.5.2. Dynamic rotational coupling ......................................................................... 168

5.5.3. Evaluation of the Peak Rotation Demand (PRD) on the building ................. 170

5.6. Evaluation of the additional shear force demands on tower walls above the TFL 176

5.7. Verification studies on case study buildings ......................................................... 178

5.7.1. High-rise building ......................................................................................... 179

5.7.2. Medium rise building .................................................................................... 188

5.8. Chapter summary .................................................................................................. 193

ix
Chapter 6. Assessment of response modification factors of podium-tower type buildings
195

6.1. Introduction ........................................................................................................... 195

6.2. Methodology ......................................................................................................... 195

6.3. Description of the 2D podium-tower building models .......................................... 198

6.4. Construction of the nonlinear numerical model of the buildings .......................... 200

6.4.1. Nonlinear modelling of the shear and flexural response behaviour of walls 200

6.4.2. Verification study of the numerical wall model ............................................ 205

6.4.3. Numerical modelling of floor slabs and transfer plate .................................. 207

6.5. Nonlinear quasi-static analyses ............................................................................. 209

6.5.1. Ultimate limit state analysis .......................................................................... 210

6.5.2. Inelastic yielding progression in the building ............................................... 211

6.5.3. Preliminary assessment of the response modification factors ....................... 217

6.5.4. Effects of the lateral (pushover) force distribution pattern on the response
factors 223

6.6. Nonlinear dynamic analyses .................................................................................. 225

6.6.1. Eigenvalue analysis ....................................................................................... 226

6.6.2. Modelling damping ....................................................................................... 227

6.6.3. Ground motions selection criteria ................................................................. 228

6.6.4. Definition of the performance levels ............................................................. 234

6.7. Assessment of the seismic performance ................................................................ 235

6.7.1. Buildings featuring a transfer structure ......................................................... 237

6.7.2. Buildings featuring a setback ........................................................................ 245

6.7.3. Comparison of results for the building models ............................................. 251

6.8. Development of fragility curves ............................................................................ 252

6.9. Refined assessment of the response modification factors ..................................... 256

6.10. Chapter summary .............................................................................................. 260

Chapter 7. Recommendations for design practice .......................................................... 263

x
 7.1. Introduction ........................................................................................................... 263

7.2. Response modification factors for the seismic design of podium-tower buildings263

7.3. Verification study .................................................................................................. 265

7.4. Determination of the additional shear force demands in the tower walls above the
interface 271

7.4.1. Podium-tower buildings featuring a setback ................................................. 271

7.4.2. Buildings featuring a transfer plate ............................................................... 273

7.5. Modelling of floor slabs ........................................................................................ 275

7.6. Code classifications for buildings with setbacks ................................................... 275

7.7. Performance levels for podium-tower buildings ................................................... 275

Chapter 8. Conclusions and recommendations for future research................................. 277

8.1. Conclusions ........................................................................................................... 277

8.2. Limitations and recommendations for future studies ............................................ 281

References ............................................................................................................................ 283

Appendices ........................................................................................................................... 295

Appendix A- Summary of the ground motion records ..................................................... 296

Appendix B- Details of the 2D and 3D case study building models ................................ 303

Appendix C- Calculation of the podium and tower lateral stiffnesses ............................. 309

Appendix D- Summary of the key parameters used in the analytical model of Chapter 5
............................................................................................................................................... 316

Appendix E- Strutting force generation above the TFL ................................................... 320

Appendix F- Summary of RC wall specimens considered in the study ........................... 327

Appendix G- Summary of the results used in constructing the fragility Function........... 329

xi
 LIST OF FIGURES
Figure 1-1 Elevation view of a multistorey podium-tower building with a setback ................ 2

Figure 1-2 Tall podium-tower building featuring a transfer plate in Hong Kong (Wong, 2013)
....................................................................................................................................................... 2

Figure 2-1 Diagrammatic illustration of vertical irregularities as defined in ASCE/SEI 7-10

..................................................................................................................................................... 13

Figure 2-2 Definition of setback irregularity according to EC8 (CEN, 2004) ....................... 14

Figure 2-3 Scaled prototype of the setback building tested by Shahrooz & Moehle (1990b)
..................................................................................................................................................... 16

Figure 2-4 Idealisation of the response behaviour of the setback buildings using a 2DOF
system (Shahrooz & Moehle, 1987) ............................................................................................ 17

Figure 2-5 Definition of the backstay effect in a podium-tower building (PEER/ATC 72-1,
2010) ........................................................................................................................................... 18

Figure 2-6 A tall building with a basement (Rad & Adebar, 2009) ....................................... 18

Figure 2-7 desired inelastic response in buildings with setbacks ........................................... 19

Figure 2-8 Image of a high-rise building featuring a transfer plate (Wong, 2013) ................ 20

Figure 2-9 Effects of local deformations of the transfer plate on the shear force demands on
the walls above (Su & Cheng, 2009)........................................................................................... 21

Figure 2-10 Distribution of seismic damages of the building model (Li, 2005) .................... 22

Figure 2-11 Building model employed in a quasi-static (pushover) study (Mwafy & Khalifa,
2017) ........................................................................................................................................... 23

Figure 2-12 Total stiffness of the podium substructure including both the lateral and flexural
stiffnesses (Su, 2008) .................................................................................................................. 24

Figure 2-13 Deformation and compatibility force trends in a dual frame-wall system (Paulay
& Priestley, 1992) ....................................................................................................................... 25

Figure 2-14 Internal forces in inelastically responding two connected wall system (Avigdor
Rutenberg, 2004) ......................................................................................................................... 26

Figure 2-15 Moment-roof displacement response of the two-wall system and the resulting in-
plane forces in the connecting floors (Avigdor Rutenberg, 2004) ............................................. 26

Figure 2-16 Flexible slab model between connected walls (total of two walls) .................... 27

Figure 2-17 Effect of in-plane slabs stiffness on the shear response of connected walls

xii
(Rutenberg & Nsierie, 2006) ....................................................................................................... 28

Figure 2-18 Deep beam analogy for determining in-plane (bending and shear) actions in a floor
diaphragm (Moehle et al., 2010) ................................................................................................. 29

Figure 2-19 Stiffness degradation of the tested slab-column ensemble (Moehle & Diebold,
1985) ........................................................................................................................................... 31

Figure 2-20 Proposed effective slab width for strength (a) and initial stiffness (b) (Dovich &
Wight, 2005) ............................................................................................................................... 32

Figure 2-21 Modelling slab-core column assemblies in buildings from (PEER/ATC 72-1,
2010) ........................................................................................................................................... 33

Figure 2-22 Illustration of the varying effective slab width using FE simulations ................ 34

Figure 2-23 Lumped moment-rotation backbone to model inelastic actions and deformations
of the slab (PEER/ATC 72-1, 2010) ........................................................................................... 35

Figure 2-24 Illustration of fibre-based beam element modelling of reinforced concrete

members (Taucer, Spacone, & Filippou, 1991) .......................................................................... 36

Figure 2-25 Variation of the softened response of a reinforced concrete beam with increasing
number of integration points (Coleman & Spacone, 2001)......................................................... 37

Figure 2-26 Analysed building using solid elements to represent slab members (Henry, 2011)
..................................................................................................................................................... 37

Figure 2-27 The rigid diaphragm constraint (CSI, 2004)....................................................... 39

Figure 2-28 Schematic definition of conditions leading in the assumption of flexible

diaphragms (from ASCE (2010) Chapter 12) ............................................................................. 39

Figure 2-29 Diagrammatic illustration of the flexible, semi-rigid and rigid diaphragms
(PEER/ATC 72-1, 2010) ............................................................................................................. 40

Figure 2-30 Shear span ratio of multistorey walls and coupled walls (Krolicki et al., 2011) 41

Figure 2-31 Strain distributions at the base section of a wall (at various drift demands)
(Greifenhagen, 2006) .................................................................................................................. 42

Figure 2-32 Illustration of typical force-displacement curves for (a) force-controlled (b)
deformation-controlled response behaviour ................................................................................ 43

Figure 2-33 Illustration of the modelling of the implicit shear-flexure interaction for walls 44

Figure 2-34 Schematic illustration of modelling the uncoupled shear and flexural response
behaviour of walls ....................................................................................................................... 45

xiii
 Figure 2-35 Shear cracking in structural walls (artwork adopted from Krolicki et al. (2011))
..................................................................................................................................................... 46

Figure 2-36 multilinear shear constitutive relationship proposed by Kelly (2004) ............... 48

Figure 2-37 The Shear force-displacement capacity envelopes recommended by the ASCE/SEI
41-13 (2013) the term ∆/ℎ is the storey drift .............................................................................. 48

Figure 2-38 Example results of a shear assessment involving a step-by-step mapping of shear
demands against degrading shear capacity (Alwaeli et al., 2017) ............................................... 50

Figure 2-39 Modelling of cracked concrete under combined in-plane and shear stresses
(Vecchio & Collins, 1986) .......................................................................................................... 50

Figure 2-40 Spatial distribution of seismic hazard in the continent of Australia (Burbidge,
Leonard, Robinson, & Gray, 2012) ............................................................................................. 52

Figure 2-41 Differences in seismic intensities at increasing return periods for regions of high,
moderate and low seismicity (figure adopted from Paulay and Priestley (1992)) ...................... 53

Figure 2-42 Response spectrum in the acceleration, velocity and displacement formats
(Lumantarna, Lam, & Wilson, 2013) .......................................................................................... 54

Figure 2-43 Spectral shape factors for different site classes as given by AS1170.4:2007 ..... 55

Figure 2-44 Description of response modification factors ..................................................... 57

Figure 2-45 MSA approach (figure adopted from Baker (2015)) ........................................... 63

Figure 3-1 Backstay actions in a podium-tower building featuring a setback ....................... 66

Figure 3-2 Example podium-tower sub-assemblages ............................................................ 67

Figure 3-3 Strutting forces in the connecting floors and beams between the tower walls in
offset and centred podium-tower sub-assemblages ..................................................................... 68

Figure 3-4 Kinematic behaviour of tower walls in podium-wall sub-assemblage ................. 69

Figure 3-5 Influence of strutting forces on wall shear forces................................................. 69

Figure 3-6 Shear force distribution in tower walls ................................................................. 70

Figure 3-7 M/V ratio in tower walls in podium-tower sub-assemblage................................. 71

Figure 3-8 Podium-tower sub-assemblages with varying podium height .............................. 71

Figure 3-9 Correlation of wall shear force ratios with podium height ratios ......................... 72

Figure 3-10 Internal forces in tower walls for the different podium-tower sub-assemblages 72

Figure 3-11 Correlation of wall shear force ratios with podium eccentricity ........................ 73

xiv
 Figure 3-12 Behaviour of tower walls at an offset from the centre of the podium ................ 73

Figure 3-13 Results of analyses of sub-assemblage building models for podium height ratio of
0.32 (hp/hb = 0.32) ....................................................................................................................... 74

Figure 3-14 Results from sensitivity study for various diaphragm extents ............................ 76

Figure 3-15 3D rendering of case study building structure .................................................... 78

Figure 3-16 Floor plans of case study building structure ....................................................... 78

Figure 3-17 Wall shear force distributions of case study building structure.......................... 79

Figure 3-18 Wall shear force ratio of elastic case study building structure ........................... 79

Figure 3-19 Comparison between parametric study and results from case study buildings
[podium height ratio]................................................................................................................... 80

Figure 3-20 Shear force distribution of Set II walls, models with rigid diaphragms assigned
..................................................................................................................................................... 81

Figure 3-21 Shear force distribution of Set II walls, models with explicit diaphragm definition
..................................................................................................................................................... 81

Figure 3-22 Base shear of the suite of artificial records ........................................................ 82

Figure 3-23 Out-of-phase vibration of case study building ................................................... 83

Figure 3-24 Reinforcement details of the C-shaped core walls ............................................ 83

Figure 3-25 Numerical simplification technique to incorporate the primary and secondary
lateral load resisting system from (PEER/ATC 72-1, 2010 ) ...................................................... 84

Figure 3-26 Slab-wall/column representation for NLTH analysis ......................................... 85

Figure 3-27 Illustration of core the inelastic modelling of the core ....................................... 86

Figure 3-28 Elevation views of the analysed building models .............................................. 87

Figure 3-29 Geometric and reinforcement details of the secondary gravity columns
(immediately above the level of the interface) ............................................................................ 87

Figure 3-30 Rayleigh damping model employed in the NLTH analyses ............................... 88

Figure 3-31 Deformed shape at onset of flexural yielding of coupling beams ...................... 88

Figure 3-32 Base shear time histories for the analysed building models ............................... 88

Figure 3-33 Wall shear force distribution .............................................................................. 89

Figure 3-34 Strutting force time histories .............................................................................. 90

xv
 Figure 3-35 Simulated and experimental response of the wall specimens............................. 92

Figure 3-36 Elevation of inelastic case study building structure (VecTor2) ......................... 93

Figure 3-37 Push over curve for the sub-assemblage (shear response included) ................... 96

Figure 3-38 Considered response phases on the push over curve .......................................... 96

Figure 3-39 Normalised in-plane backstay forces in the main backstay diaphragm (at the
podium interface level)................................................................................................................ 96

Figure 3-40 Simulated damage in the four response phases .................................................. 97

Figure 3-41 Shear demand to capacity ratio for interior and exterior walls........................... 98

Figure 3-42 Stress distribution of the transverse reinforcement of the wall after stage III .... 98

Figure 3-43 Curvature distribution of interior wall within the vicinity of interface in response
stages I and II ............................................................................................................................ 100

Figure 4-1 Internal force distribution in the tower walls above the interface level ............. 104

Figure 4-2 The effect of the relative wall-coupling beam stiffness (KW/Kb) on the axial (in-
plane) coupling beam forces ..................................................................................................... 106

Figure 4-3 Analysed tower walls without the podium structure for determining the value of 𝛿 o
................................................................................................................................................... 107

Figure 4-4 Results from the parametric study on podium-tower sub-assemblage models with
different podium heights ........................................................................................................... 108

Figure 4-5 Variation of the normalised relative wall displacement with podium height ratio
................................................................................................................................................... 108

Figure 4-6 Results from the parametric study on building models with different (KT/KP) ratios
................................................................................................................................................... 109

Figure 4-7 Analytical represntation of the podium restraint on the tower structure ............ 111

Figure 4-8 Shear force distribution and shear reversal in the propped cantilever model of the
building (based on application of a single point load) .............................................................. 112

Figure 4-9 Analytical represntation of the interfacial restraint of the podium on the tower
structure ..................................................................................................................................... 113

Figure 4-10 Comparison of the backstay (reactive) forces .................................................. 114

Figure 4-11 Comparison of backstay forces from the predictive (analytical) model to the
results obtained from FE analyses ............................................................................................. 115

xvi
 Figure 4-12 Illustration of the assumed distribution of internal forces in the vicinity of the
interface level ............................................................................................................................ 116

Figure 4-13 Comparison of the analytical strutting (in-plane) force distributions to those
obtained from the FE simulations on 2D models with various hP/hb and KT/KP ratios ............ 119

Figure 4-14 Typical tower floor plan showing the effective width of the slab connecting the
examined walls .......................................................................................................................... 121

Figure 4-15 Normalised relative wall displacement ratio for set I and II walls .................. 122

Figure 4-16 Comparison of the strutting force distributions up the height of the building. 122

Figure 4-17 Comparison of FSTRUT distributions obtained from the analytical models ....... 123

Figure 4-18 Illustration of a typical construction cycles of a podium-tower (or base-tower)

(artwork adopted from (PEER/ATC 72-1, 2010) ...................................................................... 125

Figure 4-19 2D model of the analysed podium-tower sub-assemblage model and the mode
shapes of the first and second translational modes.................................................................... 127

Figure 4-20 Modelling of the backstay slab ......................................................................... 128

Figure 4-21 In-plane modelling parameters of the backstay slab ........................................ 128

Figure 4-22 Pushover (capacity) curves from simulations................................................... 129

Figure 4-23 Shear force demands on the core walls and the in-plane (strutting) forces in the
coupling beams.......................................................................................................................... 130

Figure 4-24 In-plane force- deformation hysteretic response of the interfacial slab (Model 1)
................................................................................................................................................... 132

Figure 4-25 Pinched hysteretic model defining the in-plane response behaviour of the backstay
slab ............................................................................................................................................ 132

Figure 4-26 Modified Takeda hysteretic model defining the in-plane response behaviour of the
backstay slab (Model 2)............................................................................................................. 133

Figure 4-27 Linear model defining the asymmetric (tension/compression) in-plane response
behaviour of the backstay slab (Model 3) ................................................................................. 133

Figure 4-28 A comparison of the base shear time histories obtained from analyses ........... 135

Figure 4-29 A comparison of the backstay (reactive) podium forces .................................. 136

Figure 4-30 Core shear force time histories above the interface level ................................. 137

Figure 4-31 In-plane forces in the coupling beams above the interface level ...................... 137

xvii
 Figure 4-32 Inelastic (flexural) yielding progression in the tower cores and coupling beams
................................................................................................................................................... 138

Figure 4-33 Comparison of the roof displacement time histories of the three models ........ 139

Figure 4-34 Illustration of a podium-tower building undergoing differential settlement and

typical details of the ‘floating bay’ ........................................................................................... 139

Figure 4-35 Structural simplification of a floating bay and internal actions resulting from the
differential settlement of the support ........................................................................................ 140

Figure 4-36 Bending moment distributions in the podium slabs before and after the settlement
(differential) of the tower .......................................................................................................... 141

Figure 4-37 Analytical backbone model representing the interfacial settlement/expansion

joints .......................................................................................................................................... 143

Figure 4-38 Results of analyse of 2D podium-tower sub assemblage models with different
types of slab-wall connections .................................................................................................. 145

Figure 4-39 Comparison of (a) the shear ratio (VInt / VExt )and (b) FSTRUT obtained from analyse
of 2D podium-tower assemblages with different podium-tower connectivity .......................... 145

Figure 5-1 2D model of the building featuring a transfer plate ........................................... 149

Figure 5-2 Displacement incompatibility between connected walls and the resulting strutting
(compatibility) slab force and wall shear force distributions .................................................... 149

Figure 5-3 Tower walls shear force distribution .................................................................. 150

Figure 5-4 Introducing FSTRUT, STRUT and TP .................................................................... 151

Figure 5-5 Angle of rotation at the base of the tower walls ................................................. 151

Figure 5-6 In-plane strains between walls and differential angle of rotation at their base TP
= TP1 - TP2................................................................................................................................ 152

Figure 5-7 Time-histories of In-plane strutting forces between walls and differential angle of
rotation at their base ∆𝜃TP= 𝜃TP1- 𝜃TP2 (record no. 29) .............................................................. 152

Figure 5-8 Shear force time-histories of the walls at different levels up the height of the tower
................................................................................................................................................... 153

Figure 5-9 Effects of 𝛼 r on strutting forces in between tower walls (analyses of records 29-36)
................................................................................................................................................... 154

Figure 5-10 Correlation of Flexibility Index with 𝛼 r ........................................................... 154

Figure 5-11 2D model of a building with tower walls spaced at different intervals ........... 155

xviii
 Figure 5-12 Peak strutting strains vs. ∆𝜃TP for the case study building ............................... 156

Figure 5-13 Sample time history results for (a) FSTRUT and (b) TP (rec Nos. 29, 30, 31 &40)
................................................................................................................................................... 156

Figure 5-14 Model of a building with multiple tower walls with different stiffness ........... 157

Figure 5-15 Strutting (compatibility) force demands in the floor slabs ............................... 158

Figure 5-16 Flexibility index values for the different tower walls....................................... 158

Figure 5-17 2D model of planted tower walls employed in the parametric study ............... 159

Figure 5-18 Angle of drift at position of CM in comparison with differential angle of rotation
at the base of the walls ∆𝜃TP = 𝜃TP1- 𝜃TP2 ................................................................................. 160

Figure 5-19 Tower drift profiles .......................................................................................... 161

Figure 5-20 The effect of floor slab coupling on the tower displacement ........................... 161

Figure 5-21 Lateral deformations in transfer structures under lateral loads ........................ 163

Figure 5-22 Analytical model of the building representing the uncoupled translational response
................................................................................................................................................... 164

Figure 5-23 Illustrations of the translational displacement of the tower and the podium .... 164

Figure 5-24 Variation of hx with the relative tower-podium stiffness ratio KT/ Kp .............. 165

Figure 5-25 Schematic representation of the rigid body rotation model .............................. 166

Figure 5-26 Rigid-body rotations of the tower block imposed by the podium .................... 167

Figure 5-27 Parametric study for determining the parameter 𝑘𝜃𝑇𝑃 .................................... 168

Figure 5-28 Effects of (br) and (RSDmax) on the Peak Rotational Demand (PRD) .............. 172

Figure 5-29 Sample results on a building model with br of 2.5 based on use of record nos. 29-
35 (Table A-1) ........................................................................................................................... 172

Figure 5-30 Analytical model for the estimation of PRD on a building .............................. 173

Figure 5-31 Peak Rotational Demand as a parameter to constrain in-plane strutting forces 175

Figure 5-32 Comparison of ∆𝜃TP and the PRD values obtained from the FE analyses of the
building models with varying transfer plate stiffness ............................................................... 176

Figure 5-33 Comparison of the shear force demands on the tower walls ............................ 178

Figure 5-34 Comparison of the values of VSLAB and FSTRUT .................................................. 178

Figure 5-35 Case study building ......................................................................................... 180

xix
 Figure 5-36 Mode shapes of the case study building (in the global x-direction) ................ 181

Figure 5-37 ∆𝜃TP vs. 𝜀 STRUT and the flexibility index ........................................................ 183

Figure 5-38 Sample results of slab strutting (in-plane) force and relative base rotation trends
for the walls (Records 29-35 with mean RSDmax=107mm) ....................................................... 185

Figure 5-39 Comparison between the maximum relative transfer plate rotations (for Walls 1
and 2) with the PRD computed for individual record ............................................................... 186

Figure 5-40 Shear force distribution above TFL from ETABS (Record nos. 29-35) .......... 187

Figure 5-41 3D render and tower floor plan of the medium rise building featuring a transfer
plate ........................................................................................................................................... 189

Figure 5-42 𝜀 STRUT versus ∆𝜃TP plot of Wall 2 (considering different ground motion records)
................................................................................................................................................... 190

Figure 5-43 Shear force distribution on tower wall (Wall 2) ............................................... 191

Figure 5-44 Comparison between the maximum relative transfer plate rotations (for Walls 2
and 3) with the PRD computed for individual record ............................................................... 192

Figure 6-1 The assessment framework and the structure of Chapter 6 ................................ 197

Figure 6-2 Elevation view of the analysed building models ................................................ 199

Figure 6-3 Illustration of the reduction in the shear span ratios of the tower walls above the
TFL............................................................................................................................................ 201

Figure 6-4 Collapse drift limits for walls with low shear span ratios (M/Vlw) ..................... 202

Figure 6-5 Diagrammatic illustration of inelastic modelling of the tower walls ................. 203

Figure 6-6 Force-deformation options for the wall subjected to lateral loads ..................... 204

Figure 6-7 verification of the simulated shear force-displacement plots ............................. 206

Figure 6-8 Inelastic numerical model of the floor slab ........................................................ 208

Figure 6-9 Illustration of the horizontal construction joints in a transfer plate (Wong, 2013)
................................................................................................................................................... 209

Figure 6-10 Pushover (capacity) curve of building TS-1 ..................................................... 212

Figure 6-11 Pushover (capacity) curve of building TS-2 ..................................................... 212

Figure 6-12 Pushover (capacity) curve of building SB-1 .................................................... 212

Figure 6-13 Pushover (capacity) curve of building SB-2 .................................................... 213

Figure 6-14 Mapping the local and global responses of walls in building SB-2 (a) total and

xx
shear response of Wall 1 showing the occurrence of the ultimate limit state (Wall 1) (b) concrete
stress-strain response of the outer most fibre (c) total and shear response of the central wall (CW)
................................................................................................................................................... 214

Figure 6-15 Mapping the local and global responses of walls in building TS-2 (a) Total
response of the critical wall (wall 2) (b) Total response of the central wall (c) monitored stress-
strain curves for outer most reinforcing bar embedded in the central wall in two storeys above the
TFL (d) monitored concrete (fibre) stresses and strains in tower Wall 1 & 2 (d) FSTRUT in the floor
slab connecting the central wall to Wall 2 ................................................................................ 214

Figure 6-16 Storey ductility demand distributions for the examined sub-frame models (a) TS-
1 (b) TS-2 (c) SB-1 (d) SB-2 ..................................................................................................... 216

Figure 6-17 Distribution of inelastic damages up the height of the building ....................... 217

Figure 6-18 Implications of strength provisions in transfer structures ................................ 219

Figure 6-19 Results from the sensitivity study performed on building TS-1 employing different
lateral loading shapes (a) bilinear base shear- roof drift curves (b) comparison of the effective
stiffness (c) comparison of the ductility parameter (𝜇 = 𝛿 ULS/ 𝛿 y) ......................................... 224

Figure 6-20 A schematic illustration of the result of the dynamic pushover analysis on a
building in the form of the intensity measure (IM) vs. the EDP (maximum inter storey drift limit)
................................................................................................................................................... 225

Figure 6-21 Mode shapes of the building models ................................................................ 226

Figure 6-22 Rayleigh damping model showing mass and stiffness proportional components
................................................................................................................................................... 228

Figure 6-23 Shear wave velocities used in the study that are representative of Australian soil
and geologic conditions (for Site Class D defined in AS 1170.4:2007) ................................... 230

Figure 6-24 Simulation of accelerogram on soil using DEEPSOIL .................................... 230

Figure 6-25 Comparison of rock, soil and code displacement response spectrum .............. 231

Figure 6-26 Acceleration and displacement spectra of the ground motion records used in the
study (plotted in logarithmic scale) ........................................................................................... 233

Figure 6-27 Summary of the ground motion records presented in terms of their respective
values of (a) PGA (b) PGV, (c) RSDmax (d) ordinate of the spectral acceleration evaluated at the
first period of vibration T1 = 2.0 seconds .................................................................................. 234

Figure 6-28 Dynamic pushover envelope (TS-1) ................................................................. 237

Figure 6-29 Dynamic pushover envelope (TS-2) ................................................................. 238

xxi
 Figure 6-30 Mapping first occurrences of limit states in the building (TS-1) ..................... 239

Figure 6-31 Mapping first occurrences of limit states in the building (TS-2) ..................... 240

Figure 6-32 Building TS-1 (a) IDA curves for various IMs (b) summary of the ISD values
defining occurrences of the three limit states ............................................................................ 241

Figure 6-33 Building TS-2 (a) IDA curves for various IMs (b) summary of the maximum ISD
values defining the occurrences of the three limit states ........................................................... 242

Figure 6-34 Building TS-1 (a) Computed PRD vs. ∆𝜃TP obtained from analyses (b) in-plane
strutting strains vs. ∆𝜃TP (c) comparison of the predicted magnitudes of FSTRUT (Eq. 5-24) to those
obtained from analyses .............................................................................................................. 244

Figure 6-35 Building TS-2 (a) Computed PRD vs. ∆𝜃TP obtained from analyses (b) in-plane
strutting strains vs. ∆𝜃TP (c) comparison of the predicted magnitudes of FSTRUT (Eq. 5-24) to those
obtained from analyses .............................................................................................................. 245

Figure 6-36 Dynamic pushover envelope (SB-1) ................................................................ 246

Figure 6-37 Dynamic pushover envelope (SB-2) ................................................................ 247

Figure 6-38 Mapping first occurrences of limit states in the building (SB-1) ..................... 248

Figure 6-39 Mapping first occurrences of limit states in the building (SB-2) ..................... 249

Figure 6-40 Building SB-1 (a) IDA plots for various IMs (b) summary of the ISD values
defining occurrences of the three limit states ............................................................................ 250

Figure 6-41 Building SB-2 (a) IDA plots for various IMs (b) summary of the ISD values
defining occurrences of the three limit states ............................................................................ 250

Figure 6-42 Normalised strutting forces (FSTRUT) in building SB-2 ..................................... 251

Figure 6-43 Fragility functions for building (sub-frame) models (a) TS-1 (b) TS-2 using
RSDmax as the IM ....................................................................................................................... 254

Figure 6-44 Comparison of the fragility curves of the building models derived using RSD max,
PGV and PGA as the IM ........................................................................................................... 255

Figure 6-45 Results from the fragility assessment of buildings ........................................... 256

Figure 6-46 Quantification of the R factor by mapping the demands (RSDmax) to the response
behaviour of the building (IDA curve) ...................................................................................... 258

Figure 6-47 Comparison of the R factors obtained from this study to those proposed in some
seismic codes............................................................................................................................. 259

Figure 6-48 A comparison between the AS1170.4:2007 assumed ductility factor (and

xxii
those obtained from the assessment study................................................................................. 260

Figure 7-1 Case study building ............................................................................................ 266

Figure 7-2 Comparison of the dynamic properties of the modified and full building models
................................................................................................................................................... 267

Figure 7-3 Acceleration response spectra of the records employed in the time history analyses
................................................................................................................................................... 268

Figure 7-4 Base shear time history (record no. 40) showing five key yielding stages......... 268

Figure 7-5 Inelastic yielding up the height of the building (greed highlight indicate that yield
strains are reached in some or all reinforcement bars in the section) ........................................ 269

Figure 7-6 Distribution of the maximum storey ductility demand up the height of the case
study building ............................................................................................................................ 269

Figure 7-7 Illustration of the effect of the component yielding on the developed sectional
strength ...................................................................................................................................... 270

Figure 7-8 Comparison of the maximum storey shear force envelopes obtained from analyses
and those of the design .............................................................................................................. 271

Figure 7-9 A step-by-step procedure for determining total shear force demands on tower walls
in setback buildings (above the interface level) ........................................................................ 273

Figure 7-10 A step-by-step procedure for determining total shear force demands on transferred
walls above the TFL .................................................................................................................. 274

xxiii
 LIST OF TABLES
Table 2-1 Summary of vertical irregularity classifications in design codes .......................... 14

Table 3-1 Geometric configuration of the examined sub-assemblages ................................. 67

Table 3-2 Material properties used (structural walls) ............................................................ 67

Table 3-3 Loads in case study building.................................................................................. 77

Table 3-4 Geometric details for analysed sets of walls .......................................................... 77

Table 3-5 Shear ratios for models with explicit diaphragms ................................................. 82

Table 3-6 Geometric and sectional details of the floor slabs ................................................ 85

Table 3-7 Reinforcement and modelling parameter details of the nonlinear model .............. 93

Table 4-1 Summary of the key parameters required for calculating FSTRUT (Stiffness-based
approach) ................................................................................................................................... 121

Table 4-2 Summary of calculations of the simplified method (introduced in Section 4.3) . 124

Table 4-3 Summary of design parameters of the main elements of the examined building model
................................................................................................................................................... 126

Table 4-4 Slab (in-plane) modelling parameters adopted in the inelastic study .................. 134

Table 4-5 Summary of the typical types of connections in a podium-tower building ......... 142

Table 5-1 Sectional properties of the walls .......................................................................... 157

Table 5-2 Key dynamic coupling model parameters of the base model shown in Figure 5-1
................................................................................................................................................... 174

Table 5-3 Floor load assignment .......................................................................................... 179

Table 5-4 Translational modes of the case study building (high-rise) ................................. 181

Table 5-5 Design details of Walls 1 & 2 .............................................................................. 188

Table 6-1 Geometric and reinforcement details of the floor slabs and the transfer plate .... 209

Table 6-2 Summary of the response modification factors for the analysed buildings ......... 221

Table 6-3 Vibrational periods of the building models ......................................................... 227

Table 6-4 Performance levels adopted in this study ............................................................ 236

Table 6-5 Median ISD values for the three limit states........................................................ 252

Table 6-6 Comparison of the R factors obtained from the dynamic and quasi-static analyses
................................................................................................................................................... 259

xxiv
 Chapter 1. INTRODUCTION

1.1. Background
Many, if not most, medium-rise and tall buildings feature elements of irregularity in their
geometry. The building layout is often dictated by architectural considerations that aim to
maximise the net ‘usable floor area’ of the building. More so, the soaring land prices in most
metropolitan cities around the world necessitate the prevalence of the economic factor in the
design of a building that is fifteen storeys or higher. As such, land and property developers are
more inclined to invest in buildings that serve multiple functionalities (mixed usages). Thus, a
building morphology that features a podium at the lower storeys has become a trending type of
building construction especially in Central Business District (CBD) areas in cities. The podium
is an augmented floor plan area at the base of a medium-rise or tall tower building. The lower
(podium) levels are often utilised for commercial purposes (e.g. shopping malls) which can
generate substantial return on investment for all stakeholders. The tower of a smaller footprint,
typically hosts residential apartments or office spaces. This configuration allows mixed
functionality (commercial/residential) of the building imposing different structural requirements
in the podium and tower storeys. Essentially, the commercial feature of the podium dictates the
need for a larger floor plan area and sometimes a wider spacing between the columns and walls
to minimise space obstructions in these storeys. The differences in the structural layouts in the
podium and the tower storeys can vary quite significantly in some buildings while these
(differences) can simply be restricted to the size of the floor plan.

A setback in elevation is a feature of a building with different floorplan areas in the tower and
the podium. Conventionally, some of the columns and walls in the podium are extruded up the
height of the building to support the tower above, whereas the remaining columns and walls
terminate at the interface between the two portions of the building. An example of a multistorey
podium-tower setback building is shown in Figure 1-1.

In some extreme scenarios, the building can feature an entirely different column/wall
arrangement in the tower and the podium. In such instances, it is common for the podium to be
supported by a frame system comprising widely spaced columns or walls, whereas closely spaced
structural walls, which can conveniently outline partitions between different apartments and
corridors, supporting the tower above. This form of podium-tower building construction features
acute discontinuities in the load path between the upper and lower portions of the building as the
tower walls terminate at the level of the interface (see Figure 1-2). To accommodate these
variations, a transfer structure is introduced at the level of the podium to redistribute forces from
the discontinued tower walls (or columns) to the supporting podium below. A transfer structure

1
in the form of a deep reinforced concrete plate is commonly employed in the construction of such
buildings (see Figure 1-2).

Figure 1-1 Elevation view of a multistorey podium-tower building with a setback

Figure 1-2 Tall podium-tower building featuring a transfer plate in Hong Kong (Wong,
2013)

2
 Most seismic design standards classify the described geometry of a podium-tower building
(with a setback or a transfer plate) as one that features an irregularity in its elevation.
Consequently, the performance of the building in the event of an earthquake is a matter of serious
concern. The poor seismic performance of this form of building construction has been verified in
some experimental and analytical investigations reported in the literature. These studies have
highlighted that significant damage can potentially be developed in some of the storeys in the
building which can substantially limit the lateral deformation and strength capacity of the building
in projected earthquake attacks. Despite this, there have been only limited guidelines to practising
engineers on how to mitigate (or limit) potential seismic induced damage to the building.

1.2. Problem statement

There is a consensus amongst engineers and researchers on the high vulnerability of a podium-
tower building for experiencing substantial damages under projected earthquake events. Yet, very
limited guidance is currently provided in design standards to instruct engineers on some
methodical descriptions or techniques for mitigating such risks. This knowledge gap grows wider
in regions of low-to-moderate seismicity (e.g. Melbourne, Hong Kong and New York) where the
paucity of seismic design guidelines (for the design of such buildings) is paralleled with the
limited detailing provisions stipulated in the codes of practice.

Some ten or so articles can be found in the literature to draw attention to the potential poor
seismic performance of a building incorporating the use of a transfer plate (like the one shown
Figure 1-2). Two such studies reported the occurrences of high seismic shear force concentrations
in some tower walls that are planted atop of a transfer plate. These concentrations have been
attributed to the unfavourable slab-wall interactions that occur in some tower storeys above the
transfer plate. Under lateral loads, the podium structure, represented by the transfer plate, imposes
different boundary conditions on the supported tower walls by virtue of the uneven distortions
(deformations) of the plate. Essentially, these differences (in the plate rotations at the base of the
planted walls) restrict the tower walls in one or two storeys above the plate from laterally
displacing by equal amounts. The resulting displacement incompatibilities between the walls
impose internal deformations and forces in the floor slabs as they attempt to ‘push’ and ‘pull’ the
tower walls to restore the uneven displacements in between the walls. The resulting forces in the
floor slabs are highly complex, particularly since their development is not mandated for the global
equilibrium of the building under lateral loads.

The very high shear force concentration in some tower walls would potentially mean that the
seismic performance of the building is aggravated even further when interferences by the transfer
plate are considered in the response. Although much of the insight on such occurrences have been
established in the literature (Su & Cheng, 2009; Tang & Su, 2015), there remains a pressing need

3
to systematically quantify these complex interactions by the transfer plate in order that their
influences on the shear force demand of the tower walls can be accounted for early in the design
process.

Interferences by the podium structure in setback buildings (like the one shown in Figure 1-1)
have not received similar attention amongst the research community. Seismic design guidelines
and provisions targeting this form of buildings (i.e. setback podium-tower type) have focused
mainly on the reactions of the base structure (i.e. the podium or the basement) on the walls of the
tower at and below the podium storeys. Particularly, the backstay effect (which will be discussed
in detail in Chapters 2, 3 & 4) has been shown to result in significant reversals in the shear force
of the tower walls at and below the level of the interface. Strictly speaking, these reactions will
lower the shear force intensities in the tower walls in the storeys below the level of the interface.
Thus, their implications would not have been critical to the shear force demand on the walls
(particularly above the interface). These reactive forces are essentially developed in the podium
slabs that are connected to the tower structure to maintain an adequate lateral force transfer
between the two portions of the building. Intuitively, these (reactive) forces would have developed
proportionately to the amount of restraint (and thus the interference) imposed by the podium on
the walls (or cores) of the tower. It is also possible (as will be shown later in this study) that a
non-uniform restraint is imposed on the tower walls for the case where the tower is placed at an
offset from the centre of the podium (which is typically the building form constructed). The
differences in restraint conditions imposed by the podium onto the tower walls are also likely to
result in displacement incompatibilities in between the walls similar to those investigated for the
case of a building featuring a transfer plate. As such, the slab-wall interactions and the consequent
concentrations of shear forces in some tower walls may also be prevalent in this form of buildings
(with a setback).

Central to this interference (by the podium) is the consequential generation of the internal, in-
plane, strutting (compatibility) forces in the floor slabs connecting the tower walls above the
interface (referring to the top level of the podium or the level of the transfer). The current state-
of-practice typically limits the role of a floor slab to: (i) support and redistribute gravity loads at
a storey level, and (ii) distribute seismic induced (inertia) forces to the main lateral load resisting
elements (walls, cores and frames). These functions of the floor slab, as described, are generally
well understood and have been at the core of the design practice for a long time. Yet as it stands,
a similar understanding is not weighed into the in-plane response behaviour of the floor slab. The
earlier described strutting forces in the floor slabs are in-plane in nature and are often ignored in
most Finite Element (FE) simulations which incorporate the use of the rigid diaphragm constraint
applied on the floor slabs of the building. This common practice restricts the in-plane forces (and
strains) from developing in the floor slabs of the building. Thus, these strutting (in-plane) slab

4
forces and the resulting shear force concentrations in some tower walls may be misrepresented in
practice. The PEER/ATC 72-1 (2010) and TBI (2017) guidelines instruct engineers against the
use of the rigid diaphragm constraint in the floors at and below the level of the interface (in a
setback building). However, to the authors’ best knowledge, similar provisions are not mandated
for other storeys where the consequences of these internal actions are pronounced.

Another recurring issue in podium-tower buildings (and irregular buildings in general) is the
adequacy of use of the force reduction factor (R factor for short) in the seismic design of the
building. Most seismic design codes and guidelines adopt the force-based approach in designing
buildings for earthquake loads. This approach is founded on the assumption that the building will
trade seismic strength (forces) with ductility (displacements and deformations) as the building
responds inelastically to the earthquake. Accordingly, the seismic force demands on the building
are reduced proportionately to a code-specified R factor. These response modification factors
(which constitute the R factor) are inherently generic and highly dependent on the yielding
mechanism sustained by the building when subjected to earthquake excitations. Some design
codes recognise the potential of the poor seismic performance of irregular buildings and
accordingly mandate certain reductions to be applied to the value of the code specified R factor
for use in the design of such buildings (e.g. Eurocode 8 (2004)). Other code provisions require
the podium structure to remain virtually elastic in projected seismic events (e.g. TBI (2017)).
More notably, for the case of buildings with transfer plates (or beams), ASCE/SEI 7-10 (2010)
recommends the design of the podium structure (inclusive of all the columns, walls and the
transfer plate) for a much higher seismic design force. The objective of this provision is to ensure
that no (or very limited) nonlinear deformations are experienced in the podium storeys. This is
somewhat contrary to the observations reported in some experimental studies on buildings
incorporating the use of a transfer plate. These studies suggest that most of the seismic-induced
damages have been observed above, and not below, the podium interface level. The earlier
presented discussions on the potential concentration of shear forces in the tower walls above the
interface level (in buildings with setback and transfer plates) are in line with the experimental
observations and further suggest that the critical storeys in such buildings are in fact those
immediately above the podium level.

It is further speculated that the ASCE/SEI 7-10 provision can, in fact, aggravate the
disproportionate distribution of both the lateral strength and stiffness in the building. This is the
case since the higher design forces allocated to the podium storeys may result in proportioning
larger structural sections (larger columns and thicker walls) which in turn increase the stiffness
along with the strength of the podium storeys. It should be noted that the provision (of ASCE/SEI
7-10) is based on the capacity design philosophy which also proportions larger strength reserves
to structural components in a building that are likely to experience limited-ductile (or brittle)

5
response behaviour under earthquake excitations (e.g. shear-critical walls).

The earlier presented discussions suggest that further investigations are required to evaluate
the adequacies of use of the response modification factors (the R, ductility and overstrength
factors) for podium-tower type buildings. Particularly, the effects of the potential concentration
of the seismic-induced damages in the tower storeys on the assumed distribution (and magnitude)
of the design forces and the deformation capacity of the building will need to be explicitly
evaluated to improve the seismic resilience of the building.

1.3. Research objectives

The main thrust of this thesis is channelled towards investigating the effects of the interference
by the podium structure on the seismic shear demand and the response behaviour of the walls
supporting the tower above the podium structure. Two forms of building constructions featuring
a podium will be examined:
i- Buildings featuring setbacks
ii- Buildings featuring a transfer plate structure

The two building types constitute a large portion of the building stock of cosmopolitan cities
in regions of low-to-moderate seismicity (e.g. Melbourne and Hong Kong). Nevertheless, the
study in large, is not intended to be exclusive for a specific seismic region, since the effects of the
podium on the response behaviour of the building will be examined.

The following is a list of the main objectives of this study:

1. Investigate Podium interferences in setback buildings

A thorough investigation will be undertaken to examine the shear force demand on the tower
walls above the podium interface. Several parametric studies will be conducted on buildings
(including the use of simplified models) to investigate the parameters influencing the seismic
response behaviour of the building and the shear force demand on the tower walls. As mentioned
earlier, most studies have investigated the in-plane force transfer (backstay effect) in the lower
storeys (podium or basement), but no attention has been devoted to anomalous increase in the
shear force in the tower walls above the level of the podium. Accordingly the in-plane response
behaviour of the link elements (floor slabs or beams) will have to be monitored in the critical
storeys above the level of the podium. A special attention will be cast on the influences of the
restraint (boundary conditions) imposed by the podium on the displaced shapes of the tower walls
above the interface level. This study will also examine the applicability of the conventional rigid-
diaphragm constraint on the floor slabs. The adverse effects of the podium structure on the
inelastic response will be illustrated to signal potential anomalies in the response behaviour of the
building.

6
 2. Develop analytical modelling tools for quantifying the effects of the podium in setback
buildings

Simple analytical modelling techniques will be developed to quantify the effects of

interferences of the podium structure on the building. The primary goal here is to model the
additional shear force demands that are imposed on some tower walls (above the interface level)
as a result of the internal (strutting) force transfer in between the walls by the connecting slab (or
beam). As mentioned earlier, these interferences (by the podium) are highly complex and are in
most cases misrepresented (or ignored) in the design of the building. The developed tools will
thus allow the structural engineer to make a better-informed design decisions particularly for the
tower walls of the building.

Moreover, the effects of the inelastic yielding of the interface slab (the slab connecting the
tower to the podium) will be evaluated with specific attention cast on the potential presence of
construction joints between the podium and the tower. Similarly, the effects of settlement joints
(or shrinkage strips) at the interface between the podium and the tower on the degree of the
restraint (interference) provided by the podium structure will be investigated analytically.

3. Develop analytical modelling tools for quantifying the effects of the podium in transfer
structures

The effects of the podium structure represented by the transfer plate on the shear demand of
the tower walls will be investigated. An analytical model will be developed to quantify the
intricate ‘boundary conditions’ imposed by the out-of-plane distortions of the transfer plate on
the walls of the tower. The internal strutting forces that are developed in between the tower walls
will be correlated with the rotations of the transfer plate at the base of the walls. The development
of a closed-form solution for the maximum magnitudes of the internal strutting floor slab forces
is central to quantifying the additional shear force demand (and concentrations) that have been
observed in the previous studies (as discussed earlier). The accuracies of the developed model
will be verified by numerical simulations performed on 2D and 3D building models.

4. Assess the seismic performance and the response modification factors of podium-tower
buildings

A systematic assessment of the response modification factors will be carried out for buildings
featuring podium structures (i.e. transfer structures and buildings with setbacks). The assessment
necessitates the development of representative nonlinear numerical models of the buildings.
Particular attention will be placed on the modelling of the potential limited-ductile shear response
behaviour of the tower walls above the interface level as a consequence of the investigated
interference by the podium. Nonlinear analyses (quasi-static and dynamic) will be performed to
evaluate the failure patterns, the distributions of the seismic demands and the overall capacity of

7
the building under rare and very rare earthquake intensities. Seismic fragility curves will be
constructed for the buildings to assess the seismic vulnerability of the buildings for a wide range
of seismic demands. Results from a large number of nonlinear time history analyses performed
on the building models along with those obtained from the fragility analyses will then be
combined to evaluate the maximum response modification factors (the R, the ductility factors)
that can be sustained by podium-tower buildings. The values of these factors will also be
compared to the values recommended in some design codes to assess their appropriateness for
use in the design of the building. Based on the results of the study, practical recommendations
will be derived that aim to improve some of the inconsistencies in the current design codes
particularly those pertaining to the response modification factors.

1.4. Thesis organisation

This thesis comprises of eight chapters. The content of this research is organised as follows:

Chapter 1 is an introductory chapter that provides a general description of the buildings

incorporating the use of a podium and a background on some areas of concern in their seismic
response behaviour.

Chapter 2 provides a review of the relevant literature on the following topics: 1) the seismic
response behaviour of the buildings featuring a podium, including a review of some code
specifications for determining vertical irregularities in buildings, 2) diaphragm actions under
seismic conditions which is mostly invested in reviewing previous research on compatibility
forces in the floor slabs, 3) a brief overview on the shear response behaviour of walls and a listing
of the different modelling techniques for incorporating the shear flexibility and mechanism in the
modelling of the wall, 4) the seismic loading criteria for the study region namely Australia, 5) a
summary of seismic design requirements for buildings and the state-of-the-art approaches for
assessing their seismic performance.

Chapter 3 investigates the podium-tower interactions in setback buildings of different heights

and configurations. Notably, this chapter outlines an investigatory study on the differential
(uneven) restraints that are imposed by the podium on the walls (and cores) of the tower above.
The positioning of the tower relative to the podium is central to the determination of the extent of
the (differential) restraint by the podium. Parametric studies have been performed on 2D planar
building models to shed light on some key parameters affecting these interferences (by the
podium). The generation of the strutting (compatibility) forces in the floor slabs above the
interface will be shown to result in significant shear force concentrations in the tower walls in
these storeys. The increase in the shear force demand on the tower walls is significant and may
(in most cases) exceed the design capacity of the wall.

Importantly, shear force anomalies in the tower walls (as reported in this chapter) may not

8
have been reported in conventional design practices. Accordingly, the response behaviour of a
building (with a setback) might be misinterpreted as one that is governed by a ductile-flexural
response which is in stark contrast to the shear-governed conditions that are explored in this
chapter (above the level of the interface).

Chapter 4 introduces an analytical framework for modelling and quantifying the intricate slab-
wall interactions in the tower floors of a building featuring a setback. A stiffness-based approach
is presented to resolve the strutting (compatibility) forces that are developed in the floor slabs
above the podium interface level. The framework is based on evaluating the local force
equilibrium in the walls of the tower that are subjected to different amounts of restraint by the
supporting podium structure. Interferences by the podium structure on the shear response
behaviour of the walls are also analysed by making a propped-cantilever analogy to quantify the
restraint (backstay) forces and the consequent internal forces developed in the links above the
podium interface level. The inelastic damage localisation in the podium slabs and their modelling
at the interface between the podium and the tower are examined to illustrate the potential effects
on the local and global response behaviour of the building. Furthermore, the effects of detailed
expansion joints in these locations (interface zone) are explored in the light of the shear force
anomalies (investigated in Chapter 3) that characterise the response behaviour of this form of
building construction.

Chapter 5 investigates the adverse interferences by the transfer structure on the response
behaviour of the supported tower walls. A simplified approach is derived for quantifying the
effects of flexibility (of the transfer structure) on the adverse shear concentrations that occur
above the level of the podium. The analytical model makes use of the displacement-controlled
principle to establish a closed-form solution (using only a few of parameters) for estimating the
maximum magnitudes of the internal forces that are expected to be developed in the floor slabs
of the tower.

Chapter 6 presents the results of an assessment study to evaluate the seismic performance of
podium-tower buildings. Detailed descriptions of the numerical model and the assessment
formwork are also presented. Results from numerous nonlinear static and dynamic analyses are
examined to develop a better understanding of the seismic performance of the buildings and
particularly the performance levels that are applicable to this form of building constructions. The
results from the vulnerability assessment are also presented in the form of fragility functions (or
curves) to highlight the influences of the podium structure on the seismic vulnerability of the
building models. A framework is presented to estimate the response modification factors
(ductility and R factors) of the buildings. Results are then benchmarked against code-specified
values stipulated in some seismic standards to assess their adequacies.

9
Chapter 7 outlines the key recommendations from the study presented in chapters 3 to 6.
Particularly, a methodology is proposed for determining the response modification factors
(mainly the R factor) for a podium-tower building. An example case study is presented to
showcase the potential usage of the approach. Furthermore, frameworks are developed to assist
design engineers in estimating the additional shear forces that are imposed on the tower walls in
the case of setback buildings and buildings featuring a transfer plate. These frameworks are
presented in a simple, yet sufficiently detailed, flow charts. Recommendations for the use of
certain floor slab assumptions and performance levels are also summarised in this chapter.

Chapter 8 summarises the key findings of the study on the podium-tower structures. The
limitations of the study are explicitly stated and recommendations for future studies are
accordingly outlined.

10
 Chapter 2. LITERATURE REVIEW
This chapter presents a review of the literature on issues relating to the seismic response
behaviour, the seismic performance and the design of podium-tower type buildings (Section 2.1).
The bulk of the reviewed literature highlights the critical role of the in-plane behaviour of the
floor slabs in the seismic response behaviour of buildings with configurations similar to those
examined in this study. The key features of these responses (of the floor slab) and their modelling
are outlined in Section 2.2. Past studies also shed light on the potential occurrence of shear-critical
conditions in the walls supporting the building. As such, a review of the state-of-the-art techniques
for modelling the shear response behaviour of structural walls is presented in Section 2.3. Seismic
loading considerations for the design of buildings in Australia (which is the study region) are
specified in Section 2.4, and conventional procedures adopted for the seismic design and the
performance assessment of buildings are reviewed in Section 2.5.

2.1. Seismic response behaviour of buildings featuring podiums

As discussed earlier (in Chapter 1), a podium-tower type building is classified by most design
codes as one featuring irregularity in its elevation. The following section outlines an overview of
the general code classifications of buildings with vertical irregularities (Section 2.1.1). Particular
attention is cast on the implications of building irregularity on the seismic response behaviour of
buildings featuring a setback (Section 2.1.2) and a transfer structure (2.1.3).

2.1.1. A review of code classifications of vertical irregularities in buildings

The American seismic design standard ASCE/SEI 7-10 (2010) outlines a detailed
classification of the types of vertical irregularities that are commonly featured in buildings. There
are some five distinct clusters of vertical irregularities which are outlined in table 12.4-2 of
ASCE/SEI 7-10 standard (summarised in below and in Table 2-1). The stiffness irregularity, or
what is commonly referred to as ‘soft-storey’ irregularity, is defined when an abrupt variation in
the lateral stiffness exists between adjacent storeys in a building. Conversely, the building is
classified with features of strength irregularity (weak-storey irregularity) when differences in the
lateral strengths between adjacent storeys amount to 65%-80%. It is not uncommon for a building
to exhibit both strength and stiffness irregularity up its height (i.e. occurring simultaneously).
Mass irregularity pertains to buildings where the seismic mass distribution up the height of the
building is non-uniform. ASCE/SEI 7-10 classifies a building featuring single or multiple
setbacks in their elevation in a separate class defined as ‘geometric’ irregularity.

Buildings that incorporate discontinued columns and/or walls up their height are classified (by
default) as being vertically irregular. The latter two classes define the type of building
configurations that are examined in this study (buildings featuring setbacks and transfer

11
structures). FEMA P750 (2009) outlines a diagrammatic illustration of the various types of
vertical irregularities in a building according to the ASEC/SEI 7-10 classifications which is
reproduced herein in Figure 2-1.

The European seismic code (CEN, 2004), classifies a building as irregular in elevation when
some features of the building are mapped against a listing of descriptions that are somewhat
similar, yet broader, to those defined by the ASCE/SEI 7-10. The three main clusters of vertical
irregularity in buildings (referring to stiffness, strength and mass) are similarly described in the
European standard. Distinctively, Eurocode 8 outlines detailed requirements for the classification
of setback buildings (see Figure 2-2). Transfer structures are also categorised, by default, as
irregular per Section 4.2.3.3 of the Eurocode 8.

In the 1993 edition of the Australian seismic code AS 1170.4 (AS 1170.4, 1993) building
irregularity classifications similar to those found in ASCE 7-10 have been reported (in Section
2.9.3 of AS 1170.4:1993). This code provision did not appear to have survived the later (current)
edition of the AS 1170.4 (Standard Australia, 2007). Similar descriptions of building irregularity
(in elevation) have been stipulated in the Uniform Building Code (UBC, 1997). Table 2-1 outlines
a comparison of the general code classifications of vertical irregularities in the building.

12
Figure 2-1 Diagrammatic illustration of vertical irregularities as defined in ASCE/SEI
7-10

13
 Figure 2-2 Definition of setback irregularity according to EC8 (CEN, 2004)

Table 2-1 Summary of vertical irregularity classifications in design codes

Seismic Design Code

Type of
irregularity (in ASCE/SEI 7-10 Eurocode 8 AS 1170.4:1993
elevation)
𝑘𝑖 < 0.7 𝑘𝑖+1 or
𝑘𝑖+1 + 𝑘𝑖+2 +𝑘𝑖+3
𝑘𝑖 < 0.8 ( )
3
And for extreme cases (see Figure
Stiffness 2-1 for schematic illustrations) 𝑘𝑖 < 𝑘𝑖+1 𝑘𝑖 < 𝑘𝑖+1
𝑘𝑖 < 0.6 𝑘𝑖+1 or
𝑘𝑖+1 + 𝑘𝑖+2 +𝑘𝑖+3
𝑘𝑖 < 0.7 ( )
3

𝐹𝑖 < 0.80 𝐹𝑖+1

Strength And for extreme cases 𝐹𝑖 < 𝐹𝑖+1 𝐹𝑖 < 𝐹𝑖+1
𝐹𝑖 < 0.65 𝐹𝑖+1

𝑀𝑖 > 1.5 𝑀𝑖+1 or 𝑀𝑖 > 𝑀𝑖+1 or 𝑀𝑖 > 𝑀𝑖+1 or

Mass
𝑀𝑖 > 1.5 𝑀𝑖−1 𝑀𝑖−1 𝑀𝑖−1

14
 As shown in Figure
Geometric 𝐿𝑖 > 1.3 𝐿𝑖+1
2-2

Offset in the lateral load resisting

If discontinuity
In-plane system is greater than the length of If a discontinuity in a
exists in a
discontinuity the element above or below the member exists
moment frame
discontinuity
Legend
𝑀𝑖 , 𝑀𝑖+1 , 𝑀𝑖−1 is the mass of the storey, the storey above and below respectively.
𝐾𝑖 , 𝐾𝑖+1 is the lateral stiffness of the storey and the storey above respectively.
𝐹𝑖 , 𝐹𝑖+1 is the total lateral strength of the storey and the storey above respectively.
𝐿𝑖 , 𝐿𝑖+1 is the horizontal dimension of the floor plan (in the direction of loading) and the plan above
respectively.

2.1.2. Buildings featuring a setback in elevation

A setback is a recess in the floor plan area of one, or more, storeys in a building. The resulting
irregular distribution of lateral stiffness up the height of the building is a major cause for concern
when the building is subjected to earthquake ground excitations. Shahrooz and Moehle (1990a)
conducted a series of experimental studies on scaled prototype moment frames with a setback at
the mid-height of the building (see Figure 2-3).

15
 Figure 2-3 Scaled prototype of the setback building tested by
Shahrooz & Moehle (1990b)

The main objective of their investigation was to evaluate the appropriateness of the use of
common code descriptions for characterising a building as being irregular (e.g. the lateral stiffness
of the storeys above and below the setback). Interestingly, their observations did not find a strong
premise for some code provisions for the seismic design of the building. For instance, the use of
response spectrum analysis approach (by modal superposition) is reported (in the aforementioned
study) not to have any added value over the (much simpler) equivalent static approach in
predicting seismic demands on the building. A simple rational modelling approach has been
developed for evaluating and proportioning the extent of the seismic damages on the tower
(above) or the podium (below). As such, the response behaviour of the building has been
simplified to a two-degree-of-freedom-system (as shown in Figure 2-4). The effective lateral
stiffness of the lower (below the setback) and the upper storeys are quantified as k1 and k2
respectively. Analyses performed on the building are much simplified when this substitute
structure (the two-degree-of-freedom system) is employed to model the response behaviour of the
building. The relative displacement of the upper and lower portions of the building has been
identified as a key parameter that characterises the potential accumulation of seismic damages in
the above (or below) structure. The trends shown in Figure 2-4 indicate that the most adverse
building configuration (where significant seismic damages are to be anticipated) is when the
relative stiffness ratio (i.e. k2/k1) is low. Wood (1992) reports similar observations on the response
behaviour of setback frames.

16
 Figure 2-4 Idealisation of the response behaviour of the setback buildings
using a 2DOF system (Shahrooz & Moehle, 1987)

Interactions of flexible tower structures with a stiff (lower) sub-structure have been first
examined in early research by Baven-Pitchard et al. (1983). Almost a quarter of a century later,
Rad & Adebar (2009) further investigated the effects of these interactions particularly on the
response behaviour of the tower walls at the base storeys of the building. Although their
investigations were not invested on setback buildings, the fundamentals of the seismic
interactions between a slender tower and a stiff base have been examined (which is analogous to
the anticipated interactions in podium-tower building featuring setbacks as will be demonstrated
in Chapters 3 & 4).
The referenced study focused on the backstay forces that are generated at (and below) the
interface between the two portions of the building. Backstay forces are the reactions exerted by a
podium or an undergrounded (basement) structure on the lateral load resisting members (cores
and walls) of the building (see Figure 2-5 for illustrations). The stiffness contributions by the
basement slab, basement walls and the retained soil have been reported to be the key parameters
in determining the extent of the force reversals (reactions) experienced by the cores and walls of
the tower (refer Figure 2-6). Interestingly, their observations were similar to those reported in the
earlier study by Shahrooz and Moehle (1990b) where the relative stiffness of the lower portion of
the building was also found to be the key parameter controlling the extent of the damage
experienced by the building in an earthquake. Significant reductions in the backstay forces have
been observed when the stiffness of the tower cores degrades when subjected to higher intensity
ground shaking. This also indicates that the reaction (backstay) is a function of the relative (as
opposed to the absolute) stiffness of the tower and the basement sub-structure.

17
 Figure 2-5 Definition of the backstay effect in a podium-tower building
(PEER/ATC 72-1, 2010)

The in-plane response of the basement slab is a governing parameter for predicting the
magnitude of the shear reversal at the base of the tower. This (in-plane) response behaviour is
highly complex particularly when the effects of cracking of the basement slab are considered in
the analysis of the building. Rad and Adebar (2009) hypothesise that the transfer of forces
between the upper and lower portions of the building is initiated by the in-plane (compression-
tension) stresses that are developed in the basement floor diaphragms at the interface between the
tower and the sub-structure. The study presented in Chapter 4 provides further investigations on
the modelling of this force transfer in podium-tower type buildings.

(a) Typical building (high-rise) (b) Cross-sectional view of the building

Figure 2-6 A tall building with a basement (Rad & Adebar, 2009)
Seismic design implications of the backstay effect in podium-tower building featuring a

18
setback are outlined in the PEER/ATC 72-1 (2010) and the TBI (Tall Buildings Initiative) (2017)
technical guidelines. The provisions are mostly focused on the seismic design of buildings with a
sub-surface (basement) structure. Recommendations presented in these guidelines suggest that
the stiffnesses of the sub-soil structure (basement walls, slab and the retained soil) will need to be
varied from low to high values in order that a more realistic estimate of the shear reversal at the
base of the tower walls can be made (refer Figure 2-5). With this approach, the effects of the
basement structure would be ‘bracketed’ by finding the maximum and minimum magnitudes of
the reversal. According to the guidelines, the in-plane stiffness of the basement slabs will need to
be accurately defined into the numerical model of the building (as will be discussed in Section
2.2). Further requirements for the seismic performance of the building are stipulated in the above-
referenced guidelines. The listed recommendations are based on capacity design philosophy
wherein certain components of the building are required to yield at specific locations to dissipate
the input seismic energy. As such, the design of the building will need to ensure that yielding
progression up the height of the building follows the conventional pattern where flexural yielding
is concentrated at the base of the core walls and at the ends of the coupling beams in the tower
(as illustrated in Figure 2-7).

(a) Cantilever wall (b) Coupled walls

Figure 2-7 desired inelastic response in buildings with setbacks

More importantly, the guidelines require that flexural yielding should not occur outside these
preferred locations (highlighted in green in Figure 2-7) and that the walls (and columns) of the
podium should remain elastic throughout the response. One way to achieve this requirement is by
proportioning a large strength reserve in the design of the structural components of the podium
structure supporting the tower (as will be discussed in Section 2.5).

19
 2.1.3. Buildings featuring a transfer structure

Buildings incorporating the use of a transfer structure are commonly found in regions of lower
seismicity. A transfer structure is a structural member (a plate, girder or truss) with a function of
redistributing gravity and lateral forces from the discontinued (upper) walls and columns to the
lower supporting walls (and columns). As in the case of setback buildings, a transfer structure is
introduced at a single or multiple storeys to accommodate changes in the floor plan up the height
of the building. Example image of a high-rise podium-tower building featuring a transfer plate is
given in Figure 2-8.

Figure 2-8 Image of a high-rise building featuring a transfer plate (Wong, 2013)
Research on transfer structures in seismic loading conditions has generally aimed at
qualitatively assessing the seismic performance behaviour of the building. In seismic conditions,
the drift and the shear force demands up the height of the building can vary considerably as a
result of interactions by the transfer structure. Su and Cheng (2009) demonstrated the
unfavourable interference of the transfer plate on the shear force demands on the tower walls.
Shear concentrations in the form of spikes in the shear force diagrams of some walls have been
observed above the transfer floor level. These shear concentrations have been associated with the
local deformations of the transfer plate (Figure 2-9). The central core wall (in Figure 2-9) is
subjected to different deformations at the level of the transfer plate compared with those
experienced by exterior shear walls (on its left and right of the central wall). As a result,

20
compression (and tension) forces can be developed in the link members (slabs or beams) that
ultimately redistribute the seismic shear forces between the connected walls. Su and Cheng (2009)
expressed these shear force concentrations by proportioning the maximum wall shear forces to
the average shear force sustained by the walls above the transfer plate. This ratio was identified
as the shear concertation factor (SCF). SCF was examined in view of the deformation
incompatibilities between the central and exterior walls. As such, higher SCF is to be anticipated
as the displacement ‘gap’ between the walls is widened (Su & Cheng, 2009).

Figure 2-9 Effects of local deformations of the transfer plate on the shear force
demands on the walls above (Su & Cheng, 2009)

A thick transfer plate is sometimes assumed to be infinitely rigid against in-plane and out-of-
plane deformations. In practice, this assumption is favoured as it prompts substantial savings on
computational costs and analysis runtimes. Past studies on transfer structures like those reported
by Zhitao (2000) and Qian and Wang (2006) have adopted this assumption (of modelling the plate
as an infinitely rigid entity). However, the study by Su (2008) and Su and Cheng (2009)
emphasised some key shortcomings when making this assumption. One such issue is the loss of
the interaction of the transfer plate flexibility on the drift and shear force demands on the tower
walls (as described earlier).

Deformations at the transfer floor level are primarily attributed to the local out-of-plane
(bending) distortions of the transfer plate. However, additional distortions of the tower structure
(as a whole) have also been observed when considering the flexibilities of the supporting podium
structure. These flexibilities have been mostly manifested by the axial push-pull actions of the
podium columns and walls as the tower structure (above the podium and the transfer plate)
laterally displaces under the earthquake (Su, 2008). Interactions by the transfer plate are not
necessarily exclusive to lateral loads. Gravity-induced deformations of the transfer plate have also
been shown to result in similar shear force concentrations on the tower walls as examined by Tang

21
and Su (2015).

Interferences by the transfer plate have also been experimentally investigated by Li (2005) and
Li et al. (2006). A shake-table test on 1:20 scale model of a 34 storey building supported on a
transfer plate has given substantial insight into the seismic performance behaviour and damage
propagation in the building. Limited ductile response behaviour of the lateral force resisting
elements has exacerbated damage to the vicinity of the transfer floor (see Figure 2-10).

(a) Damage distribution up the height (b) Physical manifestation of damage at

of the building (damaged structural the vicinity of the transfer plate and up
members are highlighted in red) the height of the tower
Figure 2-10 Distribution of seismic damages of the building model (Li, 2005)

According to Li et al. (2006), the wholesale collapse of the building when subject to high-
intensity shaking (roughly corresponding to an earthquake intensity with a 2% probability of
exceedance in 50 years) is not anticipated. However significant reduction in the lateral stiffness
of the tower block is to be expected (in the order of 76% as observed from the shake-table
experiments). A recent analytical assessment study by Khalifa (2015) and Mwafy and Khalifa
(2017) revealed similar seismic damage concentrations and patterns in the walls above the transfer
floor level (see Figure 2-11(b)).

22
 (b) The analytical model showing
(a) High rise building featuring a damage distribution up the height of
transfer plate the building (note the concentration of
damages denoted by the points above
the transfer floor level)
Figure 2-11 Building model employed in a quasi-static (pushover) study
(Mwafy & Khalifa, 2017)

The deformation capacity of the building can be limited to that of the podium columns and
walls that support the transfer structure (and the tower above). This was investigated by Su et al.
(2002) by way of numerical simulations on limited ductile building models. The brittle mode of
failure in the podium columns that support the transfer plate is highlighted to be ‘a matter of some
concern’. These podium columns were found to be particularly susceptible to brittle shear failure
even under low or moderate earthquake intensities. This type of failure in the column (brittle
shear) was exacerbated by high axial loads imposed on the column and the limited confinement
provided by the reinforcement (i.e. low amount of transverse reinforcement provided).

Intuitively, abrupt changes in the lateral stiffness up the height of the building are precursors
to the development of a soft-storey mechanism in an earthquake (as discussed in Section 2.1.1).
This unfavourable mode of response has not been observed in experimental (shake-table) tests
reported by Li et al. (2006) which is contradictory to the assumed response behaviour of a building
with a similar configuration. In an attempt to ratify such inaccuracies, Su (2008) hypothesised
that the flexural stiffness of the podium structure (below the plate) significantly contributes to
‘relaxing’ the acute changes in stiffnesses up the height of the building. Accordingly, the lateral
stiffness of the building has been modified to account for the term 𝜑1 𝜃𝑏 where 𝜃𝑏 is the additional

23
flexibility of the tower ensued by the rotations of the podium structure as shown in Figure 2-12.
This modification has led to a reduction of the stiffness gradient above, and below, the level of
the transfer which was speculated to be in better agreement with the experimental observations.

(a) Total stiffness (b) Lateral stiffness (c) Flexural stiffness

Figure 2-12 Total stiffness of the podium substructure including both the lateral
and flexural stiffnesses (Su, 2008)
The effects of the positioning of the transfer plate (up the height of the building) on the seismic
response behaviour of the building have been investigated in numerous studies (e.g. Su and Cheng
(2009), Wang and Wei (2002), Xu, Wang, Hao, and Xiao (2000)). Experimental investigation by
Wu et al. (2007) on a scaled Plexiglas prototype building model reported the sensitivity of
parameters like forces, drifts and displacements to changes in the location of the transfer plate.
But no conclusive quantifications of these effects have been reported. More recently, the distinct
dynamic features of buildings with the transfer structures positioned at higher levels have been
reported (Abdelbasset, Sayed-Ahmed, & Mourad, 2014; Elawady, Okail, Abdelrahman, & Sayed-
Ahmed, 2014). These studies observed that the dynamic response behaviour (drift and force
demands) of such a building (with a transfer plate at a higher level) is dominated by the
fundamental mode of vibration. This is not surprising as a building with such features would have
larger masses positioned closer to the effective height of the building which is analogous to the
response of a single-degree-of-freedom system. Conversely, podiums at lower levels would have
larger contributions by the higher modes on the dynamic response behaviour of the building.

2.2. Diaphragm actions in seismic conditions

2.2.1. Diaphragm forces

Floor diaphragms in seismic conditions have the function of distributing the earthquake loads
to the primary lateral load resisting members (i.e. walls, cores and moment frames). This,
however, does not preclude the original functionality of a slab in redistributing the vertical
(gravity) loads to the different members (beams, columns or walls) of the storey.

An early paper by Clough (1982) investigates the forces that can be developed in floor slab

24
diaphragms. Two types of forces are distinguished. The first type is the equilibrium (inertial)
forces that are generated in the floor slabs as a result of the floor acceleration (which are
proportional to the mass of the floor). The second type is the compatibility (self-straining) force
that is internally developed as a result of deformation incompatibilities in the structure (commonly
referred to as the compatibility force). Compatibility floor forces are typically developed in the
floor slab connecting two lateral load resisting (LLR) systems with different stiffnesses. Clough
argues that extent of these forces is highly dependent on the geometry of the building and the
distributions of lateral stiffness up its height. Differences in the deformed shapes of the frame and
the walls are maintained by the internal actions (and deformations) that are developed in the links
between the walls and the frame. Paulay and Priestley (1992) presents a similar account for the
compatibility forces in dual frame-wall buildings (refer Figure 2-13).

Compression and tension (in-plane) forces in the link members (shown in Figure 2-13(d)) can
be developed to ‘push’ and ‘pull’ the two systems to achieve compatible (equal) displacements
of the frame and the wall. The magnitude of the generated forces (the amount of the ‘push’ and
‘pull’ by the diaphragm) can significantly vary depending on the relative stiffness of the
connected systems (the stiffness of the frame relative to that of the wall). Bull (2004) argues the
existence of a complex interaction between the two types of forces that develop (simultaneously)
in the floor diaphragm (i.e. inertia and compatibility). As the building responds elastically to an
earthquake, the generated compatibility forces are proportional to the intensity of the earthquake
(i.e. proportional to the inertial forces). Conversely, variation in the spread of the inelastic
demands in a building can, in some cases, exacerbate the intensity of the compatibility forces.

(a) Applied (b) Deformations of (c) Deformation of (d) Deformation of the dual
(lateral) loads the frame the wall system
Figure 2-13 Deformation and compatibility force trends in a dual frame-wall system
(Paulay & Priestley, 1992)
A study by Rutenberg (2004) highlighted the effects of floor diaphragms on the seismic shear
response behaviour of the connected walls. It has been demonstrated that the distribution of shear
forces between the walls (coupled with a floor diaphragm or a beam) is dependent on the
sequencing of the plastic hinge formation in the walls. In Figure 2-14(b), internal forces (N) are

25
shown to have developed in the link members as a plastic hinge forms at the base of the stiffer
(longer) wall (Wall 2) at a roof displacement of 0.1m (Figure 2-15(a)).

(a) Idealised system comprising two- (b) Internal (compatibility) actions in the
connected walls (thick line represents the connecting slabs at the onset of formation of
stiffer wall) a plastic hinge at the base of the stiffer wall
(on the right)
Figure 2-14 Internal forces in inelastically responding two connected wall system
(Avigdor Rutenberg, 2004)

(a) Bending moment vs. roof displacement of the (b) Reactive (compatibility) forces
two-wall system
Figure 2-15 Moment-roof displacement response of the two-wall system and the
resulting in-plane forces in the connecting floors (Rutenberg, 2004)

It should be noted that prior to this displacement limit (0.1m) the internal forces (in the links)
are zero (Figure 2-15(b)). The reported zero internal forces in the elastic range are attributed to
modelling the links with infinitely (very high) in-plane stiffnesses. Accordingly, displacement
compatibility (equal displacements) in between the walls (in the elastic range of the response)
would have been enforced purely by virtue of the high (infinite) in-plane stiffness of the links. As
such, the above-reported link forces are in fact self-equilibrating (required for equilibrium) as
opposed to self-straining (for achieving compatible wall displacements). This is the case since
these forces would have developed (in the links) to restore global equilibrium conditions by
relocating some loads to Wall 1 (flexible wall) as the nominal capacity of the stiffer wall (Wall 2)
is attained (flat line in Figure 2-15(a)).

26
 Rutenberg argues that if the in-plane flexibility of the link member has been explicitly
modelled, much lower compatibility forces would have been attained due to the in-plane
deformation of the link member. Similar observations have been reported in an analytical study
where nonlinear time-history analyses have been performed on a dual frame-wall type building
(Sullivan, Calvi, & Priestley, 2006). Beyer (2005) further examined the influences of the in-plane
flexibility of the floor diaphragms connecting the walls on their seismic response behaviour. An
analytical simplification for modelling the ensemble that comprises two walls linked by a flexible
floor is shown in Figure 2-16.

Figure 2-16 Flexible slab model between connected walls (total of two walls)

Beyer (2005) reports an extensive parameter study to outline the effects of parameters like the
stiffness of the wall, the foundation flexibility and the in-plane deformations on the generation of
the in-plane (compatibility) forces in the floor diaphragms. High sensitivity of the shear force
demands in the walls as a result of the in-plane flexibility of the connecting slab is highlighted.
The general conclusion by Rutenberg is reaffirmed (in the referenced study) in that the transfer
(compatibility) forces can adversely alter shear forces in the walls once the building (or the
system) sustains inelastic demands (i.e. in the inelastic range of the response). Differences in the
shear force intensities between connected walls have been also attributed to the transfer of internal
forces from the slab to the walls.

Rutenberg and Nsieri (2006) extended the initial study by Rutenberg (discussed earlier) to
incorporate the in-plane flexibility of the slabs connecting walls of different stiffnesses. The
combined effects of shear flexibility in the walls and the in-plane deformation of the slab were
investigated on the shear force demands in the walls up their height. As demonstrated in Figure
2-17 the shear distribution of the stiffer wall (annotated by wall 4 in Figure 2-17(a)) is highly
sensitive to the assumed in-plane stiffness of the diaphragm. It is worth noting that the force
transfer is only assumed to have occurred once a plastic hinge is formed at the base of Wall 4 (the
stiffer wall). The difference in the shear force demands on Wall 1 and 4 is at its lowest when the
in-plane stiffness of the connecting floors pertain a low value (see Figure 2-17(b). It is interesting
to note that the total base shear on the building did not feature similar dependence to the in-plane
flexibility of the floor diaphragm (the link). This is inferred from Figure 2-17(b) where smaller
differences are reported for the total base shear of the building compared to the differences in the

27
shear demands on walls 1 and 4 with the increasing in-plane stiffness of the link.

(a) Wall ensamble employed in the parametric study

(b) Maximum wall shear forec distribution (from a nonlinear time hsistory analyese)
with variable in-plane stiffness
Figure 2-17 Effect of in-plane slabs stiffness on the shear response of connected walls
(Rutenberg & Nsierie, 2006)

2.2.2. Compatibility force considerations in the structural design practice

Most seismic design codes require that the floor slab diaphragms are designed for a strength
level sufficient to ensure adequate transfer of forces to the vertical LLR members. The
PEER/ATC 72-1(2010) guideline recommends the use of an overstrength factor (in the range of
2-3) to the design diaphragm actions to ensure that floor diaphragms remain ‘elastic or near-
elastic’ in projected earthquake events. Section 12.10.1 of ASCE/SEI 7-10 also requires that
design considerations are made for both the out-of-plane (bending) and in-plane actions in a floor
slab. It is further required that the design compatibility forces to be increased by as much as 25-
35% where features of horizontal (and vertical) irregularity are manifested in the building.

In Eurocode 8 (CEN, 2004), concrete floor slabs of thickness greater than 70mm and with
reinforcement placed along the two horizontal directions are categorised as diaphragms.
Diaphragm strength verifications are required for buildings designed in high seismic regions

28
(DCH; ductility class high) particularly for those featuring vertical irregularities (e.g. setbacks)
or where large floor slab openings are positioned in the proximity of the LLR walls or cores.
Furthermore, in-plane (tensile) stresses in the diaphragms are required to be kept below 1.5fctd (fctd
is the design tensile strength of concrete) for crack control. Additional dowel bars are also be
required to enhance the transfer of forces from the diaphragm to the wall (or core) through dowel
actions (transverse shear transfer in reinforcing bars). Similar to ASCE/SEI 7-10 provisions, the
use of ‘overstrength’ factors are recommended for actions that are likely to cause brittle failure
mechanism in the floor slab (e.g. in-plane forces in the slab). A factor of 1.3 (i.e. 30% increase)
is recommended for this purpose. The technical guideline drafted by Moehle et al. (2010)
recommends the use of explicit numerical tools for determining the in-plane actions in floor
diaphragms under earthquake loads. Estimates of the in-plane slab forces transferred to the
adjoining walls/columns can be obtained by the deep beam analogy of the slab shown in Figure
2-18 or by the Strut-and-tie approach like the one proposed by Gardiner (2011).

Figure 2-18 Deep beam analogy for determining in-plane (bending and shear)
actions in a floor diaphragm (Moehle et al., 2010)

In-plane force transfer in the diaphragm can be compromised as the floor slabs crack under
the applied loads (Nakaki, 2000). Some design codes recommend the use of a reduced stiffness
(and strength) properties for diaphragms in numerical simulations to accommodate the effects of
cracking. These factors (modifiers) take on a value in the range of 0.15-0.5 for the in-plane
stiffness (TBI, 2017) and 0.25 for the out-of-plane (bending) stiffness (ACI, 2014).

29
 2.2.3. Diaphragm modelling in numerical simulations

Modelling of flat slabs has been a recurring issue amongst researchers and engineers. Flat slabs
are reinforced concrete plates supported atop an array of walls and columns in buildings. In this
form of slab construction, interior beams are omitted from the floor plan while perimeter beams
may still be required to stiffen the peripheries of the floor. This method of construction is favoured
for tall buildings as it can result in significant savings in formwork costs and the floor construction
cycles. The structural analysis and design of flat slabs involve a detailed finite element (FE)
numerical model with sufficiently fine (small) FE mesh sizes to capture the internal stresses that
are developed in the floor slab as a result of the applied loads. Openings in the slabs (often
required by the other trades; e.g. mechanical ducts) can significantly affect the distribution of
theses stresses in floor slabs (Moehle et al., 2010).

Detailed modelling of the floor slab in the numerical simulations of tall buildings is central to
achieving an accurate estimate of the seismic response behaviour and the seismic demands on the
building (Zekioglu, Willford, Jin, & Melek, 2007). The influence of the slab on the global seismic
response of the building has also been confirmed by way of experimental studies performed at
the University of California, San Diego (Waugh & Sritharan, 2010). The investigation revealed
that a high degree of coupling between the primary (wall) and secondary (frame) systems is
ensued by the floor slabs which in turn increased the lateral strength and stiffness of the building
when subjected to lateral loads.

A substantial bulk of research on floor slabs has been focused on determining the exact
contributions of the floor slab to the strength and stiffness of the building when subjected to
earthquake loads. There seems to be consensus amongst researchers (and practitioners alike)
about the importance of modelling these contributions in the analysis of a building. The simplest
rational approach of incorporating slabs in-plane stiffness into the modelling of the building is by
dividing it into several interconnected frame elements with effective widths. An early study by
Pecknold (1975) revealed that an equivalent beam with a width of approximately 0.54 of the span
(perpendicular to the loading direction) is capable of simulating the moment coupling between
columns and slabs. Moehle and Diebold (1985) argued that this value (0.54) is in fact too high
and results in an overestimation of the global stiffness of the building. Their conclusion was based
on observations from an experimental program involving a slab-column ensemble subjected to
simulated ground motions. The stiffness of the slab-column assembly has been observed to
diminish with the increase of the drift demands on the building. This is outlined in Figure 2-19
where the damage (cracking) patterns on the face of the slab are illustrated for two drift values.

30
 (a) Cracking pattern in the slab at low (b) Cracking pattern of the slab in high
drift demands intensity shaking
Figure 2-19 Stiffness degradation of the tested slab-column ensemble (Moehle &
Diebold, 1985)
Intuitively, the stiffness of the slab, and hence its width, varies throughout of the response.
Hwang and Moehle (2000) developed a numerical expression for calculating the effective width.
The expression for determining the effective width of an interior span is given (Eq. 2-1).

𝑙1
𝑏 = 2𝑐1 + (2-1)
3

where c1 and l1 is the dimension of the column and the length of the span in the direction of loading
respectively. A similar expression is proposed for the effective width of an exterior frame where
the value of b is 1/2 of that given in Eq. 2-1. These expressions were derived based upon the
elastic plate theory by making the assumption that the joint region between the slab and the
column is infinitely rigid (i.e. a full transfer of bending actions are assumed at the joint between
the slab and the column).

The effect of reduced stiffness due to cracking of the slab is accounted for by way of
parameter 𝛽. A lower-bound value of 𝛽 = 1/3 was recommended for conservative estimate of
the stiffness.

Variations in slab stiffness (i.e. varying throughout the analysis) can be incorporated explicitly
following a step-by-step updating approach (Luo & Durrani, 1995a, 1995b). In the referenced
study, the effective moment of inertia (Ie) of the slab is updated in proportion to the ratio of the
applied moment (at the interface of the column and the slab) and the cracking moment of the slab
(i.e. Ma/Mc).

Dovich and Wight (2005) observed that the effective widths of the slab for the determination
of the strength and stiffness of the slab are, in fact, different. Their study, which is also based on
experimental observations on column-slab assembly, concluded by recommending distinct

31
effective slab widths for stiffness and strength as shown in Figure 2-20. It is worth noting that
these distinctions (in the slab widths) are purely to match the observations from the experiment.

Grossman (1997) also proposed a modification to the effective slab width formulation to best
match the available results from some experimental programs on slab-column assemblies (similar
to those discussed earlier). The proposed expression incorporates stiffness degradation (at the
connection) by the use of the factor ‘Kd’ in Eq. 2-2.

𝑙2 (𝐶2 − 𝐶1 ) 𝑑
𝑏𝑒𝑓𝑓 = 𝐾𝑑 [0.3𝑙1 + 𝐶1 ( ) + ]( ) (𝐾𝐹𝑃 ) (2-2)
𝑙1 2 0.9ℎ

where C1, C2 are the dimensions of the column in the direction parallel and perpendicular to the
(lateral) loading direction respectively. The parameter d, h are the flexural depth and the gross
depth of the slab (respectively) and KFP is a location factor which takes the value of 1.0 for interior
supports and 0.8 for exterior and edge supports. The value of the drift-proportional adjustment
factor (Kd) is in the range of 1.1-0.5 depending on the storey drift demands. Robertson (1997)
proposes a simplified set of parameters for defining the effective width of the slab for two
admissible drift levels (0.5% and 1.5%). The study considers asymmetric cracking patterns as a
result of different positive (sagging) and negative (hogging) moment capacities in the slab.

Figure 2-20 Proposed effective slab width for strength (a) and initial stiffness (b)
(Dovich & Wight, 2005)
The above-described models are mostly derived for slab-column assemblies. It is worthy of
note that similar studies on the slab-wall connections have not been reported in the literature. A
study reported by Shin, Kang, and Grossman (2010) noted that in the absence of experimental

32
verification, effective slab width equations (Eq. 2-2) could conservatively be used in the
numerical modelling. Nevertheless, differences in the behaviour of a slab-column and a slab-wall
connection can be substantial. For instance, the third term enclosed by the brackets in Eq. 2-2 (i.e.
𝐶2 − 𝐶1 /2) is, typically, negative when the slab is framing in the direction that is perpendicular
to the thickness of the wall (note that C2, in this case, is the thickness of the wall). Eq. 2-2 yields
a smaller effective width value. Conversely, a larger effective width value is obtained for the slab
framing in the direction perpendicular to the length of the wall. The PEER/ATC 72-1 (2010)
guideline proposes a generic method for defining the floor slabs in the numerical modelling of
the building. The procedure is outlined in Figure 2-21. The effective slab width at (and closer) to
the core is larger than at a distance further away from the core. This is the case since the width of
the slab, at the core, is supported over a larger length (longer support). This, according to the
guideline, can be accommodated in the numerical model (adopting frame elements) by varying
the effective width along the length of the span (i.e. by the use of beams with different widths).
The choice of the non-uniform distribution of the effective width of the floor slab can be verified
(in part) by observing the internal stress distribution in the slabs close to the walls and cores of
the building. For instance, Figure 2-22 plots the bending stress distribution in the floor slab of a
wall dominated building under lateral loads (the floor plan pertains to one of the case study
buildings analysed in later chapters of this study). The flexural stresses are shown to pinch (get
narrower) away from the core (noting that the warmer colours in Figure 2-22(b) representing
negative bending stresses) which confirms the choice of having slabs of different widths along
the length of the span (as discussed earlier).

Figure 2-21 Modelling slab-core column assemblies in buildings from PEER/ATC 72-1,
(2010)

33
 (a) FE model of a floor plan in (b) Average flexural stress distribution
multistorey building in the floor slab under factored
gravity loads
Figure 2-22 Illustration of the varying effective slab width using FE simulations

2.2.4. Nonlinear response of floor slabs

Numerical analyses have shown that the inelastic response behaviour of the floor slabs in high-
rise buildings substantially contributes to the overall seismic energy dissipation of the building
(Melek, Darama, Gogus, & Kang, 2012; Yang, Hurtado, & Moehle, 2010). The inelastic response
behaviour of the slab is however not well defined (or less researched) and requires an exercise of
engineering judgement to quantify the anticipated responses. There are several ways of modelling
the inelastic response behaviour of a floor slab in a building. These modelling techniques are
discussed in the subsequent sections. It should be noted that although the main focus of this
section is in the modelling of the floor slabs, the presented modelling techniques apply to all other
components in a building (e.g. walls and columns).

2.2.4.1. Lumped plasticity models

Lumped plasticity models are used widely in practice for simulating the nonlinear response
behaviour of the structural members. The modelling approach involves ‘lumping’ structural
nonlinearity at predefined locations along the length (or height) of the member. As such, a
backbone capacity envelope in the form of ‘force-displacement’ or ‘moment-rotation’ is defined
to control the elastic and inelastic deformations and forces that are developed in the member. The
lumped plasticity models prescribe inelastic responses (in members) only in few locations that
are expected to yield in the building (e.g. slab-wall joint and the base of a wall). As a consequence,
computational runtimes (of the numerical model) can be significantly reduced when this
modelling technique is adopted. The primary drawback of this approach is such that the
interactions between the different seismic actions are typically foregone (e.g. axial-flexural or
flexural-shear) which makes the modelling technique not feasible for some simulations.
‘Interaction’ in this context refers to the effects of certain loads and deformations (e.g. axial) on

34
the other types of actions resisted by the component (e.g. shear force or bending moments).

Figure 2-23 Lumped moment-rotation backbone to model inelastic actions and

deformations of the slab (PEER/ATC 72-1, 2010)

In modelling a floor slab, the flexural-axial interactions are often neglected in practice. In other
words, only flexural actions (moments) are anticipated to occur during the response of the
building to an earthquake. As such, a moment-rotation backbone curve (like those shown in
Figure 2-23) is modelled to transcribe the effects of the inelastic bending response behaviour of
the slab at the connections with the walls and the columns. The backbones are obtained by
performing sectional analyses (e.g. moment-curvature) on the effective portions of the slab (on
either end of the slab). The curvatures are then converted to equivalent rotations at the ends of the
slab (red circles in Figure 2-23) following the procedure recommended by Priestley et al. (2007).
It should be noted that neglecting the in-plane response behaviour of a floor slab is not justified
when in-plane forces are expected to develop in the floor slab (e.g. compatibility forces).

2.2.4.2. Distributed plasticity models

Distributed plasticity models have gained popularity in the modelling nonlinear effects in
structural members. Fibre-based element modelling is arguably the most commonly used
distributed-plasticity model for reinforced concrete elements. The cross-section of a beam,
column, wall or a slab is divided into several fibres (see Figure 2-24). The nonlinearity of the
member is uniformly defined along the length of the member at the material level (Spacone,
Filippou, & Taucer, 1996b). This is achieved by assigning distinct uniaxial constitutive (stress-
strain) material properties to the fibres representing the concrete (both confined and unconfined)
and the embedded reinforcing steel. Forces and deformations (when subjected to external loads)
are computed at user-defined locations known as integration points. The obvious advantage of
this modelling approach over the lumped approach is that the response of the member need not
be defined prior to the analyses. This modelling technique is computationally exhaustive as the

35
solution algorithm (for the response of the member) requires a step-by-step integration of the
internal stresses and strains developed at the level of the section (i.e. in the fibres) and in the
integration points to ensure that the equilibrium condition is achieved throughout the response
(Spacone, Filippou, & Taucer, 1996a).

The literature has identified other drawbacks of adopting this approach. Pugh et al. (2015)
argue that the response behaviour of members modelled using the distributed plasticity approach
(e.g. fibre based elements) is sensitive to the element (fibre) discretisation and the spacing in
between the integration points along the length of the member. This sensitivity is particularly
prevalent for structural components that are expected to sustain substantial damages in the
duration of the response (Coleman & Spacone, 2001). These components (e.g. slabs, walls or
columns) potentially develop concentrations of inelastic strains which result in a non-uniform
distribution of section curvatures up the length (or height of the component). One obvious case is
when the component develops plastic hinges at its ends (Figure 2-25). The usual treatment for the
issue of localisation of sectional strains is by refining the elemental discretisation (i.e. adding
more integration points) at these locations (e.g. along the plastic hinge length as shown in Figure
2-25).

Figure 2-24 Illustration of fibre-based beam element modelling of reinforced concrete

members (Taucer, Spacone, & Filippou, 1991)

36
 Figure 2-25 Variation of the softened response of a reinforced concrete beam with
increasing number of integration points (Coleman & Spacone, 2001)

It should be noted that slab modelling using the distributed plasticity approach is favoured
when the in-plane slab forces are significant in which case the axial-flexural interactions are
central for determining accurate estimates of the internal forces and deformations of the floor slab
(as discussed in the earlier sections).

2.2.4.3. Nonlinear finite element solids

Nonlinear finite element modelling of floor slabs adopts the use of the plane stress shell
elements with prescribed material rules defining the response of concrete and steel. This method
is arguably the most accurate since it is well founded on the theory of the finite element analysis.
Notwithstanding, the modelling approach is highly complex and requires substantial expertise
and computational resources to obtain a step-by-step solution of the response behaviour of the
floor slab in a building. Given the computational rigour involved in this approach, its use has been
restricted to the modelling of the building sub-assemblies (e.g. slab wall connection). Henry
(2011) recently performed analyses on a four storey building model utilising solid elements for
the modelling of the floor slabs (see Figure 2-26). Good consistencies reported by this study from
such analyses (employing explicit FE models) with the experimental results further attest to the
superiority of this approach in modelling slab-wall interactions in the building.

Figure 2-26 Analysed building using solid elements to represent slab members (Henry,
2011)

2.2.5. Assumptions for diaphragm behaviour

In most FE program packages (e.g. ETABS), floor slabs are defined by way of rectangular or
triangular shell elements. The formulation of shell elements combines the membrane (in-plane)
and plate (out-of-plane) degrees of freedom in a floor slab (Ibrahimbegovic & Wilson, 1991).
Membrane degrees of freedom (total of three) represent the in-plane actions (and deformations)
while those of the plate quantify the out-of-plane bending and shear actions in the slab. A shell
element combines the in-plane and out-of-plane stiffnesses in the definition of the floor slab.

37
 In most buildings, the in-plane stiffness of cast-in-place floor slabs is assumed to be
substantially high (i.e. rigid compared to the supporting columns and walls). The assumption of
rigid-diaphragms thus warrants a numerical simplification to be made in proportioning the
contribution of the slabs in the lateral response of the building. Using the rigid-diaphragm
constraint, nodal in-plane displacements, at the level of the floor, are 'slaved' to those of a
predefined master node (see Figure 2-27). Strictly speaking, the walls that are symmetrically
positioned on a floor plan will displace by equal amounts under lateral loads. Numerically, this
translates into placing a very large value (close to infinity) of the in-plane stiffness of the floor
slab when the rigid-diaphragm constraint is applied on the storey. With this approach, only inertial
force (refer Section 2.2) are transferred to the walls of the floor in proportion to their respective
lateral stiffness (i.e. K1, K2 and K3 in Figure 2-29). Moreover, the compatibility forces in the slab
would not develop in the modelling that involves rigid-diaphragms (as will be thoroughly
investigated in later chapters of this thesis).

The semi-rigid diaphragm assumption proportions the actual in-plane and the out-of-plane
stiffnesses of the floor slab into the numerical model. This type of floor slab modelling is
particularly important when compatibility forces that are expected to be developed in the floor
slab (e.g. the backstay reactions in a building featuring a setback).

Flexible diaphragms are typically adopted in the modelling of concrete-topped steel decks
(thin slabs atop of corrugated metal sheets) which are used in timber-framed and steel-framed
buildings. ASCE/SEI 7-10 (2010) classifies a floor diaphragm as being flexible if the in-plane
deformation of the slab amounts to twice the storey drifts of the connecting walls and columns
(the definition is schematically presented in Figure 2-28). Highly flexible diaphragms are also
found in precast constructions. For instance, the use of precast slabs in the form of hollow-core
blocks placed between ‘ribbed’ cast-in-place beams is gaining popularity in the construction
industry. A topping slab of 50mm-80mm is cast atop the hollow core units to join the block units
and to level the floor surface. It can be seen that in such cases, the stiffness of the slab (or the
topping) is significantly lower than the (lateral) stiffness of the columns and the walls of the floor
as shown in Figure 2-29.

38
 Figure 2-27 The rigid diaphragm constraint (CSI, 2004)

Figure 2-28 Schematic definition of conditions leading in the assumption of flexible

diaphragms (from ASCE (2010) Chapter 12)

39
 Figure 2-29 Diagrammatic illustration of the flexible, semi-rigid and rigid diaphragms
(PEER/ATC 72-1, 2010)

2.3. Critical review of modelling the shear response behaviour of walls

Modelling the shear response behaviour of walls is central to analyses that involve the
assessment of the seismic performance of medium-rise or tall buildings. The shear span of a wall
is quantified as the ratio of the base bending moment to the base shear (i.e. M/V). The parameter
represents the effective height at which if an equivalent lateral load is applied; the wall would
develop the same bending moment (M) at its base. The shear span ratio, on the other hand, is the
value of M/V normalised with respect to the length of the wall (i.e. M/Vlw). This parameter has
been widely used for characterising the expected response behaviour of the wall under lateral
loads. The response behaviour of structural walls with ratios greater than 2 (i.e. M/Vlw >2) is
predominantly governed by flexural. For lower values of this ratio, the wall is classified as shear-
critical (Priestley, Verma, & Xiao, 1994). It is worth noting that in some design codes (e.g. AS
3600:2009) the shear span ratio is replaced by the ratio of the height of the wall to its length (i.e.
hw/lw). While this approach (replacing M/Vlw by hw/lw) may be appropriate for the isolated
cantilever walls, its extension for walls encased within multistorey buildings is questionable.

40
Krolicki, Maffei, and Calvi (2011) explain that considerable reductions of the shear span can be
exhibited in the walls of a building (see Figure 2-30). In coupled structural walls, the bending
moments up the height of the wall are redistributed into the connecting beams (or floor slabs) in
order that the shear span of the wall is much less than that of an isolated cantilever wall (as
diagrammatically shown in Figure 2-30(b)). Clearly these reductions (in M/V) would not have
been capture by the use of the hw/lw identity.

Figure 2-30 Shear span ratio of multistorey walls and coupled walls (Krolicki et al., 2011)

The flexural response of a wall is typically modelled either by the lumped plasticity model
(Hoult, Goldsworthy, & Lumantarna, 2017; Kazaz, Gülkan, & Yakut, 2012) or by distributed
plasticity models (as discussed in Sections 2.2.4.1& 2.2.4.2 respectively). Lumped plasticity
models are used for modelling walls that are expected to yield at certain locations along its height
(e.g. at the base). The sectional response of the wall is obtained by way of a moment-curvature
analysis which is then converted to rotations by using the well-established formulations proposed
by Priestley et al. (2007). The spread of plasticity up the height of the building is determined by
either limiting it to a specific portion of the wall (Plastic Hinge Analysis) or by analyses involving
explicit modelling of this spread (e.g. fibre model discussed in details in Section 2.2.4). These
approaches are based on the fundamental hypothesis of plane sections remaining plane which is
well placed for modelling flexural behaviour of a wall (J. N. Priestley et al., 2007; Spacone et al.,
1996b). The extension of this hypothesis to the modelling of the shear flexibility of walls has
many limitations. It has been long established that the assumption of plane-section-remain-plane
(i.e. a linear distribution of section strains) is not representative of the actual strain distribution
when a section of the wall (near its base) undergoes significant yielding. Greifenhagen (2006)
examines the distribution of strains (compressive and tensile) in a wall undergoing inelastic
deformations (artwork reproduced from the same study in Figure 2-31).

41
 Figure 2-31 Strain distributions at the base section of a wall (at various drift demands)
(Greifenhagen, 2006)
The hypothesis (of linear strain distribution) holds true prior to (or at) the first instance of
yielding of the steel reinforcement embedded in the wall (represented by the solid thick line in
Figure 2-31). At ultimate conditions, strains to the right of the neutral axis (which is at a distance
of x in Figure 2-26) are larger than those on the left of the neutral axis (compressive strains).
Clearly the assumption of strain distribution is violated. It is worth noting that significantly higher
compressive strains can be developed in walls that are subject to larger axial loads.

As discussed earlier, the lateral response behaviour of slender walls (walls with high shear
span ratios) is mostly governed by flexure. Even then, Beyer, Dazio, and Priestley (2011) argue
that significant shear deformations may still be developed at the locations close to the critical
section of the wall (at the plastic hinge). This was also observed from experimentation on slender
walls (like the ones tested by Dazio, Beyer, and Bachmann (2009)).

Paulay and Priestley (1992) show that the shear flexibility of walls (and thus shear
deformations) is predominant (to varying degrees) in walls with shear span ratios lower than 2.
In most cases, the modelling of slender walls that are governed by flexure has been widely covered
in the literature (Orakcal & Wallace, 2004) whereas the modelling of the shear response behaviour
is a complex issue and is, often, not considered in the nonlinear analyses of buildings. Subsequent
sections provide a summary of the different techniques for incorporating the shear flexibilities
(actions and deformations) in the modelling of the wall.

2.3.1. Shear force-deformation relationship for structural walls

The limited ductile nature of the response behaviour of a shear-critical wall (with shear span
ratio ≤ 2) is classified as ‘force-controlled’ response by most seismic standards and guidelines
(e.g. FEMA 356 (2000), ASCE/SEI 41(2013)). A force-controlled component (or action) is
associated with a brittle or limited ductile component response behaviour which sustains no (or

42
very limited) deformation capacity after the component had attained its nominal strength (Figure
2-32(a)). Strictly speaking, the strength of a force-controlled component is assumed to hit a value
of zero shortly after onset of the ultimate strength (FEMA 356, 2000).

(a) Force-controlled behaviour (b) Deformation-controlled behaviour

Figure 2-32 Illustration of typical force-displacement curves for (a) force-
controlled (b) deformation-controlled response behaviour

A deformation-controlled behaviour, which typically characterises the flexural response of

walls, combines a pre-capping response (prior to the attainment of nominal strength) and post-
capping response (after the attainment of nominal strength) into the modelling of the response
behaviour of the component subject to external loads (Figure 2-32(b)). Modelling deformation-
controlled response behaviour requires rigorous calibrations against the results from experimental
testing on similar components to verify the accuracies of the assumed force-deformation
backbone in capturing the variations of the strength and stiffness of the component throughout
the response. ASCE/SEI 41-13 appropriates the force-controlled behaviour for the modelling of
the shear response of walls subjected to axial load ratios larger than 0.15 (which is the axial load
normalised by 𝑓𝑐′ 𝐴𝑔 ). This provision (in ASCE/SEI 41-13) would entail that no ductility (capacity
of a component to deform beyond its yield deformation) can be attributed to some walls that
support towers in a multi-storey building where such conditions are often applied.

In a recent experimental investigation by Looi et al. (2017), reinforced concrete wall

specimens that are subject to high axial loads (axial load ratios >0.1) have been tested under cyclic
displacement excursions. The study revealed that loosening the stringent (force-controlled)
requirement of the ASCE/SEI 41-13 may be feasible for walls with axial load ratios in the range
of 0.1-0.2. In other words, these walls have been shown to be capable of developing some ductility
prior to their failure. Other experimental studies have also highlighted the appropriateness for
assigning some ductility to walls with low shear span ratios (Greifenhagen & Lestuzzi, 2005; B.
Li, Pan, & Xiang, 2015; Salonikios, Kappos, Tegos, & Penelis, 1999).

In any case, the modelling of the shear response behaviour has been mostly uncoupled from
the flexural response of the wall. The term uncoupled in this context means that the flexural-shear

43
interaction is implicitly captured. This concept (the uncoupled shear response) is clearly
demonstrated in Figure 2-33 where three distinct shear force-deformation curves of a wall are
presented. The thick blue dashed line is the assumed shear backbone of the wall which will be
thoroughly discussed in the remainder of this section. A typical shear force- deformation curve of
a flexural wall (with M/VlW > 2) is shown in Figure 2-33(a). In this case, the wall typically attains
the flexural strength (Vf) prior to the shear strength (Vn) and thus no flexural-shear interaction is
exhibited by the response of the wall. For walls with lower values of M/VlW, this interaction is
examined for two conditions. In Figure 2-33(b), the wall is shown to have developed its flexural
strength (Vf) at a shear force value lower than the nominal strength of the wall (Vn). Beyond the
intersection of the capacity curve (of the wall) with the predefined shear envelope, the shear
response behaviour of the wall will follow the trend prescribed by the blue curve (as shown by
the green arrows of Figure 2-33(b)) only if the shear response of the wall is implicitly defined in
the numerical modelling. Conversely, when no such considerations are made, the capacity curve
will proceed with the initial flexural path defined by the black arrows (also shown in the same
figure). For much lower M/VlW ratios (e.g. <1.0), the shear strength of the wall would have
developed before the flexural strength is attained (i.e. Vf > Vn).

(a) Flexural response (b) Flexural-shear (c) Pure shear

Figure 2-33 Illustration of the modelling of the implicit shear-flexure interaction for
walls

Implicit modelling of this interaction (flexural-shear) between the shear walls involved
defining the backbone shear constitutive relationship (in the form of shear force-deformation) into
a ‘zero-length’ spring element. As such, deformations imposed by the shear mechanism (defined
by the backbone) will be added to those from the flexural element that defines the wall. The two
sub-elements (shear and flexure) are analogous to two-springs-in-series system where the forces
on both the flexural and shear sub-elements are the same despite the variations in the actual
response of the sub-elements (refer Figure 2-34). The flexural response behaviour of a wall is
conveniently modelled using fibre-based elements (distributed plasticity as discussed in Section
2.2.4.2).

44
 Figure 2-34 Schematic illustration of modelling the uncoupled shear and flexural
response behaviour of walls

There are typically three distinct phases in the defining the shear backbone of a wall. The first
phase is defined by the uncracked shear response which is purely characterised by the uncracked
shear stiffness (𝑘𝑢𝑛𝑐𝑟𝑎𝑐𝑘𝑒𝑑 ) of the wall (Eqs. 2-3 & 2-4).

𝐾𝑢𝑛𝑐𝑟𝑎𝑐𝑘𝑒𝑑 = 𝐺𝐶 𝐴𝑠ℎ𝑒𝑎𝑟 (2-3)

𝐺𝑐 is the uncracked shear modulus of concrete which is approximated by

𝐸𝑐
𝐺𝑐 = (2-4)
2(1+𝜈)

where 𝐸𝑐 is the modulus of elasticity of concrete and 𝜈 is the Poisson’s ratio which is in the range
of 0.2-0.25 for concrete. 𝐴𝑠ℎ𝑒𝑎𝑟 is the shear area of a prismatic rectangular section which is
typically assumed to be in the order of 0.8 times the gross sectional area of the wall. The 20%
reduction (in the shear area) is to represent the non-uniform shear stress distribution along the
length of the section. Equation 2-3 can be further refined to approximate the uncracked shear
stiffness of the wall (expressed in Eq. 2-5).

𝐾𝑢𝑛𝑐𝑟𝑎𝑐𝑘𝑒𝑑 ≅ 0.4𝐸𝐶 𝐴𝑠ℎ𝑒𝑎𝑟 (2-5)

Some studies proposed a slightly higher value for 𝐺𝑐 𝐴𝑠ℎ𝑒𝑎𝑟 (e.g. Park and Paulay (1975).
Differences are marginal and can, in most cases, be ignored. The cracking shear force (shear force
at the first formation of the shear crack in the wall) corresponds to a shear stress ratio in the range
of 0.25√𝑓𝑐′ -0.33√𝑓𝑐′ as recommended by Park and Paulay (1975). It is worth noting that the
presence of axial loads enhances the shear transfer by concrete. This enhancement is owed in part

45
by the favourable effects of the axial loads in controlling the widths of cracks which in turn
prolongs the concrete contribution to the shear strength (see Figure 2-35). Shear capacity models
by Gérin and Adebar (2009) and Wibowo, Wilson, Lam, and Gad (2013) incorporate these
favourable effects by the axial loads on the total shear resistance of the wall.

After the formation of the shear crack (illustrated in Figure 2-35), the shear stiffness of the
concrete gradually degrades as the transfer of the in-plane shear forces becomes restricted by the
crack (which widens with the increase in the applied loads). The post-cracking stiffness of the
wall is typically found to be in the range of 1.7%-2.5% the elastic stiffness (Alwaeli, Mwafy,
Pilakoutas, & Guadagnini, 2017; Massone, 2006; Mwafy & Elkholy, 2017). Beyond this phase,
the contribution by the transverse reinforcement (horizontal ties) in transferring shear forces in
the wall becomes more crucial for its shear resistance.

Figure 2-35 Shear cracking in structural walls (artwork adopted from Krolicki et al.
(2011))

The wall shear strength equations prescribed by the design codes can approximate the nominal
shear strength of the wall. These equations typically result in a lower-bound estimate of the actual
(shear) strength that can be sustained by the wall (Gulec, Whittaker, & Stojadinovic, 2008). The
expression for quantifying the nominal shear strength of a wall as recommended by the American
Concrete Institute (ACI 318 (2014)) is given in Eq. 2-6.

𝑉𝑛 = 𝐴𝑐𝑣 (𝛼𝑐 𝜆√𝑓𝑐′ + 𝜌𝑡 𝑓𝑦 ) (2-6)

where 𝛼𝑐 takes the value of 0.25 for relatively squat walls (hw/lw ≤ 1.5) and is 0.17 for hw/lw ≥ 2.
𝜆 takes the value of one for normal weight concrete. 𝜌𝑡 is the transverse reinforcement ratio for
slender walls or the lesser of the transverse and longitudinal reinforcement ratios for walls with
low aspect ratios (following recommendations by Wood (1990)). It should be noted that the Eq.
2-6 does not take into account shear strength enhancements by axial forces applied to the wall.

46
Alternatively, the elaborate expression for predicting the shear capacity can be used (ACI, 2014).

𝑉𝑛 = 𝑉𝑐 + 𝑉𝑠 (2-7)

0.2𝑁𝑢
𝑙𝑤 (0.1𝜆√𝑓𝑐′ + )
𝑙𝑤 ℎ
𝑉𝑐 = [0.05𝜆√𝑓𝑐′ + 𝑀𝑢 𝑙 𝑤 ] ℎ𝑑 (2-8)
−
𝑉𝑢 2

𝐴𝑣 𝑓𝑦𝑡 𝑑
𝑉𝑠 = 𝑠
(2-9)

where Nu is the ultimate axial load applied to the wall and the term 𝑀𝑢 /𝑉𝑢 is the shear-span
of the wall. Av, s fyt ,and d in Eq. 2-9 is the area of the transverse reinforcing bars, the spacing (up
the height of the wall), the yield strength of the transverse reinforcement and the length of the
wall respectively. It should be noted that for purposes of assessment, the shear strength reduction
factor (∅) is typically taken as one and the expected material strengths (as opposed to the lower
characteristic strength values) are used. Various models have been presented for simulating the
post-capping (post- ultimate) strength and stiffness of the wall. Gérin and Adebar (2009) proposed
a constant (flat) backbone where the ultimate shear strain is computed based on the magnitude of
the applied shear stress. Sedgh, Dhakal, and Carr (2015) adopted a similar approach that employs
a zero post-capping stiffness. It is likely that for walls with low shear span ratios (i.e. shear-critical
walls) the shear transfer mechanisms across the section of the wall would degrade at higher
displacement demands (i.e. the post-capping shear strength reduces with higher displacement
demands on the wall). Kelly (2004) proposed a multilinear shear constitutive relationship with a
degrading post-capping stiffness (shown in Figure 2-36). The relationship recognises that a higher
contribution (to the shear capacity of the wall) can be attained from the transverse steel at peak
strength (hence the factor 1.25). This amplification is also to incorporate the potential strain
hardening of the transverse reinforcement after cracking of the concrete. It can be argued that a
lower-bound (conservative) model would employ a Vn = Vc + Vs (Eq. 2-7) instead of the Vc + 1.25Vs
(in Figure 2-36).

47
 Figure 2-36 multilinear shear constitutive relationship proposed by Kelly (2004)

The shear-displacement backbone proposed by ASCE/SEI 41-13 (2013) for walls with
substantial axial load ratio (but less than 0.15) is illustrated in Figure 2-37(a). The values on the
y-axis are the imposed shear forces normalised by the nominal shear strength of the wall (e.g. Eq.
2-7). The ASCE/SEI 41-13 backbone does not proportion residual strength at ultimate conditions
(i.e. the capacity of the wall is zero at a wall drift of 1.0%). This is only the case for walls with
high axial loads. The post-peak shear model proposed by Kelly (Figure 2-36) is in good agreement
with that of the ASCE/SEI 41-13 model for walls for lower axial load ratios (as illustrated in
Figure 2-37(a)). Interestingly, a recent experimental study by Looi et al. (2017) (discussed earlier)
reported ultimate inter-storey drift limits (ISD) in the range of 1.0% for walls characterised by
short shear span ratios and high axial loads (axial load ratios greater than 0.10). A post-capping
shear response model that is consistent with the above observations has been used in the seismic
assessment study reported in Chapter 6 (of this thesis).

(a) Shear capacity envelope for walls with (b) Shear capacity envelope for walls
axial load ratios > 0.05 but <0.15 with axial load ratios ≤ 0.05

Figure 2-37 The Shear force-displacement capacity envelopes recommended by the

ASCE/SEI 41-13 (2013) the term ∆/𝒉 is the storey drift

48
 2.3.2. A post-processing approach to the shear assessment of walls

In view of the complexity involved in simulating the flexural-axial-shear interaction in the

response behaviour of structural walls, it is common for researchers and practitioners to resort to
indirect approaches for assessing the shear capacity of the wall and for evaluating whether (or
not) shear failure of the wall had occurred in the simulations. One such method is a post-
processing technique (i.e. after running the simulation) that involves a step-by-step mapping of
the shear force demands on the wall against a predefined (shear) capacity model. With this
approach, shear failure of the wall is assumed to have occurred as the shear demands exceed the
shear capacity of the wall. The method is consistent with the ‘force-controlled’ description
outlined in the previous section in that the shear failure is assumed to occur instantaneously with
the attainment of the shear capacity. The assumption of constant shear capacity is somewhat
contradictory to the observations reported from experiments where the shear deformations (of
walls) have been observed to be proportional to the extent of the flexural deformations and
damages sustained by the wall (Beyer et al., 2011; Hines, Restrepo, & Seible, 2004). The main
premise for this proportionality is that the shear transfer mechanism in a wall (and thus its
resistance) gradually depletes as seismic damages (cracking and inelastic sectional strains)
accumulate in the wall (Beyer et al., 2011). Accordingly, flexural-shear interaction is commonly
incorporated in the shear assessment of walls by employing a degrading shear capacity model.
These models account for reductions in the nominal shear strength as the wall yields in flexural
which is, in turn, proportional to the displacement (or curvature) ductility demands imposed on
the wall. Shear strength models like the ones proposed by Orackal et al. (2009), the UCSD model
by Kowalsky and Priestley (2000) Priestley et al. (Priestley et al., 2007; Priestley et al., 1994) are
commonly employed for this purpose. This technique has been recently adopted by Alwaeli et al.
(2017) and Mwafy and Khalifa (2017) where shear demands on the core walls of a tall building
have been mapped against the capacity model proposed by Kowalsky and Priestley (2000) as
illustrated in Figure 2-38 (noting that the dashed red line in Figure 2-38 is the shear strength
envelope which is formulated to degrade with the increase in the curvature ductility demands on
the wall).

49
 Figure 2-38 Example results of a shear assessment involving a step-by-step mapping of
shear demands against degrading shear capacity (Alwaeli et al., 2017)

2.3.3. The Modified Compression Field Theory (MCFT)

MCFT is the predecessor of the Compression Field theory (CFT) originally developed by
Collins and Mitchel (Collins, 1978; Mitchell & Collins, 1974). The MCFT was incepted from
numerous reinforced concrete panel tests conducted by Vecchio and Collins (1986) in the late
1980s. The formulation of MCFT inscribes constitutive models based on average stresses and
strains that are developed in a panel subjected to in-plane stresses.

Cracking in the concrete member is explicitly accommodated in MCFT where cracked

concrete is treated as an entirely new material with distinct constitutive (stress-strain) properties
under a two-dimensional stress field. The response behaviour of reinforced concrete is thus
quantified in two principal directions (shown in Figure 2-39(c)).

(a) Shear and in-plane (b) Panel deformation under (c) Concrete principle strains
stresses on a reinforced the combined stress field (annotated by 𝜖1 , 𝜖2 )
concrete panel
Figure 2-39 Modelling of cracked concrete under combined in-plane and shear
stresses (Vecchio & Collins, 1986)

MCFT attempts to resolve (average) stresses and the resulting deformations on a reinforced

50
concrete member (wall, slab or deep beam) as a whole by enforcing conditions of compatibility
and equilibrium at the sectional level. Local crack stresses are then resolved to ensure that the
assumed shear transfer across the crack is achieved. These local crack considerations make the
MCFT framework well suited for analyses of reinforced concrete structures. It is further argued
that the MCFT formulation is ideal for modelling walls that are subject to a combination of in-
plane (shear) and normal (axial) stresses (Beyer, Dazio, & Priestley, 2008).

The program VecTor2 materialises the usefulness of the MCFT in modelling the shear
response behaviour of walls. VecTor2 is a state-of-the-art 2D nonlinear finite element program
designed specifically for reinforced concrete members (Wong, Vecchio, & Trommels, 2014).
Researchers have extensively relied on the program to simulate the response behaviour of RC
walls subject to reversed or monotonic displacement excursions (Almeida, Tarquini, & Beyer,
2016; Gulec & Whittaker, 2009); Hoult et al. (2017)).

2.4. Seismic load criteria for Australia

In this section, a review of the seismic loading considerations on buildings for Australia is
presented. In the subsequent section, a background on the seismologic features of the Australian
continent is highlighted (2.4.1). The seismic design spectrum is discussed in Section 2.4.2 and the
displacement-controlled behaviour and its relevance in regions of low-to-moderate seismicity
(e.g. Australia) are explained in Section 2.4.3.

2.4.1. Seismicity of Australia

Australia is considered to be a region of low-to-moderate seismicity with intraplate features

(Brown & Gibson, 2004; McCue, Gibson, Love, & Cranswick, 2008). The term intraplate refers
to all the regions that are situated away from major tectonic boundaries (Lam, Tsang, Lumantarna,
& Wilson, 2016). Strictly speaking, earthquakes in intraplate regions can occur anywhere in the
world regardless of their proximity to a well-defined fault. Potential earthquakes in the Australian
continent have been attributed to the compressions induced by the “northward movement of the
Indo-Australasian plate” (Brown & Gibson, 2004). The spatial distribution of the historic seismic
events in Australia exhibits several ‘hot spots’ where major (large magnitude) events cluster
within certain locations along the (Australian) continent (see Figure 2-40). Earthquakes in
Australia occur at shallow depths with a reverse faulting mechanism. These seismological
characteristics are not dissimilar to those of the destructive 2011 earthquake that hit the city of
Christchurch (in New Zealand). The frequency of occurrence of earthquakes in Australia is quite
distinct than those in other stable continental intraplate regions around the world. It is argued that
the continent of Australia is likely to observe an earthquake of a magnitude 6 (or greater) once
every five years (Wilson & Lam, 2007) which classify Australia as being one of the most active
low-to-moderate seismic regions across the globe. There had been several occurrences of major

51
earthquakes of magnitudes 5 or greater in Australia. Most notably, the 1989 Newcastle earthquake
is perhaps considered as one of the most destructive seismic events in the last few decades. The
magnitude 5.6 earthquake event has caused fatalities and economic losses amounting to more than
1 billion (Australian) dollars (Walker, 2010). Although some few hot spots have been pinpointed
in terms of relative seismic activity, subsequent events have not shown a consistent pattern (also
in terms of their spatial distribution). Moreover, these hotspots might suggest that some regions
(within the Australian continent) have heightened seismic risk compared to other more stable
regions with the continent. However, the reliance on such assumption in developing seismic
hazard maps has been criticised to result in a considerable underestimation of the seismic risk
other regions located away from these regions (Lam et al., 2016). It is further argued that for most
regions the observed rates of earthquake occurrences exceed those inferred from seismic hazard
maps (Tsang, 2011).

Figure 2-40 Spatial distribution of seismic hazard in the continent of Australia

(Burbidge, Leonard, Robinson, & Gray, 2012)

AS1170.4:2007 (2007) outlines the current seismic loading requirements for Australia. The
seismic hazard in various cities is quantified by the hazard factor ‘Z’ that was introduced in the
latest edition of the standard. This factor quantifies the peak ground acceleration (in unit of g’s)
for a rate of exceedance amounting to 10% in a design life of 50 years (approximately representing
a 500 year return period event). This recurrence rate is referred to as ‘design base earthquake’
(DBE in short) as also defined in most seismic codes around the world. Higher intensity
earthquakes (with lower recurrence rates) can be quantified by the use of the probability factor
‘kp’ which takes the value of 1 for DBE level earthquake and a value of 1.8 for seismic events
with a 2500-year return period. One of the recurring issues amongst the engineering and research
communities in Australia is a concern that the probability factor (kp) adopted in AS1170.4:2007
is low, or lower than it should be, compared to the values adopted for other intraplate regions of

52
similar rate of seismic activity (Amirsardari, Goldsworthy, & Lumantarna, 2017). Paulay and
Priestley (1992) elaborate on this issue by comparing the seismic hazard in regions of high,
medium and low seismicity (refer Figure 2-41). Differences between seismic intensities (e.g. peak
ground accelerations) in high and low seismicity regions are clearly shown when comparing the
‘gap’ between the seismic intensities for a 500 and 2500 year return period events. For low
seismicity regions, the 2500 year return period earthquake is about three folds higher than that of
the design (roughly corresponding to 500 year return period). This discrepancy is translated to
higher risks of collapses for buildings when subjected to rare or very rare seismic events (Wilson,
Lam, & Pham, 2008).

Figure 2-41 Differences in seismic intensities at increasing return periods for regions of
high, moderate and low seismicity (figure adopted from Paulay and Priestley (1992))

2.4.2. Design response spectrum for Australian seismic conditions

Seismic loads on buildings are conventionally represented by the use of a response spectrum
which can be prescribed in either the acceleration velocity or displacement formats (see Figure
2-42 for diagrammatic illustrations).

53
 Figure 2-42 Response spectrum in the acceleration, velocity and displacement
formats (Lumantarna, Lam, & Wilson, 2013)

Seismic loads on buildings vary depending on the positioning of the building along the
acceleration spectrum by virtue of its period of vibration. The value of the acceleration is then
multiplied by the effective mass of the building to determine the base shear. Once the (total) base
shear is obtained, the spatial distribution of inertial forces at the different storeys is derived
proportionally to their respective masses.
The design spectrum is also adjusted depending on the site class at which the building is
situated. This is achieved by a spectral shape factor (Ch(T)) that modifies the generic (rock)
spectrum to a site specific spectrum. A total of five distinct site classifications are described in
AS 1170.4:2007 these classes range from hard rock (Ae) to deep soft soil (De) to very soft (Ee)
sites. The corresponding spectral shape factors are plotted in Figure 2-43. As discussed earlier,
amplifications of the response spectral vales are most pronounced for the soft soil classes (Ce-Ee).
Amplifications of spectral ordinates in the longer period ranges for classes De and Ee are also
considered.

The displacement response spectrum can be directly derived from the design acceleration
spectrum by multiplying the ordinates of the latter (at each period) by the value (4 𝜋/𝑇 2 ) where
T is the period of vibration of the building. The second corner period (defined as T2 in AS
1170.4:2007), which is the period at the start of the displacement-sensitive region of the response
spectrum, takes a constant value of 1.5 seconds. In EC8 a slightly higher value is adopted (2
seconds) and the ASCE/SEI 7-10 maps different T2 values depending on the region along the

54
American continent (typically T2 > 4.0 seconds). Studies have shown that the value of T2 is
dependent on the magnitude of the earthquake (Mw) (Lumantarna, Lam, & Wilson, 2012) and the
sub-soil conditions of the site (Amirsardari et al., 2017; H. Tsang, Wilson, Lam, & Su, 2017).

Figure 2-43 Spectral shape factors for different site classes as given by
AS1170.4:2007

2.4.3. Displacement controlled behaviour of buildings

It has been long established that displacement controlled phenomenon characterises the
response behaviour of long period (flexible) structures when subjected to earthquake ground
excitations. The phenomenon suggests that the displacement demand on a building by the
earthquake is finite and, in fact, does not exceed with the increase in the buildings period. The
displacement and drift demands on the building are thus proportional to the maximum response
spectral displacement (RSDmax) which can be read directly from the displacement response
spectrum (similar to the one shown in Figure 2-42). This principle provides the theoretical premise
for the use of a bi-linear displacement spectrum in regions of low-to-moderate seismicity (like
Australia). The value of RSDmax can be obtained either from the design displacement spectrum or
by the use of stochastic simulations using seismological models that typify the seismic
characteristics in the region (e.g. Lam, Wilson, Chandler, & Hutchinson, (2000a, 2000b)).

The principle has been widely used in quantifying the response behaviour of buildings and
other structures in regions of lower seismicity. As such, seismic drift demands on tall buildings
have been parameterised proportionally to the RSDmax of the ground motion (Su, Tsang, & Lam,
2011). The use of this principle is not restricted to tall buildings. For instance, the principle has
been successfully employed in modelling the risk overturning of free-standing objects (Al Abadi,
Lam, & Gad, 2006; Kafle, Lam, Gad, & Wilson, 2011) and in the modelling of the displacement
demands on a strength degraded (Lumantarna, 2010) and torsionally unbalance buildings
(Lumantarna et al., 2013) subject to earthquake ground motions.

55
 2.5. Seismic design and performance assessment of buildings
In this section, a review of the fundamentals of the seismic design and performance assessment
of building is outlined. Conventionally, the seismic design of a building requires designers to
determine seismic actions directly from an acceleration response spectrum (defined in Section
2.4) that is representative of projected seismic hazards at the site (where the building is located)
or the region (city). The calculated seismic demands on the building are then reduced to
incorporate the likelihood of the building to develop inelastic deformations during an earthquake.
The response modification factors are thus central to the seismic design of a building and are
thoroughly reviewed in the next section (Section 2.5.1.). ‘Assessment’ in the context of
earthquake engineering, weighs into the quantification of the performance of a building when
subjected to different intensity shaking (different levels of seismic hazard). This quantification
requires rigorous analyses to be performed on detailed numerical models of the building to
represent possible component and global responses as the building progress well into the inelastic
state (severe damage). The performance of a building is assessed by benchmarking the results of
analyses (in terms of strains, rotations or drifts) against predefined limits (performance levels) to
indicate whether (or not) the building had performed within the acceptable margins set for a
particular seismic intensity (return period). A review of the basic performance objectives is
outlined in Section 2.5.2. The vulnerability assessment of a building and the development fragility
curves are reviewed in Section 2.5.3.

2.5.1. Response modification factors for seismic design of buildings

The current state-of-practice employs the classical force-based design philosophy in the
seismic design of buildings. Although this approach has been fundamentally criticised by
researchers in the field of earthquake engineering (Priestley, 1993), the force-based approach
remains central to most (if not all) seismic design codes. By the use of this approach, the effects
of the inelastic demands on a building are incorporated into the seismic design forces by the use
of the response modification factor. The methodology, in its simplest form, reduces the seismic
(lateral) load on the basis that the building will trade strength (force) with ductility (deformation)
as the response progresses into the inelastic range (Elnashai & Mwafy, 2002; Whittaker, Hart, &
Rojahn, 1999).

There are two predominant sources contributing to this reduction factor and these are the
overstrength factor (designated as the parameter (Ω𝑜 ) in the American practice and (1/𝑆𝑝 ) in the
Australian/New Zealand practices; where Sp is the performance factor) and the ductility
factor (𝜇). Sources of structural overstrength in buildings are primarily attributed to the inherent
conservatism of the design approaches adopted in practice. Elements of this conservatism are
manifested at the material level where lower-characteristic material strength values are used in

56
lieu of the expected or mean strength values, at the component strength estimation by the design
equations (which are often lower-bound strength estimates and are further reduced by the strength
reduction factor ) and in the determination of the design loads on a building. The ductility (𝜇)
on the other hand, is the capacity of the building (or a component) to deform beyond its yield
capacity. Similar to overstrength, sources of ductility in buildings are broad, and they range from
the sectional level (e.g. detailing of sections) to building geometry and the assumed (lateral) load
pattern (FEMA, 2009; FEMA, 2005). The early establishment of the response modification factor
(R- factor) was drafted in the ATC (Applied Technology Council) (1995) and is expressed in Eq.
2-10.

𝑅 = 𝑅𝜇 𝑅𝑆 𝑅𝑅 (2-10)

where 𝑅𝜇 , 𝑅𝑆 and 𝑅𝑅 are the ductility, overstrength (Ω𝑜 or 1/Sp) and redundancy factors
respectively. It should be noted that the redundancy factor (RR) takes the value of 1.0 in most (if
not all) seismic codes around the world (Whittaker et al., 1999).

There are several ways of quantifying the response modification factors of buildings.
Conventionally, a quasi-static pushover analysis is performed on a building where the lateral
forces are incrementally increased until the building attains a predefined target displacement (e.g.
roof drift) or until the ultimate state condition has been met. This approach is illustrated in Figure
2-44 where the overstrength and ductility factors of a building are shown on the capacity curve
obtained from the pushover analysis.

Figure 2-44 Description of response modification factors

The applicability of this approach (pushover) is restricted to the height of the building. For
tall buildings (h> 30.0m), static pushover analyses are generally not permitted. Researchers have
resorted to dynamic approaches to quantify the seismic response parameters for tall (and irregular

57
buildings). The Incremental Dynamic Analysis (IDA), or what is referred to as the dynamic
pushover, has been successfully employed in quantifying overstrength and ductility measures of
the building (Alwaeli et al., 2017; Elnashai & Mwafy, 2002; Kim & Choi, 2005; Mwafy, 2011;
Mwafy & Elnashai, 2002; Mwafy & Khalifa, 2017). The approach is also recommended by
FEMA-P695 (2009) guideline and will be systematically explained in Chapter 6.

The ductility reduction factor (𝑅𝜇 in Eq. 2-10 and Figure 2-44) is equal to the buildings
ductility (𝜇) for tall buildings (but not for short/stiff buildings). The physical interpretation of this
correlation (𝑅𝜇 = 𝜇) is rooted in the ‘equal-displacement’ principle which states that the
displacement of an inelastic system is equal to that of an elastic system of the same period (Fajfar,
2000; Hutchinson, Lam, & Wilson, 2003; Lam, Wilson & Hutchinson, 1996; Lam, Wilson, &
Hutchinson, 1998; Newmark & Hall, 1982).

While the response modification factors have been central to the seismic design of buildings,
several discretionary measures impose considerable limitations on their applications. For
instance, the use of a uniform (single) force reduction factor for a building requires that the
building is regular (in elevation and plan) with a well-defined load path. Furthermore, designers
will need to ensure that the response behaviour of critical components in the building is within
the acceptable strengths and ductility (deformability) limits as the building develops the
(presumed) ductility. In high seismic areas, capacity design principles are adopted to ensure that
the building is adequately proportioned (detailed) to achieve a preferred yielding mechanism as
it responds inelastically to the earthquake. Moreover, the use of this principle (capacity design)
ensures that the brittle (sudden) modes of failure will not occur during the earthquake. This is
achieved by proportioning larger strengths to elements that are expected to potentially undergo
limited-ductile response behaviour.

The response behaviour of an irregular building is distinct from that of a regular building (Lu,
Su, & Zhou, 2013). More specifically, the ductility and the overstrength factors are highly
dependent on the degree of the featured irregularity of the building (Mwafy, 2011). Even then,
direct procedures for quantifying the response modification factors for irregular buildings are not
commonly mandated by seismic design code. Eurocode 8 (CEN, 2004) requires a 20% reduction
of the behaviour factor qo (which is equivalent to the R factor in the international practice) for
irregular buildings of classes DCH or DCM (which stand for ductility class high and ductility
class medium respectively).

For the design of transfer structures, the ASCE/SEI 7-10 (2010) standard allocates higher
design forces for the structural elements that support discontinued members (i.e. podium
columns/walls and the transfer structure). The factor 2.5 (defined as the overstrength factor Ω𝑜 )
is typically used to amplify the seismic forces in these storeys (the transfer and the below storeys).

58
This amplification factor is incorporated in the design of the building (according to ASCE/SEI 7-
10) by the use of ‘special load combinations’. This provision is in line with the capacity design
philosophy discussed earlier. The implications of the use of such (and other) provisions will be
further discussed in Chapter 6.

2.5.2. Seismic performance assessment of buildings

The performance-based earthquake engineering (PBEE for short) framework has gained vast
popularity amongst researchers and practitioners in the field of earthquake engineering. The
framework is founded on the ‘displacement-based’ philosophy in the seismic design and
assessment of buildings. The displacement-based concept has gained prevalence over its ‘forced-
based’ counterpart in the past few decades. This is because of the well-established consensus
amongst researchers and practitioners on the good correlations between structural deformations
(building or component) and the degree of the seismic damages experienced during the response
of the building to an earthquake (Paulay & Priestley, 1992; Priestley et al., 2007; Priestley, 1997).
The seismic performance of the building is evaluated by benchmarking the response behaviour of
the building to predefined performance objectives. These (performance) objectives had
undergone several developments from their first inception in the early drafts of the “Blue Book”
(SEAOC, 1995) some two and half decades ago. Different performance objectives are accordingly
set for the seismic assessment of buildings when subjected the different seismic hazard intensities
(or return periods). Essentially, the building is required to remain operational (sustaining its
original functionality) when subjected to frequent earthquakes, whereas the Life Safety
performance objective would need to be met for the design level earthquake (which typically
correspond to a 500 year return period event). The Life Safety performance objective requires that
the building sustains damages (sometimes considerable) whilst some margin against collapse (full
or partial) would need to be achieved. A ‘no-collapse’ requirement is mandated for an adequate
performance of the building when subjected to ‘very rare’ earthquake events (corresponding to
approximately 2500 year return period event). The Collapse Prevention performance objective is
defined in ASCE/SEI 41-13 as a component in the building attains its collapse limit (drift or
rotation) or a certain drift limit is sustained by the building. The discussed performance objectives
are referred to as the Basic Safety Objective in FEMA 356 (2000).

For each of the seismic performance levels, a set of parameters are typically listed as measures
for quantifying the degree of damages in the structural elements. For ductile walls and columns,
distinct upper-bound plastic hinge rotation limits (inelastic rotations at the base) are specified for
different performance levels (i.e. IO, LS and CP). It should be noted that other damage measures
like concrete and reinforcing steel strain limits are more appropriate for use for structural walls
(and columns) that have poor seismic detailing. For structural walls that are dominated by shear,

59
ASCE/SEI 41 (2013) recommends the use of the inter-storey drifts (ISD) to quantify the extent of
the imposed damages to the wall. ISD limits are also used for determining potential damages for
non-structural components (e.g. cladding/glazing and partition walls). The deformation limits (as
the ones previously discussed) are only adopted for structural components that can develop some
ductility in their response (i.e. deform beyond the yield). Conversely, ASCE/SEI 41-13
recommends the use of force limits (i.e. strength capacity) for structural components (or actions)
that are brittle (or limited-ductile) in nature.

2.5.3. Seismic Fragility assessment of buildings

Seismic fragility functions quantify the probability of reaching (or exceeding) a performance
level in a building for a given seismic intensity level. The development of fragility functions
require extensive nonlinear dynamic simulations performed on a building by incrementally
scaling the input record until the capacity of the building is significantly depleted (Vamvatsikos
& Cornell, 2002). The suite of ground motion records used in the fragility assessment are typically
representative of a specific seismic hazard level (i.e. design base earthquake) or a range of
scenarios (e.g. far-field or near-field seismic events). The lognormal cumulative distribution
function is often adopted for defining fragility functions for buildings (Porter, Kennedy, &
Bachman, 2007). This choice has been confirmed by numerous studies reported in the literature
(Aslani & Miranda, 2005; Bradley & Dhakal, 2008; Bradley, Dhakal, Mander, & Li, 2008; Brown
& Lowes, 2007; Ibarra & Krawinkler, 2005). The classical format of the lognormal cumulative
distribution is expressed in Eq. 2-11.
x
ln( )
P(Limit State|IM = x) = Φ ( θ
) (2-11)
β

The term on the left-hand-side of Eq. 2-11 (P (Limit State | IM =x)) is the nomenclature
definition for the cumulative probability of a building exceeding a predefined limits stat when
subjected to a ground motion of a specific intensity. IM is the intensity measure used to
characterise the intensity of the input ground motion. Parameters like the peak ground acceleration
(PGA) or the acceleration spectral ordinate at the fundamental period of the building (i.e. Sa(T1))
are the usual candidates of the IM. Other IMs have been explored in the literature (e.g. peak
ground velocity). The annotation (𝛷( ) ) is the classical cumulative distribution function (CDF).
The parameter θ and  represents the median (50% fractile) and the statistical dispersion (standard
deviation) respectively (Vamvatsikos & Cornell, 2006).

A fragility function is constructed by fitting the mean and dispersion of the CDF to the trends
obtained from the raw (analyses) data observed from the IDA. Equations 2-12 and 2-13 provide
an estimate of the observed mean and dispersion from the NLTH analyse (Porter et al., 2007).

60
 1
ln(θ̂) = ∑ni=1 ln IMi (2-12)
n

1 IM 2
β̂ = √ ∑ni=1 (ln ( ̂ i)) (2-13)
n−1 θ

where n denotes the number of ground motions considered at an intensity IMi at which a damage
limit state of a certain performance level (IO,LS & CP) is observed from the analyses. While this
method is conventionally used for the development of fragility curves, uncertainties associated
with adopting large (record) scaling factors for simulating collapse can be substantial.

The Multiple Stripe Analysis (MSA) is one of the ‘fitting’ techniques used for estimating the
fragility functions for buildings. Jalayer (2003) laid the basis of the MSA framework which has
since been under continued development. Baker (2015) discusses that using the MSA framework
the fragility functions can be developed from analyses of records (scaled or unscaled) with IMs
corresponding to a specific seismic hazard level (e.g. return period) or a scenario (near-field/far-
field). The novelty of this method (MSA) is such that fragility functions may still be constructed
even when a ground motion (or a subset thereof) does not, strictly speaking, cause the attainment
of a damage limit state in a building. An example to illustrate the use of the MSA approach in
determining the fragility function for a building is given in Figure 2-45 (which is originally
adopted from Baker (2015)). In Figure 2-45(a) results from analyses performed on a structure are
presented by the IM (intensity measure) of the ground motion and the peak storey drift ratios
observed in the building from the analyses (the dots at each IM showing the maximum ratios
obtained from the analyses). The building is assumed to have reached its collapse state when the
storey drifts exceed a threshold value of 8.0% (shown by the dashed line in Figure 2-45(a)). The
triangles shown in Figure 2-45(b) represent the fraction of the records that have caused this
damage limit (peak storey drift of 8.0%) to be achieved in the building to the total number of
records evaluated at a particular IM (e.g. PGA). The fragility curve is constructed using the MSA
approach (the mathematical framework of which will be discussed later on) for the entire range
of IM values shown in the figure. It can be shown that the probabilities of collapses at higher IM
values (IM> 1 in this example) are extrapolated from the fragilities at lower IM values (up to an
IM value of 1.0).

The fragility function (Figure 2-45(b)) can be constructed once the values of the median and
dispersion values (θ and  respectively) have been evaluated. A robust optimisation methodology
may then be adopted for estimating these values (θ and ). In this approach, a binomial
distribution is used to quantify the probability of obtaining 𝑧𝑗 exceedances (of a limit state) from
performing analyses using a 𝑛𝑗 number of ground motions at a specific IM (Eq. 2-14).

61
 𝑧 𝑛𝑗 −𝑧𝑗
𝑃(𝑧𝑗 | 𝐼𝑀) = (𝑛𝑧 𝑗) 𝑝𝑗 𝑗 (1 − 𝑝𝑗 ) (2-14)
𝑗

The likelihood of observing a limit state for an IM is mathematically evaluated as the product
of the individual probabilities over the entire set of ground motions (Eq. 2-15 is directly obtained
by substituting 𝑝𝑗 in Eq. 2-14 by the lognormal cumulative distribution function given in Eq. 2-
11)

x zj x
nj −zj
nj ln( ) ln( )
Likelihood = ∏m θ
j=1 [( zj ) Φ ( β ) (1 − Φ ( θ
)) ] (2-15)
β

where m in Eq. 2-15 is the total number of IM levels n and z are the total number of records (for
each IM) and the fraction of records having caused (or exceeded) a limit state (respectively). It
follows (from statistics as argued by Baker) that a fragility function is mathematically equivalent
to the maximum likelihood of an event causing the occurrence of a damage limit state.
Accordingly, the median (𝜃) and the standard deviation (β) can be estimated (for the entire set
of simulations) by maximising the likelihood function given in Eq. 2-15. To do so, the natural
logarithm of Eq. 2-15 is maximised and the corresponding estimates of the median and standard
deviation are expressed (Eq. 2-16).

x
nj ln( )
{θ, β} = 𝑎𝑟𝑔𝑚𝑎𝑥(θ, β) ∑𝑚
𝑗=1 ln ( z ) + 𝑧𝑗 ln Φ (
θ
) + (nj − zj ) (1 −
j β
(2-16)
x
ln( )
θ
Φ( β
))

Eq. 2-16 provides a simple numerical procedure to compute the values of and which
completely define the fragility curve (similar to the one shown in Figure 2-45(b)).

62
 (a) Discrete results from analyses using an (b) Full fragility function using MSA showing
intensity subset extrapolated fragilities (at higher IM values)

Figure 2-45 MSA approach (figure adopted from Baker (2015))

2.6. Chapter Summary

The potential poor performance of podium-tower buildings lends its weight to the irregular
distributions of the buildings stiffness and strength up the height of the building. Most design
codes classify a podium-tower type building as having irregularity in elevation. The reviewed
literature on the seismic response behaviour of a setback building is mostly concerned with
moment frames. Seismic design guidelines (e.g. PEER/ATC 72-1 and TBI) typically provide
qualitative treatment for the adverse effects of the backstay phenomenon in setback buildings.
The bulk of the work tends to focus on the effects of the shear reversals on the lower portion of
the building (on basement walls and diaphragms or podium walls and columns). Similar issues
have been highlighted in the works of Rad and Adebar (2009) and Bevan-Pritchard et al. (1983).
The key (take-home) conclusion of their work was that the basement diaphragm at the base of the
building redistributes significant forces away from the tower walls to the perimeter foundation
walls by way of in-plane actions in the basement floor slabs. Such interactions have been
identified to be important for determining the seismic demands on the tower walls of the building
(close to their bases).

The detrimental interactions of the podium in buildings featuring a transfer plate have been
highlighted in the works of Su and Cheng (2009) and Tang and Su (2014) which discuss the
occurrences of anomalous shear force concentrations in the tower walls that are planted on a
transfer plate. The reported interactions were attributed to the local deformations of the transfer
plate supporting the tower walls which ultimately imposes unequal displacements of the tower
walls above the transfer floor level. The resulting slab-wall interactions that occur as a result of
the displacement incompatibilities between the walls resulted in the shear force concentrations
above the transfer floor. These interactions were reported in a qualitative manner by observing

63
the various shear force concentrations as the building is subjected to lateral loads. Quantifying
these interactions (by the transfer plate) together with their effects on the response behaviour of
the tower walls above is central for the seismic design of the building.

In the two forms of the podium-tower buildings (setback and transfer structures), the literature
put substantial emphasis on in-plane interactions of the floor slabs with the connecting walls.
These effects (slab-wall/core interactions) are often overlooked in practice. The paucity of
research on this topic warranted a thorough review of the numerical tools for modelling such
interferences (outlined in Section 2.2). The bulk of the research has been focused on simple frame-
wall (dual) systems linked by floor diaphragms. The in-plane (compatibility) forces were mostly
reported for when the building responds inelastically to the input motion. As such, differences in
the yielding patterns (drifts) between interconnected members were shown to have been
maintained by the in-plane forces in the floor diaphragms. These forces were also required to
ensure equilibrium conditions are maintained throughout the response of the building (Beyer,
Simonini, Constantin, & Rutenberg, 2014; Rutenberg, 2004; Rutenberg & Nsieri, 2006). In-plane
(compatibility) forces may also be developed even when the building is well within the elastic
range of the response. The generation of these forces is prominent in buildings that feature
elements of irregularity in their elevation. Very limited literature focused on investigating such
occurrences (and their effects) in buildings (Gardiner, 2011; Su & Cheng, 2009). Moreover, most
of the reported studies on this topic have taken a ‘qualitative’ approach where compatibility forces
are observed, rather than quantified, in the simulations.

The potential of some tower walls to experience shear-critical response behaviour is typically
ignored in tall buildings. Nevertheless, the studies reviewed in this chapter suggest that such
conditions (shear-governed) can be manifested in some tower walls planted on a transfer plate.
These observations motivated a thorough review of the different ways the shear response
behaviour of a wall can be accounted for in numerical simulations of building particularly those
employed in seismic performance assessment studies (discussed in detail in Section 2.3).

There is sufficient evidence in the literature to suggest the limited-ductile seismic performance
of irregular podium-tower buildings. Observations from the experimental and analytical studies
highlighted the poor performance of the building when subjected to rare and very rare earthquake
intensities. It is possible that such issues can be mitigated in the early stages of the design.
However, a review of the response modification factors (used in the design of the building) in
some codes and standards revealed that there is (typically) limited guidance on the applicability
of such factors (the force reduction (R) and the ductility factors) in the design of a podium-tower
building. At the moment, these issues bear much weight on discretions and judgment of the design
engineer.

64
Chapter 3. PODIUM INTERFERENCES IN SETBACK BUILDINGS

3.1. Introduction
This chapter outlines an in-depth study on the effects of podium interferences on the seismic
response behaviour of the tower walls in buildings featuring setbacks. The podium structure at
the base of a medium and a high-rise building imposes a non-uniform distribution of lateral
strength and stiffness up the height of the building. Not surprisingly, this building configuration
is classified by most design codes as a prominent feature of vertical irregularity in buildings (as
discussed in Section 2.1). Nevertheless, the effects of the podium-tower interactions subject to
earthquake loads have not been thoroughly investigated in the literature (as also pointed out in
Section 2.1.2).

The backstay force (as described in Section 2.1.2) is generated at the interface level to resist
the overturning actions from the lateral forces imposed on the tower (refer Figure 3-1 for a
schematic illustration). The described (backstay) forces can induce high shear forces in the lower
podium levels of the building. It is demonstrated in this chapter that high shear force intensities
are also induced in the tower walls above the level of the interface. This has been shown to be
manifested in buildings where the tower is positioned at an offset from the centre of the podium.
The structural wall which is closer to the centre of the podium (referred herein as the interior
wall) is subject to higher moment restraints from the podium structure than the exterior wall. As
a result, strutting (compatibility) forces have been observed in the connecting floor structure
(beam and slab) above the interface level to restore displacement compatibility (i.e. equal lateral
displacements). This strutting action can only be modelled accurately if the horizontal in-plane
deformation of the floor diaphragm has been incorporated into the modelling. Thus, the extent of
such actions can be misrepresented by analyses in which the (usual) rigid floor diaphragm
assumption has been made.

Parametric studies based on quasi-static analyses have been first conducted on two-
dimensional (2D) planar podium-tower sub-assemblages to demonstrate shear anomalies that can
be developed in structural walls that are resulted from interferences from the podium structure
(Section 3.2). The cause of a significant increase in the shear intensity of the structural wall above
the podium level is to be illustrated. Results from the parametric studies have been used to track
the trends of the increase in the shear force intensities in order that conditions deserving special
attention in design can be identified. Findings from parametric studies of the sub-assemblages
have been verified by static and dynamic analyses of example 3D finite element (FE) models of
existing buildings assuming linear elastic behaviour and non-linear inelastic behaviour. The
content of this chapter has been reported in the following publication by the author (Yacoubian
et al., 2017a).

65
 Figure 3-1 Backstay actions in a podium-tower building featuring a setback

3.2. Analyses on 2D planar podium-tower sub assemblages

The first part of the study to be reported in this chapter is based on 2D planar sub-assemblages
of a podium-tower building in which tower walls are offset from the podium (Figure 3-2(a)).
Results obtained from the analysis of the sub-assemblages are compared with that of centrally
configured sub-assemblages as control (shown in Figure 3-2(b)). This comparison is essential for
highlighting the effects of the positioning of the tower with respect to the supporting podium.

For both building models (shown in Figure 3-2(a)-(b)), the podium structure was simplified to
a series of walls connected by floor slabs. The later methodology was chosen such that it can
simulate real building conditions wherein the lateral stiffness of the tower is typically 20%-60%
of the lateral stiffness of the podium (the building models used in the first part of the study pertain
to the lower-bound value). Details of structural elements making up the subassemblies are
summarised in Table 3-1 & Table 3-2.

66
 (a) Tower walls offset from the centre (b) Centrally positioned tower walls.
Figure 3-2 Example podium-tower sub-assemblages

The occurrence of significant strutting (in-plane) forces that are generated in the
interconnecting floor beams and slabs in between the pair of shear walls are first illustrated
(Figure 3-3). The magnitude of the strutting force is shown to be highest in situations where the
structural tower walls are offset significantly from the centre of the building. Interestingly, a
significant amount of strutting force is shown to occur above the podium level and up the height
of the wall for as much as 10 m. In contrast, no strutting force is generated in the centrally
positioned tower walls. The observed anomalies are best illustrated by showing the deflection of
the tower walls in a hypothetical configuration in which the floor beams connecting the walls
have been intentionally removed (Figure 3-4(a)). The deflection of the interior wall normalised
with respect to that of the exterior wall is defined herein as the deflection ratio 𝛿𝑟 (Figure 3-4(b)).
Table 3-1 Geometric configuration of the examined sub-assemblages

Dimension, indication Units in [mm]

Length of coupling beam, a 2000
Tower walls length, a & c 6000
Tower wall thickness 300
Podium wall length, d (Typical) 6000
Coupling beam depth 1000
Clear podium span, e (Typical) 6000
Effective slab width (podium) 3100
Podium wall thickness (Typical) 600

Table 3-2 Material properties used (structural walls)

Material Properties
fC′ 40 MPa
Poission′s Ratio 0.2
Ec 31.6 GPa

67
 Figure 3-3 Strutting forces in the connecting floors and beams between the tower walls
in offset and centred podium-tower sub-assemblages

The ingression of the exterior wall onto the interior wall (Figure 3-3(a)) sheds light on the
differential restraints exerted by the podium structure on the tower walls. It is shown that δr <
1 in the lower (podium) portion of the model (meaning that wall lateral deformations at these
levels are not equal).

This trend extends to one-to-two storeys above the level of the podium beyond which the ratio
(𝛿𝑟 ) tends to unity (suggesting that compatible wall displacements are achieved above this level).
These strutting (in-plane) forces generated in the connecting beams can result in a localised
increase in the shear intensity of the interior wall, whereas the shear intensity is decreased in the
exterior wall. This mechanism results in an asymmetric distribution of shear demands in the
connected walls as shown in Figure 3-5. The resulting distribution of shear forces in the twin
walls above, and below, the podium-tower interface of the example structure is shown in Figure
3-6(a).

68
 (a) Deflection of hypothetical structure (b) Deflection Ratio
without CB
Figure 3-4 Kinematic behaviour of tower walls in podium-wall sub-assemblage

Figure 3-5 Influence of strutting forces on wall shear forces

The moment-shear ratio (M/V) has been used to characterise the relative likelihood of a
structural wall, or a column, to experience shear critical, or flexural shear critical, mode of failure
in ultimate conditions (as discussed in Section 2.3). In a coupled structural wall, bending moments
up the height of the wall are redistributed into the connecting beams in order that the M/V ratio
is much less than that of an isolated cantilever wall (refer Figure 2-30). The M/V ratios of both

69
the interior and exterior tower walls that are offset from the buildings centre are shown in a similar
manner (Figure 3-7(a)). Compared with the M/V ratio of the control model (Figure 3-7(b)), the
much lower ratio predicted for the interior wall is consistent with the shear anomalies illustrated
in the schematic diagram of Figure 3-5.

(a) Shear force distribution in tower walls (b) Externally applied lateral forces
Figure 3-6 Shear force distribution in tower walls

Parametric studies have been undertaken on a group of building models featuring variable
heights of the podium. The elevations of the six planar models of podium-tower building systems
are shown schematically in Figure 3-8. Arbitrary lateral quasi-static loads were applied to each of
the considered models. Their respective maximum shear force (VInt /VExt ) ratios are shown for
comparison in Figure 3-9 (noting that the subscript i+1 refers to the maximum shear force ratios
evaluated at the storey immediately above the level of the interface). It is shown that the most
acute shear anomalies occur when the height of the podium is about 1/4 - 1/3 of the height of the
building (i.e. building designated by: 3P9T, 0.25). Further details of the strutting forces and the
resulting wall shear forces in two of the example models are also shown in Figure 3-10(a)-(c).
Further parametric studies on more podium-tower structural systems have been undertaken to
investigate the sensitivity of the maximum shear intensity to changes in the position of the tower
(Figure 3-11), wall length ratio (Figure 3-12(b) & Figure 3-12(c)) and clearance between the twin
walls (Figure 3-12(d)). It is shown that sub-assemblages with high eccentricity, low wall length
ratio (Lint /Lext ) and high clearance between the walls pertain to the development of high shear

70
intensities.

(a) Tower walls at offset from centre (b) Tower walls centrally positioned
of podium
Figure 3-7 M/V ratio in tower walls in podium-tower sub-assemblage

Figure 3-8 Podium-tower sub-assemblages with varying podium height

71
Figure 3-9 Correlation of wall shear force ratios with podium height ratios

Figure 3-10 Internal forces in tower walls for the different podium-tower sub-
assemblages

72
(a) Effect of eccentricity on the shear force (b) Schematic representation of podium-
ratio tower eccentricity calculation
Figure 3-11 Correlation of wall shear force ratios with podium eccentricity

(a) Elevation (b) Effect of the relative wall length on

the shear force ratio

(c) Wall deflection ratio (d) Strutting (in-plane) force

distributions
Figure 3-12 Behaviour of tower walls at an offset from the centre of the podium

73
 The effect the relative lateral stiffness of the tower (and the podium) has upon shear
force distributions up the height of the tower wall is next examined. The effective stiffness
of the tower (𝐾𝑇 ) was computed by independently analysing the tower structure (which
was fixed at its base) by the use of the classical modelling technique recommended by
Priestley et al.(2007). The stiffness of the podium (KP) was also found by employing the
same technique. The relative stiffness ratio is henceforth defined as the lateral stiffness of
the tower normalised with respect to the lateral stiffness of the podium (𝐾𝑇 /𝐾𝑃 ) (the
procedure for calculating the lateral stiffness is outlined in Appendix A-3 for some of the
building models investigated in this study). The typical range of this ratio (for the building
examined in this chapter) was found to be in the range 0.2-0.6. Figure 3-13 plots the results
of the analyses performed on four different sub-assemblage models (similar to that shown
in Figure 3-2(a)) with varying the values of KP in order that distinct stiffness ratios can be
achieved 𝐾𝑇 /𝐾𝑃 = 0.2, 0.48, 0.64 and 0.85. It can be shown that as the value of the relative
stiffness parameter is increased, the restraint of the podium on the tower walls in the form
of shear reversals at the interface level is reduced (refer Figure 3-13(b)). Interestingly the
strutting (compatibility) forces in the coupling beams were accordingly reduced (Figure
3-13(c)). Consequently, the most adverse shear force concentration was manifested in the
tower walls which were supported by a stiff podium structure. This observation is similar
to that reported in an early publication by Shahrooz & Moehle (1990b) on setback frame
buildings as discussed in Section 2.1.2.

(c) Strutting force

(a) Externally applied (b) Tower wall shear
profile in the
lateral loads on all distribution (interior
connecting coupling
building models. and exterior walls)
beams
Figure 3-13 Results of analyses of sub-assemblage building models for podium
height ratio of 0.32 (hp/hb = 0.32)

74
 3.3. Effects of diaphragm modelling on shear force demands on the tower walls
In the seismic design and assessment of high-rise buildings rigid diaphragm constraints are
often utilised at floor levels up the height of the building. This assumption is warranted given that
floor slabs often high in-plane stiffness. ASCE/SEI 7 (2010) and ASCE/SEI 41(2013) recommend
modelling slabs that are likely to be subject to high transfer and inertia forces with semi-rigid, or
flexible, characteristics (refer Section 2.2.5 and Figure 2-29 for more details). This can be
achieved by distinctly modelling the building floor based on its sectional and material properties.
Similarly, PEER/ATC 72-1(2010) recommends against the use of “rigid diaphragm” assumptions
in the modelling of the podium-tower interface slabs (main backstay slabs). The effects of
imposing such numerical constraints on the structural wall are examined by means of a parameter
study in which planar sub-assemblage models similar to that shown in Figure 3-2(a) were
analysed. In a control model, rigid diaphragms were assigned to all storeys above and below the
podium level. Results by the computations were substantially different when compared with a
model featuring explicitly defined floor slabs up the height of the building (refer Figure 3-14(a)
& Figure 3-14 (b)). The comparison is not surprising given that rigid diaphragm constraints
“slave” all in-plane degrees of freedom on the floor to a master node (refer Figure 2-27).
Consequently equal wall displacements are numerically enforced at each floor level (Shin et al.,
2010; Zekioglu et al., 2007).

To investigate trade-offs between practicality and accuracies with the modelling, sensitivity
analyses are conducted to determine the portion of the height of the building that requires explicit
modelling of the floor slabs. The extent of the shear reversal (the reaction of the podium) is
significantly exaggerated when the rigid-diaphragm constraint is imposed on the floor slabs
(compare the shear reversals for the different modelling approaches plotted Figure 3-14(b)).
However, the shear demands on the tower walls were misrepresented above and below the podium
interface level. Particularly the shear force intensities on the tower walls were higher (for the
interior) and lower (for the exterior) compared with the control model that utilises the rigid-
diaphragm constraint.

It is shown that sufficiently accurate results are obtained when the floor slabs in between these
height limits are explicitly modelled: (i) two storeys above and (ii) 60% of the height of the
podium (designated by: 2T+0.6HP in Figure 3-14(a)-(b)).

75
 (a) Wall deflection ratio (b) Wall shear force distribution

Figure 3-14 Results from sensitivity study for various diaphragm extents

3.4. Verification studies on 3D FE model of buildings

A 3D FE model of an existing building was employed in static and dynamic analyses to verify
findings reported in Sections 3.2-3.3. The case study building is a podium-tower structure wherein
the tower is offset from the centre of stiffness of the podium (refer to Figure 3-15). The medium
rise ten-storey reinforced concrete structure has been designed and detailed for gravity and wind
loads, but without taking into account the potential occurrence of seismic actions. The building
features vertical irregularity above the 4th floor with a setback from the 5th floor up the height of
the building. The lateral load resisting system consists of dual moment frames-wall systems
spanning in both directions.

The FE model of the building has been assembled using program ETABS (Habibullah, 1997)
with shell elements to represent the floor slabs and the structural walls, and frame elements to
represent the beams and the columns. Loads were uniformly distributed on the floor slabs based
on values listed in Table 3-3. Seismic mass was assigned as per the estimated self-weights of the
structural elements, the permanent floor loads and 30% of the imposed floor loads in accordance
with the Australian earthquake loading standard, AS 1170.4 (2007). Two modelling approaches
were employed for the analysis of the structure when subject to lateral seismic actions. In the first
model, rigid floor diaphragms (as commonly assumed in the structural design practice) were

76
adopted. In the second model, explicit modelling of the floor structure was adopted.
Table 3-3 Loads in case study building

Podium storeys Tower storeys

[kPa] [kPa]
Imposed (live) loads 4.00 3.00
Electromechanical loads* 3.00 2.00
Superimposed dead loads (including floor cover) 1.00 1.00
* The case study building functions as a hospital, the Electromechanical loads are categorised as
additional loads (dead) applied to the floor slabs by virtue of the piping, heavy duty HVAC ducts and
other equipment.

The extent of podium interference on wall behaviour (tower walls) was predicted to be
substantial as shown in the 2D planar sub-assemblages parametric studies (Section 3.2). Tower
walls I and II (as shown in Figure 3-16) were positioned very differently on the floor plan: with
the Set I walls being close to the centre of stiffness of the podium whereas the Set II walls were
much further away from the centre. Table 3-4 contains a listing of the dimensions of the walls in
both wall sets.
Table 3-4 Geometric details for analysed sets of walls

Set I Walls
Wall thickness 300 [mm]
Interior wall length 6100 [mm]
Exterior wall Length 7700 [mm]
Set II Walls
Wall thickness 350 [mm]
Interior wall length 7700 [mm]
Exterior wall Length 7700 [mm]

Lateral loads that were applied to the structure were assigned in accordance with the equivalent
static force procedure in AS 1170.4:2007. Shear force distribution on both the interior wall (the
wall which was closer to centre of stiffness of the podium) and the exterior walls have been
analysed. Results of the estimated shear forces up the height of the walls in both models based on
the rigid diaphragm assumption and explicit modelling are shown in Figure 3-17.

The backstay mechanism is shown to have resulted in reversal of the shear force below the
level of the podium-tower interface. Conservative predictions were obtained from analyses based
on the rigid diaphragm assumption at, and below, the podium level. However, the highest shear
force intensity is shown to occur above (and not below) the level of the podium-tower interface
when explicit modelling of the in-plane action of the building floor was adopted. The interior

77
walls were subject to much higher shear intensities than the interior walls. In Set I, the shear force
in the interior wall was 4850 kN compared to only 3800 in the exterior wall. In Set II the wall
shear force was 4050 kN and 2990 kN for the interior and exterior walls respectively. The shear
force (VInt /VExt ) ratio was accordingly 1.28 and 1.35 for the two wall sets. Results of correlation
between the (VInt /VExt ) ratio and the podium height ratio, and the wall length ratio are consistent
with observations from the parametric studies of the podium-tower sub-assemblages (refer Figure
3-18(a)-(b)).

Figure 3-15 3D rendering of case study building structure

(a) Plan view of typical tower floor (b) Plan view of typical podium floor
showing the examined walls showing the examined walls

Figure 3-16 Floor plans of case study building structure

78
 (a) Set I (b) Set II

Figure 3-17 Wall shear force distributions of case study building structure

Results derived from the 3D FE analyses of the building model considered in this section
(Figure 3-17) have been used to provide further support of the predictive relationships derived in
the previous sections as to how hp/hb and LExt/LInt parameters control shear anomalies in the
structural walls (see Figure 3-18(a)-(b)).

(a) Podium height effect (b) Wall length effects

Figure 3-18 Wall shear force ratio of elastic case study building structure

To expand the scope of the study, additional case study buildings featuring comparable

79
configurations have been examined further. Details of the case study buildings are provided in
the Appendix A-2. Results from linear lateral analysis were superposed on results obtained from
the parametric study (Figure 3-19). It is shown that results from the latter provided an upper bound
approximation to the detrimental shear demand localisation (shear ratio) above the podium
interface.

Figure 3-19 Comparison between parametric study and results from case study
buildings [podium height ratio]

3.4.1. Linear time history analysis

To further investigate the uneven share of shear distribution between the twin tower walls,
linear time history analyses were performed using ETABS on the case study building (Figure
3-15). SeismoArtif (2012) was used to generate a suite of artificial time history records based on
the AS 1170.4:2007 design spectrum evaluated on site class De. Details of the records are outlined
in Table A-1 of Appendix A-1 (Record nos. 29-35). For each record, two sets of analysis were
performed: one with rigid diaphragms assigned at each floor level and one with diaphragms that
have been explicitly modelled.

Consistent with the earlier findings, both shear force magnitude and maximum location varied
between the different modelling approaches (with rigid diaphragms imposed and explicit
diaphragm modelling). The 5% damped shear response of Set II walls (refer to Figure 3-16) are
plotted in Figure 3-20 and Figure 3-21. Table 3-5 summarises maximum shear ratios obtained
from the analyses. On average, the interior wall experienced 30% higher shear demands than the
exterior wall when in-plane actions of the floors have been explicitly modelled. The observed
behaviour was in close agreement with results obtained from the equivalent static analyses.
Conversely, no shear force anomalies were found in models with rigid diaphragms assigned for
the collectors. Generally, higher shear demands in the tower walls were found from analyses in
which the explicit diaphragm definition has been adopted (similar to earlier findings). Time
histories of the base shear are shown in Figure 3-22.

80
(a) Maximum shear response: -ve Shear (b) Maximum shear response: +ve
Shear
Figure 3-20 Shear force distribution of Set II walls, models with rigid
diaphragms assigned

(b) Maximum shear response: +ve

(a) Maximum shear response: -ve Shear
Shear
Figure 3-21 Shear force distribution of Set II walls, models with explicit
diaphragm definition

81
 Figure 3-22 Base shear of the suite of artificial records

Table 3-5 Shear ratios for models with explicit diaphragms

[VInt ⁄VExt ]𝑖+1 -ve Shear [VInt ⁄VExt ]𝑖+1 +ve Shear

Rec. 29 1.36 1.27

Rec. 30 1.35 1.27
Rec. 31 1.25 1.29
Rec. 32 1.36 1.27
Rec. 33 1.34 1.33
Rec. 34 1.25 1.28
Rec. 35 1.27 1.34

Mean 1.31 1.29

STD (standard
deviation) 0.05 0.03

COV (%)
(coefficient of 3.84% 2.12%
variation)

Considerable interferences of the higher modes contributed to further anomalies in the shear
distribution trends up the height of the wall. This can be seen by comparing shear reversal in
Figures 3-20 and 3-21. The backstay effect for the models with rigid diaphragms would only be
evident at a level below the interface. This phenomenon relates to the out-of-phase vibration of
the tower and podium portions of the building as depicted in Figure 3-23.

82
 Figure 3-23 Out-of-phase vibration of case study building

3.5. Inelastic shear behaviour of coupled tower walls

In view of limitations with analyses based on the assumption of linear elastic behaviour, non-
linear time history (NLTH) analysis is performed on a case study building (building 3). The 56
storey building features a podium in the lower 9 storeys supporting a 47-storey tower in the upper
levels (refer Figure B-2 in Appendix B). The lower (podium) levels of the building serves as a
commercial lot and a mix of residential and office spaces occupy the upper (tower) levels. The
building is laterally supported by continuous coupled C-shaped walls that form the primary lateral
load resisting system of the tower. The geometric and reinforcement details of the C-shaped walls
are shown in Figure 3-24. Coupling (link) beams of 500mm x 900mm connect the two sides of
the C-shaped walls. The reinforcement content of these beams is also outlined in Figure 3-24.

Figure 3-24 Reinforcement details of the C-shaped core walls

The floor slabs (flat reinforced concrete slabs) in the tower and the podium levels are 200mm

83
and 250mm thick respectively. A 25mm expansion joint is positioned at the interface between the
podium and the tower (shown in Figure B-2(b) in Appendix B). This form of separation is
typically adopted for buildings of this configuration. The effects of the expansion (or in the
context of podium-tower buildings sometimes referred to as ‘settlement joints’) on the shear force
redistributions in the tower walls will be examined in details in Chapter 4. For simplicity, the
simulations presented in this section do not consider the effects of the expansion joint.

3.5.1. Inelastic model development

The program SeismoStruct (2016) is used to construct the nonlinear numerical model of the
case study building. The program is a state-of-the-art numerical tool with wide capabilities of
modelling the inelastic response of buildings subject to dynamic and quasi-static loading
(Almeida et al., 2016; Pinho & Antoniou, 2005; Pinho, Bhatt, Antoniou, & Bento, 2008). A
reduced sub-frame building model has been constructed to illustrate the shear force anomalies in
the tower cores that have been investigated in the earlier sections of this chapter. The portion of
the building spanning in between axes 4-7 (in Figure B-2 in Appendix B) has been considered in
this study. The secondary (gravity) frames have been incorporated into the numerical model
following the procedure recommended in the PEER/ATC 72-1(2010) guideline (an excerpt from
the guideline is shown in Figure 3-25). The stiffness and strength contributions of the secondary
(gravity) frames surrounding the cores are thus explicitly defined into the numerical model.

(a) actual floor plan (b) simplification used for nonlinear

modelling
Figure 3-25 Numerical simplification technique to incorporate the primary and
secondary lateral load resisting system from (PEER/ATC 72-1, 2010 )

The gravity frames (parallel to the long direction of the podium) are connected to the cores by
inelastic fibre-based frame elements representing the effective width of the slab as recommended
by Grossman (1997), Shin et al. (2010) and Zekioglu et al.(2007). The slab modelling approach
is demonstrated in Figure 3-26, and the geometric details and reinforcement content of the floor

84
slabs are summarised in Table 3-6. The effective width at the core equals to the width of the core
(length of the web). This width is reduced at the mid-span between the column and the core as
shown in Figure 3-26 to emulate the ‘pinching’ of the stress contours (flexural) in the vicinity of
the columns (as discussed in Section 2.2.3 and demonstrated in Figure 2-22). The value of the
reduced effective width of the floor slab has been calculated using Eq. 2-2 reported in Section
2.2.3.

Figure 3-26 Slab-wall/column representation for NLTH analysis

Table 3-6 Geometric and sectional details of the floor slabs

𝑓𝑐′ , 𝑀𝑃𝑎 𝜌, (%)† Width x depth, (mm ×

mm)
Tower 24.0 0.678 % (N12 @ 200mm). Additional N12 7200x200 at the core
Slab bars at the columns (in both directions) (1800x200 at the column)
Podium 32.0 0.693 % (N12 @ 150mm). Additional N12 7200x200 at the core
Slab bars at the columns (in both directions) (2200x200 at the column)

† The reported ratios are for the reinforcement that is parallel to the direction considered in the
analyses. Similar reinforcement ratios are adopted in the transverse (perpendicular) direction

The continuous core walls (spanning the entire height of the building) have been modelled as
wide columns where the frame elements representing the web and the flanges are connected by
horizontal rigid links following recommendations by Bayer et al. (2008). This is illustrated in
Figure 3-28 where the vertical lines representing the web and flanges of the core walls are
connected (at the nodes that are shown in blue) by the horizontal rigid links and coupling beams.
The vertical spacing of the elastic rigid links (the frequency of links with height) is selected as
the height of the storey following recommendations by Bayer et al. (2008) where a spacing of
half the length of the wall (or 1/5 of the shear span) is recommended.

At the cross-section of the core wall, each element was divided into fibres with distinct uniaxial
material properties which were assigned to both the concrete and the reinforcing steel (refer

85
Section 2.2.4.2 for more details). The Mander concrete material model was assigned for both the
confined and unconfined concrete (Mander, Priestley, & Park, 1988). The Meneggotto-Pinto
model was adopted for modelling the steel reinforcement (Menegotto & Pinto, 1973).
Reinforcement details and geometries of the cores and the gravity columns are outlined in Figure
3-24 and Figure 3-29 respectively.

Masses of the structural members along with the tributary weights from the imposed loads are
manually calculated and lumped at the nodes. Rayleigh (modal) damping was used in the
nonlinear time history analyses. The value of the mass and stiffness proportional damping
parameters were determined by assigning a 2.5% damping at 6.2 seconds, and 1.16 seconds,
which correspond to 1.1T1 , and 0.2 T1 respectively (refer Figure 3-30). The natural period of
vibration of the first mode (T1 ) was found to be 5.79 seconds. This approach is consistent with
recommendations given in PEER/ATC 72-1 (2010) guide for tall buildings. Two sets of
simulations were considered: first with the secondary columns (for resisting gravity loads) were
included in the numerical modelling, and second a more conservative model with only the tower
core walls considered in the analysis (see Figure 3-28). The main purpose of conducting the
analyses was to demonstrate the extent of podium interferences on the shear demands of the
primary and secondary lateral systems in podium-tower buildings.

Figure 3-27 Illustration of core the inelastic modelling of the core

86
(a) case with only the core walls are modelled in (b) Tower core walls and gravity
the tower frames are incorporated in the
modelling.
Figure 3-28 Elevation views of the analysed building models

Figure 3-29 Geometric and reinforcement details of the secondary gravity columns
(immediately above the level of the interface)

87
 Figure 3-30 Rayleigh damping model employed in the NLTH analyses

3.5.1. Shear demands in nonlinear time history analyses

The models were subjected to excitations by ground motions as recorded from the 1989 Loma
Prieta earthquake event (record no. 37 outlined in Table A-1 in Appendix A). It is shown in Figure
3-31 that the model with secondary columns exhibited significant higher mode contributions
compared with the model with only the tower cores. The base shear time histories for both models
are plotted in Figure 3-32 which shows that comparable base shear was observed for both models.

Figure 3-31 Deformed shape at onset of flexural yielding of coupling beams

Figure 3-32 Base shear time histories for the analysed building models

88
 Shear demands on the structural walls were assessed for both models. Figure 3-33 plots the
shear force distribution within the vicinity of the podium interface at the onset of flexural yielding
of the coupling beams and maximum base shear (in Figure 3-33(a) and Figure 3-33(b)
respectively). Shear localisation and asymmetric demands were observed in the tower core which
is consistent with earlier findings. The (VInt /VExt ) ratios for the core walls in the model which
did not include the secondary (gravity) columns (designated as: Only Cores) are reported to be
higher when compared with the model with secondary columns included in the analysis (1.4 and
1.3 respectively). Results obtained from linear elastic analysis (Section 3.2 and Figure 3-9) as the
upper bound estimate (of 1.39 for ℎ𝑝 /ℎ𝑏 = 0.16) suggests that contributions from the higher
modes (and thus the height of the building) on the presented shear distribution anomalies are
minimal.

As stated in Section 3.2, strutting forces in the connecting slabs and beams were found to be
the primary contributor to the presented wall shear demand anomaly. Interestingly, pronounced
strutting forces were observed in the slabs which connect the tower cores to the secondary
(gravity) columns (Figure 3-34). These forces caused an unfavourable increase in the shear
demand of the interior secondary columns just above the level of the podium (see Figure 3-33(a)
and Figure 3-33(b)). It should be noted that a more detailed assessment of the inelastic response
behaviour of buildings with similar configurations will be presented in Chapters 4 and 6.

(a) at first yield of coupling beams (b) at maximum base shear

Figure 3-33 Wall shear force distribution

89
 Figure 3-34 Strutting force time histories

3.6. Discrete modelling of shear failure progression in podium-tower buildings

The NLTH analysis discussed in Section 3.5 provided further verification to issues pertaining
to shear anomalies in the tower walls featuring setbacks as presented earlier. The direct
consequences of these anomalies are to put forth in a discrete non-linear sub-assemblage model
that is capable of modelling shear-critical behaviour and axial-flexure-shear interaction in the
planar walls. VecTor2 (Wong et al., 2014) has been employed for this purpose. The theoretical
framework and the element formulation of VecTor2 have been already outlined in Section 2.3.3.
The nonlinear FE model of the sub-assemblage was first calibrated against results reported from
experimental programs on limited-ductile walls representative of the walls commonly found in
low-to-moderate seismic regions, such as Australia.

3.6.1. Calibration and validation of the numerical model

Wall specimens from the literature have been selected to validate the modelling parameters to
be adopted for the 2D sub-assemblage model (in Section 3.6.2). Publications by Bing et al. (Li et
al., 2015) wall LW2, Griefenhagen et al. M4 specimen (2005) and Looi et al. specimens C30-N-
ALR01 and C30-N-ALR02 (2017) have been selected for this purpose (details of the wall
specimens are summarised in Table F-1 in Appendix F). The set-out of the specimen including
the wall, the stiff distribution beam (at the top of the wall) and the foundation block have been
modelled using the rectangular plane stress concrete elements readily available in Vector2. A fine
mesh has been adopted for modelling the walls between the beam and the foundation block (the
maximum aspect ratio-width/height of these elements has been set to 1.5). Smeared reinforcement
was used to model the embedded reinforcing. The material properties of concrete and reinforcing
steel (both transverse and longitudinal) have been directly taken from the information provided
in the respective publications (and summarised in Table F-1). The Popovics (normal strength
concrete model) is adopted to define the pre-peak and post-peak concrete compressive stress-
strain relationship. The Modified Bentz 2003 (Wong et al., 2014) tensioning stiffening model is
adopted to define the tensile behaviour of concrete. The Seckin reinforcement model is used to

90
define the hysteretic response of the reinforcement under cyclic load reversals.

Results from the numerical simulations on VecTor2 were superimposed on the experimental
data points and backbone curves representing experimental data (Figure 3-35). The simulated
cracking patterns (in VecTor2) are compared to damage patterns that are obtained from the
experimental programs in Figure 3-35. It is shown that the numerical simulations were capable of
simulating displacements, stiffness and ultimate strength to a reasonable degree of accuracies
(Figure 3-35). Cracking patterns of the walls were also adequately captured in the numerical
simulations.

91
 (a) Force-displacement response of (b) Simulated (c) Actual cracking
the specimens cracking pattern pattern
Figure 3-35 Simulated and experimental response of the wall specimens

3.6.2. 2D planar model of the sub-assemblage

The results outlined in the previous section validate the use of VecTor2 in modelling shear-
critical walls with low shear spans and high axial loads. It is speculated that these conditions are
prevalent in tower walls supporting setback towers (as discussed earlier). Numerical simulations
on 2D nonlinear finite element models that incorporate the likelihood of walls undergoing shear
failure and axial collapse are presented in this section.

The building model in Figure 3-36 comprises of two tower walls that are connected by 2m
long coupling beams. At the lower levels, the existence of the podium is simulated by a single
podium wall that is positioned at an 8m offset from the tower walls.

92
 Figure 3-36 Elevation of inelastic case study building structure (VecTor2)

The backstay floor slabs (connecting the podium to the tower as shown in Figure 3-36) were
numerically modelled by assigning an effective width based on recommendations given by
Grossman (1997). The slab elements were modelled as quadrilateral (4-nodes) plane stress
concrete elements which are readily available in VecTor2. The material (constitutive) models
assigned for concrete and the smeared reinforcement meshes are similar to those discussed in
Section 3.6.1. It is worth noting the effect of lap-splicing (of the reinforcement) was not modelled
explicitly. The tower walls of the sub-assemblage were numerically constructed on VecTor2 and
were assigned a calibrated material (as demonstrated earlier). Geometric and sectional properties
of the various components making up the building model are given in Table 3-7.

Table 3-7 Reinforcement and modelling parameter details of the nonlinear model

Tower Wall
Vertical web reinforcement ratio ρv [%] 1.5
Transverse web reinforcement ρt [%] 0.25
ratio
Axial load ratio P/fc′ Ag 0.08

Wall thickness t wall 200 [mm]

Coupling Beams

Longitudinal reinforcement ratio ρl [%] 1.5

Transverse reinforcement ratio ρt [%] 0.86
Depth of section dCB 850 [mm]
Backstay Diaphragms

93
 Effective slab width beff [𝑚] 4.10
Longitudinal reinforcement ratio ρ[%] 0.75
Slab thickness t slab 250[mm]

Modal pushover analysis was performed on the 60 m sub-assemblage based on the

fundamental modal behaviour of the building. The choice of this loading pattern is warranted
given that the intent of the analysis was to demonstrate relative behaviour between the exterior
and interior tower walls and to encapsulate failure mechanism under extreme conditions, and
not to accurately capture period shifts and variation in the dynamic properties of the structure
in the inelastic range. Furthermore, the effect of the higher modes on the shear distribution
anomalies presented earlier in this chapter was shown to be minimal (refer to Section 3.5.2).
The adequacy of this assumption has been examined towards the end of this section.

Two sets of simulations were conducted. In the first set of simulations, discrete modelling
of the podium based on equivalent stiffness and strength was constructed (case 1). In the
second set of simulations elastic material properties of the podium elements were specified
(case 2). The latter was conducted for the purpose of examining the influence of yielding at
the podium level and the implications on the overall seismic performance behaviour of the
sub-assemblage. In both sets of simulations, the response behaviour of the tower walls was
governed by a brittle shear failure mechanism which is characterised by a prominent diagonal
crack forming in the interior wall leading to a sliding mode of failure (along diagonal crack)
followed by axial failure of the tower. With reference to Figure 3-37, when inelastic behaviour
is prompted in the podium, the ultimate shear capacity was found to be approximately 12%
higher compared to the case of an elastic podium. The roof displacement in the former case
is about 30% higher when compared with the latter implying that drift demands are less
restraint with inelastic podiums. These trends in the response behaviour are not surprising as
higher force redistributions between the tower and the podium are prompted in the case where
the inelastic response behaviour of the podium is incorporated into the modelling.
Furthermore, higher deformation capacities are exhibited in case 1 which is associated with
the yielding displacements at the podium level (prompted by the softened stiffness of the
podium at the onset of yielding and the consequent partial “relief” of the restraint on the
interior wall).

In analysing the failure progression in the sub-assemblage (case 1), four different damage
states are identified: the flexural yielding of coupling beams, the onset of shear yielding in
the interior wall, the formation of diagonal shear crack and the sliding shear failure along the
prominent diagonal crack. The stages are numbered as: I, II, II and IV respectively. As shown

94
in Figure 3-38 and Figure 3-40, the first stage (I) is mainly characterised by flexural yielding
of the coupling beams. However, the effects of the strutting actions in the coupling beams can
implicitly be inferred from the yielding pattern up the height of the tower. For levels in the
immediate proximity of the podium-tower interface, the observed yielding of the coupling
beams (as inferred from cracking patterns at the ends of the beam) range from mild to non-
existent. This implies that the additional in-plane stresses (in the form of strutting forces)
augment the flexural strength of the beams (i.e. the moment capacity) and reduce the cracking
exhibited at their ends. The lag extends up until the occurrence of stage (II), where flexural
hinging is pronounced at the ends for all of the coupling beams above the podium-tower
interface. Upon the onset of shear yielding, cracking in the wall was found to occur in the
interior wall and specifically in the level just above the interface. As the response ingress into
the inelastic range, most of the successive inelasticity was channelled into the widening of
the prominent diagonal shear crack. In stage (III), the shear mechanism was prompted along
the diagonal and in the subsequent displacement steps. Stage IV culminates in abrupt shear
and axial failure in the interior wall (which was characterised by a rapid degradation in
strength as shown in the force-displacement curve). Figure 3-39 shows the value of the
backstay force in the main backstay diaphragm at the interface level. The values shown have
been normalised with respect to fc′ Aeff (where Aeff is the gross sectional area of the slab). An
abrupt increase in the in-plane forces of the backstay diaphragm is shown in between stages
II and III. This is attributed to the difference in the yielding pattern between the podium and
the tower structures. In the case of an elastically responding podium structure, the reported
increase in the backstay forces is pronounced given that large forces are redistributed from
the yielding (softened) tower to the elastic (stiffer) podium structure. The magnitudes of the
normalised (backstay) forces are shown to be lower than 0.15 (with a maximum value at
0.12), which is deemed to be not high enough to cause premature failure of the backstay
diaphragm. Further investigations on the in-plane yielding of the backstay diaphragm will be
presented in Chapter 4.

Damage to the shear walls in Stage I to IV is depicted in Figure 3-40. Clearly, the interior
wall is shown to be much more susceptible to shear-critical failure than the exterior wall.

95
Figure 3-37 Push over curve for the sub-assemblage (shear response included)

Figure 3-38 Considered response phases on the push over curve

Figure 3-39 Normalised in-plane backstay forces in the main backstay

diaphragm (at the podium interface level)

96
 Figure 3-40 Simulated damage in the four response phases

The response behaviour of the 20 storey sub-assemblage model of the building typifies the
force-controlled behaviour (discussed in Section 2.3.1 and illustrated in Figure 2-32(a)) which is
often associated with a shear-dominated response of the structural members (walls). FEMA 356
(2000) characterises a member as ” force-controlled” when the response of the member exhibits
elastic behaviour prior to reaching the ultimate limit state which can be followed by a sudden loss
in the gravitational load carrying capacity of the member at the collapse-limit state. This is
consistent with response behaviour observed for shear-critical non-seismically detailed walls. The
shear force in the walls at stages I-IV have been normalised with respect to the predicted shear
strength (capacity) of the walls as recommended in ACI 318 (2014) concrete design standard (Eq.

97
2-7 –Eq. 2-9) . The shear force demand-to-capacity ratio of the interior wall in stage III (at the
onset of shear yielding) was approximately equal to 1.0 whereas the ratio was approximately
equal to 0.61 for the exterior wall. Subsequent shear demands on the interior wall resulted in
significant “shear yielding” in the wall which was characterised by the widening of the diagonal
cracking and the yielding of the transverse reinforcement up until the collapse limit state is
reached in stage IV.

Figure 3-41 Shear demand to capacity ratio for interior and exterior walls

Furthermore, following the formation of the prominent diagonal crack (between stages III and
IV), stresses in the transverse reinforcement in the interior walls are shown to have reached their
characteristic yield capacity (approximately 410 MPa). This indicates the existence of substantial
stress transfer from the cracked concrete to the transverse bars prior to axial failure which was
observed in the interior wall at stage IV (see Figure 3-42).

Figure 3-42 Stress distribution of the transverse reinforcement of the wall after stage
III

98
 Figure 3-43 presents the curvature distribution up the height of the interior wall in both elastic
and inelastic podium models for two distinct stages: I and II. Figures for inelastic podium model
for both stages exhibit a larger spread of inelastic behaviour compared with the elastic podium
model as is reflected in the curvature distribution profiles. Curvatures below the interface level
are higher in the inelastic podium model compared to the elastic model implying a lower degree
of inelastic distribution occurring in the former model. Synonymously, comparing the area
enclosed by the curvature profiles the higher inelastic deformation capacity maintained by the
interior wall in the inelastic podium model (see Figure 3-37) is further highlighted.

The lateral load profile (based on the first mode shape) is valid provided that the following
conditions are met:
1- No significant stiffness variation (tangent stiffness) is shown in the force-displacement
behaviour of the building (see Figure 3-37). Analogously, no significant variation in the
fundamental mode shape or period lengthening can be associated with the response
behaviour of the building which is in strike contrasts to buildings exhibiting large ductility
demands.
2- The brittle (non-ductile) response behaviour of the interior wall in a coupled shear wall is
primarily caused by differential boundary conditions imposed by the adjoining podium
structure. The strutting (compatibility) forces generated in the coupling beams are
proportional to the relative drifts in the tower and the podium structure. In ultimate
conditions, the first mode of vibration (with the largest tower drift) is deemed most critical
to the shear response behaviour of the tower walls.

The simplified sub-assemblage model presented herein has provided better insight into the
consequence of localised shear concentration in the tower walls. As discussed earlier, these
localised effects imposed on the offset tower wall model might not have been evident in the
simplified analyses of the building structure (based on rigid diaphragm assumptions) as is
commonly adopted in design practices.

99
 Figure 3-43 Curvature distribution of interior wall within the vicinity of
interface in response stages I and II

3.7. Chapter Summary

This chapter sheds light on the unfavourable interference of the podium structure on the walls
of the tower. It was found that podium structures can impose significant differential restraints on
the connected tower walls. Diaphragm-wall interactions in the form of (strutting) compatibility
forces were mobilised in the storeys immediately above, and below, the podium level. These
forces are shown to redistribute internal actions between the interior and the exterior wall, and
consequently offsetting the equilibrium shear force distribution above the level of the podium.
Significant shear concentration can result from this internal force redistribution. A parametric
study was devised to quantify factors affecting these shear anomalies. The rigid diaphragm
assumption which is typically enforced in the design and assessment of medium-rise and high-
rise buildings was assessed in the light of the podium-tower interferences as described. It was
found that such (commonly used) assumptions may lead to unconservative representation of the
shear demands in the tower walls of a building.

Results from analyses of the case study buildings provided further validation to the initial
parametric study. The ramifications of shear distribution between interior and exterior walls on
failure progression in the inelastic realm were examined by means of nonlinear pushover analysis
on a simplified sub-assemblage. The shear-critical (brittle) response behaviour of the interior

100
tower wall (within the direct proximity of the interface) characterised the response of the building.
These conditions have been attributed to the interference by the podium which was demonstrated
to impose low shear spans (M/V) and high shear force demands on the interior wall. Further,
secondary (gravity) walls and columns connected to primary walls or cores were also shown to
exhibit high localised shear demands above the level of the podium.

101
102
 Chapter 4. ANALYTICAL MODELLING OF THE EFFECTS OF
PODIUM INTERFERENCES IN SETBACK BUILDINGS

4.1. Introduction
The main objective of the first part of this chapter is the development of simple analytical tools
for quantifying the extent of interferences by the podium structure on the response behaviour of
the tower walls in a building featuring a vertical setback. Two analytical models are presented to
estimate the maximum values and the distribution of strutting (compatibility) forces that are
developed in between two tower walls above the interface level. An explicit stiffness-based
approach is introduced to derive solutions for these internal (strutting) forces by evaluating the
deformation trends of the tower walls and the internal force equilibrium conditions at and above
the level of the interface. Results from the numerous sensitivity studies that have been presented
in Chapter 3 are employed to evaluate the applicability of the approach in predicting the intricate
slab-wall interactions occurring above the interface. A refined approach is also developed, and
verified, as a rapid estimation tool for use in a structural design office. The developed tool
simplifies the response behaviour of the building under lateral loads to establish a closed-form
estimate of the restraint of the podium and its effects on the tower walls above.

It is also the objective of this chapter to investigate influences of the in-plane (and out-of-
plane) yielding of the main podium diaphragm on the response behaviour of the building.
Particular attention is cast on the force transfer in between the tower and the podium (by the
interface slab) and the local redistribution of forces (above the interface) as the podium slab
experiences inelastic deformations. Similarly, the effects of construction and expansion joints at
the junction between the podium and the tower on the interferences imposed by the podium
structure are investigated in Section 4.6. Some sections of this chapter have been reported in the
following publication by the author (Yacoubian et al., 2017b).

4.2. Explicit Stiffness-based approach

In this section, discrete modelling of strutting force distribution in podium-tower buildings
and the force transfer between the connecting slab and the wall is to be developed. Deformations
of the connected walls above the interface level are schematically shown in Figure 4-1. As
demonstrated in Chapter 3, the lateral displacement of the interior wall is smaller compared with
the displacement of the exterior wall above the podium level. The coupling beam (or connecting
slab) is subject to in-plane strains (deformations) as a result of the displacement incompatibility
between the two walls. Internal (compatibility) forces in the coupling beams are generated with
magnitudes proportional to the axial (in-plane) stiffness of the beam (𝐾𝑏 ). From the free body
diagram shown in Figure 4-1 the local shear force equilibrium of the interior and exterior walls
immediately above the interface level (i = 1) are given in Equations 4-1 and 4-2.

103
 (4-
𝑉𝐼𝑛𝑡 = 𝑓𝑊𝐼 + ∑ 𝑓𝑊𝐼,𝑖 + 𝐹𝑏
𝑖=2 1)

(4-
𝑉𝐸𝑥𝑡 = 𝑓𝑊𝐸 + ∑𝑖=2 𝑓𝑊𝐸,𝑖 − 𝐹𝑏
2)

The terms 𝑓𝑊𝐼 , 𝑓𝑊𝐸 , ∑𝑖=2 𝑓𝑊𝐼,𝑖 and 𝐹𝑏 are the interior (and exterior) wall forces from the
applied lateral loads at the level above the podium interface, the total wall forces transferred from
above storeys and the resultant in-plane strutting forces at the same floor level respectively. The
terms 𝑉𝐼𝑛𝑡 , 𝑉𝐸𝑥𝑡 represent the total shear force intensities in the interior and exterior walls at the
interface level respectively. Equations 4-1 and 4-2 are rearranged and the increment in wall shear
force immediately above the podium level is defined as ∆𝑓𝑖𝑛𝑡 = 𝑉𝐼𝑛𝑡 − ∑𝑖=2 𝑓𝑊𝐼,𝑖 . These forces
are explicitly defined as the product of the stiffness and wall displacement (Eq. 4-3 & Eq. 4-4).

Figure 4-1 Internal force distribution in the tower walls above the interface level

∆𝑓𝐼𝑁𝑇 = 𝛿𝐼𝑁𝑇 𝐾𝐼𝑁𝑇 + 𝛿𝐼𝑁𝑇 𝐾𝑏 + (𝛿𝐸𝑋𝑇 − 𝛿𝐼𝑁𝑇 )𝐾𝑏 (4-3)

∆𝑓𝐸𝑋𝑇 = 𝛿𝐸𝑋𝑇 𝐾𝐸𝑋𝑇 − 𝛿𝐸𝑋𝑇 𝐾𝑏 − (𝛿𝐸𝑋𝑇 − 𝛿𝐼𝑁𝑇 )𝐾𝑏 (4-4)

(δEXT − δINT ) in Eq. 4-3 and 4-4 amounts to the extension or the contraction of the
coupling beam connecting the walls. K INT , K EXT are the total (flexure and shear) lateral
stiffness of the interior and exterior walls respectively. Equations 4-3 and 4-4 can also be
presented in the matrix format (Eq. 4-5).

104
 𝐾𝐼𝑁𝑇 +𝐾𝑏 𝛿 ∆𝑓
[ ] [ 𝐼𝑁𝑇 ] = [ 𝐼𝑁𝑇 ] (4-5)
+𝐾𝑏 𝐾𝐸𝑋𝑇 − 2𝐾𝑏 𝛿𝐸𝑋𝑇 ∆𝑓𝐸𝑋𝑇

In this way, increment in the shear intensity of the walls can be computed given that the value
of the relative wall displacement (𝛿𝐼𝑁𝑇 − 𝛿𝐸𝑋𝑇 ) is known in priori. The total internal forces
(𝐹𝑆𝑇𝑅𝑈𝑇,𝑡𝑜𝑡 ) in the connecting beams above the podium level can also be computed as:

𝐹𝑆𝑇𝑅𝑈𝑇,𝑡𝑜𝑡 = ∑ 𝐹𝑆𝑇𝑅𝑈𝑇,𝑖 (4-6)

𝑖

where the term ‘i’ indicates the number of storeys above the podium interface where significant
strutting forces are expected. It has been demonstrated in Chapter 3 that the compatibility forces
in the connecting floor slabs or beams are prominent in the storeys close to the interface level. In
the tower above the interface, these internal forces are significant in only two storeys above the
podium interface level (i.e. i =1→ 2). It can be seen that the magnitude of 𝐹𝑆𝑇𝑅𝑈𝑇,𝑖 (strutting force
at storey i) is composed of the strutting (restoring) forces generated in the coupling element (beam
or slab) and the force in the wall due to the reaction of the coupling beam on the
wall (𝑓𝑏/𝑊𝐼 , 𝑓𝑏/𝑊𝐸 ). Forces 𝑓𝑏/𝑊𝐼 , 𝑓𝑏/𝑊𝐸 are attributed to local deformation of the wall in
reaction to in-plane strutting forces transferred from the connecting beams. These deformations
are shown to be proportional to the relative stiffness of the wall and the coupling beam (Eq. 4-7
& 4-8).

𝐾𝑊 − 2𝐾𝑏
𝐹𝑆𝑇𝑅𝑈𝑇,𝑖 = (𝛿𝐼𝑁𝑇 − 𝛿𝐸𝑋𝑇 )𝑖 (𝐾𝑏 )𝑖 ( ) (4-7)
𝐾𝑊 𝑖

𝑖=𝑛
𝐾𝑊 − 2𝐾𝑏
𝐹𝑆𝑇𝑅𝑈𝑇,𝑡𝑜𝑡 = ∑(𝛿𝐼𝑁𝑇 − 𝛿𝐸𝑋𝑇 )𝑖 (𝐾𝑏 )𝑖 ( ) (4-8)
𝐾𝑊 𝑖
𝑖=1

where 𝐾𝑊 is the total (shear and flexure) lateral stiffness of the most flexible (i.e. shorter) wall
(interior or exterior). KW can be estimated by analysing the wall in isolation (i.e. fixed at its base).
𝐾𝑏 is the gross in-plane stiffness of the coupling beam:
𝐸𝐶 𝐴𝑒𝑓𝑓
𝐾𝑏 = (4-9)
𝐿

Eq. 4-9 can be further evaluated when the slab or beam are subject to tensile or compressive
strains in order that a more rigorous estimate of the in-plane stiffness can be made (Beyer et al.,
2005). In the expanded form, contributions from both the reinforcing steel and the concrete jointly
define the gross beam stiffness under compression (Eq. 4-10). On the other hand, the uniaxial
stiffness of the reinforcing steel can be used to define the in-plane stiffness of the beam (or the
slab) under tension (Eq. 4-11).

(𝐾𝑏 )𝐶𝑜𝑚𝑝 = 𝐸𝐶 𝐴𝑒𝑓𝑓 [ 1 + 𝜌𝑟𝑒𝑜 (𝑛 − 1)]/𝐿 (4-10)

105
 (𝐾𝑏 ) 𝑇𝑒𝑛𝑠𝑖𝑜𝑛 = 𝑛𝐸𝑐 𝐴𝑒𝑓𝑓 𝜌𝑟𝑒𝑜 /𝐿 (4-11)

where n in Eq. 4-10 and 4-11 is the modular ratio (i.e. 𝐸𝐶 /𝐸𝑆 ), 𝜌𝑟𝑒𝑜 is the longitudinal
reinforcement ratio (parallel to the direction of the applied loads) and 𝐿 is the length of the slab
or the beam between the examined walls.

Eq. 4-8 suggests that an upper bound estimate of the strutting in-plane forces can be made by
considering only the in-plane stiffness of the connecting link element (e.g. beam). This assumes
that lateral stiffness of the wall (KW) at a storey is larger than the in-plane stiffness of the
connecting beam (i.e. 𝐾𝑊 ≫ 𝐾𝑏 ). As such, a conservative estimate of FSTRUT is obtained by
making the term (KW-2Kb/Kw) in Eq. 4-8 (and Eq. 4-7) equal to unity (as also demonstrated in
Figure 4-2).

Figure 4-2 The effect of the relative wall-coupling beam stiffness (KW/Kb) on the axial
(in-plane) coupling beam forces

Analyses results obtained from the parametric study described in Section 3.2 on 2D building
models with podiums at different heights (see Figure 3-8) are examined herein to verify the
developed analytical approach. The differences in wall displacements between the interior and
the exterior walls (𝛿𝐼𝑁𝑇 − 𝛿𝐸𝑋𝑇 ) are normalised with respect to the displacements of coupled
structural walls that are not connected to a podium (as shown in Figure 4-3) annotated as δo (at
each storey level).

106
 Figure 4-3 Analysed tower walls without the podium structure for determining the value
of 𝜹o

The strutting force profiles computed by Eq. 4-7 are benchmarked against results obtained
from the FE analyses (Figure 4-4(b)). The corresponding shear forces up the height of the
buildings have already been presented in Figure 3-10. It is demonstrated that conservative
estimates of the strutting forces in the beams (or slabs) connecting the walls can be obtained by
the use of the analytical model presented in Eq. 4-7 (see Figure 4-4). Figure 4-5 further
demonstrates that the ratio (𝛿𝐼𝑁𝑇 − 𝛿𝐸𝑋𝑇 )⁄𝛿𝑜 exhibits large values when the podium is positioned
in the lower 20%-30% of the height of the building (Figure 4-5). Moreover, displacement
incompatibilities between the tower walls are much reduced when the podium is positioned at a
higher level.

107
 (a) Relative displacement (b) Analytical vs. FE strutting force
distribution distribution
Figure 4-4 Results from the parametric study on podium-tower sub-assemblage
models with different podium heights

Figure 4-5 Variation of the normalised relative wall displacement with podium
height ratio

In Section 3.2, the relative podium-tower lateral stiffness ratio (KT/Kp) has been demonstrated
to be a key parameter for characterising the restraint of the podium on the tower walls. The
magnitudes of the strutting forces (in the link elements between the walls) have been shown in
Figure 3-13 to be particularly sensitive to this ratio (KT/Kp ).

Figure 4-6 outlines a comparison of the strutting force distributions predicted by Eq. 4-7 and
those from FE analyses on 2D podium-tower buildings with different values of KT/Kp (similar to

108
those outlined in Section 3.2). The stiffness-based approach as presented in this section is shown
to be well capable of estimating, to a reasonable degree of accuracy, the distribution and the
magnitude of FSTRUT that is developed in between the tower walls. Further verifications are
presented in Section 4.4.

(a) Strutting force distributions (b) Normalised relative displacement

of connected walls
Figure 4-6 Results from the parametric study on building models with different
(KT/KP) ratios

4.3. A simplified approach for rapid estimation of the strutting force distributions
In this section, a simplified tool is introduced for a rapid estimation of the strutting forces
developed in between the tower walls. The calculation procedure is introduced in two steps. First,
the restraint of the podium on the tower walls is quantified as a function of the magnitude of the
backstay (reaction) forces in the podium slab at the level of the interface (Section 4.3.1). Second,
the critical strutting (compatibility) forces in the few storeys above the interface are then
calculated for given magnitude of the backstay force found from the first step.

4.3.1. Estimate of the backstay podium forces

Backstay forces are the reactions of the podium structure on the walls of the tower under lateral
loads (see Figure 4-7(a)). The backstay effect, as described, has been shown to result in high shear
force reversals on the lateral load resisting member (walls and cores) of the tower at and below
the level of the interface. These reactions (by the podium) have also been demonstrated (in
Sections 3.2 and 4.2) to be proportional to the stiffness of the podium structure. As such, stiffer
podiums (with lower KT/KP ratios) impose larger differential restraints and higher shear force
reversals on the tower walls at, and below, the level of the interface (compare the shear force
profiles in Figure 3-13). The (differential) restraint by the podium has been identified (in Chapter
3) to be the main attributor to the generation of the strutting (compatibility) forces in the links
between the connected tower walls. It has been demonstrated (in Section 4.2) that the magnitude
of these strutting forces are proportional to differences in the wall displacements (𝛿𝐼𝑁𝑇 −

109
𝛿𝐸𝑋𝑇 ) above the interface level which in turn are proportional to the degree of the restraint by the
podium structure. Clearly, the magnitude of the backstay forces and the strutting forces (FSTRUT)
in the links between the tower walls are directly correlated.

The backstay forces (the reaction of the podium) developed in the interface floor slab of a
building with features similar to those shown in Figure 4-7(a) can be idealised by two analytical
models (bounds) as shown in Figure 4-7(b) & (c).

In Figure 4-7(b) the floor slabs at the podium-tower interface are replaced by a set of springs
with in-plane stiffness values (i.e. spring constants) equal to the in-plane stiffness of the floor
slab. A similar analogy is adopted by Rad and Adebar (2009) for modelling interactions between
the basement sub-structure and the supporting towers in a high-rise building (as discussed in
Section 2.1.2). Assuming linear elastic behaviour of the floor slabs, it can be shown that these
backstay (reactive) forces are proportional to the in-plane deformations of the interface slabs. This
analogy assumes that the connection between the tower and the podium structure is monolithic
(i.e. continuous with no expansion/settlement joints present).

An upper-bound estimate of the reactive forces (generated in the interfacial podium slabs) is
obtained when the podium structure is assumed to be infinitely rigid (i.e. 𝐾𝑇 ⁄𝐾𝑝 ≈ 0). This is
shown in Figure 4-7(c) where an idealised ‘pin’ is positioned at the interface between the podium
and the tower. The analytical model (of Figure 4-7(c)) is further simplified in Figure 4-8. The
reactive force (at the level of the interface) for a single (externally) applied point load (𝑉𝑖 ) at a
height hi (above the interface level) can be resolved by employing the principle of virtual work
(Figure 4-8). Similarly, the reaction (at the pin) of the indeterminate structure (propped cantilever)
for multiple lateral loads applied at various heights (total of ‘n’ loads; where n is the number of
storeys above the interface) is expressed in Eq. 4-12.
n
3 hi
Reaction (at the pin support) = ∑ Vi [( )+1] (4-12)
2 hp
i=1

110
(a) 2D model of the building (b) Simplified analytical model (c) Upper-bound model (propped
cantilever)
Figure 4-7 Analytical represntation of the podium restraint on the tower structure

111
 Figure 4-8 Shear force distribution and shear reversal in the propped cantilever
model of the building (based on application of a single point load)

Following the above description (propped cantilever model in Figure 4-8), an estimate of the
maximum shear reversal experienced at the level of the interface (backstay effect) is expressed in
below (Eq. 4-13).
n
3 hi
FBS,pinned = ∑ Vi ( ) (4-13)
2 hp
i=1

where hi is the height to the location of the externally applied load Vi measured above the interface
level (all other parameters are consistently defined in Figure 4-8). The ‘upper-bound’ building
idealisation as described assumes that the podium structure is infinitely stiff and thus the relative
podium drifts (flexibilities) are assumed to be negligible when compared to the drifts of the tower.
Strictly speaking, this idealisation can be overly conservative in most cases since the podium
structure also deforms (laterally) when the building is subjected to lateral loads. The flexibilities
of the podium structure are next examined to refine the ‘pin’ assumption at the interface in order
that a realistic estimate of the reactive backstay forces can be obtained. At the podium-tower
interface, the backstay forces that are developed in the main backstay diaphragm are proportional
to (a) the in-plane stiffness of the main backstay diaphragm (annotated by 𝐾𝐵𝑆 and not to be
confused with Kb in Eq. 4-9) or (b) the total lateral stiffness of the podium (Kp) or (c) a combination
of (a) and (b).

For a building with a rigid podium structure where the lateral stiffness of the podium (𝐾𝑃 ) is
much larger than that of the tower (𝐾𝑇 ), the shear reversal experienced at the interface is mainly
dependent on the in-plane flexibility (or stiffness) of the backstay diaphragm (KBS in Figure 4-9).
This is schematically presented in Figure 4-9(a) where the stiffnesses of the podium-tower

112
assembly are formulated analogously to a system of springs-connected-in-series. The theoretical
basis of this analogy is such that resulting reaction of the podium (on the tower) is proportional
(to varying degrees) to the lateral stiffness of the podium and the in-plane stiffness of the
interfacial slab. Accordingly, the total stiffness of the podium structure KP,total can be estimated
by the use of Eq. 4-14.
−1
1 1
𝐾𝑃,𝑡𝑜𝑡𝑎𝑙 =( + ) (4-14)
𝐾𝐵𝑆 𝐾𝑝

It can be shown that for a hypothetical building scenario where 𝐾𝑝 → ∞, the deformations
(and forces) in the lower podium levels are proportional to 𝐾𝐵𝑆 . In this case, the stiffness of the
‘restraint’ is approximately identical to the ‘pin’ support idealisation (discussed above) given that
the in-plane stiffness of the interface diaphragm is considerably high (which is commonly the
case in buildings). Also in this case, KP,total takes a very high value. Conversely, when the stiffness
of the podium is comparable to the stiffness of the tower (typical 𝐾𝑇 /𝐾𝑃 values are in the range
of 0.2-0.6 as discussed in Chapter 3), the restraint of the podium is proportional to the relative
lateral stiffness ratio (𝐾𝑇 /𝐾𝑝 ) as shown in Figure 4-9(b) (in which case KBS >> Kp).

(a) building model with high relative (b) building model with comparable
podium-tower stiffness ratio relative podium-tower stiffness ratio

Figure 4-9 Analytical represntation of the interfacial restraint of the podium on the
tower structure

113
 A series of parametric studies have been performed on several 25 storey sub-assemblage
building models (similar to those shown in Figure 3-2(a)) to investigate the effects of the
value of 𝐾𝑇 /𝐾𝑃 on the reactive (backstay) forces at the interface. Buildings with podium
height ratios of hp/hb= 0.2, 0.23, 0.34 and 0.4 have been investigated since buildings with this
range of ratios (hp/hb) have been shown to exhibit significant interferences by the podium
structure (in Section 3.2). For each of the building models the flexural and shear stiffnesses
of the podium structure (primarily walls) were varied in order that different relative stiffness
ratios can be obtained (𝐾𝑇 /𝐾𝑃 ratios in the range of 0.1 to 0.8 have been examined as shown
in Figure 4-10(b)). The results of the parametric study are summarised in Figure 4-10(a)
where the backstay forces obtained from the FE simulations are normalised with those
obtained from the hypothetical ‘pin’ idealisation (i.e. Eq. 4-13 which is illustrated
diagrammatically in Figure 4-8).

(a) A comparison of the backstay forces obtained (b) building models employed in the
from the FE simulations to the forces generated parametric study
assuming a ‘pinned’ connection

Figure 4-10 Comparison of the backstay (reactive) forces

It is demonstrated in Figure 4-10(a) that the values of the maximum reactive (backstay)
forces retrieved from FE analyses (FBS, FE) can be significantly lower than those predicted by
the analytical ‘pin’ model (FBS,pinned). Considerable reductions in the podium restraint with
increasing values of (KT/Kp) are demonstrated in Figure 4-10(a). This observation is consistent
with that made earlier (in Section 3.2) where reductions in the magnitude of the strutting
forces (in the tower floor slabs) have been reported for buildings with high values of
𝐾𝑇 /𝐾𝑃 (Figure 3-13). The good consistencies indicate that a direct correlation exists between
the backstay (reactive) podium forces and the strutting forces in the floor slabs. The results
of the parameter study are next employed to refine Eq.4-13 by incorporating the flexibility of

114
the podium. A closed-form solution is developed by regression analysis of the results
presented in Figure 4-10(a) (Eq. 4-15)

3h
FBS = [∑ni=1 Vi (2 h i )] × 𝛾𝐵𝑆 (4-15a)
p

where

𝛾𝐵𝑆 = −0.1 × ln(𝐾𝑇 ⁄𝐾𝑃 ) (4-15b)

The first (bracketed) term of Eq. 4-15a quantifies the maximum magnitude of the shear
reversal for an idealised “pin” support at the interface. The second term (𝛾𝐵𝑆 ) incorporates
contributions of podium flexibility (and thus stiffness) on the reaction (backstay) force FBS.
The values of Vi are the externally applied lateral loads which in turn are obtained by
performing response spectrum analysis (or the equivalent lateral force method) on the
building. The values of 𝐾𝑇 and 𝐾𝑃 can be obtained by analysing the podium and tower
structures in isolation (as discussed in Section 3.2 and further details provided in Appendix
A-3). Consistencies in the comparison of the values of the backstay forces obtained from FE
analyses (described earlier) and those predicted by Eq. 4-15 is demonstrated in Figure 4-11.

Figure 4-11 Comparison of backstay forces from the predictive (analytical) model to
the results obtained from FE analyses

It should be noted that the prominent assumption of the described analytical model is that
the in-plane stiffness of the main backstay diaphragm (interfacial slab) is much larger than
the lateral stiffness of the podium (i.e. 𝐾𝐵𝑆 ≫ 𝐾𝑝 ). This assumption may not be accurate for
quantifying the interactions between a building and the basement sub-structure (similar to
those reported by Rad and Adebar (2009)). The lateral stiffness of an underground sub-
structure comprises of the stiffness of the basement walls and the surrounding (retained) soil
mass. This total stiffness (structure + soil) is typically much larger than the in-plane stiffness
of the basement diaphragms. In this case, using the terms adopted in this study, the value of

115
 𝐾𝑃 is much larger than the value of 𝐾𝐵𝑆 and consequently the backstay forces are dependent
on the in-plane stiffness of the basement diaphragm (this case is schematically shown in
Figure 4-9(a)).

4.3.2. Strutting (compatibility) force distribution

The backstay forces that have been determined in the previous section are employed herein to
estimate the maximum magnitude of the strutting (compatibility) forces and their distribution up
the height of the tower. It was demonstrated in Chapter 3 (and in the earlier sections of this
chapter) that the strutting (compatibility) forces in the connecting beams/slabs are proportional to
the extent of the restraint by the podium on the tower walls (e.g. Figure 3-13). The distribution of
the strutting forces in the (tower) floor slabs is approximately linear in the storeys above the
interface (as consistently demonstrated in Figure 3-3, Figure 3-10, Figure 3-12 & Figure 3-13).
Hence, the distributions (and values) of the strutting forces are known once the value (of FSTRUT)
at the interface level is determined. It can be shown (by equilibrium) that the maximum (upper-
bound) value of the strutting force (in the tower) at the interface level is equal to the magnitude
of the backstay force (FBS) in the podium slab. By the use of this assumption, the force equilibrium
in the tower floors (above the interface) necessitates the full transfer of these (internal) forces to
the connected walls (as “additional shear force” as shown in Figure 4-12.) This argument is
equivalent to having the parameter (KW -2Kb/KW) in Equations 4-7 and 4-8 equated to unity (as
discussed in Section 4.2).

Figure 4-12 Illustration of the assumed distribution of internal forces in the vicinity of
the interface level

The triangular profile of FSTURT (in Figure 4-12) defines the distribution of these internal forces
with the maximum value equal to the value of FBS at the level of the interface and a zero value
(i.e. no force) in three storeys above the interface levels. A simple linear formulation of the
maximum values of FSTRUT in the links (slabs or beams) above the interface is expressed in Eq. 4-

116
16.
F
BS
Fstrut,i = −FBS + 3h (hi ) (4-16)
i

where FBS is the magnitude of the backstay force at the interface level (Eq. 4-15), 3ℎ𝑖 is the height
(level) at which compatibility forces are assumed to be insignificant (close to zero), ℎ𝑖 is the floor-
to-floor height in the tower and 𝐹𝑠𝑡𝑟𝑢𝑡,𝑖 is the maximum compatibility forces that are developed
in level (i) above the interface. It should be noted however the strutting forces computed in this
manner (i.e. use of Eq. 4-16) assumes that displacement incompatibilities in between the tower
walls occur only above the interface. Strictly speaking, considerable interferences (by the podium)
are also manifested in several levels below the interface. However, this study focuses on the
occurrences of shear anomalies above the interface which is expected to be more critical than
those occurring below the podium. Shear force demands on the walls in these (podium) storeys
are typically offset by shear reversals (on the walls) that occur at and below the level of the
interface. Thus shearing forces developed in the walls are somewhat less critical to the design of
the wall in the podium storeys than the tower storeys as demonstrated in Chapter 3 (e.g. Figure
3-6(a) and Figure 3-10).

The strutting force distributions (above the podium) calculated using the simplified analytical
approach (Eq. 4-16) are compared with the distributions obtained from FE analyses on 2D
building models as described in the earlier section (Figure 4-13(a)-(d)). Good consistencies have
been demonstrated when the simple analytical tool (which is based on the propped-cantilever
analogy of the building) is employed to estimate the magnitude of the internal forces that are
generated locally in the floor beams/slabs above the interface level. This force is equal (in
magnitude) to the additional shear force (annotated by ∆𝑉𝑠𝑙𝑎𝑏 ) that is transferred to the interior
tower walls, or columns. Further verifications of the use of this model are presented in the next
section.

117
 (a) ℎ𝑝 ⁄ℎ𝑏 = 0.20

(b) ℎ𝑝 ⁄ℎ𝑏 = 0.24

Figure 4-13 continued on next page

118
 (c) ℎ𝑝 ⁄ℎ𝑏 = 0.34

(d) ℎ𝑝 ⁄ℎ𝑏 = 0.40

Figure 4-13 Comparison of the analytical strutting (in-plane) force distributions to
those obtained from the FE simulations on 2D models with various hP/hb and
KT/KP ratios

119
 4.4. Verifications using a case study building
The 3D case study building introduced in Section 3.4 is employed in a verification study of
analytical models (introduced in the previous sections). The stiffness based approach (Section
4.2) and the simplified approach (Section 4.3) are employed to verify their accuracies in
estimating the internal strutting forces in the floor slabs of the tower.

For Set I & II walls (details of the walls have been already presented in Section 3.4), the
stiffness parameter (𝐾𝑊 ) has been determined by analysing the wall in isolation (and fixed at its
base). The flexible (shorter) wall has been analysed in the program ETABS where the
displacement at the effective height (eff) of the wall is used to determine the total stiffness of the
wall (i.e. KW = Fbase / eff; Fbase is the shear force at the base of the wall). This procedure (for
determining KW) for determining the stiffness of a wall is recommended by Priestley et al. (2007).
Parameter 𝐾𝑏 has been computed for the connecting floor slab (using Eq. 4-10) with effective
section dimensions as shown in Figure 4-14.

In-plane forces in the floor slabs from the FE model can be evaluated by means of section cuts
at three distinct locations along the length of the span (shown by the green lines in Figure 4-14).
The described approach is commonly adopted for determining the value of internal forces when
the shell elements are employed in modelling the floor slabs (see Section 2.2.3). The in-plane slab
forces are determined by integrating the in-plane stresses at the section cuts (as done automatically
in the program ETABS). The internal strutting slab force distributions (up the height of the tower)
are plotted in Figure 4-16. Predictions of the strutting forces in the floor slabs have been computed
using Eq. 4-7 and superimposed in Figure 4-16 for direct comparison. A listing of the values of
the main parameters in Eq. 4-7 is outlined in Table 4-1. The maximum predicted value of the ratio
𝛿𝐼𝑁𝑇 −𝛿𝐸𝑋𝑇
( 𝛿𝑜
) for the case study building (with ℎ𝑝 /ℎ𝑏 = 0.4) is 0.05 (Figure 4-5) which is in good

agreement with the maximum values of (𝛿𝐼𝑁𝑇 − 𝛿𝐸𝑋𝑇 / 𝛿𝑜 ) obtained from the FE analyses (Figure
4-15). The incompatible (unequal) wall displacements are also shown to be most pronounced in
the vicinity (at and above) of the interface. These incompatibilities are reduced to zero at locations
away from the interface level. The estimated and observed (from FE analyses) strutting forces
follow a similar trend (refer Figure 4-16).

120
Table 4-1 Summary of the key parameters required for calculating FSTRUT (Stiffness-
based approach)

Set I Set II
𝐾𝑤 (Interior wall), kN/m 1.03 × 107 1.10 × 107
𝐾𝑏 , kN/m (using Eq. 4-10) 2.23 × 106 2.10 × 106
𝑘w − 2𝑘𝑏 / 𝑘𝑤 0.57 0.62
hp /hb 0.4 0.4
δINT − δEXT / δo , (Figure 4-5) 0.05 0.05
δo , mm (evaluated at the level of the
10.2 8.6
interface)

Figure 4-14 Typical tower floor plan showing the effective width of the slab
connecting the examined walls

121
 Figure 4-15 Normalised relative wall displacement ratio for set I and II walls

(a) Set I (b) Set II

Figure 4-16 Comparison of the strutting force distributions up the height of the
building

The simplified approach introduced in Section 4.2 (for the rapid estimation of 𝐹𝑆𝑇𝑈𝑅𝑇 ) is also
employed as part of the verification study. The effective lateral stiffness ratio of the building
K T /K p is found to be 0.26 (refer to Table C-1 in Appendix C for detailed calculations) and the
applied forces (Vi) at each storey level have been determined by the program ETABS. The

122
reaction of the podium on the tower walls has been determined using Eq. 4-15 whereas Eq. 4-16
was employed to compute the distribution (of the strutting forces) three storeys above the level of
the interface (Table 4-2 summarises the calculation procedure). The FSTRUT distributions obtained
from the simplified tool is compared with those obtained from the FE analyses and the explicit
model (described earlier) in Figure 4-17. The use of the much simpler approach, which is
formulated by adopting the propped cantilever analogy of the building, is demonstrated to report
slightly higher estimates of FSTRUT in the tower storeys. Nevertheless, the differences are only
marginal.

(a) Set I (b) Set II

Figure 4-17 Comparison of FSTRUT distributions obtained from the analytical models

123
Table 4-2 Summary of calculations of the simplified method (introduced in Section 4.3)

hi 3 hi
𝑉𝑖
ℎ Vi ( )
(height 2 hp
(Applied loads
(height above the above the
obtained from program
base) , [m] interface [kN]
ETABS), [kN]
level), [m]

30 18 665.41 1497.17

27 15 652.25 1222.97

24 12 579.78 869.67

21 9 507.31 570.72

18 6 434.83 326.13

15 3 362.36 135.89

3 hi
∑ Vi ( )
2 hp
12 0 899.65 4622.55
(Eq. 4-13 , [kN]

714.05 𝛾𝐵𝑆
9 (Eq. 4-15b)
0.13
n
476.33 3 hi
[∑ Vi ( )] × 𝛾𝐵𝑆 622.69
6 2 hp
i=1
(Eq. 4-15a) , [kN]

3 237.65

0 0.00

4.5. Effects of in-plane yielding of the backstay diaphragm

The restraint of the podium is governed by the capacity of the main backstay diaphragm in
transferring (redistributing) loads from the above tower structure. Under high-intensity loading,
the main backstay diaphragm (which is attached to the tower structure) is likely to experience
significant cracking and flexural yielding at the interface zone (where high internal forces are
developed). As a direct consequence, the force transfer by the backstay slab can be considerably
reduced. Intuitively, occurrences of inelastic damages at the interface slab can potentially relax
the restraint of the podium (on the tower) which in turn reduce the strutting forces (𝐹𝑆𝑇𝑅𝑈𝑇 ) that
are developed in the floor slabs (or beams) connecting the tower walls. The severity of the
damages experienced at the interface zone is exacerbated when construction joints (also referred
to as cold joints) are present at the junction between the tower and the podium. These joints are
typically manifested as a consequence of construction practice wherein the tower cores are first
erected and the floor slabs are subsequently ‘stitched’ to the cores/walls by the use of dowel
(reinforcement) bars anchored in the cores of the tower (see Figure 4-18). With this form of
construction, early age cracking at the wall-slab connections (joints) are inevitable as these cracks
are attributed to differences in the shrinkage (drying) rates between the floor slabs (that have been

124
cast at a later stage) and the (already set) tower cores. Consequently, these joints introduce
potential ‘weak’ points at the junction between the podium and the tower where severe cracking
is to be anticipated under reversed loading excursions. In practice, the performance of these
(construction) joints is enhanced by the roughening (scabbling) of concrete surfaces at either side
of the joint to increase the area of contact between the core and the attached floor slab. A definitive
assessment of the performance behaviour of these joints in seismic conditions has not been
explored. In this section, a sensitivity study is presented to investigate the effects of the in-plane
yielding of the interface slab on the global and local response behaviour of the building. The next
section presents a description of the analytical model of the building for use in nonlinear quasi-
static analyses (Section 4.5.3) and dynamic analyses (Section 4.5.4).

Figure 4-18 Illustration of a typical construction cycles of a podium-tower (or base-

tower) (artwork adopted from PEER/ATC 72-1 (2010))

4.5.1. Numerical model definition

The numerical model of the podium-tower sub-assemblage as shown in Figure 4-19 is

constructed in program SeismoStruct (2016). The building comprises coupled C-shaped concrete
cores which form the primary lateral system of the tower structure. A 6 storey podium
structure (ℎ𝑝 /ℎ𝑏 = 0.3 ) is connected to the tower structure by means of interface (backstay)
floor slabs. The vertical elements making up the cores have been modelled as inelastic fibre-
based elements similar to the procedure described in Section 3.5.1 (and in Figure 3-27). The cores

125
have been modelled as inelastic frame elements that are positioned in the geometric centre of the
webs and flanges (following the wide-column technique described in Section 3.5). Rigid links
have been introduced to connect the webs and flanges to the coupling beams. At the sectional
level, the elements have been discretised into several ‘fibres’ with distinct uniaxial constitutive
material models similar to those described in Section 3.5.1. The podium structure has been
modelled using an equivalent elastic element with a stiffness chosen to simulate a constant
ratio 𝐾𝑇 /𝐾𝑃 of 0.2. The choice for the elastic modelling of the podium elements is based on the
observations reported in Section 3.6.2 where more adverse podium interferences have been
reported when a similar assumption is made. The low value of KT/KP = 0.2 is also intended to
maximise restraints by the podium on the tower walls (as discussed in the earlier sections). As the
seismic response of the building progresses into the inelastic range, the backstay (reactive) forces
in the interface floor are solely governed by in-plane characteristics of the backstay slab. This is
the case since the stiffness of the elastic podium would be much higher than the degraded stiffness
of the backstay slab (as schematically illustrated in Figure 4-9(a)).

It should be noted that analyses presented in this section are intended to only demonstrate the
effects of the inelastic behaviour of the diaphragm on the restraining forces imposed by the
podium in severe seismic conditions. A detailed listing of the geometries and reinforcement
content of the floor slabs and the tower cores is presented in Table 4-3.

Table 4-3 Summary of design parameters of the main elements of the examined building
model

𝑏𝑒𝑓𝑓 × 𝑡 𝑓𝑐 ′,
[m x m] [MPa] 𝜌𝑙𝑜𝑛𝑔 , [%] 𝜌𝑡𝑟𝑎𝑛𝑠 , [%]

C-Shaped Wall

Flange 2.95 × 0.500 40.0 1.10 0.27

Web 6.60 × 0.500 40.0 0.74 0.27
Floor slabs and coupling beams

Backstay 6.60 × 0.20 32.0 1.2 (two curtains of 1.2 (two curtains of
(interface) slab N16 @ 200 centres N16 @ 200 centres on
on each face) each face)

Coupling beam 0.90 × 0.500 40.0 1.00 (6N25 top and 0.42 (4 legs of N10
(depth × width) bottom) @ 150 mm centres)

126
 (a) 2D model of the sub-assemblage (b) Translational model shapes

Figure 4-19 2D model of the analysed podium-tower sub-assemblage

model and the mode shapes of the first and second translational
modes

4.5.2. Inelastic features of the backstay slab

Elastic frame elements have been utilised for modelling the behaviour of backstay slabs in the
podium storeys. The width of the frame element is equal to the length of the web of the tower
core that connects to it (see Figure 4-20). The in-plane and out-of-plane (bending) response
behaviour of the slab has been lumped into two phenomenological axial and rotational springs
(respectively) that are positioned at either ends of an elastic frame member (as demonstrated in
Figure 4-20). The rotational spring represents the moment-curvature response behaviour of the
slab which is determined by way of a sectional analysis using the program Response2000 (Bentz
& Collins, 2000). Results from the sectional analyses of the slab are plotted in Figure 4-21(a) for
two axial load ratios (which is the axial (in-plane) force divided by the product of the effective
slab area (Aeff) and the characteristic strength of concrete fc’). Not surprisingly, higher flexural
strength is developed in the backstay slab when considerations are made for the in-plane (axial)
forces compared to the conventional case where no in-plane forces are considered. Nevertheless,
the nominal yield curvatures (defined as the curvature at the start of the horizontal portion of the
idealised bilinear curve) are somewhat identical.

The in-plane stiffness of the slab (shown in Figure 4-20) has been defined based on Eq. 4-10.
In view of the very limited experimental verifications on the in-plane characteristics of a floor
slab, a simple bilinear backbone (capacity) curve has been defined to simulate the in-plane force-

127
displacement response behaviour of the floor slab. Two distinct upper-bound strength values have
been selected based on observations reported earlier (in Section 3.6.2) regarding the expected
magnitude of the backstay forces at ultimate conditions (refer Figure 3-39). The adopted strength
‘caps’ are intended to simulate the degradation of the force transfer that would occur at the
interface as a result of the accumulation of inelastic damage at this location. The backbone in-
plane force-deformation responses with maximum (in-plane) strength limits of 10.0% and 5.0%
of the gross in-plane capacity of the slab are demonstrated in Figure 4-21(b).

Figure 4-20 Modelling of the backstay slab

(a) Moment-curvature relationship (b) In-plane backbone response

Figure 4-21 In-plane modelling parameters of the backstay slab

128
 4.5.3. Quasi-static pushover analyses on podium-tower sub-assemblage models

The building model described in Sections 4.5.1-4.5.2 has been employed in a series of quasi-
static pushover analyses to illustrate the effects of yielding of the backstay diaphragm on the
response behaviour of the building. Quasi-static (pushover) analyses have been performed using
a load distribution that is proportional to the first mode shape (see Figure 4-19(b)). The choice of
this distribution is based on observations reported in the earlier part of the study (Section 3.5.1).
It should be noted that this part of the study investigates the sensitivity of the response of the
building to the in-plane (and out-of-plane) deformations of the backstay diaphragms. The results
reported herein are to compare various anticipated trends of backstay responses. An elaborate
performance assessment studies on this type of building configuration will be presented in
Chapter 6.

In this part of the study, three sets of analyses have been performed. The building models are
identical across the three sets except that the in-plane feature of the backstay (interface) slab
differs. Building models with ultimate in-plane strength caps (of the backstay diaphragms) equal
to 0.1𝑓𝑐′ 𝐴𝑒𝑓𝑓 and 0.05𝑓𝑐′ 𝐴𝑒𝑓𝑓 in conjunction with a building model where no such strength caps
placed have been analysed. The pushover (capacity) curves obtained from the analyses are plotted
in Figure 4-22. The initial stiffness is shown to be virtually the same for the three building models
whereas slight variations are observed at ultimate conditions. The building with the lowest in-
plane ‘strength cap’ develops 8.2% lower base shear strength compared with the elastic ‘no-cap’
model. Nevertheless, the observed variations are only marginal and may well be ignored.

Figure 4-22 Pushover (capacity) curves from simulations

The effects of in-plane yielding of the backstay slab are more pronounced when examining the
shear force demands on the tower cores. Figure 4-23(a) plots the core shear forces immediately
above the interface against the roof displacement of the building. The larger shear force demands

129
experienced by the interior core (closer to the podium) compared with the exterior core are clearly
demonstrated for all the building models. Significant increases in the strutting forces (in Figure
4-23(b)) are observed as the exterior core reaches its ultimate lateral capacity at a roof
displacement of 0.15m (at the start of the descending branch of the force-displacement curve).
The consequential surge of the strutting forces (above the interface) is partly attributed to the
redistribution of forces to the interior core following the loss of lateral strength of the exterior
core (i.e. to restore equilibrium conditions above the interface as discussed in Section 2.2). A
comparison of the response trends of the three building models reveals that the reductions in the
backstay forces (due to the imposed strength caps) are paralleled with reductions in the
magnitudes of the 𝐹𝑆𝑇𝑅𝑈𝑇 in the coupling beams (compare the values in Figure 4-23(b)). These
trends are consistent with earlier findings reported in this chapter where the strutting forces in
between the tower walls have been associated with the amount of restraint imposed by the
podium. Lower strutting force values reported for the building model with the lowest strength cap
(0.05f’cAeff). This is attributed to the higher ‘relaxation’ of the restraint (by the podium on the
cores) in the early stages of the response.

(a) Core shear force demands (b) In-plane strutting force demands in the
coupling beams connecting the cores

Figure 4-23 Shear force demands on the core walls and the in-plane (strutting) forces
in the coupling beams

130
 4.5.4. Nonlinear time history analyses

The hysteretic response behaviour of the backstay slab between the podium and the tower is
next examined. The quasi-static response behaviour of the building (with different in-plane
flexibilities assigned to the backstay slab) has given substantial insight on the effects of these
flexibilities on the restraint imposed by the podium on the tower cores. With dynamic analyses,
the hysteretic response (reversed cyclical response) of the interfacial zone will need to be defined
to accurately simulate the relaxation of the restraint (by the podium) that occurs as a result of in-
plane yielding of the backstay slabs. In most cases, hysteretic response of a component (e.g. a
wall, column or beam) is simulated by the use of calibrated elements (e.g. zero-length spring).
This calibration requires well detailed experimental data to ensure that the numerical model can
accurately capture the strength and stiffness under cyclic load excursions (similar to those
presented in Section 3.6.1). However, the lack of experimental programs reported in the literature
on the in-plane flexibility of the floor slabs renders the described approach difficult to adopt. In
the subsequent sections, the in-plane hysteretic response behaviour of the backstay slab is
bounded by the use of several numerical models with distinct hysteretic features to represent the
anticipated behaviour of the slab (at the joints) under cyclic load reversals.

4.5.4.1. In-plane hysteretic response of the backstay diaphragm

The three approaches that are adopted for the modelling of the in-plane response behaviour of
the backstay slab (at the interface) are: the pinched hysteretic model, degraded stiffness model
and elastic model.

The pinched hysteretic model (designated as Model 1) as illustrated in Figure 4-24 is

characterised by low energy dissipation capabilities due to the limited “area” enclosed by the
load-displacement loops (Figure 4-24). This behaviour is typically associated with cracked
reinforced concrete members subjected to reversed loading cycles (Ozcebe & Saatcioglu, 1989;
Priestley et al., 2007; Zhao & Sritharan, 2007). The flat-line (zero stiffness) between the
unloading and reloading (which is the load excursions in the reverse direction) represents the
closing of the crack which would have been widened in the previous displacement cycles. The
resulting slip response is manifested by the closing of the gap formed by the crack which is
associated with a negligible stiffness (depending on the width of the crack). This phenomenon is
illustrated in Figure 4-25(a)-(b) where a single unloading and reloading branch of the force-
displacement curve is presented along with the physical manifestation of the interfacial cracking
when subjected to cyclic loading (Figure 4-25(b)). The existence of ‘cold’ construction joints (at
the interface between the podium and the tower) increases the likelihood of cracking formation at
the podium-tower interface even in the early stage of the response to earthquake excitation (as
described earlier in Section 4.5).

131
 Figure 4-24 In-plane force- deformation hysteretic response of the interfacial slab
(Model 1)

(a) Unloading and reloading response of the (b) Physical explanation of the
backstay (in-plane) in the presence of interfacial ‘slip’ occurring at
interfacial cracks (only a single unloading- the interface (at the
reloading branch is shown for clarity) construction joint)

Figure 4-25 Pinched hysteretic model defining the in-plane response behaviour of the
backstay slab

The second hysteric model (Model 2) also captures the effects of stiffness degradation as a
result of accumulation of inelastic strains at the connection. The simplified Takeda model has
been adopted to simulate the (in-plane) response behaviour of the slab (see Figure 4-26). The
difference between this model and the previous model (Model 1) is in modelling the slip that
occurs at the interfacial zone. The response behaviour of the slab defined by Models 1 & 2 is also
compared to an elastic model with distinct in-plane compression/tension stiffness properties
(Model 3 shown in Figure 4-27). The definition of the model has been presented in Section 4.5.2

132
wherein the stiffness in tension is quantified by Eq. 4-11. The properties of the in-plane response
defined by the various models (Model 1-3) are summarised in Table 4-4.

Figure 4-26 Modified Takeda hysteretic model defining the in-plane response
behaviour of the backstay slab (Model 2)

Figure 4-27 Linear model defining the asymmetric (tension/compression) in-plane

response behaviour of the backstay slab (Model 3)

The flexural response behaviour of the interface slab is modelled as per results from moment-
curvature analyses performed on the floor slabs (presented in Figure 4-21(a)). The resulting
curvature predictions are then converted to rotation predictions by multiplying their respective
values by half the span length as recommended by Priestley (2007) and FEMA 356 (2000). The
hysteretic rules of the flexural response have been defined similarly to the in-plane responses (of
the slab). Unlike in-plane slab properties, flexural properties are identical for both the
negative/positive moments as similar reinforcement content (mesh) is provided at the top and
bottom face of the slab. In all models, dowel actions of the reinforcement bars in transferring

133
shear forces are neglected.

It should be noted that the above described slab modelling approach is different to that used in
Chapter 3 (and Chapter 6) wherein fibre-based inelastic element was used to represent flexural-
axial interactions in the floor slabs. The lumped plasticity approach adopted in this section is
intended to showcase (by way of a sensitivity study) some potential response trends of the
interface slab and their respective effects on the response behaviour of the tower.
Table 4-4 Slab (in-plane) modelling parameters adopted in the inelastic study

Slab modelling
Formulation Value
parameters

General parameters

Stiffness in compression,
Eq. 4-10 3779218.13 kPa
(Kb)Comp
Stiffness in Tension,
Eq. 4-11 211200.00 kPa
(Kb)Tension

10% of the in-plane compressive strength

𝐹𝑐 3379.20 kN
(using a depth of 0.8 × thickness of the slab)

Assumed to take a low value

𝐹𝑇† 𝐹𝑇 = 0.1 × 𝑓𝑦 × 𝜌𝑙𝑜𝑛𝑔 × 𝐴𝑒𝑓𝑓 532 kN

Model 1 (pinched)

Using Sivaselvan and Reinhorn hysteretic rule

Slip parameter (2000) available in SeismoStruct. The slip 0.25
parameter (governing pinching)

Model 2 (modified Takeda)

Using a modified bilinear Takeda model available in SeismoStruct program

(based on the formulation initially proposed by Otani (1981)). The illustrative
Figure below is obtained from SeismoStruct manual

Modified bilinear
Takeda model

134
 The degrading (unloading) stiffness relationship
proposed by the above referenced work is given by
Stiffness the expression:
degradation factor 𝐷𝑦 𝛽0
𝑘 = 𝑘𝑦 ( ) 0.60
(outer hysteretic 𝐷𝑚
loop) (𝛽𝑜 )
where 𝐷𝑦 , 𝐷𝑚 is the yield and previous maximum
displacement and 𝑘𝑦 is the yield stiffness

The unloading (post elastic) stiffness degradation is

Stiffness computed by:
degradation factor
𝐷𝑦 𝛽0 0.50
(inner hysteretic 𝑘 = 𝑘𝑦 ( ) × 𝛽1
loop) (𝛽1 ) 𝐷𝑚

† The low value for the tensile limit is to recognise that cracking in the concrete under tensile forces can
significantly reduce the tensile backstay force transferred to the podium at the level of the interface.
Thus, the upper limit is used for capping the maximum compression force in the backstay slab is also
used to cap the maximum tensile forces that can be developed at the connection between the tower and
the podium.

4.5.4.2. Global and local responses of the building models

The building models have been subjected to ground accelerations as recorded in the 1994
Northridge earthquake (record no. 40 details of which are given in Table A-1 in Appendix A).
Rayleigh (modal) damping has been assigned to model the equivalent viscous damping similar to
the procedure described in Section 3.5.1. The first and second modes of vibration are
demonstrated in Figure 4-19(b). Seismic masses were calculated manually based on the tributary
areas and were lumped at the nodes. A comparison of the base shear time histories obtained from
the analyses of the three building models is given in Figure 4-28. Differences in the base shear
time histories were only marginal and generally occurred only after the building had sustained
flexural yielding. Even then, the extent of discrepancies (in the base shear time histories) was not
significant. Not surprisingly, similar observations have been reported from quasi-static analyses
(Section 4.5.3).

Figure 4-28 A comparison of the base shear time histories obtained from analyses

The backstay forces of the elastic model (Model 3) are shown to be significantly larger
compared with those obtained from the inelastic slab models (Models 1 and 2 in Figure 4-29). It

135
is further shown in the figure that the elastic backstay force demands exceed the capacity of the
slab particularly in the positive (tensile) load excursions of the backstay slab.

Figure 4-29 A comparison of the backstay (reactive) podium forces

The reduction in the extent of the restraint (interference) by the podium on the tower structure
is further shown when comparing shear forces of the cores (interior and exterior) supporting the
tower. Differences in the shear demands on the interior (C-shaped wall connected to the backstay
slab) and exterior cores are most pronounced for model 3 (with elastic in-plane slab properties
Figure 4-30(d)). In contrast, the relaxation of the restraint by the podium upon yielding of the slab
(in-plane) resulted in comparable shear force demands on the cores (exterior and interior as
demonstrated in Figure 4-30(a)). This confirms the earlier stated description of the effects of in-
plane yielding on the redistribution of internal actions above the interface.

A comparison of the strutting in-plane force time histories in the beams three levels above the
level of the podium is outlined in Figure 4-31. For all three modes, the magnitudes of the forces
are shown to decrease with increase in height above the level of the interface. The elastic slab
model (Model 3) is demonstrated to exhibit highest compatibility force demands in the beams
connecting the tower cores (Figure 4-31), which further explains the occurrence of high shear
offsets in the cores of this building model (large differences between the interior and exterior
cores as shown in Figure 4-30(c)). The time trace of 𝐹𝑆𝑇𝑅𝑈𝑇 (of Model 1) at the level immediately
above the interface is shown to ‘shift’ upwards (into the tension range as shown in Figure 4-31(c)).
This effect is attributed primarily to differences in the flexural yielding between the interior and
exterior core where high irrecoverable deformations are sustained by the exterior core (but not
the interior) above the interface level. These deformations (of the exterior core as shown in in
Figure 4-32(c)) impose tensile forces onto the connecting beams. Similar findings have been
reported in Beyer (2005) and Gardiner (2011).

136
 (a) Model 1 (b) Model 2 (c) Model 3 (d) Definition of the
cores
Figure 4-30 Core shear force time histories above the interface level

(a) Model 1 (b) Model 2 (c) Model 3 (d) Definitions

Figure 4-31 In-plane forces in the coupling beams above the interface level

137
 (a) Model 1 (b) Model 2 (c) Model 3
Figure 4-32 Inelastic (flexural) yielding progression in the tower cores and coupling
beams

In-plane yielding of the main backstay diaphragm has a negligible influence the global
response behaviour of the building. Particularly, the base shear (Figure 4-28) and roof
displacement (Figure 4-33) time histories are shown not to be affected significantly by the
response behaviour of the interfacial zone between the tower and the podium.

Locally, shear demands on the tower cores and the generation of the strutting forces in between
the cores are highly sensitive to the accumulation of inelastic damage at the interface. Striking
differences in the inelastic demand distributions up the height of the tower are also observed from
analyses of the building in which the backstay slabs feature distinct inelastic features (Figure
4-32). In Model 1, the inelastic demands on the tower cores are much less than those observed for
Model 3. These differences are attributed to the localisation of the inelastic demands at the
interfacial (weak) zones in the building. In other words, most of the seismic (input) energy would
have been dissipated by the yielding of these interfacial zones (in-plane and in bending) whereas
only marginal seismic demands are imposed on the tower cores. Strictly speaking, this response
behaviour is analogous (to some degree) to buildings where external seismic energy dissipaters
have been installed.

Although only three different models have been examined, this study highlights the high
sensitivity of the strutting forces and the shear force distributions to the degree of flexibility
prompted at the interfacial zone. It is acknowledged that further studies are required to verify the
modelling assumptions that have been adopted in this study. A more elaborate study with

138
considerations made for the shear stiffness and strength degradations of walls in podium-tower
buildings is presented in Chapter 6.

Figure 4-33 Comparison of the roof displacement time histories of the three models

4.6. Effects of expansion/settlement joints at the podium-tower interface

In the design of a podium-tower building separation joints are often positioned at the interfacial
zone between the podium structure and the tower. These joints are required to reduce the degree
of (structural) indeterminacy in the building undergoing differential settlements (wherein the sub-
soil directly under the tower is subjected to larger settlements than those supporting the podium).

Figure 4-34 Illustration of a podium-tower building undergoing differential

settlement and typical details of the ‘floating bay’

An illustration of the expansion joints (in this context they are often referred to as settlement

139
joints) at the podium-tower interface is outlined in Figure 4-34. It should be noted that a full
separation of the tower and the podium (including the foundation) may not be practically feasible
especially for buildings that are supported on a raft foundation (a continuous slab foundation).
Alternatively, ‘floating’ bays are introduced at the podium-tower interface. The term ‘floating’
refers to backstay slabs that rest atop of two ledges at either end of the span (between the
supports). This form of slab detailing is favoured as it provides in-plane continuity of the floor
between the tower and podium whilst relieving internal forces and moments that can be generated
in the building undergoing differential (uneven) settlement. A schematic representation of the
structural system of a floor slab incorporating the use of floating bays is shown in Figure 4-35.

Figure 4-35 Structural simplification of a floating bay and internal actions resulting
from the differential settlement of the support

The internal actions that are developed in the interface slab as a result of differential support
settlement (𝛿𝑑𝑖𝑓𝑓 ) is presented in Figure 4-35. The differential settlement of the building would
not result in additional bending moment actions at the podium (M=0 in Figure 4-35). In other
words, bending moments in the slabs (due to gravity loads) remain virtually unchanged before
and after the settlement (as shown in Figure 4-36). These joints are typically positioned at
locations of minimum internal forces (along the length of the span); such conditions are typically
achieved at a distance of 1/5th of the length of the span measured from the end.

140
 Figure 4-36 Bending moment distributions in the podium slabs before and after the
settlement (differential) of the tower

A 20mm-40mm separation gap is typically accommodated at the locations of the joints. The
interface (backstay) slab rests on a corbel or a short cantilever ledge as shown in Figure 4-34 it
should be noted that the differences between the two types are only in their design; corbels are
designed for direct shear transfer by the strut-and-tie approach. Vertical actions (due to gravity
loads on the slabs) are transmitted across the joint either by direct bearing on an elastomeric
neoprene pad (Figure 4-34) or by flexure (bending) when the ledges are designed as short
cantilevers. The vertical (differential settlement) and the lateral (expansions/contraction of the
slab relative to the supporting members) movements are accommodated at the gaps. Alternatively,
a partial moment release (where M at the joint is very low but ≠ 0) can also be detailed at the
podium-tower junction during construction. This is achieved by casting low-strength concrete
(shrinkage strips) in the portion of the slab closest to the tower or by detailing the steel
reinforcement of the interface slab to not transfer bending moments to the supporting tower wall
(i.e. reducing the top slab reinforcement at the connection with the wall).

Conversely, a full moment connection between the podium slab and tower is achieved when
the main backstay slabs are cast monolithically with the adjacent tower structure. The reinforcing
bars of the slab (at the top face) are developed into the tower structure to ensure a full transfer of
the internal actions between the two portions of the building.

Table 4-5 provides an outline of the typical types of slab-wall connections and their respective

141
analytical representations in a building.

Table 4-5 Summary of the typical types of connections in a podium-tower building

Connection Typical detailing of the interfacial Analytical Representation/ support

type zone between the tower and the podium reactions
Monolithic connection (full
continuity)
backstay slabs supported on a
Joint connection (type 1:

cantilever ledge)
backstay slabs supported on a
Joint connection (type 2:

corbel)

Static lateral analyses on building models similar to those outlined in Figure 3-2(a) are
performed to illustrate the effects of these forms of construction on the shear force and bending
moment demands of the tower walls. The ‘gap’ element, which is readily available in the program
ETABS (Habibullah, 1997), has been calibrated to model the properties of the joint. Essentially,

142
the compressive in-plane stiffness of the slab is mobilised upon the complete closure of the gap
as illustrated in Figure 4-37. The in-plane force transfer is assumed to occur instantaneously once
the gap is closed. The in-plane stiffness of the slab (adjacent to the tower) is computed by the use
of Eq. 4-10. The initial branch of the bilinear backbone curve represents a ‘zero-stiffness’ slip
occurring within the separation (which is typically in the order of 20mm-40mm as discussed
above). The modelling procedure as described does not consider the effects of surface pounding
on either side of the joint under dynamic loads. This pounding can result in high (contact) forces
as the two surfaces come in contact with one another. Such local forces are not considered in this
study. Nevertheless, the global force transfer (with or without considerations of the dynamic
pounding) is approximately the same since this (transfer) is required for global equilibrium.

Figure 4-37 Analytical backbone model representing the interfacial

settlement/expansion joints

Four building models featuring distinct types of joints in between the podium and the tower
have been examined. Buildings incorporating joints with a gap (separation) of 20mm and 40mm
have been analysed in parallel to a building model where no such joints are present (i.e.
monolithic) and to a building model that features a full axial release (i.e. gap opening >> 40mm).
The shear force, bending moment distributions in the tower walls are plotted in Figure 4-38 along
with the strutting force distributions developed in between the tower walls. Not surprisingly, the
model with a full axial release exhibits almost no restraint by the podium consequent strutting
force (Figure 4-38(d)). The efficiency of the joint (in reducing the restraint of the podium) is
considerably reduced when the gap opening is reduced from 40mm to 20mm (compare strutting
forces and the shear force distributions in Figure 4-38(a) and Figure 4-38(b)). Where a full axial
release is modelled, the shear force anomalies (offset in the shear demands between the walls) are
not manifested in the tower walls (Figure 4-38(d)).

A comparison of the shear force ratios defined as (𝑉𝐼𝑛𝑡 /𝑉𝐸𝑥𝑡 ) and the maximum 𝐹𝑆𝑇𝑅𝑈𝑇 for
different the case with 20mm joint, 40mm joint, monolithic connection and a full axial release
(i.e. gap opening >> 40mm) are outlined in Figure 4-39. The significant reduction in both shear

143
force intensity ratio (𝑉𝐼𝑛𝑡 /𝑉𝐸𝑥𝑡 ) and the strutting force demands in the connecting beams in
buildings with interfacial joints (releases) are presented.

Wall shear force Wall bending moment FSTRUT distribution in Externally applied
distribution distribution the connecting beams load profile

(a) 20mm joint

(b) 40 mm joint

Figure 4-38 continued on next page

144
 Wall shear force Wall bending moment FSTRUT distribution in Externally applied load
distribution distribution the connecting beams profile

(c) Monolithic connection

(d) Full axial release (large gap)

Figure 4-38 Results of analyse of 2D podium-tower sub assemblage models with

different types of slab-wall connections

(a) (b)
Figure 4-39 Comparison of (a) the shear ratio (VInt / VExt ) and (b) FSTRUT obtained from
analyse of 2D podium-tower assemblages with different podium-tower connectivity

145
 4.7. Chapter summary
In this chapter, analytical solutions are presented for quantifying strutting forces that are
generated in the connecting elements (beams or slabs) spanning in between tower walls. An
explicit stiffness-based approach is introduced in the form of a 2x2 matrix to solve for the internal
forces (and their distribution) up the height of the tower. The solution (for the strutting force in
the connecting elements) is complete once the displacement profiles of the tower walls (subject
to lateral loads) are known. An alternative rapid estimation technique has also been introduced
for approximating the magnitude of these critical strutting (compatibility) forces above the
interface level. This approach adopts a propped-cantilever idealisation of the building which
involves incorporating a ‘hypothetical’ pin support at the level of the interface. Using this
analogy, the reaction forces of the podium (on the tower structure) can be solved using the
classical approach of analysing a structurally indeterminate system. The resulting upper-bound
value of this reaction force is further refined to account for the effects of flexibility of the podium.
The extent of the restraint imposed by the podium has been shown to be highly dependent on the
stiffness of the podium structure. Larger (strutting) forces are anticipated in buildings with a low
tower-podium stiffness ratio (i.e. KT/Kp). The larger forces are attributed to the larger lateral
restraint by the podium on the tower structure. Verifications of the simplified and the stiffness-
based model have been demonstrated on 2D and 3D building models, and good consistencies
have been demonstrated.

The in-plane yielding of the main backstay slab has been investigated to illustrate the potential
reductions of the restraining forces (by the podium) in the inelastic range. The paucity of
information available on the interactions between the podium slab and the tower structure
necessitated the use of various approaches for modelling the effects of cracking and yielding that
can occur at these locations. It is shown that the global behaviour of the building (base shear and
roof displacements) are only marginally sensitive to the local response behaviour of the backstay
slab. However, internal force redistributions occurring above the interface level can be highly
sensitive to the degree of (in-plane) stiffness degradation of the backstay slab. The additional
shear force demands on the tower and core walls imposed by this force redistribution are generally
lower when considerations are made for the inelastic in-plane response behaviour of the interface
slab. The effects of the structural separations between the podium and the tower have also
revealed similar trends. Particularly, a well-detailed expansion joint at the interfacial zone
between the podium and the tower (at the level of the slab) can potentially reduce the shear force
concentrations in the tower walls significantly.

146
 Chapter 5. ANALYTICAL MODELLING OF THE EFFECTS OF
TRANSFER PLATE FLEXIBILITY IN BUILDINGS

5.1. Introduction
Buildings featuring a transfer structure can be commonly found in metropolitan cities situated
in regions of lower seismicity. A transfer structure can be in the form of a rigid plate or an array
of deep girders positioned at the podium level of the building to support the tower structure of the
building. A transfer structure has the function of redistributing gravitational loads where there is
a discontinuity in the load path to accommodate a major change in the architectural layout of floor
spaces in the building. This form of building construction is regarded as a major feature of vertical
irregularity which is expected to compromise the seismic performance of the building in projected
seismic conditions. Even then many such buildings have been designed and built in regions of
lower seismicity (e.g. Hong Kong and Australia). Seismic design provisions that address these
issues are scarce. The reviewed literature highlights major issues relating to shear force
concentrations in the tower walls supported on the transfer plate (refer Chapter 2). The anomalous
increase in the tower wall shear force demands has been identified as a ‘major cause for concern’
in the seismic response behaviour of the building. Yet design guidance or formulations for
estimating these adverse effects are not available.

In this chapter, a simplified method for quantifying the increase in the shear force demand on
the tower walls is presented. Analogies are drawn from the seismic response behaviour of
torsionally unbalanced buildings to develop an analytical formulation for quantifying the effects
of transfer plate distortions in the building. Accordingly, expressions are derived to estimate the
magnitudes of the in-plane strains (and forces) that are developed in the floors slabs between
assemblies of connected tower walls supported on a transfer plate. The next section presents
examples of the anomalous increase in the shear force demands on the tower walls which are
supported by a transfer structure. In subsequent sections of the chapter, the simplified method is
introduced to quantify the adverse effects of the transfer structure. In the final section of the
chapter, accuracies of the introduced model are illustrated by way of worked examples on case
study buildings. The study presented in this chapter is reported in the following publication by
the author (Yacoubian et al., 2018).

5.2. Shear force anomalies in tower walls above the TFL

As discussed in Section 2.1.3 previous investigations on buildings featuring a transfer plate

have highlighted the onset of high shear force concentrations occurring in the tower walls above
the transfer floor level (TFL) as the building is subjected to lateral loads (Su, Chandler, Li & Lam,
2002;Tang & Su, 2014). These shear force concentrations are resulted mainly from the out-of-

147
plane bending of the transfer structure supporting the tower walls. Differential rotation of the
transfer structure along its length results in displacement incompatibility in between adjacent
tower walls. The incompatibility induces high in-plane forces in elements (link beams and slabs)
connecting the tower walls thereby resulting in a significant increase in the shear force above the
TFL. Similar shear force distribution anomalies have been reported in Chapter 3-4 for tower walls
in podium-tower buildings featuring a setback.

The 2D model of an example building (Figure 5-1) was employed to illustrate such shear
force anomalies. The numerical model (of the building) is constructed on program ETABS
(Habibullah, 1997) where 2D shell elements are adopted for the modelling of the tower walls and
the transfer plate. Elastic frame elements have been employed for the modelling of the stiff
podium column. The 1500mm thick transfer plate supports the tower structure which is planted
at the TFL. The floor slabs connecting the tower walls are modelled as equivalent frame elements
with an effective width (beff) assigned following recommendations by Grossman (1997) and the
PEER/ATC guideline (2010). The building was subjected to lateral loads (the distribution of
which is shown in Figure 5-2(d)) in accordance to the equivalent static force procedure stipulated
in the Australian Standard (AS 1170.4:2007).

The displacement ratio (∆𝑟 ) is introduced herein as the ratio of the lateral displacement of
wall no. 1 (δ1 ) , or wall no. 3(δ3 ) , divided by the respective displacement (δo ) of a control model
wherein the transfer plate was modelled as infinitely rigid when bending out-of-plane. In the
control model all three walls have identical deflection profile (above the TFL level).

The increase in the shear force demand on Wall no. 3 (which is paralleled by a reduction in
the shear force demand on Wall no. 1) was the result of incompatible displacements between the
tower walls (∆r ≠ 1) at the level immediately above TFL (refer Figure 5-3(a)). These displacement
incompatibilities are the direct result of the out-of-plane bending of the transfer plate as shown in
Figure 5-2(e). Strutting forces that are developed in the connecting elements in between adjacent
tower walls is shown in Figure 5-2(c). Interestingly, the amount of increase in the shear force
demand as described diminishes rapidly up the height of the buildings and is negligible at only
two-to-three storeys above the TFL (refer Figure 5-3a). The shear force redistributions are not
observed in the control model where the transfer plate is rigid in bending (see Figure 5-3(b)). It
should be noted that these shear anomalies might not have been identified in a FE analysis of the
building should the usual “rigid diaphragm” constraint be adopted (similar to discussion presented
in Chapter 3). Thus, shear demands on the tower walls might have been misrepresented in certain
FE analyses of the building.

148
 Figure 5-1 2D model of the building featuring a transfer plate

(a) 2D planar (b) (c) Strutting force (d) Applied (e) Deformed shape
model of the Displacement distribution in the lateral load and out-of-plane
building ratio connecting slabs profile bending of TP

Figure 5-2 Displacement incompatibility between connected walls and the resulting
strutting (compatibility) slab force and wall shear force distributions

149
 (a) with flexible TP (in bending) (b) with rigid TP

Figure 5-3 Tower walls shear force distribution

5.3. In-plane strains between tower walls above TFL

Strutting forces occurring in between adjacent tower walls have resulted from differential wall
deflections together with in-plane strutting actions of the connecting slabs ( Figure 5-4). The
magnitude of the strutting force developed in a connecting member (e.g. link beam and slab) is
expressed as the product of the in-plane strains that are associated with the differential deflection
of adjacent tower walls and the axial stiffness of the member. Thus, the expression to solve for
the strutting force (FSTRUT) is shown by equation (5-1)

FSTRUT = εSTRUT × EC Aeff (5-1)

where STRUT is the in-plane strain in between two adjacent tower walls above the TFL and
effective slab area (Aeff ) is the product of the width of the column-strip and the gross thickness
of the slab.

150
 Figure 5-4 Introducing FSTRUT, STRUT and TP

Figure 5-5 Angle of rotation at the base of the tower walls

The amount of in-plane strain (STRUT) is in turn dependent on differences in the angle of
rotation of two adjacent tower walls at their base. Parameter TP (which is TP1 - TP2 or TP2 -
TP3) is introduced herein as the parameter characterising the differential wall rotation (refer
Figure 5-4 and Figure 5-5 for graphical illustrations). Should Wall nos. 1 and 3 have identical
dimension then the values of TP1 - TP2 and TP2 - TP3 are dependent on the positioning of the
wall along the supporting span. To further illustrate the correlations between STRUT and TP,
more analyses have been performed on the case study building (of Figure 5-1) using higher lateral
force intensities (by simply scaling up the applied forces). In Figure 5-6, the STRUT profiles (from
analyses) are plotted for the different values of TP at the walls bases. The observed trends (of
Figure 5-6) confirm the proportionality between the values of TP and STRUT. This correlation
has been confirmed for a building which is subjected to earthquake ground accelerations. The
time histories of FSTRUT (hence strut) and TP obtained from the analysis (also performed on
ETABS) of record no. 29 (defined in Table A-1 in Appendix A) are plotted in Figure 5-7 on the

151
same graph. Clearly the two quantities are directly correlated. Wall shear force time histories
shown in Figure 5-8 confirm the observations reported from the static analyses on the building
(Section 5.1). The notable differences in the shear force intensities of the tower walls immediately
above the TFL reduce as the displacement incompatibility (between the walls) is restored at higher
levels.

Figure 5-6 In-plane strains between walls and differential angle of rotation
at their base TP = TP1 - TP2

Figure 5-7 Time-histories of In-plane strutting forces between walls and

differential angle of rotation at their base ∆𝜽TP= 𝜽TP1- 𝜽TP2 (record no. 29)

152
 Figure 5-8 Shear force time-histories of the walls at different levels up the height of
the tower

5.4. Analytical modelling of the Flexibility Index

In section 5.3, the strutting force of a link element connecting two adjacent tower walls has
been shown to be correlated with the value of TP. To further illustrate these dependencies, more
linear time history analyses have been performed on several building models with different
transfer plate rigidities (thicknesses). The parameter 𝛼𝑟 is introduced as a measure to quantify the
relative flexural stiffness of the transfer plate (EcI)TP to that of the tower wall (EcI)wall (defined in
Eq. 5-2 where Ec is the concrete modulus of elasticity and I is the moment of inertia of the wall
and the transfer plate). The buildings have been analysed (on ETABS) using records (no. 29-36
in Table A-1 of Appendix A). For each simulation, the maximum strutting strains in the floor
slabs (STRUT) are plotted (in Figure 5-9) against corresponding values of TP evaluated at the
base of walls 1 & 3. The correlation (STRUT-TP) is consistently linear as shown in Figure 5-9
(a)-(f). The slope of the correlation is defined herein as the Flexibility Index (FI). Importantly,
this slope (or value of FI) is shown to be dependent on the value of 𝛼𝑟 . Thus, the in-plane strain
(STRUT) may be expressed as the product of FI and TP as shown by Eq. 5-3.

(𝐸𝑐 𝐼) 𝑇𝑃
𝛼𝑟 = √ (5-2)
(𝐸𝑐 𝐼)𝑊𝑎𝑙𝑙

𝜀𝑆𝑇𝑅𝑈𝑇 = 𝐹𝐼 × ∆𝜃𝑇𝑃 (5-3)

The value of FI (the slope of the correlation between TP andSTRUT) reduces as the stiffness
of the transfer plate increases (increase in the value of αr ). Considerable reductions in the values
of TP and STRUT are observed when the rigidity of the transfer plate is increased (with increasing
values of αr in Figure 5-9). The combined effects of TP and 𝛼𝑟 on the magnitude of the strutting

153
forces 𝐹𝑆𝑇𝑅𝑈𝑇 between two adjacent tower walls are well demonstrated in Figure 5-10.

It is shown further in Figure 5-10 that higher values of FI (meaning higher strutting
forces) can be developed in between very stiff tower walls in which case the value of r is less
than 0.4. Interestingly a similar observation has been reported in Chapter 4 where larger values
of FSTRUT were determined in between stiffer tower walls (above the level of the podium). On the
other hand the magnitude of the strutting forces can be reduced by as much as 60% with relatively
flexible tower walls in which case the value of r is greater than one. In cases with multiple tower
walls above the TFL αr is calculated based on the sectional properties of the most flexible wall
in the assembly (maximum of three walls for interior bays and two walls for the exterior bays).
Where a stiff continuous core shaft is present, αr is evaluated based on the properties of the
transferred tower wall connected to the core (verifications are presented in Section 5.7).

Figure 5-9 Effects of 𝜶r on strutting forces in between tower walls (analyses of

records 29-36)

Figure 5-10 Correlation of Flexibility Index with 𝜶r

154
 5.4.1. FI for connected tower walls with different span lengths

In the most conventional building set-out, the tower structure caters residential apartments or
office spaces that are supported by transferred structural walls typically spaced at 4m-8m
intervals. Accordingly, the in-plane stiffness of the connecting slabs (defined as 𝐸𝑐 𝐴𝑒𝑓𝑓 /𝐿𝑠𝑙𝑎𝑏 )
varies among the different walls in a storey depending on the span length of the slab (𝐿𝑠𝑙𝑎𝑏 ).
Intuitively this variation (in the in-plane stiffness) entails that the strutting forces in the floor slabs
also vary depending on their aspect ratio (slenderness). The 2D model (in Figure 5-11) is
employed in a series of linear time history analyses to identify possible dependencies of the
magnitude of 𝜀𝑆𝑇𝑅𝑈𝑇 and the length of the floor slab. The building model is similar to the model
shown in Figure 5-1 with the exception of the unequal span lengths between the transferred walls.
Interestingly, the extent of the in-plane slab strains (𝜀𝑆𝑇𝑅𝑈𝑇 ) has been found to be independent of
the length of the slab’s clear span. This is clearly illustrated when plotting the peaks of 𝜀𝑆𝑇𝑅𝑈𝑇
along with the corresponding values of ∆𝜃𝑇𝑃 between walls 1|2 and 2|3 in the 2D case study
building shown in Figure 5-12. A sample comparison of the time histories of FSTRUT and ∆𝜃𝑇𝑃 of
walls 1|2 and 2|3 is shown in Figure 5-13.

It is illustrated in Figure 5-11 and Figure 5-12 that larger in-plane deformations (δ𝑠𝑙𝑎𝑏 ) are
imposed on the longer slab by the tower walls in order that the magnitude of STRUT remains
virtually unchanged for a given value of TP and 𝛼𝑟 as inferred from Eq. (5-4).

𝛿𝑠𝑙𝑎𝑏 𝐹𝑆𝑇𝑅𝑈𝑇
𝜀𝑆𝑇𝑅𝑈𝑇 = = (5-4)
𝐿𝑠𝑙𝑎𝑏 𝐸𝐶 𝐴𝑒𝑓𝑓

Figure 5-11 2D model of a building with tower walls spaced at different intervals

155
 Figure 5-12 Peak strutting strains vs. ∆𝜽TP for the case study building

(a) 𝐹𝑆𝑇𝑅𝑈𝑇 (b) TP

Figure 5-13 Sample time history results for (a) FSTRUT and (b) TP (rec Nos. 29, 30,
31 &40)
In summary the value of 𝐹𝑆𝑇𝑅𝑈𝑇 can be estimated using Eq. (5-5) which is derived by
combining Eqs. (5-1)- (5-2).

𝐹𝑆𝑇𝑅𝑈𝑇 = 𝐹𝐼 × ∆𝜃𝑇𝑃 × 𝐸𝑐 𝐴𝑒𝑓𝑓 (5-5)

156
 5.4.2. Validation studies on 2D model of a building with multiple (connected) tower
walls

The robustness of the flexibility index which is defined diagrammatically in Figure 5-9 is next
examined for a tower structure comprising multiple connected structural walls. The building
model shown in Figure 5-14 is employed in linear time history analyses using record nos. 29-36
(in Table A-1). The values of parameter 𝛼𝑟 for successive walls have been computed based on
the sectional properties of the flexible wall in the assembly and were found to be 0.3, 0.6, 2.2 and
2.2 for slabs connecting walls 1|2, 2|3, 3|4 and 4|5 respectively (dimensions of the tower walls are
summarised in Table 5-1).

Figure 5-14 Model of a building with multiple tower walls with different
stiffness

Table 5-1 Sectional properties of the walls

Wall 1 6000 × 600 𝛼𝑟 = 0.3
Wall 2 † 6000 × 300 𝛼𝑟 = 0.6
Wall 3† 3600 × 300 𝛼𝑟 = 2.2
Wall 4† 1500 × 300 𝛼𝑟 = 2.2
Wall 5 1900 × 150
†
Sectional properties of the wall are used for calculating the value of αr

157
 It is demonstrated in Figure 5-15 that larger compatibility forces (in the connecting slabs) are
required to restore displacement incompatibilities between the stiffer walls (i.e. walls 1|2)
compared with the in-plane force demands in the floor slab connecting walls 3|4 (440kN and
220kN respectively). The proposed analytical model is therefore well capable of capturing the
interdependency between the strutting force (and strain) as a function of the relative stiffness of
the connected walls (as characterised by parameter 𝛼𝑟 ). This is achieved by the use of the
flexibility index (defined in Figure 5-9) where higher values of FI are assigned to the stiffer tower
walls (see Figure 5-16(a)-(b)).

Figure 5-15 Strutting (compatibility) force demands in the floor slabs

(a) Sample results (peaks) (b) Comparison of predicted and observed

FI values
Figure 5-16 Flexibility index values for the different tower walls

5.5. Analytical modelling of Peak Rotational Demand on the building

It was demonstrated in Sections 5.3 & 5.4 that the strutting forces that are developed in

158
between adjacent tower walls depend on the value of TP which characterises the differential
tower wall deflections. The calculation of the TP parameter can be computationally intensive.
Thus, other displacement related parameters have been explored to obtain approximate estimate
for this parameter. In a proposed simplified procedure, the value of TP is approximated by the
angle of drift of the building tower at its mid-height level (i.e. at the level of the centre of mass of
the tower block). To illustrate this, a series of parametric studies employing 2D building models
with varying heights have been undertaken. The numerical models of the wall-plate
subassemblies have been constructed on program ETABS (Figure 5-17). The connected tower
walls have been subjected to plate rotations at their base and the CM drift of the tower (𝜃𝐶𝑀
evaluated at the med-height level of the tower) has been examined in relation to the differences
in transfer plate rotations (i.e. ∆𝜃𝑇𝑃 = 𝜃𝑇𝑃1 − 𝜃𝑇𝑃2). The good agreement between the two
parameters (TP and mid-height drift of the tower) is well demonstrated in Figure 5-18 (for two
tower heights: 75m and 225m). The vertical line showing the angle of differential rotation
between two adjacent tower walls at their bases is superimposed onto a graph showing the variable
angle of drift at the centre line of the building up its height.

Figure 5-17 2D model of planted tower walls employed in the parametric study

159
 Figure 5-18 Angle of drift at position of CM in comparison with differential
angle of rotation at the base of the walls ∆𝜽TP = 𝜽TP1- 𝜽TP2

The resulting storey displacement profiles of the tower walls (Figure 5-19(a)) show straight
lines (i.e. constant drifts) with the value of the slope approximately equal to the value of ∆𝜃𝑇𝑃
evaluated at the base of the walls (Figure 5-19(b)).

The first storey (immediately above the TFL) is subject to higher drifts compared with the
storeys above (this is evident when comparing the different slopes above and below this level in
Figure 5-19(a)). The reduction in the tower drift above the first level is imposed by the link
elements connecting the tower walls. These elements (beams or slabs) deform and bend to restore
displacement compatibility between the connected walls. These contributions (by the connecting
elements) are demonstrated in Figure 5-20 where the tower drifts (obtained from the analyses) are
compared to a case where the in-plane and out-of-plane (bending) stiffnesses of the connecting
element are considerably reduced. The shaded area in Figure 5-20 represents the reduction of
drifts by the coupling of the connecting elements.

160
 (a) Displacement profiles (b) Comparison of 𝜃𝐶𝑀 and ∆𝜃𝑇𝑃
Figure 5-19 Tower drift profiles

Figure 5-20 The effect of floor slab coupling on the tower displacement

The good consistencies between the two angles of rotation (as demonstrated earlier) mean that
the amount of in-plane stresses, and strains, developed between two adjacent tower walls can be
predicted conveniently by referring to the global displacement (and rotational) demand of the
tower block as a whole. Equation (5-5) for estimating the value of 𝐹𝑆𝑇𝑅𝑈𝑇 may therefore be written
as follows:

𝐹𝑆𝑇𝑅𝑈𝑇 = 𝐹𝐼 × 𝜃𝐶𝑀 × 𝐸𝐶 𝐴𝑒𝑓𝑓 (5-6)

161
 The concept of linking in-plane strains between adjacent tower walls to the (global) angle of
drift of the tower wall as a whole (e.g. CM) is motivated by the development of a simplified hand
calculation procedure which enables strutting forces in the link elements to be predicted with good
confidence (provided that the structural configuration and gross dimensions of the tower walls
and that of the podium are known, and the stiffness properties of the transfer plate and that of the
tower walls are also known). Estimates of the strutting forces can therefore be made without the
need to execute memory intensive finite element analysis of the complete model of the tower -
podium assemblage. Contributors to the global displacement (and rotation) of a building
incorporating a transfer plate are next examined to estimate the amount of centre of mass rotations
of the tower block (𝜃𝐶𝑀 ) when the building is subjected to earthquake ground excitations.

5.5.1. Deformation modes of buildings featuring a transfer plate

The configuration of a building featuring a transfer plate can be viewed as one composed of
three sub-structures: the podium, the transfer plate and the tower. The lower podium portion
comprises widely spaced stiff columns (mega columns) and eccentrically positioned core walls.
The tower structure (above the TFL) typically accommodates different column/wall arrangements
that are planted on the transfer plate. As such, the total deformation of the building when subjected
to ground shaking can be decomposed into three modes: translational displacements, rotations
(drifts) imposed by the distortion of the plate between the supports and rotations of the tower
block ensued by the podium structure (Figure 5-21 (a), (b) and (c) respectively).

162
 (a) Translational (b) Rotations of the (c) Rotational motion of the
motion tower imposed by the podium
distortions of the transfer
plate
Figure 5-21 Lateral deformations in transfer structures under lateral loads

The translational component of the response (Figure 5-21(a)) is represented (mainly) by the
translational stiffnesses (flexural and shear) of the lateral load resisting system in both the tower
and the podium. As these systems (and their respective stiffnesses) can be notably distinct in the
podium and the tower and the tower, a single degree of freedom (SDOF) representation of the
translational response behaviour of the building is analogous to the response of a two spring
system that are connected in-series (as demonstrated in Figure 5-22). The described simplification
of the buildings response behaviour is widely in practice for determining seismic demands on
buildings (e.g. the substitute structure analogy (Priestley et al., 2007)).

163
 (a) SDOF (inverted (b) lateral stiffness (c) Stiffness representation of
lollipop) model of the contributions of the the lateral load resisting elements
translational component podium and the tower in the podium and the tower
(connected in series)

Figure 5-22 Analytical model of the building representing the uncoupled

translational response

By way of the described simplification, the value of the elastic translational stiffness of the
building (of Figure 5-22) is defined by Equation 5-7 and the terms of the Equation are illustrated
in Figure 5-23.

1 1 −1 (5-7)
𝐾𝑥 = (𝐾 + 𝐾 )
𝑇 𝑃

where 𝐾𝑇 and 𝐾𝑃 are lateral stiffness values for tower and podium structure respectively
(schematic illustration of the translational response is outlined in Figure 5-23). The theoretical
basis of Eq. 5-7 is founded on the spring-in-series system discussed earlier (and demonstrated in
Figure 5-22).

(a) Podium sturcture (b) Tower structure (assumed fixed at TFL)

Figure 5-23 Illustrations of the translational displacement of the tower and the

164
 podium
The seismic displacement demands on a building are commonly evaluated at the building’s
effective height (hx). The effective height (hx ) can be evaluated by adopting the above
representation of the translational system (see Figure 5-22(a-c)). The total translational drift of
the SDOF model of the building is approximated by the sum of the translational drifts of the tower
and the podium structures (Eq. 5-8).
𝐹 𝐹 𝐹
𝐾𝑥 ℎ𝑥
=𝐾 +𝐾 (5-8)
𝑇 ℎ𝑇 𝑃 ℎ𝑃

where hP and hT are the heights of the podium and the tower structures respectively. The term
F/𝐾𝑥 ℎ𝑥 is the effective drift of the equivalent translational SDOF system (noting that F/K is a
measure of displacement). The terms of Eq. 5-8 are next rearranged to solve for the effective
height (Eq. 5-9).

1 𝐾 ℎ 𝐾 ℎ𝑃
ℎ𝑥 = 𝐾 (𝐾 𝑇ℎ 𝑇+𝐾𝑝 ) (5-9)
𝑥 𝑃 𝑃 𝑇 ℎ𝑇

Eq. 5-9 expresses the effective height of the translational system proportionally to the ratio of
the effective lateral stiffness of the tower (𝐾𝑇 ) to that of the podium (𝐾𝑃 ). Interestingly for
buildings with a range of 𝐾𝑇 /𝐾𝑃 ≤ 0.2, the value of hx is in the order of 0.70 of the total height
of the building (refer Figure 5-24). This is consistent with effective height range for the dual
frame-wall buildings (Priestley et al., 2007). Conversely, as the stiffness of the tower (relative to
the stiffness of the podium) increases, the effective height (ℎ𝑥 ) asymptotically approaches the
height of the podium structure (ℎ𝑃 ) (see Figure 5-24). Similar observations have been reported in
the works of Lee et al. (Lee & Ko, 2004, 2007) on low-rise “piloti-type” buildings that feature of
highly stiff tower block. It is worth noting that for buildings with high 𝐾𝑇 /𝐾𝑃 ratios, the
translation stiffness (𝐾𝑥 in Eq. 5-7) is solely represented by the stiffness of the podium
(i. e. for 𝐾𝑇 ≫ 𝐾𝑃 → 𝐾𝑥 ≈ 𝐾𝑃 ).

Figure 5-24 Variation of hx with the relative tower-podium stiffness ratio KT/ Kp

The rotational response behaviour of the building (illustrated in Figure 5-21(b)-(c)) subject to

165
earthquake ground motions is next examined. The ‘rotational’ model to be presented herein is
based on simplifying the tower walls collectively as a rectangular "rigid-body" which is supported
on the podium structure (Figure 5-25). The elastic stiffness of the translational spring which is
positioned at the base of the rectangular rigid body is to emulate influences by the elastic
deflection behaviour of the podium structure (Figure 5-23(a)) and that of the tower block
experiencing sway (Figure 5-23(b)). The elastic stiffness of the rotational spring (also positioned
at the base of the rectangular rigid body) is to emulate effects of the shortening-lengthening (push-
pull) actions of the columns in support of the podium and the building tower above it (Figure
5-26). The rotational spring is also to incorporate the effects of the transfer plate flexibility on the
rotations evaluated at the centre of mass (𝜃𝐶𝑀 ) of the rigid block (Figure 5-25). It should be noted
that the sway deflection of the tower as described is actually contradictory to the rigid-body
assumption of the simplified model which is, strictly speaking, not representing real behaviour.
Thus, results so derived from the simplified calculation method (assuming the rigid-body
behaviour of the tower block of Figure 5-25) need to be verified to demonstrate that errors
incurred by the idealisation are within acceptable limits in the practical context (as demonstrated
by results presented later in this section and in Section 5.7).

Figure 5-25 Schematic representation of the rigid body rotation model

166
 Figure 5-26 Rigid-body rotations of the tower block imposed by the
podium
The value of the elastic rotational stiffness of the tower block undergoing a rigid-body rotation
is expressed in Eq. 5-10
−1 (5-10)
1 1
𝐾𝜃 = ( 𝐾 + K𝜃,𝑝𝑜𝑑𝑖𝑢𝑚
)
𝜃𝑇𝑃

where meaning of K 𝜃,𝑝𝑜𝑑𝑖𝑢𝑚 is as indicated schematically in Figure 5-26; 𝐾𝜃𝑇𝑃 is the rotational
stiffness parameter which is defined as the aggregated bending moment at the base of all the
walls (∑ M) divided by the rotation at the CM of the tower structure θCM (Figure 5-5). The value
of K θTP can be estimated using Eq. (5-11).

2 𝐷 3
𝐾𝜃𝑇𝑃 = 𝐸𝑐 𝑡𝑇𝑃 𝑊 √[ℎ ] (5-11)
𝑇

where D is gross dimension of the building tower (as a whole) in the direction of loading, tTP and
W is the thickness and width of the transfer plate, and 𝐸𝐶 is the modulus of elasticity of the
concrete which the transfer plate is built of.

Equation 5-11 has been derived empirically based on results obtained from a sensitivity study
on the wall-plate assemblages similar to those shown in Figure 5-17. The connected tower walls
have been subjected to rotations at their base and the CM drift of the tower (𝜃𝐶𝑀 at mid-height of
the tower) has been examined in relation to the differences in transfer plate rotations (i.e. ∆𝜃𝑇𝑃 =
𝜃𝑇𝑃1 - 𝜃𝑇𝑃2 ). The net overturning moment at the base of the walls (∑ 𝑀) is evaluated for different
scenarios and the rotational stiffness ( 𝐾𝜃𝑇𝑃 ) has been computed as ∑ 𝑀 /∆𝜃𝑇𝑃 which is also
equivalent to ∑ 𝑀 /𝜃𝐶𝑀 as demonstrated earlier. It is worth noting that the 𝐾𝜃𝑇𝑃 parameter is not
just representing the flexural stiffness of the transfer plate on its own as it is also representing the

167
degree of restraint on the rotation of the building tower at the centre of mass. Thus, the aspect
ratio of the building tower: D/hT is a parameter in Eq. (5-11) for estimating the value of 𝐾𝜃𝑇𝑃 .
Values of 𝐾𝜃𝑇𝑃 so obtained from the analyses have been found to be equal to unity when
normalised with respect to the product of the square of the transfer plate thickness(t 2TP ), the width
𝐷 1.5
of the transfer plate (W), elastic modulus of concrete (EC ) and (ℎ ) as shown in Figure 5-27(b)-
𝑇

(c). Equation (5-11) has therefore been verified.

(a) Sample results of 𝑘𝜃𝑇𝑃 /Ec d2TP W (b) Correlation between analytical 𝑘𝜃𝑇𝑃 and the
empirical formulation (Eq. 5-11)

Figure 5-27 Parametric study for determining the parameter 𝒌𝜽𝑻𝑷

The good agreement between the analytical results and results obtained from the empirical
expression (Eq. 5-11) contributes to validating the model for predicting the intricate interferences
of the transfer plate on the response behaviour of the tower walls of the building. Further
validations on a 3D case study building are presented later on in this chapter.

5.5.2. Dynamic rotational coupling

The spring-connected rigid-body (“rocking”) model of the building as depicted in Figure 5-25
is complete when the value of Kx and Khave been determined using Eqs. (5-7) – (5-11) and
distribution of mass up the height of the tower block is also known. The original computer model
of the building may have thousands degrees-of-freedom (DOFs). The use of the rocking model
has the number of DOFs reduced to just two: a translational and a rotational DOF. The problem
is essentially one of dynamic rotational coupling. A similar approach of modelling has been
adopted for analysing dynamic coupling of torsionally unbalanced buildings featuring plan
irregularities. The building models employed in those studies have also been reduced to 2 DOFs

168
(Lam, Lumantarna, & Wilson, 2016).

The dynamic rotational coupling analysis of the rigid-body model of the building towers (of
Figure 5-25) provides predictions for the value of the Peak Rotational Demand (PRD) of the
building as a whole which is defined herein as the maximum rotations experienced at the CM of
the tower imposed primarily by the local distortions of the transfer plate. The analytical details of
the rotational coupling problem are next examined.

The dynamic equilibrium equations (Eqs. 5-12 – 5-13) are employed to solve for the coupled
dynamic properties of the building and the displacement/rotation response behaviour of the tower
when subjected to earthquake ground accelerations.

mẍ + K x (x + eθ) = 0 (5-12)

Jθ̈ + K x (x + eθ)θ + K θ θ = 0 (5-13)

m in equation 5-12 represents the translational mass of the building, J (in equation 5-13) is the
mass moment of inertia of the tower, 𝐾𝜃 is the total rotational stiffness of the tower structure
above the TFL (defined in Eq. 5-10) and 𝐾𝑥 is the equivalent translational stiffness of the building
(Eq. 5-7).

Eqs. 5-12 & 5-13 are next normalised with respect to mr and mr2 respectively. The parameter
r is the radius of gyration of the tower block undergoing rigid body rotations (Eq. 5-14).

h2T + D
𝑟= √ (5-14)
12

The parameter 𝑏 2 is the ratio of the total rotational and translational stiffness (Eq. 5-15).

Kθ
b2 = (5-15)
Kx

In Equation (5-16) the normalised equations of dynamic equilibrium are presented in the
matrix format. The “eccentricity” e defining the rigid body rotation of the tower block is
representing the distance between the TFL and the CM of the tower (assumed at mid-height of
the tower) as shown in Figure 5-25.

1 0 xr̈ 1 er xr 0
[ ] ( ) + ω2x [ 2 )] ( ) = ( ) (5-16)
0 1 θ̈ er (b2r + er θ 0

x e ẍ Kx
where xr = r ; er = r ẍr = r , ω2x = m
and

Kθ 12
b2r = (
Kx h2T +D2
) (5-17)

The coupled Eigen solution for conditions of free vibration is obtained for the coupled dynamic

169
properties of the building. Parameters Ωi and λi are introduced as the coupled angular velocity
and the angular frequency ratio for the i-th mode of vibration respectively (as shown by Eq. 5-
18).

Ωi = λi ωx (5-18)

where ωx is the translational angular velocity of the building. Full details of the derivations have
been presented elsewhere (Lam et al., 2016; Lumantarna et al., 2013). The two coupled angular
frequency ratios λi are obtained by solving the 2x2 matrix expressed in Eq. 5-19.

2
1 + (b2r + e2r ) 1 − (br2 + e2r )
λ2i =( ) ∓ √[ ] + e2r (5-19)
2 2

The introduced analytical model prompts a framework for estimating the rotation of the tower
at the CM (θCM (t)) which is primarily imposed by the distortions of the transfer plate as shown
by Eq. 5-20.
i=2 (5-20)
θi uΩi,ζ (t)
θCM (t) = ∑ ( )
1 + θ2i r
i=1

where

λ2i − 1 (5-21)
θi =
er

uΩi,ζ (t) in Eq. 5-20 is the damped single-degree of freedom displacement response of an
equivalent system with an angular velocity of Ωi . The peak rotation demand (PRD) is defined
herein as the maximum value of θCM so obtained from the dynamic analysis (as shown by Eq. 5-
22).

𝑃𝑅𝐷 = max(θCM (t)) ≈ max(∆θTP ) (5-22)

5.5.3. Evaluation of the Peak Rotation Demand (PRD) on the building

Building models similar to that shown in Figure 5-1 were employed in a parametric study to
investigate trends of PRD on the building structure for varying intensities of ground shaking. The
building models have been proportioned to achieve a range of br values varying from 0.1 to 8.
This is done by incrementally increasing the rigidity (flexural and shear) of the transfer plate.

For all the models the value of K θ was approximately equal to the value of K θTP . This was the
case since the amount of deformation that was associated with the flexing of the transfer plate
was by far higher than that of the (much more rigid) podium structure in below. Thus, the
computed value of br (Eq. 5-17) is dependent on the flexural rigidity (hence thickness) of the

170
transfer plate. A MATLAB script has been developed to solve for the dynamic coupling equation
(Eq. 5-16) and the translational SDOF response (uΩi,ζ (t)). The values of 𝐾𝑇 and 𝐾𝑃 are computed
from the FE model (on ETABS) following recommendations by Priestley et al. (2007). The
procedure for computing the values of the lateral stiffness is outlined in Appendix C.

The building models have been analysed using three suites of synthetic ground motion records
matching the Australian Standard code spectrum (AS1170.4: 2007). The considered
accelerograms are representative of three distinct site classes (Ae, Ce & De as defined in the
AS1170.4). The choice of these site classes has been such that different intensity input ground
motions could be evaluated to highlight potential dependencies of some of the response
parameters (of the introduced model) with those of the ground motion (as will be shown later).
Details of the accelerograms used in this study are summarised in Table A-1 (Appendix A). The
fundamental period of vibration of the building has also been varied to produce a wide spectrum
of building responses to earthquake excitations. Results of analyses of buildings with increasing
fundamental periods of vibration (Tx ) and br values are presented in Figure 5-28.

The PRD (evaluated based on Eq. 5-22) trends shown in Figure 5-28 suggest that the parameter
exhibits displacement-controlled behaviour (discussed in Section 2.4.3) where the values of PRD
are insensitive to the period of the building (i.e. approximately constant for a given value of br ).
The value of PRD is also shown to decrease with an increase in the transfer plate rigidity (the
value of br) and a decrease in the RSDmax of the ground motion (compare Figure 5-28 (a-c)).
Sample results from the analytical modelling of a building with a br value of 2.5 are plotted in
Figure 5-29.

171
 (a) Record Nos. 29-35 with mean
RSDmax =107mm

(b) Record Nos. 41-47 with mean

RSDmax =67mm

(c) Record Nos. 48-54 with mean

RSDmax =38mm

Figure 5-28 Effects of (br) and (RSDmax) on the Peak Rotational Demand (PRD)

Figure 5-29 Sample results on a building model with br of 2.5 based on use of record
nos. 29-35 (Table A-1)

The displacement controlled phenomenon which has been shown to govern the PRD on the
building prompted the development of a deterministic solution for predicting the value of the PRD
given that the values of 𝑅𝑆𝐷𝑚𝑎𝑥 and 𝑏𝑟 are known (Eq. 5-23 (a)-(b) and Figure 5-30(a)).
𝑃𝑅𝐷
̅̅̅̅̅̅̅
𝜑
= −0.2 × 𝑙𝑛(𝑏𝑟 ) + 0.6 (5-23a)
𝑎𝑣𝑒

172
where br is a dimensionless parameter representing the stiffness properties of the rotational spring
which restrains the rigid-body from rotation (Eq. 5-17).
RSDmax
φave =
̅̅̅̅̅̅ hx
(refer Figure 5-30(b) for the diagrammatic (5-23b)
illustration)
where ℎ𝑥 is the effective height of the building computed using Eq. 5-9 and may be taken as 0.7
times the total height of the building (hb ); and RSDmax is the highest response spectral
displacement value.

(a) Normalised PRD values obtained from the parametric study

(b) Schematic description of the parameter ̅̅̅̅̅̅

φave

Figure 5-30 Analytical model for the estimation of PRD on a building

The results from the FE model of the case study building models analysed using ETABS are
next compared with the analytical model introduced in the this section. The maximum magnitudes
of TP obtained from the linear time history analyses performed on the base model (with
geometries defined in Figure 5-1) are plotted against the in-plane strain demands in the floor
above the TFL (in Figure 5-31). The key parameters for used in the introduced model are
summarised in Table 5-2.

173
Table 5-2 Key dynamic coupling model parameters of the base model shown in Figure 5-1

𝐊 𝛉𝐓𝐏 , kNm/rad (Eq. 5-11) 2.81E+07

𝐊 x kN/m (Eq. 5-7) 9.19E+04
𝐫, m (Eq. 5-14) 19.75
𝐛𝐫 (Eq. 5-17) 0.89
PRD, rad (Eq. 5-23a- 5-23b)- RSDmax =107mm 0.00122
PRD, rad (Eq. 5-23a- 5-23b)- RSDmax =67mm 0.000766
PRD, rad (Eq. 5-23a- 5-23b)- RSDmax =38mm 0.000435

The linear trends validate further the observations reported earlier (in Section 5.4.1). More
importantly, the PRD values (Eq. 5-23(a)-(b)) are well capable of approximating the maximum
values of TP experienced at the TFL (black dotted lines resemble PRD values that are
summarised in Table 5-2). The robustness of the analytical model for estimating the maximum
values of TP at the TFL has been tested by more FE time history analyses on building with
increasing values of br (Figure 5-32; the same set of records discussed earlier have been employed
in the analyses).

It is demonstrated further in Figure 5-31 & Figure 5-32 that PRD as an output parameter from
the dynamic rotational coupling analysis (of the simplified rigid-body model of the building) can
be used to constrain the values of the differential wall rotation (TP ) and in-plane strain (STRUT)
developed in between adjacent tower walls of the building.

174
Figure 5-31 Peak Rotational Demand as a parameter to constrain in-plane strutting
forces

175
 Figure 5-32 Comparison of ∆𝜽TP and the PRD values obtained from the FE
analyses of the building models with varying transfer plate stiffness

Essentially, PRD is used as the parameter to approximate the value of TP and CM in order
that Eqs. (5-5) & (5-6) can be replaced by Eq. (5-24) as shown in below.

FSTRUT = FI × PRD × EC Aeff (5-24)

The simple expression of Eq. (5-24) for estimating FSTRUT provides a conservative estimate
for the additional shear forces on the tower walls transferred from the connecting slabs |∆VSlab |.

5.6. Evaluation of the additional shear force demands on tower walls above the TFL
It has been demonstrated in earlier sections that the detrimental increase in the shear force
demands on the tower walls above the TFL has been attributed to the generation of strutting forces
(𝐹𝑆𝑇𝑅𝑈𝑇 ) in between the connected tower walls. In Section 5.5 a simple (yet robust) analytical
formulation is introduced for estimating these forces (𝐹𝑆𝑇𝑅𝑈𝑇 ). Essentially, the additional shear
forces (in the tower walls above the TFL) are equal to the magnitude of the compatibility forces
(FSTRUT) that are transferred by the floor slab.

The results from the time history analyses performed on the 2D building model introduced in
Section 5.4.2 have been further explored to evaluate the capability of the developed analytical
model in estimating the complex slab-wall interaction above the TFL. A control model with an

176
infinitely rigid transfer plate element (in bending) has also been analysed in parallel. It should be
noted that the displacement incompatibilities (and the consequent in-plane strains) in between
tower walls are not expected to occur in the control model (as discussed in Section 5.2). The
negative and positive maximum wall shear force distributions three storeys above the TFL are
plotted in Figure 5-33(a) & (b) respectively (for brevity, only results from analyses of records
nos. 29 and 30 are shown). The differences in the shear force values between the control and the
actual model (annotated by | ∆VSlab |) are attributed to the internal force transfer by the floor slab
(FSTRUT) in the actual model (but not in the control model). In Figure 5-34 a comparison of the
maximum additional shear forces obtained from the FE analyses with the values of FSTRUT (Eq.
5-24) is presented in a bar chart. Not surprisingly, higher values (with the exception of wall 3|4)
are predicted by the analytical model compared to those (VSlab ) observed from the FE analyses.
This is attributed to the use of the PRD as an upper-bound estimate of the actual ∆𝜃𝑇𝑃 experienced
at the bases of the connected walls.

(a) Maximum negative shear force distribution

Figure 5-33 is continued on next page

177
 (b) Maximum positive shear force distribution

Figure 5-33 Comparison of the shear force demands on the tower walls

Figure 5-34 Comparison of the values of VSLAB and FSTRUT

5.7. Verification studies on case study buildings

The study so far presented an analytical approach for the modelling the effects of the transfer
plate flexibility on the shear response behaviour of the supported tower walls. In this section, two
3D case study buildings featuring transfer plates are examined to illustrate the application of the

178
simplified analytical procedure introduced in this chapter and to verify its accuracies of
quantifying the strutting force developed in between adjacent tower walls and the anomalous
increase in the shear force demand on tower walls above the TFL. The two case study buildings
are of different heights (one high-rise and the other medium-rise) and feature a transfer plate of
different thicknesses (rigidities).

5.7.1. High-rise building

The 118.8m reinforced concrete building shown in Figure 5-35 features a four-storey podium
and a 2.7m thick transfer plate. The 102m tower structure comprises structural walls and a central
core that are coupled by 200mm thick reinforced concrete flat slabs. The reinforced concrete
building has been designed in accordance with design guidelines stipulated in Australian Standard
(AS3600:2009). The lateral load resisting system of the building is designed for wind loads but
without taking into account the occurrence of seismic actions. Floor loads were uniformly applied
following Australian Standard recommendations (AS 1170.0, 2002) and are summarised in below
(Table 5-3).
Table 5-3 Floor load assignment

Tower floors
Additional dead loads 𝒌𝑵/𝒎𝟐 1.0
Live loads 𝒌𝑵/𝒎𝟐 2.0
Podium floors
Additional dead loads 𝒌𝑵/𝒎𝟐 1.2
Live loads 𝒌𝑵/𝒎𝟐 5.0

The numerical (FE) model of the building was developed on program ETABS (Habibullah,
1997). The seismic response behaviour of the building was analysed using a suite of
accelerograms records the details of which are summarised in Table A-1 (record nos. 29-35; 41-
47; 48-54). The Hilber-Hughes-Taylor direct linear integration scheme was adopted to solve for
the response behaviour of the building using a time step of 0.005 seconds. Rayleigh damping
model is adopted for assigning equivalent viscous damping to the building. A damping value of
1.8% (as a % of critical damping) has been computed following recommendations given by TBI
(2017) and has been assigned for modes 1 and 4. The mode shapes and the periods of the
translational modes (1-4) in the x-direction are summarised in Table 5-4 and Figure 5-36. The
magnitude of the strutting forces in the floor slabs (modelled as thin shell elements) was obtained
by integrating in-plane slab stresses at different locations along the slab for each time step (as
discussed in Section 4.4).

179
(a) Elevation view of the case study (b) 3D render of the FE model of the
building showing the analysed walls case study building

Figure 5-35 Case study building

180
 (a) Mode 1 (x- (b) Mode 2 (x- (c) Mode 3 (x- (c) Mode 4 (x-
direction) direction) direction) direction)

(e) Mode shapes of the building (only modes participating in the x-direction are shown)

Figure 5-36 Mode shapes of the case study building (in the global x-direction)

Table 5-4 Translational modes of the case study building (high-rise)

Modes Description, direction Period, sec

1 Translational, x 4.371
2 Translational, x 1.193
3 Translational, x 0.581
4 Translational, x 0.443

181
 As explained in Sections 5.4 & 5.5 the dynamic rotational coupling behaviour of the building
tower can be approximated by the use of the 2DOF model the schematic diagram of which is
shown in Figure 5-25. This coupling analysis provides predictions for the Peak Rotational
Demand (PRD) which is central to the analytical methodology introduced in this chapter. The
verification of the accuracies of results from the dynamic coupling analysis is therefore important
in terms of verifying the simplified analytical methodology as a whole. A summary of the input
parameters and their use in calculations as per the simplified calculation procedure are presented
in Table D-1 (Appendix D). The MATLAB code (discussed in Section 5.5) is employed to obtain
estimates of the CM rotations (𝜃𝐶𝑀 ) of the building.

The proportionality between the two parameters: (a) relative transfer plate rotations TP
(which is equal to θTP1 − θcore , θTP2 − θcore) and (b) strutting forces FSTRUT developed in the
connecting slabs in between two adjacent tower walls (Wall nos. 1 and 2) have been well
demonstrated in Figure 5-38(a)-(b) in which the time-histories of both parameters are
superimposed on the same graph (refer to Figure E-1 in Appendix E for results obtained from
analyses of record nos. 41-54)

The flexibility index (FI) for the wall/plate assembly took the value of 0.4 as per
recommendations of Figure 5-9 (refer also Table D-1 which has every design parameters listed).
This value of FI which defines the slope of linear correlation between ∆𝜃𝑇𝑃 and 𝜀𝑆𝑇𝑅𝑈𝑇 is
consistent with results of FE analysis by program ETABS of the 3D model of the building (noting
that Figure 5-37(a)-(c) also show slope of 0.4 in the linear correlation of data retrieved from the
FE analysis). Good consistencies between the value of ∆θTP at the base of wall nos. 1 and 2 as
derived from FE analyses by program ETABS and the value of PRD (as defined by Eqs. 5-23(a)-
(b)) have also been demonstrated in Figure 5-39 in the form of bar charts. Verification of the FI
and PRD parameters are essential as the value of FSTRUT is controlled by the two parameters as
given in Eq. 5-24.

The force transfer by the connecting floor slabs in between the tower walls (which has been
shown to be the primary contributor to the shear force anomalies in the tower walls) was examined
next. Figure 5-40 presents the anomalous increase in the shear force demand on the tower wall
nos. 1 and 2 at the storey above TFL (the design details of the tower walls 1 and 2 are summarised
in Table 5-5).

It is shown that the anomalous increase in the shear force demand on the tower walls above
TFL was not revealed by program ETABS when “rigid-diaphragm” assumptions were specified.
When the “semi-rigid diaphragm” constraints were applied on the floor slabs the additional shear
force demand on the walls was found to be of the order of 500 kN (as shown in Figure 5-40 which
reports mean results from analysis using record nos. 29-35). This prediction from ETABS analysis

182
is also in good agreement with the prediction of 557 kN as per the simplified method introduced
in this chapter (as listed in Table D-1).

(a) Records with mean RSDmax = 𝟑𝟖 𝒎𝒎 (Rec nos. 48-54)

(b) Records with mean RSDmax= 𝟔𝟕 𝒎𝒎 (Rec nos. 41-47)

(c) Records with mean RSDmax= 𝟏𝟎𝟕 𝒎𝒎 (Rec nos. 29-35)

Figure 5-37 ∆𝜽TP vs. 𝜺STRUT and the flexibility index

183
 (a) Wall 1
Figure 5-38 continued on next page

184
 (b) Wall 2
Figure 5-38 Sample results of slab strutting (in-plane) force and relative base rotation trends for Walls 1 & 2 (Records 29-35 with mean
RSDmax=107mm)

185
(a) Records Nos.48-54 (𝑅𝑆𝐷𝑚𝑎𝑥 = 38𝑚𝑚 ) (b) Records Nos.41-47 (𝑅𝑆𝐷𝑚𝑎𝑥 = (c) Records Nos. 29-35 (𝑅𝑆𝐷𝑚𝑎𝑥 =
67𝑚𝑚 ) 107𝑚𝑚 )
Figure 5-39 Comparison between the maximum relative transfer plate rotations (for Walls 1 & 2) with the PRD computed for individual
record

186
Figure 5-40 Shear force distribution above TFL from ETABS (Record nos. 29-35)

187
 Table 5-5 Design details of Walls 1 & 2

Design parameter Value Section detail

lw , 𝐦𝐦 2000
lw , 𝐦𝐦

t w , 𝐦𝐦 300

fc′ , 𝐌𝐏𝐚 40.0

ρv , % 0.80

ρt , % 0.38

Shear capacity ∅𝑉𝑛 (AS 3600, 2009),𝒌𝑵 862

5.7.2. Medium rise building

The extent of the interference by the transfer structure is further examined for a medium rise
building (72.5m in height). The purpose of this investigation is to illustrate the sensitivity of the
magnitudes of 𝐹𝑆𝑇𝑅𝑈𝑇 and ∆𝜃𝑇𝑃 to varying building characteristics (namely building height and
thickness of the transfer plate).

The numerical model of the building is constructed on ETABS following the approach
described in Section 5.7.1. The same set of ground motion records were used in Section 5.7.1
have been utilised for the time history analyses presented in this section. The primary load
resisting system of the tower building also comprises a continuous reinforced concrete core while
tower walls and columns that are planted on the 600 mm transfer plate form the secondary
(gravity) system. The 3D render of the FE model along with the tower floor plan showing the
examined walls are illustrated in Figure 5-41(a)-(b).

(a) 3D render of the FE model

188
 Figure 5-41 continued on next page

(b) Typical tower floor plan showing the planted (secondary) walls and the continuous
core

Figure 5-41 3D render and tower floor plan of the medium rise building featuring a
transfer plate
A step-by-step calculation summary of the parameters adopted by the analytical model is
outlined in Table D-2 (Appendix D). The value of flexibility index (FI) is estimated by evaluating
the slope of the correlation between the strutting slab strains (𝜀𝑆𝑇𝑅𝑈𝑇 ) and ∆𝜃𝑇𝑃 as shown in
Figure 5-42. The value of FI is in good agreement with the value predicted by the use of the
analytical model introduced in this chapter (FI= 0.389 from analyses and 0.4 predicted by the
analytical model as shown in Table D-2). The PRD (calculated by Eqs. 5-23(a)-(b)) are also
shown to predict well the maximum magnitude of ∆𝜃𝑇𝑃 . For brevity, only results for Wall 2 are
reported in this section (results from analyses of Wall 3 are outlined in Figure E-2 in Appendix
E).

189
 Figure 5-42 𝜺STRUT versus ∆𝜽TP plot of Wall 2 (vertical lines represent PRD values for the
different ground motion intensities)

The force transfer (by the floor slab) results in high shear force demands on the planted walls
above the TFL (refer). These tower walls have not been designed to resist lateral shear forces as
their stiffness is substantially lower than the central core (and as also suggested by their
reinforcement detailing). It thus follows that the internal slab-wall force transfer is the sole
contributor to the local shear force spikes that are observed immediately above the TFL. The good
correlation is illustrated by superimposing the predicted 𝐹𝑆𝑇𝑅𝑈𝑇 values on the shear diagram of
Figure 5-43 (predicted shown in red lines are the computed 𝐹𝑆𝑇𝑅𝑈𝑇 values; i.e. 𝐹𝑆𝑇𝑅𝑈𝑇 = 𝑉𝑆𝑙𝑎𝑏 ).

The results reported here also demonstrate that the generated strutting forces are more
profound in medium-rise buildings compared to high rise ones. This is illustrated when comparing
the maximum values of 𝐹𝑆𝑇𝑅𝑈𝑇 for the two case study buildings analysed for the same set of
records (e.g. 557kN and 1132kN for the high-rise and medium-rise buildings respectively). This
is not surprising since the simplified analytical model reports that the magnitude of FSTRUT is
highly dependent on the value of 𝑏𝑟 which typically takes lower values for medium-rise building
(Eq. 5-17). The value of 𝑏𝑟 can thus be regarded as a controlling parameter along with the
flexibility index (FI) which are both dependent on the thickness of the transfer plate (tTP). The
maximum values of ∆𝜃𝑇𝑃 also follow a similar trend (i.e. exhibiting a larger value for medium-
rise buildings). The good agreement between the values obtained from the FE analyses (TP) and
those predicted by the analytical model (the value of PRD computed by Eqs. 5-23(a)-(b)) is further
demonstrated in Figure 5-44.

190
Figure 5-43 Shear force distribution on tower wall (Wall 2)

191
(a) Records Nos.48-54 (𝑅𝑆𝐷𝑚𝑎𝑥 = (b) Records Nos.41-47 (𝑅𝑆𝐷𝑚𝑎𝑥 = (c) Records Nos. 29-35 (𝑅𝑆𝐷𝑚𝑎𝑥 =
38𝑚𝑚 ) 67𝑚𝑚 ) 107𝑚𝑚 )

Figure 5-44 Comparison between the maximum relative transfer plate rotations (for Walls 2 and 3) with the PRD computed for
individual record

192
 5.8. Chapter summary
This chapter addresses the adverse effects of the flexing of the transfer structure on the shear
force demand on the tower walls above the TFL. The high shear force concentrations in the tower
walls (supported on the transfer plate) are attributed to the internal (strutting) forces that are
developed in the connecting slabs (or beams) in between the adjacent tower walls. The actual
modelling of the interference by the transfer structure is highly complex and bears much weight
on the rigour employed in developing the numerical (FE) model of the building. In the absence
of detailed guidelines on modelling such effects, the proposed simple analytical tool (which can
be programmed conveniently on a single spreadsheet or a MATLAB script) is intended to assist
the design engineer in quantifying the adverse slab-wall interactions (occurring above the TFL)
which are imposed by the transfer plate. The magnitude of the strutting force (and in-plane strain:
STRUT) has been shown to be linearly correlated with the differential rotation at the base of the
adjacent tower walls (TP). The slope of the correlation is the Flexibility Index (FI) which is a
function of the relative stiffness of the transfer structure and the tower wall. The differential wall
rotation is in turn correlated with the angle of drift of the building at mid-height level. A 2DOF
model of the building tower has been developed to provide predictions for the value of the Peak
Rotational Demand (PRD) of the building tower as a whole which can be used as a conservative
(upper-bound) estimate of the value of TP. The value of FSTRUT may then be expressed as the
product of FI, PRD and the value 𝐸𝐶 𝐴𝑒𝑓𝑓 of the connecting floor slab. Equation 5-24 is central to
the simplified calculation procedure which is introduced in this chapter for predicting the value
of FSTRUT which is responsible for the increase in shear force demand above the TFL. The
parameter (br) which controls the 2DOF dynamic rotational-coupling model is a function of the
thickness of the transfer plate (tTP). The applicability of the model has been tested on a number of
buildings of various heights and configurations. Good agreement with predictions from the
analytical model (in terms of quantifying the interferences by the transfer plate) has been
consistently demonstrated across all the models that have been investigated. These interferences
(by the transfer plate) are most adverse in medium-rise buildings where the value of parameter br
is low. It has also been shown in this chapter that significant additional shear forces are imposed
on the planted tower walls particularly in the secondary (gravity) walls that are not designed to
resist seismic shear forces. These forces are not reported in conventional FE analyses where the
rigid-diaphragm constraint is imposed on the floor slabs between the walls. The adverse effects
of the transfer structure as described can result in poor seismic performance of the building when
subjected to rare (and very rare) seismic events. From a broader perspective, the developed
technique sheds light on the appropriateness of the use of displacement-based principles for
simplifying highly complex interferences by the transfer structure on the response behaviour of
the building.

193
194
 Chapter 6. ASSESSMENT OF RESPONSE MODIFICATION
FACTORS OF PODIUM-TOWER TYPE BUILDINGS

6.1. Introduction
The objective of this chapter is to systematically quantify the response modification factors
that are applicable to the seismic design of podium-tower buildings (of the form investigated in
this study). The response modification factors obtained from this study will be benchmarked
against values recommended by some seismic design standards to evaluate their adequacies of
use in the design of the building. The state-of-the-art PBEE approach is adopted to assess
performance levels of the building when subjected to a wide range of seismic intensities. Both the
global (building) and local (component) response behaviour is examined for various seismic
intensities with particular attention cast on the shear demands on tower walls above the interface
level.

It is also the objective of this chapter to assess the seismic performance behaviour of buildings
featuring podiums (setback buildings and transfer structures). It has been demonstrated in earlier
chapters that significant internal force redistributions occur above the level of the podium. These
redistributions of internal actions have been attributed to the development of compatibility
(strutting) forces in the floor slabs connecting the tower walls above the level of the interface.
Seismic assessment studies on buildings of this feature (podium-tower) have not been widely
covered in the literature (as noted in Chapter 2) particularly where considerations are made for
interferences by the podium structure in the building. The examined building models represent a
distinct subset of vertical irregularity in buildings that represent a substantial portion of the
medium and high rise building stock in regions of lower seismicity.

6.2. Methodology
Details of the framework adopted to assess the response modification factors and the seismic
performance of the podium-tower building are described in this section. The first part of the study
involves a series of quasi-static (pushover) analyses performed on the podium-tower building
types (introduced in the next section). Results from the quasi-static analyses will be used to
investigate the yielding mechanism and the distribution of the inelastic demand up the height of
the building. At this stage, preliminary estimates will be made of the response modification factors
(ductility, overstrength and force reduction factors). The observed mechanisms will then be
evaluated to identify the key performance levels that correlate well with the degree of damage
sustained by the building. A refined evaluation of the response modification factors will be
conducted by performing incremental dynamic analyses (IDA) on the building models. The first
occurrences of each of the performance levels will be monitored to identify the ground motion
intensity (the seismic demand) and the observed deformations (building capacity) corresponding

195
to each performance level. These are obtained by conducting two auxiliary studies which are
listed in below.
1- The performance levels of the investigated buildings (podium-tower type) will be
evaluated particularly for three performance levels: IO- Immediate occupancy, LS-
Life Safety and CP-collapse prevention (described in detail in Section 6.6.4 and
Section 6.7). The first occurrences of the damage limit states will be monitored
and the deformation capacities of the buildings at these instances will be
statistically evaluated.
2- Fragility function will be developed to determine the median values of the seismic
intensities corresponding to the occurrences of the different performance levels
outlined in item 1.

A descriptive organisation of the different stages of the study to be presented in this chapter is
summarised in Figure 6-1.

196
Figure 6-1 The assessment framework and the structure of Chapter 6

197
 6.3. Description of the 2D podium-tower building models
This study is performed on two groups of building models. The first group comprises two
buildings that feature a setback in the floor plan above the podium level (buildings designated by
SB-1 and SB-2). The second group includes two building models incorporating a transfer plate at
the level of the podium (designated by TS-1 and TS-2). For all 2D sub-frame models, the tower
structure comprises of three walls connected by floor slabs.

In the first group (setback buildings) building model SB-1 comprise a tower structure that is
centred on the supporting podium. Whereas the tower walls in building model SB-2 are positioned
at an offset from the centre of the building. Anticipated differences in the response behaviour of
the two buildings (SB-1 & SB-2) have already been examined in detail in Chapter 3. Building
model TS-1 features a continuous central wall which forms the primary LLR system of the
podium and the tower structure. Tower Walls 1 & 2 are transferred (i.e. discontinued) at the
transfer floor level (TFL). Conversely, in building TS-2, all three tower walls (including the
central wall) are discontinued beyond the transfer floor level. In TS-2, Walls 1, 2 and the central
wall (CW) are a part of the LLR system of the building. The podium structure of TS-2 comprises
the stiff columns that are also coupled by floor slabs.

All seismic design codes classify both building sets as being irregular in elevation (as
discussed in Section 2.1). It has been shown in Chapter 3 that, for buildings featuring setbacks,
the positioning of the tower (relative to the podium) is central to the extent of the interferences
(restraints) imposed by the podium on the response behaviour of the tower walls. Accordingly,
building SB-2 (tower at an offset from the podium) is expected to experience a much higher
degree of interference by the podium than building model SB-1. It should be noted that most
seismic design codes offer no distinction in the irregularity classification of the two buildings. In
other words, both building types are treated equally by the design codes despite the anticipated
differences in their response behaviour in an earthquake (as described in Chapter 3 and will be
illustrated further in this chapter). The degree of vertical irregularity is more pronounced in
building TS-2 compared to the TS-1 counterpart. As in the case of setback buildings, TS-1 and
TS-2 are also grouped in the same category of building irregularity (by the design codes). It can
be argued that the differences in building geometry (of TS-1 & TS-2) may result in striking
differences in the response behaviour of the building (as will be elaborated later in this chapter).
The components making up the 2D building models are designed and proportioned following the
recommendations of the Australian concrete design standard AS 3600 (2009). The elevation
details of the buildings are presented in Figure 6-2. Geometric and reinforcement details of the
key walls and columns of the building models are given in Figure B-4 (Appendix B) whereas
those of the floor slabs and transfer plate are outlined in Section 6.4.3 (Table 6-1).

198
(a) SB-1 (b) SB-2 (c) TS-1 (d) TS-2
Figure 6-2 Elevation view of the analysed building models

199
 6.4. Construction of the nonlinear numerical model of the buildings
The inelastic numerical models of the buildings described in Section 6.3 have been constructed
on the SeismoStruct program package (SeismoSoft, 2016). The Fibre-based inelastic frame
element has been adopted in the modelling of the flexural response of the walls and the floor
slabs. This distributed plasticity modelling approach (details of which are discussed in Section
2.2.4.1) is favoured for use in numerical simulations involving a step-by-step solution of the
response behaviour of a building in an earthquake. More importantly, the fibre-based approach
preserves the axial-flexural interaction in the response behaviour of the members throughout the
response (as also discussed in Section 2.2.4.2). This interaction is particularly important for
modelling the floor slabs and the tower walls that are subject to appreciably high axial loads. The
inelastic frame elements employed for modelling the walls and the floor slabs have been
discretised at the sectional level (into various fibres) and at the elemental level where different
integration points have been defined along the length/height of the element. The approach adopted
in the modelling of the shear response behaviour of the wall is outlined in the subsequent section.

6.4.1. Nonlinear modelling of the shear and flexural response behaviour of walls

As mentioned earlier, the inelastic fibre-based elements are commonly suitable for modelling
the flexural response behaviour of the walls (Section 2.2.4.2). Although it is conventionally
argued that shear deformations of the walls may be negligible in tall buildings by virtue of their
high shear span ratios, occurrences of significantly low shear spans have been demonstrated in
earlier chapters of the thesis. These reductions have been attributed to the strutting slab forces that
are transferred locally (by the floor slab) in some storeys above the interface level. This
phenomenon is presented in Figure 6-3 where reduction of the shear span ratios (M/Vlw) is
schematically illustrated in the building incorporating the use of a transfer plate. Thus, the explicit
modelling of the effects of the shear response behaviour is central to the assessment of the seismic
performance of this form of building construction.

200
 Figure 6-3 Illustration of the reduction in the shear span ratios of the tower walls
above the TFL
The lumped shear model (described in Section 2.3.2) is adopted for modelling the uncoupled
shear response behaviour of the wall. With this approach, the shear capacity backbone curve in
the form of shear force- inter-storey drift is defined by way of a ‘zero-length’ element which is
connected in series with the flexural (fibre-based) element. Merits of this approach have already
been discussed in Chapter 2. It should be noted that analysis of the shear-flexural interactions of
the wall can be uncoupled whilst having the shear strength degradation behaviour explicitly
accounted for.

The shear backbone curve is consistent with the tri-linear model recommended in ASCE 41-
13 (2013) and FEMA 356 (2000) (see Figure 2-37(a)). It should be noted however that according
to the referenced guidelines, the shear response behaviour of the tower walls under high axial
loads is classified as a force-controlled behaviour (refer Section 2.3.2 and Figure 2-32 for details).
Nevertheless, recent observations by Looi et al. (2017) revealed a limited amount of deformation
capacity despite having a high axial load ratio. The modelling approach for the uncoupled shear
response behaviour of the wall described in below is founded on these observations.

The trilinear shear response backbone comprises three main phases (which are described in
Section 2.3.1). The cracking strength is the shear force at a shear stress level corresponding to
0.33√𝑓𝑐′ . The drift ratio at this phase (i.e. cracking) is computed by dividing the cracking force
(FCR) by the uncracked stiffness of the wall (Eq. 2-5 normalised by the shear ratio of the wall –
M/V). The nominal shear strength of the wall is the un-factored (i.e.  shear strength
formulation recommended by ACI 318-14 (2014) standard which accounts for the effects of the
axial loads and the shear span ratio on the ultimate shear strength of the wall (refer Equations 2-
7 - 2-9 in Chapter 2). The adopted post cracking shear stiffness value corresponds to 2% of the
elastic (uncracked) shear stiffness of the wall (as per the discussion presented in Section 2.3.1).
The point of “zero-shear strength” (annotated by (s)collapse in Figure 6-5) is defined as per
recommendations presented by Looi et al. (2017). The referenced study proposes a formulation
for determining the collapse drift limit for walls characterised by a low shear-span ratio (M/V)
and high axial loads (typifying some tower walls above the podium level). The ultimate drift ratio
(at full collapse i.e. F=0) is given herein by Eq. 6-1a. It should be noted that Looi et al. (2017)
propose a more generic formulation for the axial load ratio on a wall that incorporates
contributions (to axial strength) by the longitudinal reinforcement bars (Eq. 6-1b)

𝐴𝐿𝑅′ (6-1a)
𝐷𝑅𝑐𝑜𝑙𝑙𝑎𝑝𝑠𝑒 (%) = ln ( 0.85 ) /−1.8

201
 𝑃𝑜
𝐴𝐿𝑅 ′ = ′ ]𝐴
(6-1b)
[𝜌𝑣 𝑓𝑣𝑦 +(1−𝜌𝑣 )𝑓𝑐𝑚 𝑔

′
where Po is the axial load applied to the wall, 𝜌𝑣 , 𝑓𝑣𝑦 , 𝑓𝑐𝑚 is the longitudinal reinforcement ratio,
yield strength of the longitudinal reinforcement and the mean characteristic strengths of the
concrete respectively. The term DRcollapse (%) expressed in Eq. 6-1a is the ultimate drift ratio
corresponding to the axial collapse of the wall which would have occurred after the wall had
failed in shear (which is commonly evaluated at 20% loss of lateral strength, as stated in Looi et
al. (2017)). The drift limits (at axial collapse) estimated by expression 6-1a are plotted against the
axial load ratios on the wall (per Eq. 6-1b) in Figure 6-4. Not surprisingly, the collapse drift ratio
reduces with increasing value of the axial load resisted by the wall. Interestingly, the collapse
drift ratios for walls examined in this study (characterised by their respective axial load ratio) are
in the range of 1.0% (see the shaded region in Figure 6-4). This is also in good agreement with
the shear backbone capacity curve for shear-critical walls recommended by ASCE/SEI 41 (2013)
which also allocates a drift ratio of 1.0% based on the condition of zero residual lateral strength
of the wall (see Figure 2-37(a)). Given the inherent uncertainties of modelling shear-critical walls,
the lower-bound (conservative) collapse drift ratio of 1.0% is adopted in this study. It should be
noted that the drift ratio represents the amount of displacement normalised by the effective height
of the wall. For an isolated wall, the effective height corresponds to the geometric height of the
wall. In multistorey buildings the effective height is approximately equal to the shear span (i.e.
M/V) of the wall which in turn is approximated by the slope of the bending moment diagram up
the height of the building (as demonstrated in Figure 2-30). A schematic illustration of the
detailed wall model adopted in this study is schematically presented in Figure 6-5).

Figure 6-4 Collapse drift limits for walls with low shear span ratios (M/Vlw)

202
 Figure 6-5 Diagrammatic illustration of inelastic modelling of the tower walls

It is possible that some walls develop their flexural strength without the co-existing shear
forces reaching their limiting value. In such cases, flexural behaviour (and thus flexural stiffness)
will dominate (refer Figure 2-33(a)). This is illustrated in Figure 6-6 where the different ‘paths’
in the response of the wall are presented. The shaded region is to illustrate the additional shear
forces transferred onto the wall by the internal (strutting) forces in the slab. These additional shear
forces (|∆VSLAB|) meant that the limiting value for shear strength may have been reached prior to
the attainment of the ultimate flexural strength of the wall (i.e. at lower drift limits).

203
 Figure 6-6 Force-deformation options for the wall subjected to lateral loads

The cyclic (hysteretic) behaviour of the shear sub-element (tri-linear shear backbone) has been
defined by the model developed by Sivalsevan and Reinhorn (2000). This hysteretic model
incorporates lateral stiffness degradation on unloading and reloading (i.e. loading in the reverse
direction). The pinching effect which commonly characterises the shear response behaviour of
shear-critical walls (Mergos & Beyer, 2014; Ozcebe & Saatcioglu, 1989) is modelled by the slip
parameter (a value of 0.3 is recommended in the above-referenced studies). The Mander concrete
constitutive model (Mander et al., 1988) has been adopted for modelling the uni-axial response
behaviour of confined and unconfined concrete (fibres) whereas the modified Menegotto-Pinto
(1973) constitutive relationship is adopted for the reinforcing steel. Expected material strength
values are used instead of the design lower characteristic strength for both the concrete and steel.
The values of the expected strengths are 1.3𝑓𝑐′ and 1.17𝑓𝑦 for concrete and steel respectively as
recommended by the LATBSDC (2005) and PEER/ATC 72-1 (2017) guidelines.

204
 6.4.2. Verification study of the numerical wall model

In this section, a verification study of the numerical wall model (introduced in Section 6.4.1)
is presented. A total of 15 wall specimens have been selected from some experimental programs
reported in the literature to validate the appropriateness of the wall modelling approach adopted
in this study. The shear span ratios (M/Vlw) of the specimens ranged from 0.69 to about 4.5. It is
acknowledged that the response behaviour of the wall varies quite significantly within the
reported range. Nevertheless, the intention here was to demonstrate the accuracies of the hybrid
flexural-shear wall model in simulating the different trends of wall response behaviour and failure
mechanism (i.e. flexural, flexural-shear or pure shear failure). This variability has been observed
previously in earlier studies on tower walls supported on a transfer structure (Kuang & Li, 2001;
Li, Lam, Chen, & Wong, 2008; Li et al., 2006; Su & Cheng, 2009). A summary of the wall
specimens used in the verification study is presented in Table F-1 of Appendix F.

205
 Figure 6-7 verification of the simulated shear force-displacement plots

It is demonstrated in Figure 6-7 that the numerical wall model is capable, with an acceptable
degree of accuracy, of simulating the force-displacement response behaviour of walls subjected

206
to reverse displacement cycles. Higher consistencies between the experimental and simulated wall
responses are generally demonstrated for slender walls with higher shear span ratio (e.g. walls
WSH4-WSH6 in Figure 6-7). The stiffness, strength degradation and the pinching effect have
also been adequately simulated for most specimens. For wall specimens under high axial loads
the accuracies of the numerical simulations are slightly reduced. This is not surprising since the
failure mechanism typifying the response of such wall specimens is complex and in most cases
difficult to capture by fibre-based model simulations (e.g. out-of-plane buckling of the wall or
boundary reinforcement). Nevertheless, conservative estimates of the ultimate (shear) strength
and the deformation capacity of the wall are still reported in the numerical simulations.

6.4.3. Numerical modelling of floor slabs and transfer plate

The fibre-based inelastic frame elements are employed for modelling the floor slabs and the
transfer plate of the case study buildings. Only the effective widths (𝑏𝑒𝑓𝑓 ) of the floor slabs have
been modelled (refer to Section 2.2.3). The effective width formulae proposed by Grossman
(1997) has been adopted for determining these widths. The computed effective width values have
been found to be approximately 40% of the length of the span in the transverse direction (which
is 6.0m and 8.0m in the tower and the podium respectively). Similar observations on the effective
width of the slab have been reported in previous studies that employed 2D building models
(Mwafy & Elnashai, 2001; Mwafy & Khalifa, 2017). By the use of this approach, the stiffness
and strength contributions of the floor slab element are explicitly defined into the numerical
simulation.

For floor slabs, their inelastic response behaviour of the element is characterised by the
uniaxial (material) model of the individual fibres. Where concentrations of inelastic demands are
anticipated to occur in certain locations along the length of the span, a finer fibre section meshing
of the element is adopted. It was estimated that 20% of the length of the span (between walls or
columns) from either end would need to have such refinements (see Figure 6-8 for illustrations).
Conversely, a coarse element discretisation in the rest of the slab (i.e. in the middle 0.6 Ls shown
in Figure 6-8) is used to reduce computational demands (by assigning fewer integration points in
this portion of the slab). The merits of the use of the fibre-based approach for modelling the floor
slabs are such that the axial-flexural interaction is preserved throughout the response. Influences
of the strutting (in-plane) forces on the flexural strength of the slab are explicitly accounted for in
the numerical simulations.

The floor slabs are linked to the walls by means of rigid-arms (Figure 6-8(a)) as recommended
by PEER/ATC 72-1 (2010), noting that this approach is almost always adopted in the modelling
of wall-slab or wall-beam connections.

207
 (a) Plan view showing the incorporation of the slab in the numerical model

(b) Fibre-based meshing of critical zones along the length of the span

Figure 6-8 Inelastic numerical model of the floor slab

The transfer plate (for buildings TS-1 and TS-2) is similarly modelled. The discretisation of
the frame element representing the transfer plate is further refined at the location of the planted
wall (where inelastic yielding is expected). In this study, the punching shear behaviour of the
transfer plate is not incorporated into the modelling of the building TS-1 and TS-2. It can be
argued that the punching strength capacity of the transfer plate is substantially higher owing to its
depth. Additionally, the presence of transverse reinforcement bars (parallel to the depth of the
plate) can substantially increase the shear strength capacity of the plate. These transverse steel
bars are required when the plate is cast in segments to ensure adequate stress transfer across the
horizontal construction joint (see Figure 6-9 for illustrations). Waiving away the need to check
punching failure of the plate can be justified in view of large podium columns supporting the
plate. The large cross-sections of the podium columns increase the width of the ‘shear-critical’
perimeter (for resisting transverse shear) which in turn increases resistance to punching. The
reinforcement and geometric details of the floor slabs and the transfer plate are outlined in Table
6-1.

208
 Figure 6-9 Illustration of the horizontal construction joints in a transfer plate (Wong,
2013)

Table 6-1 Geometric and reinforcement details of the floor slabs and the transfer plate

d× beff 𝜌𝑙𝑜𝑛𝑔 , [%] Section and reinforcement details

Tower 200 ×
0.60%
Slabs 2400

Podium 250 ×
0.537%
Slabs 3200

1750
Transfe
× 1.2%
r Plate
3200

6.5. Nonlinear quasi-static analyses

The 2D building sub-frame models (introduced in Section 6.3) have been employed in quasi-
static (pushover) analyses using SeismoStruct program package (SeismoSoft, 2016). Static loads
have been assigned to the floor slabs to simulate the gravity forces on the walls and columns of

209
the sub-frame building models. The magnitude of the gravity loads can be manually calculated
based on typical span lengths in the perpendicular direction (6m and 8m for the tower and podium
storeys respectively) as summarised in Table B-1 (Appendix B). The primary objective of this
part of the study is to obtain preliminary estimates of the response modification factors of the
buildings. Furthermore, the extent and the distribution of the seismic-induced damages are
examined to conclude potential failure mechanisms of the building subject to earthquake loadings.

A uniform loading pattern that is proportional to the mass distribution up the height of the
building has been adopted in the analyses. It is acknowledged that most seismic design codes do
not permit the use of quasi-static analysis for tall and irregular building (typically taller than 60m).
Nevertheless, studies have shown that reasonably conservative estimates of the stiffness and
ductility of a building can still be obtained when the building is ‘pushed’ using a mass
proportional load distribution up the height of the building (Alwaeli et al., 2017; Beyer et al.,
2014; Chintanapakdee & Chopra, 2004a, 2004b; Mwafy, 2013; Mwafy & Elkholy, 2017; Mwafy
& Khalifa, 2017). The referenced studies demonstrate that the quality of the assessment is not
compromised significantly by employing a uniform load distribution in the pushover analysis. It
should be noted however that conclusions obtained from this part of the study are only preliminary
in nature and will be refined further by the use of the more elaborate dynamic analyses (Section
6.6). The lateral forces up the height can be increased monotonically until the building sustains
significant damage and where one, or more, ultimate limit states have been reached (as defined in
the next Section).

6.5.1. Ultimate limit state analysis

In the quasi-static analyses of the building models, the maximum displacement is evaluated
once one of the following conditions (or ultimate limit states) is reached:
1- Concrete fibre reaching the crushing strain value in the walls (at a compressive strain
limit 𝜀𝑐𝑢 = −0.003)
2- Reinforcing steel bars embedded in the walls and columns reaching their ultimate
tensile strain (𝜀𝑠𝑢 ).
3- 50 % loss of lateral (shear) strength of the tower walls (Walls 1 &2)
4- Shear force in the central wall (CW) attaining the nominal strength capacity.

The concrete crushing strain limit (item 1) is consistent with observations reported in the recent
study conducted by Hoult et al. (2017) on the response of walls which are commonly found in
regions of lower seismicity (study region). The little-to-no confinement provided at the boundary
regions of the walls restricts higher concrete strains from developing in the wall. This strain limit
is also enforced by the Australian concrete standard (AS 3600, 2009) which mandates a cap on
the maximum allowable compressive strains in the concrete member (𝜀𝑐𝑢 = −0.003) at the

210
ultimate limit state. It is worth noting that for some of the walls the presence of boundary
reinforcement (e.g. U-bars or ligatures) can provide confinement to the end regions of the wall.
The resulting enhancement of the concrete strength from the confining effect was in the order of
4-6% (calculated as per the procedure presented in Priestley et al. (2007)) which is considered to
be marginal.

The ultimate steel tensile strains (item 2) have been limited to values below that of their
characteristic strains as recommended by Priestley et al. (2007). The lower strain limit is to
account for the likelihood of the reinforcement bars experiencing local buckling prior to reaching
the ultimate strain limit (under cyclic load reversals). It should be noted that higher strain limits
may still be adopted provided that sufficient confining reinforcement is present (to restrain bars
from buckling). Such detailing practises are uncommon in regions of low-to-moderate seismicity.
Recently, Menegon et al. (2017) recommended an ultimate strain limit 𝜀𝑠𝑢 = 0.03 based on
experimental observations on bar buckling behaviour in walls typifying Australian constructions.
This limit has been adopted in the assessment study. Item 3 (above) caps the maximum
permissible loss of shear strength in the secondary tower walls (1&2) to 50%. This criterion is
deemed appropriate as it is likely that the walls are subject to high shear force demands in seismic
conditions. Given that structural redundancy is an issue of concern (referring to the ability of the
building to redistribute inertia loads), a force-controlled (ULS) is imposed on the central wall for
when the wall (CW) reaches its nominal shear strength.

6.5.2. Inelastic yielding progression in the building

Pushover analyses have been performed on the four building models to obtain a preliminary
estimate of the response modification factors of the buildings. The investigation of the inelastic
damage distribution up the height of the building models is a central component of this part of the
study. The buildings have been subjected to an incrementally increasing load vector with a shape
proportional to the mass distribution up the height of the building (as discussed earlier).

The pushover (capacity) curves of the building obtained from analyses are also presented in
an idealised bilinear format following the methodology outlined in FEMA 356 (2000) prestandard
(as schematically illustrated in Figure 2-44). An iterative MATLAB routine is developed to
construct the bilinear curves for each building.

The capacity curves for the building models featuring a transfer structure, TS-1 & TS-2, are
shown in Figure 6-10 and Figure 6-11 respectively. The much more inferior performance of TS-
2 is demonstrated when comparing the roof drift limits (roof displacement divide by the height of
the building) at the occurrence of the ULS which has been observed at 0.76% and 0.5% for TS-1
and TS-2 respectively. Noting that the term ‘Notional yield’ refers to the drift (and base shear) at
the first occurrence of yielding in the walls (or columns) of the building.

211
Figure 6-10 Pushover (capacity) curve of building TS-1

Figure 6-11 Pushover (capacity) curve of building TS-2

Figure 6-12 Pushover (capacity) curve of building SB-1

212
 Figure 6-13 Pushover (capacity) curve of building SB-2

The crushing (compressive) strain limit (ULS) of Wall 2 characterised the ULS of building
TS-1. In contrast, the shear force limit state (item 3 listed in Section 6.5.1) of Wall 2 has defined
the ULS of building TS-2. Differences in the response behaviour of buildings SB-1 and SB-2 are
somewhat less pronounced. In both building models (SB1 & SB2), the concrete compressive
strain limit of the wall 1 dominated the lateral response of the buildings. This, however, occurred
at a lower roof drift level in SB-2 (0.87%) compared to SB-1 (1.16%) (See Figure 6-12 and Figure
6-13).

A step-by-step mapping of the response behaviour of the critical walls is performed to obtain
accurate estimates of the first occurrence of the ultimate limit state in the building. For SB-2, the
stresses and strains in the concrete fibre (highlighted in Figure 6-14) are plotted in Figure 6-14(b)
where it is shown that the concrete fibre in wall 1 (the interior wall of building SB2) has reached
its predefined ULS strain limit (a value of -0.003). The roof drift limit corresponding to the strain
limit is illustrated by the vertical dashed red line in Figure 6-14(a). The blue line in Figure 6-14(a)
represents the uncoupled shear response of the wall.

213
 Figure 6-14 Mapping the local and global responses of walls in building SB-2 (a) total
and shear response of Wall 1 showing the occurrence of the ultimate limit state (Wall 1)
(b) concrete stress-strain response of the outer most fibre (c) total and shear response of
the central wall (CW)

Figure 6-15 Mapping the local and global responses of walls in building TS-2 (a) Total
response of the critical wall (wall 2) (b) Total response of the central wall (c) monitored
stress-strain curves for outer most reinforcing bar embedded in the central wall in two
storeys above the TFL (d) monitored concrete (fibre) stresses and strains in tower Wall
1 & 2 (d) FSTRUT in the floor slab connecting the central wall to Wall 2

214
 It is shown that the wall experiences flexural yielding, and eventually compression failure,
prior to the development of any shear degradation (i.e. the shear response behaviour of the wall
is in the pre-capping portion of the response). It should be noted that the uncoupled shear sub-
element (the phenomenological spring) is in-series with the flexural fibre based sub-element. This
entails that deformations of the two sub-elements are additive while forces are the same on both
elements. Consequently, the degradation of strength of the flexural sub-element, following the
attainment of the compressive strain limit, would also reduce the amount of shear forces on the
shear sub-element (as illustrated by the sudden reduction in the shear strength in Figure 6-14(a)).

The response behaviour of the critical tower wall (wall 2) of building TS-2 is shown in Figure
6-15(a). High strutting forces in the floor slab element (in Figure 6-15(e)) resulted in a premature
shear yielding of the wall. In other words, the internal force demands in the initial stages of the
response are not attributed to the flexural (fibre-based) element unlike the case of Wall 1 in SB-
2. The stresses and strains in the outer confined concrete fibre (shown in Figure 6-15(d))
demonstrate that the flexural response of the wall is, strictly speaking, within the elastic range.
The ULS corresponding to 50% loss of the lateral strength of the wall is observed at a roof drift
of about 0.5% (see Figure 6-11). Subsequent to this, lateral forces are redistributed to the adjacent
(CW) thereby increasing the shear force and bending moment demands on the central wall (as
illustrated in Figure 6-15(b)). The higher bending moment (and shear force) demands on the CW
resulted in yielding of the embedded reinforcement bars (of the CW) in the two storeys
immediately above the TFL. This is demonstrated in Figure 6-15(c) where the normalised stress
(fs/fy) of the reinforcing bar (marked by the magenta line), which is the uniaxial stress in the
reinforcing bar (fs) normalised to its expected yield strength value, is shown to have reached (and
exceeded) the value of unity (i.e. yielding has been activated).

Storey ductility demand is defined in this study as the maximum storey drift (at ULS) divided
by the storey drift at the first occurrence of flexural yielding in the walls/columns of the storey.
Essentially, this approximately quantifies the localisation of the seismic damages up the height of
the building on the basis that these damages are correlated to the drift demands imposed on these
storeys. It should be noted however that this quantity is an indirect measure of the seismic damage
and is sometimes associated with checks for non-structural elements (rather than structural
components). Nevertheless, owing to its simplicity and the ease with which it can be calculated,
the parameter is heavily used in most performance-based design guidelines (e.g. ASCE 41-13).
In this context, the storeys that experience larger inter-storey drift demands (ISD) exhibit a larger
storey ductility demand. Distributions of storey ductility demands for all the building models are
outlined in Figure 6-16.

215
 Figure 6-16 Storey ductility demand distributions for the examined sub-frame models
(a) TS-1 (b) TS-2 (c) SB-1 (d) SB-2

As demonstrated in Figure 6-16, the response behaviour of all the sub-frame models is
characterised by significant concentrations of inelastic damage above the level of the podium.
The trends presented for building SB-1 (Figure 6-16(c)) are somewhat more favourable compared
to the other building models. This is inferred from the large and wider distribution of inelastic
demands (ISD) up the height of the building with higher demands imposed at the lower (podium)
levels. The preferable consequence of this distribution (in SB-1) is manifested in the larger drift
capacity (ductility) of the building at ULS as shown in Figure 6-12.

Building TS-1 exhibits a slightly better redistribution of inelastic demands than TS-2. The CW
in TS-1 sustains flexural yielding at its base and above the podium level which allows the building
to achieve higher drift capacities (Figure 6-10). In contrast, the lower storey ductility demands
sustained by building TS-2 are common for when the brittle (or limited-ductile) shear-critical
response of walls governs the ULS of the building (e.g. building TS-2). Similar lateral response
trends have been reported in a recent study by Mwafy and Khalifa (2017) for this type of
buildings. Figure 6-17 presents a graphical illustration of the damage state of the building at the
ULS.

216
 (a) TS-1 (b) TS-2

(c) SB-1 (d) SB-2

Figure 6-17 Distribution of inelastic damages up the height of the building

6.5.3. Preliminary assessment of the response modification factors

As discussed in Chapter 2, most seismic design codes adopt the force-based approach for
determining the seismic demands (actions and deformations) on a building. Accordingly, the base
shear demand on a building is reduced by the response modification factor (referred to as the R-
factor, μ/Sp or the behavioural factor (𝑞0 ) depending on the adopted seismic design code) to
determine the design base shear demands on of the building. This reduction is intended to
incorporate the effects of the inelastic response behaviour of the primary lateral load resisting

217
elements (shear walls and columns) when the building is subjected to earthquake loads. For each
of the four building models (TS-1, TS-2, SB-1 and SB-2), the overstrength factor (Ω𝐨 or 1/Sp ),
the ductility (𝜇) and the ductility reduction factor (𝑅𝜇 ) have been computed following the
procedure discussed in Chapter 2 (Section 2.5.1). The design base shear force of the building (𝑉𝐷 )
has been obtained by performing response spectrum analyses (RSA) using the AS1170.4:2007
design spectrum.

The overstrength factors (Ωo , 1/𝑆𝑝 ) which quantify the “reserve” strength of the building have
been computed as the ratio of the ultimate lateral strengths of the buildings at the onset of ULS
(from the pushover analyses) to the design base shear demand (𝑉𝐷 ). The ductility factor (𝜇) is the
ratio of the drift of the building at the ULS to the drift at effective yield (i.e. 𝛿𝑈𝐿𝑆 ⁄𝛿𝑦 ). These
parameters have already been illustrated in Figure 6-10 - Figure 6-13. The values of the R have
been computed as the product of the overstrength factor and the ductility factor (Eq. 2-10). It
should be noted that the ductility factor R is taken to be equal to the value of following the
discussion presented in Section 2.5.1.

Table 6-2 summarises the computed values of the various components of the response
modification factors. The reported value of the response modification factor is, in fact, the
maximum value that can be sustained by the building. This is the case since the value of ductility
has been determined at the ultimate conditions where the building is presumed to have sustained
a considerable amount of damage. Certain reservations will need to be employed when
considering the reported ductility values for design purposes (as will be discussed later). In Table
6-2 a listing of the response modification factors (Rand the overstrength) recommended by
some seismic design codes are also presented for direct comparison.

For building models TS-1 & TS-2, the maximum values of the calculated ‘R’ factor are lower
than ASCE/SEI 7-10 value (R=4.0). The proposed overstrength value (given as 2.5 in the
ASCE/SEI 7-10) slightly overestimates the inherent overstrength for TS-2, but it predicts well the
overstrength developed by the other buildings.

For building featuring a transfer structure, the ductility factor recommended by AS

1170.4:2007 (𝜇 = 2.0) overestimates the deformation capacity of the building (compare the
values given in Table 6-2). This is partly due to the premature shear-critical conditions occurring
above the TFL (see Figure 6-15 and Figure 6-11). The limited structural redundancy (or the ability
of the building to redistribute forces) in this form of constructions also contributes to this
deficiency (compare the distributions of inelastic demands in Figure 6-17).

As discussed in Section 2.5.1, the ASCE/SEI 7-10 code recommends the use of special
strength provisions for the design of building featuring transfer structures. The provision requires

218
larger lateral forces (as much as 2.5 x the actual demand) to be considered for the design of the
structural members of the lower podium levels. Evidently, this requirement ‘shifts’ the strength
hierarchy of the building to the lower (podium) storeys. As also discussed, the requirement is
founded on ‘safeguarding’ the critical storeys and structural components that support transferred
walls or column. However, this ‘strength’ requirement does not necessarily take into account the
occurrence of concentrated inelastic demands in the upper (tower) storeys (as demonstrated in
Figure 6-17). A common lateral ‘strength’ distribution in a building with transfer structures is
schematically illustrated in Figure 6-18. Typically, the gravity (and sometimes wind) loads govern
the design strength of the podium structural members in regions of low-to-moderate seismicity.
This is demonstrated by the red line in Figure 6-18. Considerations for the design level earthquake
(DBE) for regions of lower seismicity will only marginally (if at all) increase the strength
requirement up the height of the building (shown by the green line in the same Figure). It can be
argued that the strength provision of ASCE/SEI 7-10 (shown in dashed-blue line), though only
intended for strength, will also increase the stiffness of the lower (podium) levels (e.g. when the
dimensions of podium walls and columns are increased to meet this strength requirement). It is
speculated that the resulting distribution of strength and stiffness can aggravate local ductility
demands on the storeys above this level. This may be a design issue of particular concern in
regions of lower seismicity where the design codes do not mandate local ductility checks in the
seismic design of the building.

Figure 6-18 Implications of strength provisions in transfer structures

For setback building models (SB-1 and SB-2), the computed 𝑅 factors are higher than those
recommended by AS 1170.4 and ASCE/SEI 7-10. The pronounced enhancement of the lateral
response for this type of building is also owed to the higher degree of redundancy of the structural
load path in the building. This is despite having a stiffness (and strength) discontinuity above and
below the podium level.

219
 The key observation from the reported study relates to the ductility factor (𝜇). Although, code
specified (ductility) values are often viewed as being inherently conservative (Elnashai & Mwafy,
2002), this is shown not to be the case for the examined buildings. In particular for transfer
structures (TS-1 & TS-2) where the reported values are lower than most code requirements (with
the exception of Eurocode 8 provision for DCL). The low values of ductility for TS-1 & TS-2
lends much weight to the irregular distribution of local (storey) ductility demands up the height
of the building (as illustrated in Figure 6-16 (a)-(b)). It is however acknowledged that although
the local storey ductility demands and global (𝜇) ductility demands are only sparsely correlated,
the former is shown (in this study) to be a limiting factor on the overall performance of the
building. These contingencies highlight the importance of exercising caution when adopting code
response modification factors for this form of buildings.

Of all code factors, the reduction factor recommended by Eurocode 8 (for ductility class low
– DCL buildings) is by far the most conservative (q0 =1.5). The value essentially accounts for
overstrength of the building without giving any allowance for ductility (i.e. 𝜇 = 1), which is in
good agreement with the trends of building response observed in this study.

Another important observation is that while some force reduction factors (code factors) may
well be conservative for use in the design of the building, the recommended values for ductility
can overestimate the actual deformation capacity of the building. This is consistent with the main
argument against the ‘force-based’ seismic design approach where the ‘trade-off’ between the
strength and ductility of a building is assumed to be valid. It is speculated that the ‘assumed’ value
of 𝜇 (recommended by the design code) can only be achieved when design considerations are
specified for the local ductility (deformation) and strength demands in the critical storeys above
the interface. This issue will be further explored in the subsequent sections of this chapter.

220
 Table 6-2 Summary of the response modification factors for the analysed buildings

Seismic
Description of ASCE/SEI EC 8§ EC 8§§
response SB-1 SB-2 TS-1 TS-2 AS 1170.4
parameter 7-10 (DCL) (DCM)
factor

𝟏/𝐒𝐩 (Ω𝐨 ) Overstrength (𝑉𝑦 / 2.92 2.72 2.55 2.18 1.30∗ 2.5 1.5 -
𝑉𝐷 )
𝛍 Ductility (𝜹𝑼𝑳𝑺 /𝜹𝒚 ) 2.64 2.07 1.43 1.32 2.0 4.0∗∗ 1.0 -

𝐑 Ω𝐨 × 𝛍 7.70 5.63 3.66 2.86 2.6† 4.0 1.5 3.6ˆ

𝐑𝛍 See Figure 2-47 2.30 1.87 1.43 1.32

𝐤 𝐞 , kN/m Effective stiffness 9224.33 8487.09 8012.41 9848.76

(idealised response)
𝜹𝒚 Effective yield drift 0.0044 0.0042 0.0053 0.0038
ratio
𝜹𝑼𝑳𝑺 Ultimate drift ratio 0.0116 0.0087 0.0076 0.0050
(at ULS)
+
𝐕𝐧𝐨𝐭𝐢𝐨𝐧𝐚𝐥 , kN See note 1528 1313 1625 1430

𝐕𝐲 , kN Ultimate lateral 3215 2718 2935 2610

strength
𝛂𝐮 /𝛂‡𝟏 See note 2.10† 2.07† 1.81† 1.83†

Footnotes of Table 6-2 continued on next page

221
* Overstrength in AS1170.4 is defined as the reciprocal of the structural performance factor (i.e. 1/𝑆𝑝 )
† Response modification factor adopted in the Australian Standard (𝛍/𝐒𝐩 ) is taken equivalent to the R factor adopted in ASCE/SEI 7-10.
+ base shear corresponding to the notional yield (i.e. first occurrence of yielding in any (primary lateral load resisting) member)
‡ 𝛼𝑢 /𝛼1 is a factor defined in Section 5.2.2.2 and Table 5.1 of Eurocode 8. The value is taken as the ratio Vy to Vnotional. The recommended maximum value
of 1.5 is use for calculating the value of q0 .
ˆ The value of 𝑞0 is calculated as 𝑞0 = 1.5 × 3 × 0.8, the 0.8 is the reduction factor to account for vertical irregularity (the value 3 is obtained from Table 5.1
for ductility class medium type buildings).
** 𝐶𝑑 defined in code implies system ductility, in the case of the building modes examined ASCE/SEI 7-10 (2010) adopts the ‘equal-displacement’ principle
where 𝑅 = 𝜇 = 𝐶𝑑
§ DCL (ductility class low) is recommended for the design of buildings in regions of lower seismicity. A reduction factor (1.5) is recommended which
represents the overstrength (i.e. 𝜇 = 1)
§§ DCM (ductility class medium) is also used for comparison. It is to note that EC 8 allows for some ductility level of design making the comparison with
the other codes more direct. It is however not common to use a high behaviour factor in the study region (lower seismicity), the value is thus for comparison
only.

222
 6.5.4. Effects of the lateral (pushover) force distribution pattern on the response factors

The use of quasi-static (pushover) analysis for estimating response modification factors is
highly sensitive to the adopted loading pattern. Previous studies have appropriated the use of
uniform (mass-proportional) loading pattern for obtaining conservative estimates of the buildings
stiffness and the yielding pattern sustained by the building. However, the ultimate capacity and
the system ductility of the building can vary considerably because of changes to the shape of the
loading pattern. To illustrate this point, the 2D sub-frame model TS-1 is employed in a series of
pushover analyses under different lateral load patterns (shapes). Lateral load patterns that are
proportional to the first, second and modal (adaptive) have been compared to the uniform (mass-
proportional) pattern discussed earlier.

The base shear- roof drift response the building employing the different load distribution
patterns are outlined in Figure 6-19. Of all the four force distribution patterns, the modal
(adaptive) pushover updates the shape of the load throughout the response (Chopra & Goel, 2001;
Ji, Elnashai, & Kuchma, 2007). This is achieved by performing modal analysis at each loading
step using the tangent stiffness at that specific load step (i.e. slope of the capacity curve). Hence,
the method modifies the load distribution to correspond to the softened state of the building as the
analysis commences into the inelastic range. Contributions of different modes (higher modes) on
the load shape are inherently accounted for by combining the shapes of the different modes
proportionally to their respective participation factors.

The sensitivity of the response of the building to the applied loading pattern is clearly
demonstrated when comparing the different capacity curves (in the bilinear format) outlined in
Figure 6-19(a). The ultimate lateral capacities vary considerably with the choice of the loading
pattern. The lower lateral strength observed from analysis using a loading pattern proportional to
the first mode shape suggests that this loading pattern is not representative of the true strength
and stiffness of the building. The effective stiffness (i.e. the slope of the linear branch of the
bilinear curve) is the largest for the second-mode shape proportional load distribution. This is not
surprising since higher modes are associated with higher values of stiffness as inferred from Eq.
6-2 which expresses the classical formulation of the period of vibration (Chopra, 1995).

𝑀 (6-2)
𝑇𝑖 = 2𝜋√ 𝐾 𝑖
𝑖

Interestingly, the differences in the results are much reduced when comparing the global
ductility value (𝜇) of the building (Figure 6-19(c)). This is perhaps the case since the value of
ductility is typically controlled by the local response of the storey directly above the TFL (as
previously discussed) which governs the ultimate condition of the building. As such, the building
would have developed the limited-ductile mechanism (above the interface) shortly after the

223
effective yield drift.

The good agreement between the results from analyses involving the uniform and the modal
(adaptive) load further suggests the adequacy of the former (uniform) in prescribing accurate
stiffness, strength and ductility trends in the response behaviour of the building.

Figure 6-19 Results from the sensitivity study performed on building TS-1 employing
different lateral loading shapes (a) bilinear base shear- roof drift curves (b) comparison
of the effective stiffness (c) comparison of the ductility parameter (𝝁 = 𝜹ULS/ 𝜹y)

224
 6.6. Nonlinear dynamic analyses
In view of restrictions imposed on the use of quasi-static techniques for analyses of tall and
irregular buildings, incremental dynamic analysis (IDA- also referred to as the dynamic pushover)
is performed on building models to obtain a refined estimate of the seismic response modification
factors (namely the force reduction and ductility factors). IDA is a popular analysis tool that is
heavily used in the PBEE framework particularly for seismic assessment of buildings (as
discussed in Sections 2.5.2 & 2.5.3). This method was first incepted in the work of Vamavtsikos
and Cornell (2002) where the method (IDA) has been introduced as a viable substitute of the
quasi-static (pushover) analysis of buildings (Luco & Cornell, 1998). Incremental dynamic
analysis (IDA) involves the use of a suite of ground motion records that are systematically scaled
to cover a wide range of seismic demands on the building (from elastic to inelastic yielding to
collapse as illustrated in Figure 6-20).

Figure 6-20 A schematic illustration of the result of the dynamic pushover analysis
on a building in the form of the intensity measure (IM) vs. the EDP (maximum inter
storey drift limit)

A scalable intensity measure (IM) is selected to quantify the intensity of the ground motion
demand on the building whereas the capacity (or the damage state) of the building is quantified
by the engineering demand parameter (EDP). The maximum inter-storey drift (ISD) is a common
choice for the EDP in tall buildings and has been extensively used to evaluate seismic damages
on shear-critical walls and non-structural components in the building. In this study, the EDP has
been specified as the maximum inter-storey drift ratios (ISD) occurring above the transfer floor
or the podium interface levels. This choice of the EDP with the main focus on the storeys above
the level of the interface is founded on the observations reported in Section 6.5 where most of the
critical seismic damages have been reported in the storeys above (and not below) the level of the

225
podium. The choice of the parameter does not preclude explicit checks that are made for potential
damages occurring in the lower podium storeys. Damages in the building have been monitored
by mapping (step-by-step) the component response parameters (e.g. concrete and steel strains and
shear force intensities in walls) and the global response parameters (e.g. ISDs and base shears)
throughout the analyses (similar to the approach discussed in Section 6.5). The remainder of this
section outlines a description of the dynamic properties of the building models, the ground
motions employed in the analyses and the definitions of the performance levels that will be
adopted in this study.

6.6.1. Eigenvalue analysis

Eigenvalue analyses have been performed on the four building models to determine the modal
properties of the building. Despite having distinct features, the modal properties are comparable
(see Table 6-2). The mode shapes are presented in Figure 6-21wherein the product of the mode
shape vector (𝜑𝑖 ) and the participation factor (𝛤𝑖 ) are plotted for all the storeys up the height of
the building. Table 6-3 outlines a summary of the first, second and third periods of vibration of
the buildings.

(a) SB-1 (b) SB-2 (c) TS-1 (d) TS-2

Figure 6-21 Mode shapes of the building models

226
Table 6-3 Vibrational periods of the building models

First Mode, seconds Second Mode, seconds Third Mode, seconds

SB-1 1.96 0.51 0.25

SB-2 2.02 0.51 0.23
TS-1 2.01 0.48 0.25
TS-2 1.93 0.40 0.27

6.6.2. Modelling damping

Damping in numerical simulations comprises of the hysteretic and the viscous damping. The
hysteretic damping is implicitly accounted for in the simulations as the building components
(walls, slabs and columns) yield and sustain damages in their response to the earthquake. The
degree of hysteretic damping depends on the area enclosed by the force-displacement curves of
the components (walls, columns and slabs). Generally flexural responses of walls, columns and
slabs are associated with high damping (high energy dissipation) whereas the pinched response
typifying the response behaviour of shear-critical walls is associated with a limited damping. The
viscous damping accounts for other phenomenon such as friction between structural and non-
structural elements (e.g. partition walls) as the building undergoes seismic load reversals. To
account for the viscous damping in the simulations, the Rayleigh model has been utilised. The
Rayleigh damping comprises of two main components: the mass proportional component and the
stiffness proportional component (Chopra, 1995; Spence & Kareem, 2013). The formulations of
these two components are given as follows
𝑎𝑚 𝑇𝑛 𝑎𝑘 𝜋
𝜉𝑛 = + (6-3)
4𝜋 𝑇𝑛

1 (6-4)
𝑎𝑀 = 4𝜋𝜉 (𝑇 +𝑇 )
𝑖 𝑗

𝜉 𝑖 𝑗𝑇𝑇 (6-5)
𝑎𝑘 = 𝜋 (𝑇 +𝑇 )
𝑖 𝑗

where am , ak are the mass and stiffness proportionality factors, 𝜉 is the nominated damping (as
% of critical damping (𝜉𝑐𝑟𝑖𝑡 )), 𝑇𝑛 , 𝑇𝑖 and 𝑇𝑗 is the n-th vibration period and the two periods (with
subscripts i and j) that correspond to a constant value of the assumed damping (see Figure 6-22).

227
 Figure 6-22 Rayleigh damping model showing mass and stiffness proportional
components

The first component (mass proportional) is argued to result in unrealistically high damping
values as the building experiences inelastic demands (Priestley & Grant, 2005; PEER/ATC 72-1,
2010). This is also demonstrated in Figure 6-22 by the spurious increase in the value of the mass-
proportional damping with the increase in the period of the building. On the other hand, the
stiffness proportional component attributes lower viscous damping values as the building
undergoes period lengthening with higher deformation demands (see Figure 6-22). As the
inelastic demands on the building increase, the hysteretic damping component becomes more
representative of the amount of the seismic energy dissipation experienced by the building. Thus,
in this study, only the tangent (or instantaneous) stiffness proportional component of the viscous
damping is considered in the numerical simulations (i.e. am = 0) as recommended by Priestley and
Grant (2005).

6.6.3. Ground motions selection criteria

For the purpose of conducting the IDA simulations, a suite of 40 ground motion records has
been considered to represent different possible seismic demands and intensities on the buildings.
The ground motion records adopted in this study are categorised into three bins: stochastic,
artificial and historical records. For the first bin, the program GENQKE (Lam, 1999) is employed
for generating earthquake ground motions on rock (outcrop) that are representative of Australian
seismic and geologic conditions. The program uses a seismic source model that is compatible
with intraplate conditions proposed by Atkinson (1993) to estimate attenuation properties of the

228
ground motion in the upper crustal layers (Lam et al., 2000a). The magnitude (M) and distance
from source (R) combinations (M-R for short) that are representative of rare and very-rare
Australian seismic scenarios are adopted following previous studies and publications relevant to
the study region (Amirsardari et al., 2017; Kafle et al., 2011; Elisa Lumantarna et al., 2010; Shahi,
Lam, Gad, Wilson, & Watson, 2017). Input M-R combinations are varied to obtain records with
peak ground acceleration (PGA) values (or Z in AS 1170.4:2007) in the range of 0.08g-0.3g which
roughly represents seismic events with recurrence intervals ranging from 500 to 2500 year return
period respectively.

The ground motion records simulated on rock outcrop (from program GENQKE) are then fed
into the program DEEPSOIL (Hashash et al., 2015) to simulate accelerograms atop of soil sites.
The shear wave velocity profiles from the borehole logs from different sites across Australia
reported in Kayen et al. (2015) have been adopted to simulate surface accelerograms on soil sites.
Specifically, the profiles that are representative of site class De (defined as deep or soft soil sites)
based on the AS 1170.4:2007 classification have been used for simulating the soil accelerograms
for use in the numerical simulations (sample profiles are shown in Figure 6-23). The choice of
site class De (as defined in AS 1170.4:2007) for simulating soil accelerograms is based on
observations reported in a recent study by Hoult (2017) where it is reported that ground motion
records evaluated on soft soils (site classes De and Ee) are most damaging to buildings situated in
major metropolitan cities in Australia. This observation is attributed to the tendency for soil sites
to attenuate displacement demands of the ground motion. The procedure for obtaining the site
response is schematically illustrated in Figure A-1 (Appendix A)

To further illustrate the effect of soil sites on the ground motion characteristics, the rock and
soil accelerograms of a sample record (record no. 10 in Table A-1 of Appendix A) along with
their acceleration and displacement response spectra are compared in Figure 6-24. The rock (base)
record has been simulated on GENQKE whereas the soil response of the same record is obtained
by analysis using the program DEEPSOIL (with considerations made for the nonlinear soil
response characteristics). It is demonstrated that the amplifications in the spectral ordinates of the
records are mainly manifested in the short period range (T<1.0 seconds) (see Figure 6-24(b)).
Significant amplifications of the displacement demands are also observed in the longer period
range (as displayed on the displacement response spectrum of the same figure). This is particularly
important for tall buildings where significantly higher displacement demands are anticipated on
soil sites compared to those on rock. These amplifications are speculated to impose high ductility
demands on buildings of periods > 2.0 seconds.

229
 Figure 6-23 Shear wave velocities used in the study that are representative of Australian
soil and geologic conditions (for Site Class D defined in AS 1170.4:2007)

(a) Comparison of accelerograms on (b) Comparison of the displacement and

rock and on soil acceleration spectrum

Figure 6-24 Simulation of accelerogram on soil using DEEPSOIL

The second bin of ground motion records have been generated using the program SeismoArtif
(SeismoSoft, 2014). These records (total of 7) are compliant to the bilinear Australian code
spectrum (AS1170.4:2007) evaluated on site class De. It should be noted that the current AS
1170.4 design spectrum does not reflect the effects and the dependencies of the site period or the
intensity of the shaking on the shape of the spectrum. For the case with tall buildings, the high
fundamental periods of vibration position the response of the building well within the

230
‘displacement-controlled’ region of the response spectra (refer Section 2.4.3). As such, the
current bilinear design spectra is likely to overestimate displacement demands on the building
(refer Figure 6-25). This overestimation may be warranted given that more (site specific)
information will need to be examined before a refined estimate of the demand can be made.

Figure 6-25 Comparison of rock, soil and code displacement response spectrum

The last bin comprises of historical records obtained from the PEER ground motion database
(PEER, 2013). The selection criteria adopted incorporates the following features: faulting
mechanism, the average shear wave velocity to a depth of 30 m (i.e. 𝑉𝑠,30 ) and limits on the
source-to-site distance (>20 but less than 50 kM). The shallow, reverse faulting mechanism has
been chosen as the key seismological features of the records which also characterise earthquakes
in Australia (Amirsardari et al., 2017; A. Brown & Gibson, 2004b). Average shear wave velocities
similar to those obtained from analyses of soil columns have been specified in the search (a range
of 𝑉𝑠,30 270m/s-380m/s). The response spectra of the ground motion records in both the
acceleration and the displacement formats are presented in Figure 6-26(a)-(b). A full summary of
the records is outlined in Table A-1 of Appendix A.

In Figure 6-27, the key ground motion parameters and the possible IM (intensity measure)
candidates (PGA, PGV, 𝑅𝑆𝐷𝑚𝑎𝑥 and Sa (T1)) are outlined to clarify the range of ground motion
intensities considered in the study.

The record-to-record variability demonstrated in Figure 6-26(a)-(b) is primarily attributed to

the choice of seismic scenarios that have been discussed earlier. Nevertheless, good consistencies
are observed between the mean spectrum ordinates and that of the code (AS1170.4:2007)
particularly in the range of the fundamental periods of the building models (approximately T=2.0
seconds). Differences between the two spectrums (mean and design) are pronounced in the low
period range (T< 1.0 seconds) where the spectral ordinates of the mean are higher than that of the
code. This is not surprising since the code spectrum specifies the second corner period at 1.5

231
seconds (see Figure 6-25) whereas records simulated on soil show a somewhat lower value of the
second corner period (in the range of 0.8-0.95 seconds). This discrepancy is discussed in depth in
recent studies by Amirsadari et al. (2017) and Tsang et al. (2017) where similar observations in
regards to the corner period on soft soils have been noted.

This variability (in the ground motions) is an important source of uncertainty in regions of
low-to-moderate seismicity where definitive knowledge of seismic sources (e.g. faults) or
historical records are often in paucity. More importantly, the dispersion in the input motion is
central for singling-out a possible intensity measure (IM) from the pool of candidates discussed
above (as will be illustrated in Section 6.8). In the IDA study to be presented in the later sections,
each record (total of 40) will be systematically scaled to simulate distinct response characteristics
of the building (e.g. yielding patterns and ultimately collapses).

232
 (a) Acceleration spectra (total of 40 records)

(b) Displacement spectra (total of 40 records)

Figure 6-26 Acceleration and displacement spectra of the ground motion records
used in the study (plotted in logarithmic scale)

233
 Figure 6-27 Summary of the ground motion records presented in terms of their
respective values of (a) PGA (b) PGV, (c) RSDmax (d) ordinate of the spectral
acceleration evaluated at the first period of vibration T1 = 2.0 seconds

6.6.4. Definition of the performance levels

The performance levels that have been defined to evaluate the seismic response behaviour of
the buildings are next examined. Three performance levels have been considered in this study and
these are the immediate occupancy (IO), Life safety (LS), and collapse prevention (CP) limit state.
As discussed earlier, the maximum inter-storey drift ratio above the level of the interface (ISD)
has been defined as the EDP in this seismic performance assessment study.

The immediate occupancy (IO) limit state is defined as the maximum ISD ratio corresponding
to the first occurrence of flexural yielding in the walls (primary or secondary) of the building.
This description roughly corresponds to the immediate occupancy (IO) limit state defined in
FEMA 356 (2000) and ASCE/SEI 41-13 (2013).

The life safety limit state (LS) is the maximum ISD ratio for when all the tower walls above
the interface level have yielded in flexure. This limit, as shown in Section 6.5.2, might occur
before, or after, yielding of the walls and columns of the podium. It has also been demonstrated
earlier that shear force demands on the walls above the interface level can be significant. Thus,
another criterion is set for this limit state (LS) for when the tower walls (Walls 1 & 2) attain their
nominal shear strength. It should be noted that flexural yielding in this context is determined (for
any component) once the reinforcing steel reaches its yielding strain limit or when concrete (fibre)
strength reach their ultimate value (not to be confused with the ultimate strain value of -0.003
discussed in Section 6.5.1). For the buildings featuring a transfer plate (i.e. TS-1 and TS-2), the
flexural yielding of the plate is also considered as a criterion for the building to achieve the LS
performance level.

234
 The collapse prevention (CP) performance limit is defined as the maximum ISD ratio in the
tower corresponding to the first occurrence of the ultimate limit states that have been defined in
Section 6.5.1. This consistency (in defining the CP performance level) will facilitate the
comparison between the results obtained from the IDA (to be reported in the subsequent section)
and those from the quasi-static analyse (reported in Section 6.5). A descriptive summary of the
performance limits adopted in this study is outlined in Table 6-4 along with a listing of the
performance levels that have been proposed in previous studies on similar building
configurations.

6.7. Assessment of the seismic performance

A large number of NLTH are performed on the building models employing the ground motion
records discussed in Section 6.6.3. Each of the 40 records has been systematically scaled to
produce a spectrum of building responses ranging from linear elastic to ultimately the collapse of
the building (defined at the onset of the dynamic instability of the model as illustrated in Figure
6-20). Both the local (component level) and the global (building as a whole) assessment have
been carried out for the four building models. The local assessment includes monitoring the
seismic response behaviour of the components making up the building (strength and deformation
capacity) at each time increment. The strutting forces generated in the floor slabs above the
interface level are also monitored to examine the extent of the podiums interference at different
intensity shaking. The parameters that have been shown to govern these forces (e.g. relative
transfer plate rotation ∆𝜃𝑇𝑃 ) are also monitored in order that accuracies of the analytical
modelling approaches (introduced in Chapters 4 & 5) can be further assessed. The global
assessment of the response involves monitoring the deformation capacity of the building as a
whole (roof drifts) and the shear forces evaluated at the base of the building sub-frame models.
The performance levels (yielding, shear strength and crushing) are explicitly simulated by the
numerical model and are monitored throughout the duration of the response. The scaling of
records is terminated once dynamic instability (or numerical divergence) is reached. The
subsequent sections summarise the results obtained from the analyses performed on buildings TS-
1 & TS-2 (Section 6.7.1) and SB-1 & SB-2 (Section 6.7.2).

235
 Table 6-4 Performance levels adopted in this study

Khalifa (2015)
Performance ASCE/SEI Khalifa Looi et al.
Adopted in this study Li et al. (2006) with shear
level 41-13* (2015) (2017)
assessment ‡
IO/First
ISD corresponding to the first occurrence of flexural yielding in the
indication of 0.4% 0.1%† 0.27% 0.27% 0.25%
RC walls making up the building (in the tower or the podium).
yielding.
Life safety limit state is defined as the ISD corresponding to:
1- Flexural yielding of all the tower walls above the podium
interface level (or TFL)
2- Onset of nominal shear force capacity in the tower walls
LS 0.75% 0.24%† 0.59% 0.32% 0.5%
3- Flexural yielding of the transfer plate (in building models TS-
1 and TS-2)
Whichever occurs first

Collapse prevention limit state is defined as the ISD corresponding

to:
1- Concrete fibre reaching the crushing strain value in the walls
(at a compressive strain limit 𝜀𝑐𝑢 = −0.003)
2- Reinforcing steel bars embedded in the walls and columns
CP reaching their ultimate tensile strain(𝜀𝑠𝑢 ). 1.0% 0.875%† 1.18% 0.64% 0.75%
3- 50 % loss of lateral (shear) strength of the tower walls
(Walls 1 &2)
4- Shear force in the central wall (CW) attaining the nominal
strength capacity
Whichever occurs first
General note: the reported values represent the maximum inter-storey drift (ISD) limits for each of the performance levels.
*
Table 10-20 in ASCE/SEI 41-13 specifies the acceptance criteria for the performance levels for RC shear walls that are controlled by shear
†
Li et al. (2006) provides a range of maximum ISD values for each performance level, values for LS and CP reported herein correspond to the average ISD values.
‡
Post-processing technique (described in Section 2.3.2) was used to evaluate shear demands on walls (i.e. shear response behaviour of the walls was not explicitly modelled)

236
 6.7.1. Buildings featuring a transfer structure

Figure 6-28 and Figure 6-29 plot the dynamic pushover (IDA) results obtained from the IDA
performed on building models TS-1 and TS-2 respectively. The blue circles in the figures
represent the maximum base shear and the corresponding roof drift values obtained from the time
history analyses. The quasit-static responses of the buildings (in the idealised bilinear format)
have been superimposed on the plots for direct comparison. Interestingly, the data points obtained
from the dynamic pushover (quantified by the mean) are in good agreement with their quasi-static
counterparts for the initial part of the response of the building up to the roof drift corresponding
to the nominal yield drift (the ascending portion of the bilinear curve). Beyond this roof drift ratio,
the building models analysed under dynamic conditions are shown to sustain larger base shear
and roof displacement capacities compared to those from the quasi-static analysis.

The higher base shear capacities under dynamic conditions are attributed to the larger mass
participations in the response behaviour of the building. Moreover, the interferences of higher
modes (under dynamic conditions) allow more members in the building (mainly floor slabs and
walls) to yield and develop higher strengths and thereby allowing the building to sustain higher
roof drifts and base shear capacities. The phenomenon as described has been widely researched
in earlier studies (Khalifa, 2015; Vamvatsikos & Allin Cornell, 2006; Vamvatsikos & Cornell,
2002).

Figure 6-28 Dynamic pushover envelope (TS-1)

237
 Figure 6-29 Dynamic pushover envelope (TS-2)

The results outlined in the above figures reaffirm the previous observation in terms of the
relative enhancement of the seismic performance of building TS-1 compared with TS-2. Building
TS-1 is shown to sustain higher ultimate base shear capacity and approximately twice the
deformation capacity compared with TS-2 (referring to the roof drifts at which the mean dynamic
pushover curves plateaus roughly at values of 2.0 and 1.00 for building models TS-1 and TS-2
respectively).

The local response behaviour of the building components has been mapped following an
approach similar to that presented for the quasi-static analyses (Section 6.5). This step-by-step
mapping procedure has been adopted to pinpoint the first occurrence of a damage limit state
(defined in Section 6.6.4). The procedure is presented in Figure 6-30 and Figure 6-31 for the
building sub-frame models TS-1 and TS-2 respectively where key wall and slab responses that
have caused the building to achieve (or exceed) a performance limit state (LS or CP) are mapped
on the dynamic pushover (global) response of the same record.

For building model TS-2, the CP and LS performance levels are shown to have occurred at
lower roof drifts compared with the TS-1. This has also been observed from results of quasi-static
analyses presented in Section 6.5.2. Importantly, the margin between the two limit states (CP and
LS) and that between CP and dynamic instability are significantly lower in TS-2 (compared to
TS-1). This is not surprising since it has been shown earlier that the response behaviour of
building TS-2 is predominantly governed by shear-critical response of the tower walls above the
TFL which in turn is associated with a marginal ductility (i.e. failure occurs shortly after the
attainment of the shear strength of the wall as demonstrated in Figure 6-31).

238
Figure 6-30 Mapping first occurrences of limit states in the building (TS-1)

239
Figure 6-31 Mapping first occurrences of limit states in the building (TS-2)

240
 The maximum inter-storey drift ratios (ISD) obtained from IDA analyses are plotted against
the intensity measure of the corresponding ground motion record in Figure 6-32 & Figure 6-33
(for building models TS-1 & TS-2 respectively). The values of ISD that correspond to the first
occurrence of a performance level are superimposed on the figures. The three IM candidates that
have been considered to characterise the intensity of the input motion are the RSDmax, PGV and
PGA. The parameter RSDmax is the maximum ordinate of the displacement spectrum of the ground
motion record (as illustrated diagrammatically in Figure 6-24). The parameter PGA is the
maximum value of the ground acceleration time trace and PGV is the maximum value of the
ground velocities. It should be noted that the maximum ISD values obtained from the analyses
are the same for any choice of IM (i.e. variations between the different IMs occur only in the y-
axis of the IDA curves).

Figure 6-32 Building TS-1 (a) IDA curves for various IMs (b) summary of the ISD
values defining occurrences of the three limit states

241
 Figure 6-33 Building TS-2 (a) IDA curves for various IMs (b) summary of the
maximum ISD values defining the occurrences of the three limit states

The IDA curves (of Figure 6-32 Figure 6-33) show a larger variability for the CP performance
level. This dispersion of the maximum ISD values (for CP) is primarily attributed to the
description of the performance levels. Generally lower drift limits (ISD) are sustained by the
building when the shear strength of the tower walls dictate the CP limit state (i.e. ultimate shear
strength in the central wall or 50% reduction in the shear capacity of the tower walls). Conversely,
a slightly higher ISD values are reported for when the flexural criterion governs this limit state
(i.e. concrete compressive strain or reinforcement ultimate tensile strain limits). The dispersion
(in the value of ISD) is much reduced for LS and IO limit states. It is worth noting that high and
unrealistic ISD values (up to 10 %) have been observed when the building reached dynamic
instability this is however not shown in the IDA curves which are truncated at an ISD value of
0.03 (3.0%). The IM causing higher performance limits to occur (e.g. CP) are lower for building
model TS-2 than those for TS-1. The relatively poor seismic performance of building TS-2
compared to TS-1 is further demonstrated.

Results from the NLTH analyses performed on building models TS-1 and TS-2 are examined
further to evaluate consistencies of the analytical model introduced in the earlier chapter (Chapter
5) for evaluating the extent of the interference by the transfer plate on the response behaviour of
the planted tower walls. The main parameters that comprise the analytical model (defined in
details in Chapter 5) for building TS-1 are outlined in Figure 6-34. The values of the parameter
PRD (peak rotation demand) have been obtained by employing Eq. 5-23(a)-(b) for the values of
RSDmax of the ground motion records used in the study (about 400 records for each building
model). Good correlations between the computed (PRD) values and those of ∆𝜃𝑇𝑃 (from analyses)

242
are well demonstrated in Figure 6-34(a) up to values of ∆𝜃𝑇𝑃 of about 0.004rad. For higher
intensity ground motions (which result in higher value of ∆𝜃𝑇𝑃 exceeding the limit as indicated
by the broken line in the figure) the analytical model based on linear elastic behaviour predicts
lower PRD values. This is the case since the flexural yielding of the transfer plate at the location
of the planted walls (location of maximum bending plate moments) has increased the magnitude
of ∆𝜃𝑇𝑃 above those predicted by the analytical model (PRD). Interestingly, this increase in ∆𝜃𝑇𝑃
(above the predicted PRD values) is not associated with an increase in the strutting floor forces
which is contrary to the linear proportionality (between 𝜀𝑆𝑇𝑅𝑈𝑇 and ∆𝜃𝑇𝑃 ) that has been
demonstrated in Chapter 5 (refer Figure 6-34(b)). The values of 𝜀𝑆𝑇𝑅𝑈𝑇 (and thus FSTRUT) remain
constant for higher values of ∆𝜃𝑇𝑃 where the data points of the figure deviate to the right of the
predicted trend as shown by the red line in Figure 6-34(b). Reductions of the strutting (in-plane)
strains are attributed to flexural yielding of the transfer plate and that of the central wall which in
turn relaxes the degree of displacement incompatibility in between the connected walls above the
TFL. The values of FSTRUT predicted by the model developed in chapter 5 (Eq. 5-24) are therefore
conservative (i.e. result in higher predictions) even for higher intensity shaking as demonstrated
in Figure 6-34(c). Similar trends have been observed for building TS-2 (presented in a similar
manner in Figure 6-35).

243
Figure 6-34 Building TS-1 (a) Computed PRD vs. ∆𝜽TP obtained from analyses (b) in-
plane strutting strains vs. ∆𝜽TP (c) comparison of the predicted magnitudes of FSTRUT
(Eq. 5-24) to those obtained from analyses

244
 Figure 6-35 Building TS-2 (a) Computed PRD vs. ∆𝜽TP obtained from analyses (b) in-
plane strutting strains vs. ∆𝜽TP (c) comparison of the predicted magnitudes of FSTRUT
(Eq. 5-24) to those obtained from analyses

6.7.2. Buildings featuring a setback

Results from dynamic pushover analyses for SB-1 and SB-2 are plotted in Figure 6-36 and
Figure 6-37 respectively. The response behaviour of the two buildings further demonstrates good
agreement between the dynamic and quasi-static pushover analyses for roof drifts below the yield
drift. The ultimate base shear capacities of the two building models (SB-1 and SB-2) are
comparable despite the slight reduction in strength for SB-2 for higher roof drifts (beyond
approximately 3.0%). The slightly higher roof drifts sustained by SB-1 are consistent with the
observations from the results of the quasi-static analyses (Section 6.5). The mapping of local
(component) responses to determine the limit states is illustrated (for a single record) in Figure
6-38 & Figure 6-39 for buildings SB-1 and SB-2 respectively. For building model SB-1, the LS
limit state is achieved when flexural yielding commenced in all the tower walls above the podiums

245
interface (see Figure 6-38) whereas the attainment of the nominal strength of Wall 1 marks the
LS limit state for building SB-2 (See Figure 6-39). These differences are consistent with the
observations reported in Chapter 3 where the interferences by the podium structure have been
shown to be more pronounced in buildings where the tower is positioned at an offset from the
centre of the podium. The increase in shear demand on the interior tower wall (Wall 1 in SB-2) is
directly attributed to these interferences imposed by the podium structure. The CP limit state for
SB-1 is governed by the flexural response of the central wall at the base where the limiting tensile
strain value (0.03) has been reached in the outer most reinforcement. This is in stark contrast to
SB-2 (for the same record) where nominal shear strength of the central wall was first reached at
this performance level.

Differences in the response behaviour of the buildings (SB-1 and SB-2) are further illustrated
when comparing their respective IDA curves along with the distributions the performance limits
(Figure 6-40 & Figure 6-41 for building SB-1 and SB-2 respectively). Building model SB-2 is
demonstrated to have sustained lower maximum ISD values and higher dispersions (scatter) for
the CP performance level compared to building SB-1. These trends suggest that building SB-2
experiences a mixed mode of failure where lower ISD values are attributed to shear controlled
CP criteria whereas higher ISD values are attributed to the flexural response of the walls
governing the CP performance level (as per earlier discussion). Since the strutting (compatibility)
floor forces are not expected to occur in building SB-1 (for reasons discussed in Chapter 3), the
shear-critical response behaviour of the tower walls is somewhat less likely. Thus, the CP limit
state is primarily controlled by the flexural criterion (i.e. strain limits) which in turn is associated
with lower dispersion as shown in Figure 6-40.

Figure 6-36 Dynamic pushover envelope (SB-1)

246
 Figure 6-37 Dynamic pushover envelope (SB-2)

Interferences of the podium structure are next examined for building model SB-2. As
demonstrated in Chapter 4, the strutting forces in the floor slabs (above the interface level) are
highly dependent on the backstay forces developed at the interface level and the restraint imposed
by the podium structure on the tower walls. The analytical models proposed (Eqs. 4-15- 4-16)
have been employed to calculate the predicted values of FSTRUT above the level of the interface.
The value of (KT/Kp) which has been identified as the key parameter for quantifying these
interferences by the podium structure and was found to be 0.22 (calculated based on the procedure
described in Chapter 4). The maximum shear forces on the tower walls have been obtained
directly from the analyses (by summing the shear forces on the tower walls immediately above
the interface) to determine the expected magnitudes of the shear reversal at the level of the
interface based on the ‘pin’ analogy described in Chapter 4. The computed values of FSTRUT are
compared with those obtained from the analyses in Figure 6-42. It is demonstrated that the
predicted values are generally in good agreement with the values from the analyses. For higher
intensity shaking the analytical model predicts higher strutting forces than those observed from
the analyses (as demonstrated by the deviation of the data points away from the 1:1 line in Figure
6-42). This is the case since under higher intensity ground motion, the stiffness of the podium
reduces (softens) which in turn reduces the restraint (or interference) imposed by the podium on
the tower walls. Accordingly, the magnitude of FSTRUT (from analyses) decreases as the building
softens. The predictive model, which assumes linear elastic behaviour and a constant value of
KT/KP , is thus conservative when subject to higher intensity ground motions. Similar observations
have been observed from analyses performed on building models TS-1 & TS-2 (discussed in
Section 6.7.1).

247
Figure 6-38 Mapping first occurrences of limit states in the building (SB-1)

248
Figure 6-39 Mapping first occurrences of limit states in the building (SB-2)

249
 Figure 6-40 Building SB-1 (a) IDA plots for various IMs (b) summary of the ISD
values defining occurrences of the three limit states

Figure 6-41 Building SB-2 (a) IDA plots for various IMs (b) summary of the ISD values
defining occurrences of the three limit states

250
 Figure 6-42 Normalised strutting forces (FSTRUT) in building SB-2

6.7.3. Comparison of results for the building models

Table 6-5 summarises the median ISD values (designated by ISD50%) together with their
standard deviations (𝜎𝐼𝑆𝐷 ) for the three performance levels (CP, LS and IO) from the IDA
analyses performed on the four building models. It is illustrated further that building model SB-1
exhibits better seismic performance with higher maximum ISD values for all the performance
levels (CP, LS and IO). Conversely, building model TS-2 ranks the lowest (compared to other
buildings) in terms of seismic performance. Generally lower ISD values are reported for buildings
featuring transfer structures (TS-1 and TS-2) compared with the other buildings (SB-1 & SB-2)
which is consistent with observations that have been reported by Khalifa (2015). The limiting ISD
values obtained from this study (particularly for TS-1, TS-2 and SB-2) are also in good agreement
with those proposed by Looi et al. (2017) (see Table 6-4). For building models TS-1 & TS-2, the
ISD value corresponding to the CP performance limit are significantly lower than those
recommended by the ASCE/SEI 41-13 (2013) (compare values in Table 6-4 with those presented
in Table 6-5). The relatively poorer seismic performance of buildings TS-1, TS-2 and SB-2, are
attributed to adverse interferences by the podium structure on the response behaviour of the
building (which has been discussed in Chapters 3-5).

It is worth noting that strutting internal forces developed above the level of the podium are not
associated with ductility demands (inelastic deformation) imposed by the earthquake. The
resulting redistributions of internal actions in some tower storeys above the interface must be the
cause of the observed limited-ductile response behaviour of the building (which explains the
lower values of the ISD particularly for the LS and CP performance levels).

Discrepancies in the ISD values corresponding to the CP performance limit state highlight the
importance of incorporating shear response of the walls in the assessment studies involving

251
podium-tower buildings. This also confirms the apparent reduction in the ISD limits (particularly
for the CP limit state) reported by Kahlifa (2015) based on considerations of the shear response
behaviour of the walls (compare the different columns in Table 6-4). It should be noted that a
post-processing approach (described in details in Section 2.3.2) to shear assessment has been
employed by Khalifa (2015) to detect the occurrence of a shear-critical behaviour of the walls (a
description of this approach has been presented in Section 2.3.2). Uncertainties with ISD values
are much reduced with the LS and IO limit states (see Table 6-4).

Table 6-5 Median ISD values for the three limit states

CP limit state LS limit state IO limit state

ISD50% σISD ISD50% σISD ISD50% σISD
SB-1 1.14% 0.14% 0.52% 0.06% 0.25% 0.03%
SB-2 0.93% 0.20% 0.41% 0.05% 0.23% 0.13%
TS-1 0.87% 0.14% 0.46% 0.07% 0.23% 0.03%
TS-2 0.69% 0.16% 0.46% 0.19% 0.22% 0.03%

6.8. Development of fragility curves

Results obtained from IDA analyses (dynamic pushover) performed on the case study
buildings are statistically evaluated to obtain estimates of the median and the standard deviation
of the IM (intensity measure) for the different performance levels considered in this study (a total
of three). To do so, seismic fragility functions (or fragility curves) are constructed for buildings
SB-1, SB-2, TS-1 and TS-2 to systematically evaluate the correlations of seismic damage
(capacity) with ground motion intensities IM (demand). The Multi Stripe Analysis (MSA)
approach (introduced in Section 2.5.3) has been employed to construct fragility functions and to
obtain estimates of the median IM values and their respective dispersions (𝛽). Dispersion in the
context of this study is central to the selection of which IM to adopt in deriving the response
modification factors for the buildings considered in this study. Essentially, the IM that results in
the lowest dispersion (value of 𝛽) is selected.

Results from the numerous IDAs together with the median ISD values for each limit state were
input to a MATLAB script that has been written to fit a cumulative probability distribution
function (CDF) based on the actual observed data (following the approach referenced in Section
2.5.3). Sample results of the fragility curves for building models TS-1 and TS-2 are presented in
Figure 6-43. In these figures, the square symbols shown on the scattered plots represent the
fraction of records at a given intensity level (e.g. RSDmax) that have caused the building reaching
a target limit state (i.e. the number of records at a given intensity where a limit state is observed

252
divided by the total number of analyses performed at this specific intensity). The solid curves (in
Figure 6-43) are the fitted CDFs for each intensity level. The y-axis label (which reads ‘P (CP|
RSDmax)’) is the conventional nomenclature for the probability of exceedance of a limit state (CP,
LS or IO) for a given intensity measure (e.g. RSDmax).

A comparison of the fragility functions of all the building (sub-frame) models using the three
IM parameters is presented in Figure 6-44. The full distribution of the input data (in terms of the
actual number of performed analyses and the fractions of exceedances observed for each IM) is
presented in Appendix G (Figures G-1 – G-4) in the form of bar charts.

Consistent with earlier observations, building model TS-2 exhibits the ‘poorest’ performance
behaviour given that lower IM values infer considerably high exceedance percentages. This is
consistently shown in Figure 6-44(a)-(c) irrespective of the choice of IM (the magenta colour
curve lags behind the others for all the IMs). On the other hand, the more superior seismic
performance of building SB-1 is also illustrated in view of the lower fragilities for higher intensity
ground shaking. Not surprisingly, differences in the building performance observed for the CP
limit state are less pronounced for LS and IO limit states (note that IO is the first occurrence of
flexural yielding in the walls/columns of the building). The median IM (designated by  for
consistency with the original annotation of the MSA frameworkvalue at 50% probability of
exceedance and its standard deviation (designated as also for consistencyas obtained from the
fragility assessment are presented in Figure 6-45 in the form of bar charts.

The dispersion ( is the lowest for when the RSDmax is utilised as the IM. This is consistently
demonstrated for all the building models and for all of the considered limit states (Figure 6-45(b)).
Interestingly, this parameter (RSDmax) has also been utilised as a key determinant for quantifying
interferences by the transfer structure on the shear demands of the walls above the TFL (Chapter
5). The RSDmax is thus selected as the intensity measure for quantifying ground motion intensities.

253
Figure 6-43 Fragility functions for building (sub-frame) models (a) TS-1 (b) TS-2 using RSDmax as the IM

254
 (a) RSDmax as IM (b) PGV as IM (c) PGA as IM
Figure 6-44 Comparison of the fragility curves of the building models derived using RSDmax, PGV and PGA as the IM

255
 (a) Median IM values (𝜃) (b) Dispersion (𝛽) values

Figure 6-45 Results from the fragility assessment of buildings

6.9. Refined assessment of the response modification factors

The final stage of this study involves an assessment of the design response modification factors
as applied to the form of building constructions investigated. Results from the assessment study
(in Section 6.7) using the numerous NLTH (IDA) analyses performed on the four building models
are examined in parallel with those from the fragility assessment (Section 6.8) to determine
estimates of the force reduction and ductility factors of the building. The force reduction factor is
evaluated as the product of two sets of ratios following recommendations by Mwafy et al.
(Mwafy, 2011; Mwafy & Elnashai, 2002).

R = R c × Ωd (6-6)

where R c is the ratio of (ac ) to (ay ) which are defined (in the above referenced study) as the PGA
values of the ground motions where the first condition of “collapse” and “significant damage” of
the building is first reached (respectively). Parameter Ωd is the design overstrength factor.
Significant damage limit state is assumed to be coincidental with the LS performance limit defined
in this study. This analogy is warranted given that most (if not all) seismic induced damages have
been observed to occur above the podium (or transfer) level. The incorporation of the design
overstrength (Ωd ) is to recognise that the design strength of the building is typically lower than

256
the notional yield strength of the structure (which has also been confirmed through the quasi-
static analyse presented in Section 6.5). It should be noted that the original format of Eq. 6-6
considers the PGA as the IM to characterise ground motion intensity. However, the fragility study
(presented in Section 6.8) has shown that RSDmax can more effectively characterise the ground
motion intensities that cause (or exceed) a performance level (i.e. lower 𝛽 values). Eq. 6-6 has
been modified in Eq. 6-7 to incorporate the RSDmax as the IM instead of the PGA.

RSD𝑚𝑎𝑥 (CP)
R = (RSD ) × (Ωd ) (6-7a)
𝑚𝑎𝑥 (LS)

where

RSD𝑚𝑎𝑥 (LS)
Ωd = RSD𝑚𝑎𝑥 (Design)
(6-7b)

where RSD𝑚𝑎𝑥 (LS) , RSD𝑚𝑎𝑥 (CP) is the median (50% fractile) RSD𝑚𝑎𝑥 values for life safety and

collapse prevention performance levels respectively. The values (of RSDmax(CP) and RSDmax(LS)) are
obtained from the fragility study presented in the earlier section (and summarised in Figure 6-45).

The force reduction factor defined in Eq. 6-7 is in fact the maximum reduction factor that can
be adopted for a building since it considers the CP performance level as the ultimate state of the
building. This is contrary to the design objectives defined in most seismic codes which aim to
satisfy the Life Safety performance objective for the design level earthquake (Whittaker et al.,
1999). It should be noted that CP performance levels may still be considered in assessment studies
involving longer return period events (e.g. MCE level earthquakes). To distinguish the different
performance objectives of the seismic design and assessment of the building, two response
modification factors have been considered in this study. The design R factor (RD) adheres to the
LS performance objective for design purposes (Eq. 6-8) and Rmax (Eq. 6-9) which evaluates the
response behaviour of the building at the onset of collapse (i.e. CP limit state). A graphical
illustration of the meaning of RD and Rmax is presented in Figure 6-46.

RSD𝑚𝑎𝑥 (LS) (6-8)

𝑅𝐷 = (RSD )
𝑚𝑎𝑥 (Design)

RSD𝑚𝑎𝑥 (CP) (6-9)

𝑅𝑚𝑎𝑥 = (RSD )
𝑚𝑎𝑥 (Design)

where RSD𝑚𝑎𝑥 (Design) is the maximum ordinate of the code (design) displacement response

spectrum. It should be noted that Eqs. 6-8 & 6-9 differ from Eq. 6-7 in that the design overstrength
(Ωd ) is explicitly considered when RSD𝑚𝑎𝑥 (Design) is used in the numerator (as shown in Figure

6-46 and expressed Eq. 6-7b). The global ductility parameter (𝜇) is defined in Eq. 6-10.

257
 𝐼𝑆𝐷 (6-10)
𝜇 = ( 𝐼𝑆𝐷𝐶𝑃 )
𝐿𝑆

where ISDCP and ISDLS are the median values of the inter-storey drifts at the first occurrence of
CP and LS performance levels respectively (summarised in Table 6-5 in Section 6.7). It is noted
that alternative definitions of the global ductility parameter can be found in the literature. Fanaie
and Shamlou (2015) derived the ductility factor based on the ‘equal-displacement’ principle
which essentially equates the value of 𝝁 to the reduction factor (Rµ in Figure 2-47). Mwafy and
Khalifa (2017) adopted a similar definition to that of Eq. 6-10 for calculating the value Cd (which
represents the displacement amplification factor defined in the ASCE/SEI 7-10). It is also noted
that the value of 𝝁 as defined in Eq. 6-10 represents the maximum ductility (deformation capacity)
that the building can sustain without reaching the CP limit consistent with the definition used for
quasi-static analyses (Section 6.5.3). A summary of the R factors obtained following the above
discussed formulations is presented in Table 6-6. In the same table the R factors obtained from
quasi-static analysis (reported in Section 6.5.3) are also outlined for direct comparison (the
notation RS is adopted for differentiating the quasi-static R factor from that of the dynamic R
factor). These results further demonstrate the hierarchy of the seismic performance of the
buildings where, strictly speaking, larger R factors represent better potential seismic capacity of
the building (with more ‘reserve’ strength). The ductility (𝜇 ) values obtained from dynamic
analyses are generally lower than those specified in most design codes (with the exception of EC8
as outlined in Table 6-2). Not surprisingly, building models TS-1 and TS-2 sustain the lowest
ductility values whereas a slight enhancement is reported for buildings SB-1 and SB-2. The
contrast between the values of 𝜇 has been thoroughly investigated earlier where such differences
have been associated with yielding mechanisms developed in some buildings at the occurrence of
the CP limit state (i.e. shear-critical vs. flexural wall response behaviour).

Figure 6-46 Quantification of the R factor by mapping the demands (RSDmax) to the
response behaviour of the building (IDA curve)

258
Table 6-6 Comparison of the R factors obtained from the dynamic and quasi-static analyses

Quasi-static
Dynamic (IDA)
𝑅𝑆𝐷𝑚𝑎𝑥 (𝐶𝑃) 𝑅𝐷 𝑅𝑚𝑎𝑥 𝜇 𝑅𝑆 𝜇𝑆
𝑅𝑆𝐷𝑚𝑎𝑥 (𝐿𝑆) (Eq. 6-8) (Eq. 6-9)
SB-1 2.12 4.78 10.15 2.20 7.70 2.64
SB-2 1.77 3.74 6.64 2.07 5.63 2.07
TS-1 1.50 4.13 6.18 1.90 3.66 1.43
TS-2 1.54 3.02 4.66 1.50 2.86 1.32

Figure 6-47 Comparison of the R factors obtained from this study to those proposed in
some seismic codes

The fraction (𝑅𝑆𝐷𝑚𝑎𝑥 (𝐶𝑃) ⁄𝑅𝑆𝐷𝑚𝑎𝑥 (𝐿𝑆) ) in Table 6-6 defines the margin of safety for a

building against collapse. This ratio takes a lower value for when the building response is
governed by a limited-ductile mode of failure (i.e. shear failure in walls) which are associated
with limited (if any) ductility (as observed for building TS-2). The computed R factors (in

Table 6-6) are compared with those recommended by some design codes in Figure 6-47. The
conservatism of these factors varies considerably between the design codes. The value
recommended by EC 8 (for DCL class buildings) is by far the most conservative. The lowest
value of the computed RD (for building TS-2) is close to the AS 1170.4:2007 recommended value
(Sp =2.6). Nevertheless, the recommended ductility value (in
AS1170.4:2007overestimates the deformation capacity of buildings featuring transfer structures
(refer Figure 6-48). Thus, the assumed ‘trade-off’ between strength and ductility which is the
basis of the force-based approach, does not seem to hold true for buildings with features similar
to those examined in this study. The results suggest that a conservative force reduction factor of
3.0 for podium-tower buildings may still be adopted for the design of the building. On the other
hand, the use of ductility factors higher than 1.5 would result in an unconservative design of the
building (particularly for a building featuring a transfer plate).

259
 Figure 6-48 A comparison between the AS1170.4:2007 assumed ductility factor
(and those obtained from the assessment study

6.10. Chapter summary

A thorough investigation of the seismic performance of buildings featuring podiums (setbacks
and transfer structures) is presented in this chapter. The building sub-frame models that have been
considered in this study feature a podium with varying degree of irregularity. The limited ductile
response behaviour of the building has been quantified in view of the adverse interferences of the
podium structure on the seismic performance of the tower walls immediately above the interface
level. Local concentrations of inelastic demands (seismic damages) have been observed in the
direct vicinity of the podium interface level. Such concentration can restrict the building from
achieving larger deformation capacities for higher intensity shaking. This has been consistently
shown to be the case in view of results from quasi-static and dynamic analyses.

A large number of incremental dynamic analyses have been performed to estimate the limits
(inter-storey drifts) for different seismic performance levels (IO, LS and CP). Generally, lower
ISD values have been observed when shear-critical conditions were manifested in the tower walls
(above the interface). These conditions were most pronounced for building models SB-2, TS-1
and TS-2 where adverse interferences by the podium were anticipated. Seismic fragility curves
of the building models have further demonstrated the heightened vulnerability of podium-tower
buildings incorporating a transfer plate compared with the other (podium-tower) buildings
particularly in high-intensity earthquakes. The maximum displacement response spectrum
ordinate (RSDmax) has been observed to result in the lowest dispersion values (of the IM) which
had led to its selection as an intensity measure (IM) in this study. The limiting ISD values
(representing the capacity of the building) obtained from results of the IDA and the fragility
analyses (representing the seismic demand) have been combined to estimate the force reduction
factors applicable for the podium-tower type of buildings. For most cases, the observed reduction
(R) factors were higher than code recommended values (with varying margins). However the

260
reported ductility factors (from this study) were substantially lower than those recommended by
some design codes. This observation is critical as it demonstrates the limited ductile performance
of the buildings despite them having high force reduction factors. This issue will be addressed in
the next chapter.

The assessment study also highlights the adverse effects that the podium imposes on the
response behaviour of the building. The relatively poor performance of buildings incorporating
transfer plates has been consistently observed through the numerous parts of the reported study.
This has been determined to be the case in terms of the lower seismic intensities that the building
can sustain (lower IM) and the lower deformation capacities developed by the building (ISD) at
the various performance levels. The ratio of the 𝑅𝑆𝐷𝑚𝑎𝑥 at CP to that of LS showed a lower safety
margin for transfer structures (compare values in Table 6-6) which further highlights the
(relatively) poor performance of transfer structures under a larger-than-expected earthquake
intensity. For podium-tower buildings featuring setbacks, distinct differences in the response
behaviour of the building have been reported between cases with centrally positioned and offset
towers. These distinctions have been investigated in depth in Chapter 3 and also outlined in this
chapter where lower performance levels were reported for building (SB-2; with offset) compared
to building SB-1(centred towers). Such distinctions (in the seismic response behaviour) are not
currently transcribed in seismic design codes or guidelines.

The reported analyses have also verified the analytical models introduced in Chapters 4-5 for
quantifying the amount of strutting forces generated in the floor slabs. Predictions by the
analytical models of the magnitude of FSTRUT (which have been developed based on linear elastic
behaviour) have been shown to be conservative for higher intensity earthquakes. Accordingly, the
observed magnitudes of FSTRUT (above the interface level) and the additional shear forces
transferred to the tower walls reduce as the storeys above the interface level yield and respond
inelastically to the earthquake. The conservatism in the formulations of the analytical models may
well be justified for practical structural design purposes given the complexity of the podium-tower
interactions when subjected to earthquake loads.

261
262
 Chapter 7. RECOMMENDATIONS FOR DESIGN PRACTICE

7.1. Introduction
This chapter provides an outline of key recommendations to guide design practice based on
observations from the numerous studies presented in this thesis. The quantification of
interferences by the podium structure on the shear force demands of the tower walls has been the
main thrust of the study presented in Chapters 3-5. At the core of these interferences is the
development of the internal strutting forces (FSTRUT) in floor slabs connecting the tower walls
above the interface and the consequent additional shear forces transferred to the tower walls.
These forces are particularly complex and in most cases are not reported from the conventional
analyses techniques employed in practice. In Chapters 4 & 5, simple analytical tools have been
introduced for quantifying the values of FSTRUT in setback buildings and transfer structures
respectively. In both building configurations, these forces (FSTRUT) have been attributed to the
extent of interferences by the podium structure which are manifested by the differential restraint
imposed by the podium on the tower walls (buildings with a setback) or by differences in the
transfer plate rotations (∆𝜃𝑇𝑃 ) at the base of the connected tower walls (building featuring a
transfer plate). In Chapter 6, a systematic assessment is conducted on response modification
factors that are used in the seismic design of buildings. The limited-ductile response behaviour of
the building in the immediate proximity to the interface level limits the building from sustaining
larger deformation. A design seismic shear force envelope is proposed (in Section 7.2) to
safeguard the ‘critical’ storeys from experiencing limited-ductile response behaviour in seismic
conditions. A verification study is presented to demonstrate the accuracies of the proposed
methodology (Section 7.3). A framework for quantifying the maximum magnitude of FSTRUT is
presented in the subsequent section to guide design (Section 7.4). Key proposals are outlined for
potential enhancements to design code provisions and the overall understanding of the
interferences by the podium structure in the seismic design and assessment of podium-tower
buildings (in Sections 7.6 -7.7).

7.2. Response modification factors for the seismic design of podium-tower

buildings

In regions of high seismicity, stringent detailing provisions are mandated by seismic design
codes to ‘justify’ the use of relatively ‘high’ response modification factors (e.g. R factors) for
quantifying seismic demands on buildings. These requirements are set to ensure that critical
components in a building (e.g. walls and columns) are capable of deforming within a prescribed
(and acceptable) limits as the building is subject to the design (or higher) intensity earthquake.
The adopted capacity-design philosophy (in regions of high seismicity) also aims to offer
protection against limited-ductile mechanism (e.g. shear failure in walls) from forming during the

263
response of the building to an earthquake. This is typically achieved by proportioning a ‘larger-
than-expected’ strength reserve for some potentially vulnerable components (Paulay & Priestley,
1992; Priestley et al., 2007).

Similar requirements are often not mandated in seismic design codes in regions of lower
seismicity (e.g. Australia). Observations presented in Chapter 6 suggest that the unfavourable
response behaviour of the tower walls (above the podium) restricts the building from achieving
larger deformations when subjected to ground motion excitations. Despite this, the study (in
Section 6.9) still reported considerable force reduction factors to the buildings by virtue of their
inherent overstrength and the low (seismic) design forces typifying the seismic loading criteria in
regions of low-to-moderate seismicity. Evidently, the low-ductile response behaviour of some
tower storeys is attributed to the high inelastic demands developed in the (tower) storeys above
the podium interface level. It can be argued that the most intuitive approach would be not to use
a reduction factor in the first place (i.e. adopt R=1). This, however, is not justified true for the
examined buildings as significant R factors could still be applied (see Table 6-6). Furthermore,
adopting a lower force reduction factor (or even R=1) for the entire building may not be
economical since the building is well capable of developing a portion of the assumed ‘trade-off’
between strength and ductility (i.e. reported 𝜇 values always well exceeded unity).

In the absence of capacity-design checks or provisions in regions of lower seismicity, local

concentrations of inelastic demands (immediately above the interface) will need to be
accommodated (in the design of the building) by providing higher strength reserves in these
storeys. This approach will ensure that the component strength levels of the walls, cores and
columns in these critical storeys will not be exceeded in an earthquake. A similar approach is
adopted in ASCE/SEI 7-10 where amplification factors are introduced to the seismic design forces
at the podium levels and the levels below (as discussed earlier). However, no specific attention is
cast on the potentially ‘much worse’ performance of the above tower storey. To address this issue,
a local amplification factor is introduced to increase the design forces in certain storeys in the
building which would have been determined by the use of the code reduction factor. The force
amplification factor is derived based on the equal-displacement principle where the maximum
value of the ductility factor)is equated to the actual reduction factor in the storeys above the
podium. For buildings featuring transfer structures or a setback, a conservative ductility value of
1.5 is deemed appropriate based on results from the extensive numerical study presented in
Chapter 6 (refer Table 6-6). Accordingly, the amplification factor ( 𝛽𝐷𝑇 ) for use in the design is
expressed in Eq. 7-1.

𝑅𝐶𝑜𝑑𝑒
𝛽𝐷𝑇 = 1.5
(7-1)

264
where RCode refers to the code reduction factor that is uniformly applied (by default) to all parts of
the building. As noted in Chapter 6, for a conservative design of a podium-tower building a
maximum value of RCode =3.0 should be adopted (the amplification factor 𝛽𝐷𝑇 is accordingly equal
to 2.0). Seismic design storey shear forces are computed as the product of 𝛽𝐷𝑇 and the design
storey shear force intensity 𝐹𝑖,𝑑𝑒𝑠𝑖𝑔𝑛 in the two storeys above the interface level (where the limited
ductile response behaviour is anticipated).

𝐹𝑖 = 𝐹𝑖,𝑑𝑒𝑠𝑖𝑔𝑛 × 𝛽𝐷𝑇 (7-2)

Design combinations incorporating 𝛽𝐷𝑇 will ensure that the tower walls are designed and
proportioned for much higher level of strength (particularly shear strength) whilst still
incorporating force reduction factors for the design of the rest of the building (in the tower and
the podium). It should be noted that the design of the tower wall in these storeys (above the
podium) will also need to accommodate the additional shear forces that are transferred from the
floor slabs following the design recommendations of earlier chapters (and outlined in Section
7.4). The applicability of this approach is examined in the subsequent section by way of nonlinear
time history analyses employing a case study building.

7.3. Verification study

The proposed seismic design envelope introduced in Section 7.2 is verified herein by way of
analyses performed on the case study building described in Chapter 5 (Section 5.7.2). A fibre-
based 3D model of the building has been constructed on program SeismoStruct (2016). The tower
floor plan has been slightly adjusted in order that a feasible numerical model can be constructed.
The tower walls (Wall 1, 2 & 3 presented in Figure 5-41) are planted on the transfer plate (forming
the gravity system of the tower in the global x-direction). The transfer plate has been idealised as
an equivalent inelastic frame element with effective width assigned as the width of the column
strip along the supporting podium columns and walls (refer Figure 7-1(a)). The floor slabs have
been similarly defined in the numerical model (as inelastic frame elements) following the
approach recommended by the PEER/ATC 72-1 guideline. In the transverse direction (parallel to
the global y-direction), the tower walls have been lumped into equivalent elastic frame elements
following the PEER/ATC 72-1(2010) and the TBI (2017) guidelines (similar to the approach
discussed in Section 3.5). By use of this approach, both the mass and the stiffness distributions
up the height of the building are modelled despite the adopted simplification. Fibre-based inelastic
frame elements have also been employed in the modelling of the walls, columns, slabs, the
transfer plate and the core. Nonlinearity has been defined at the material level by proportioning
distinct uniaxial constitutive material properties to the reinforcing steel and concrete (confined
and confined) following the procedure outlined in Section 6.4.1 (see Figure 7-1(b)). The

265
uncoupled shear backbone of the wall has been incorporated using a ‘zero-length’ spring model
(as described in Section 6.4). The continuous core shaft has been modelled using the wide-column
analogy which has been also adopted in previous studies (Beyer et al., 2008; Mwafy & Khalifa,
2017; Stafford-Smith, 1986).

(a) 3D render of the nonlinear model of the building

(b) Inelastic modelling approach adotped for the walls of the building
Figure 7-1 Case study building

To demonstrate the accuracies of the adopted simplification, dynamic modal analysis is

performed on the building model of Figure 7-1. Results are compared against those obtained from
the full building model constructed on ETABS (described in Section 5.7.2). The comparison of
the first three translational modes (in the x-direction) is presented in Figure 7-2(a)-(b) where good
consistencies in the mode shapes (weighted by the modal participation factor i.e. ii) and periods

266
of vibration are demonstrated.

(a) Mode shapes

(b) Fundamental periods of vibrations

Figure 7-2 Comparison of the dynamic properties of the modified and full building
models

The 3D case study building is employed in a series of nonlinear time history analyses using a
suite of 20 ground motion records described in Table A-1 of Appendix A (record nos. 20-40). The
subset (of records) has been selected such that the mean spectrum (of the records) matches with
the AS 1170.4:2007 code spectrum at the first period of vibration of the building (see Figure 7-3).
Rayleigh modal damping has been adopted by assigning a critical damping value of 2.5% to the
first and third modes of vibration (the procedure has been discussed in Section 6.6.2).

267
 Figure 7-3 Acceleration response spectra of the records employed in the time history
analyses

The main objective of this part of the study is to validate the application of the recommended
amplification factor on the code shear force demands (Section 7.2). For brevity, only results from
analysis of a single record (record no. 40 in Table A-1) are presented in detail. The five
predominant yielding stages in the building together with their respective instances of occurrence
are superimposed on a graph showing the base shear time history in Figure 7-4. The
concentrations of inelastic demands in the vicinity of the TFL are confirmed in Figure 7-5. The
distribution of the storey ductility demand (refer Chapter 6 for exact definition) of Figure 7-6 is
consistent with the observations reported from the analyses of a similar building model (in Section
6.5.2).

Figure 7-4 Base shear time history (record no. 40) showing five key yielding stages

268
 Figure 7-5 Inelastic yielding up the height of the building (greed highlight indicate
that yield strains are reached in some or all reinforcement bars in the section)

Figure 7-6 Distribution of the maximum storey ductility demand up the height of the
case study building

The maximum seismic storey shear force distribution envelopes resulting from the analyses of
the 20 records are plotted in Figure 7-8. The values of the envelopes at each storey level (the grey
lines in Figure 7-8) have been computed by summing the maximum shear forces developed in the

269
components of each storey level (not to be confused with the instantaneous shear forces at a given
time). The design force distribution recommended by the AS 1170.4:2007 which incorporates the
use of a response modification factor (μ/Sp ) of 2.6 is superimposed for direct comparison (dashed
red line in Figure 7-8). This design shear force distribution is obtained by performing an elastic
response spectrum analysis (RSA) using the AS1170.4:2007 code spectrum. The mean base shear
force obtained from the NLTH is in good agreement with the code estimate of the base shear
(which utilises a force reduction factor of 2.6). This further confirms the results reported earlier
(in Chapter 6) where force reduction factors (in the order of 3.0) were observed in the response
of the building (of this configuration i.e. with a transfer plate). Nevertheless, the maximum shear
force demands immediately above the podium are much higher than those specified by the design
code. The higher seismic demands in the two storeys above the interface level are attributed to
the local concentrations of inelastic seismic demands above the TFL (shown in Figure 7-6).
Specifically, as the tower walls and the core in this level yield and respond within the inelastic
range, larger shear forces and bending moments would have been developed in the different
components of the storey. This phenomenon is schematically presented in Figure 7-7 where
discrepancies between the design and the actual component strength are clearly attributed to the
inelastic yielding of the component.

Figure 7-7 Illustration of the effect of the component yielding on the developed sectional
strength

The proposed shear force demand envelope (in Section 7.2) is obtained by multiplying the
design (code) shear forces in the two (tower) storeys above the TFL by the value of 𝛽𝐷𝑇
(determined by Eq. 7-2 as 2.6/1.5 =1.73). The proposed distribution is shown to be in a good
agreement with the shear demands obtained from the NLTH analyses in these storeys. By
proportioning a larger strength in these storeys, the limited-ductile response behaviour of the walls
and cores is postponed or even avoided. It should be noted that the lateral stiffness of these storeys
(where higher design strengths are used) would also be increased. The increase in stiffness is
manifested by the potential increase in the section sizes to accommodate the higher (design)
strengths demands. This increase in storey stiffness will ultimately reduce the stiffness gradient

270
in the lower and upper storeys in the building which in turn reduces the degree of irregularity up
the height of the building.

Figure 7-8 Comparison of the maximum storey shear force envelopes obtained from
analyses and those of the design

7.4. Determination of the additional shear force demands in the tower walls above
the interface

7.4.1. Podium-tower buildings featuring a setback

Design shear forces in the tower walls above the podium interface level have been shown to
be understated when the conventional building analyses methods are employed (as demonstrated
in Chapter 3). Internal (strutting) forces are transferred to the tower walls and columns above the
podium level by the connecting floor slab (or beam) as a result of the displacement incompatibility
between the connected walls. These forces are manifested when the tower is positioned at an
offset from the centre of the podium (refer to Chapter 3 for a detailed definition of the building).
The resulting shear force redistributions are most pronounced in the interior walls that are closer
to the centre of stiffness of the podium. In Chapter 4, an analytical approach has been introduced
for estimating the maximum strutting floor slab force intensities (and their respective
distributions) in the two storeys above the podium interface. It should be noted that in the absence
of tower walls, the tower columns (closer to the centre of stiffness of the podium) would be
exposed to these additional forces. Thus, the same model will need to be applied in determining

271
the shear forces demands on these secondary (gravity) columns (verifications have been already
presented in Section 3.5). An estimate of the maximum backstay reactive forces at the level of the
interface can be obtained by the use of the idealised building model. To achieve this, the applied
forces at the tower level will first need to be determined by either the response spectrum analysis
(RSA) or the equivalent static force method. An upper-bound estimate of the backstay (reaction)
force can be obtained by the use of Eq. 4-13 (which is based on an idealised ‘pin’ support at the
level of the podium). This reaction (backstay force) is further refined to consider the flexibility of
the podium which also deforms (laterally) as a result of the applied loads. Eqs. 4-15 (a)-(b) are
employed for this purpose. Strutting forces in the floors slabs can then be estimated by Eq. 4-16
which assumes a linear distribution of forces above the interface. It should be noted the resulting
force (Eq. 4-16) is the maximum force that can be generated in the floor slab. Accordingly, a
conservative shear design envelope of the interior wall is obtained by summing the strutting (slab)
forces to the (equilibrium) design shear forces (imposed by the externally applied loads). A
summary of the above-described procedure is outlined in Figure 7-9.

272
 Figure 7-9 A step-by-step procedure for determining total shear force demands on
tower walls in setback buildings (above the interface level)

7.4.2. Buildings featuring a transfer plate

A step-by-step flow-chart of the developed analytical model for estimating the magnitude of
𝐹𝑆𝑇𝑅𝑈𝑇 in the floor slabs connecting the tower walls that are supported on a transfer plate is
presented in Figure 7-10. First, the value of PRD is determined based on the translational and
rotational stiffness of the building and the value of RSDmax of the design level earthquake
(obtained from the design displacement response spectrum). The relative stiffness ratio (𝛼𝑟 ) is
computed based on the sectional properties of the transferred wall (most flexible wall in an

273
assembly) and the flexural rigidity of the transfer plate (Eq. 5-2). The additional seismic shear
force demand on the wall can then be determined by the use of Eq. 5-24. The calculated force
(𝐹𝑆𝑇𝑅𝑈𝑇 ) will need to be added, arithmetically, to the design shear force of the transferred wall
obtained from conventional analyses of the building in which the slab-wall interactions are
(typically) ignored.

Figure 7-10 A step-by-step procedure for determining total shear force demands on
transferred walls above the TFL

274
 7.5. Modelling of floor slabs
The study reported in Chapters 3-6 highlights the adverse interferences by the podium
structure on the tower above. These interferences have primarily resulted from the different
boundary conditions imposed by the podium structure on the tower walls. In-plane forces are
generated in the floor slabs as a result of the ‘push’ and ‘pull’ actions between the tower walls.
This is contrary to the behaviour conventionally assumed in practice for reinforced concrete floor
slabs with high in-plane stiffnesses. It has been illustrated in Chapters 3 & 5 that the use of the
rigid-diaphragm constraint numerically suppresses the generation of internal (in-plane) forces and
deformation in the floor slabs which in turn leads to unconservative estimates of the seismic shear
demands on the tower walls to be made. For buildings featuring setbacks, the current requirements
in the PEER/ATC 72-1 (2010) and the TBI (2017) guidelines which are used for modelling the
backstay forces (as discussed in Section 2.2.2) will need to be expanded to include the tower
storeys above the podium interface level (or TFL). Particularly the requirement for the use of
semi-rigid diaphragm will need to be extrapolated to the two storeys above the level of the podium
(refer Section 3.3 for details). Similar requirements will need to be put forth for buildings
featuring transfer plates where such internal strutting forces in the floor slabs are anticipated.

7.6. Code classifications for buildings with setbacks

There is sufficient evidence to suggest that code classifications for vertical irregularities
targeting setbacks buildings need to be reviewed. For instance, results of the assessment study
reported in Chapter 6 have demonstrated striking differences in the response behaviour of building
model SB-1 and SB-2 despite that the buildings are clustered in the same category (of building
irregularity) in most design codes. Discrepancies in the response behaviour are primarily
associated with non-uniform distribution of inelastic demands up the height of the building and
the shear-critical conditions occurring above the interface level in SB-2 (i.e. low shear span ratios
and high shear force intensities). These conditions (shear-critical) have been attributed to slab-
wall interactions (above the level of the podium) as a result of interferences by the podium
structure. A possible enhancement would be to introduce a refined description for this form of
irregularity. As such, structural design engineers will be required to carry out additional shear
design checks for tower walls above the podium level (Section 7.4) in conjunction with the use
of the proposed design seismic storey shear envelope outlined in Section 7.2.

7.7. Performance levels for podium-tower buildings

Another outcome of the assessment study presented in Chapter 6 is the evaluation of the
performance levels that are applied to podium-tower buildings. A summary of these performance
limits has been presented in Table 6-5 which demonstrates low inter-storey drift ratios for
buildings that incorporate a transfer plate for the ‘collapse-prevention’ performance level. The

275
reported performance levels are lower than those outlined in the ASCE/SEI 41-13 (outlined in
Table 6-4) but are (generally) in good agreement with the limited number of studies that had been
reported in the literature (also outlined in Table 6-4). These reported discrepancies are believed
to be associated with the adverse increase in the shear and bending moment intensities in the
tower walls (beyond the values required for equilibrium) even at lower deformation demands.
Strictly speaking, these additional actions are not associated with substantial deformation
demands on the tower walls. In other words, shear strength in the walls might have been exceeded
early in the response even before the tower wall attains its nominal flexural capacity (as observed
in the case of building TS-2). Thus, a direct extrapolation of these ‘generic’ performance levels
(outlined in design codes) can lead to an overestimation of the damage state and the deformation
capacity of a building. Designers will need to be informed of such discrepancies when assessing
the seismic performance of building featuring transfer plates. It is however acknowledged that
more analyses/verifications will need to be put forth to ascertain such discrepancies.

276
 Chapter 8. CONCLUSIONS AND RECOMMENDATIONS FOR
FUTURE RESEARCH

8.1. Conclusions
The study presented in this thesis aimed to identify, and quantify, the adverse effects of the
interference by the podium structure on shear force intensities in the tower walls of a building.
The two types of podium-tower buildings examined in this study are buildings featuring a setback
in elevation and buildings featuring a transfer plate. These building types comprise the vast
majority of podium-tower type building constructions in many metropolitan cities around the
world. Differences between the two building forms are primarily attributed to the geometry and
configuration of the two portions of the building (referring to the podium and tower). Despite
these inherent differences, the effects of podium interferences when the building is subjected to
earthquake excitations have been shown to be somewhat similar. Practical design
recommendations based on the results and observations of this study have been presented in
Chapter 7. The following is a listing of the main findings reported in this thesis which are mapped
against the research objectives presented in Chapter 1.
1. Investigate Podium interferences in setback buildings
In buildings featuring setbacks, interferences by the podium structure have been found to be
most pronounced when the tower structure is not centred about the supporting podium (i.e.
positioned at an offset). In a building of this form, the podium has been shown to impose
differential restraint on the tower walls at and below the podium interface level. As such, interior
walls (closer to the centre of the podium) are subject to a larger restraint by the podium compared
to exterior walls. Differences in the boundary conditions imposed on the tower walls by the
podium (at and below the interface level) have in turn resulted in a considerable displacement
incompatibility (uneven wall displacement) in some tower storeys above the level of the interface.
In-plane restoring forces developed in the connecting links (floor slabs or beams) between the
tower walls have been observed in the storeys below and few storeys above the podium interface
level. These strutting forces contributed to the development of a much higher shear force
intensities in the interior tower wall (located closer to the centre of the podium) compared with
the shear force intensities in the exterior wall (further away from the centre of the podium). Shear
force anomalies in tower walls were most pronounced in buildings with podium storeys in the
lower 25%-35% of the total building height. The restraint by the podium on the tower walls was
found to be proportional to the relative lateral stiffness of the podium and the tower. Essentially,
a stiffer podium has been shown to impose higher (differential) restraint on the tower walls which
resulted in larger shear force redistribution between the connected walls (above the interface
level). Shortcomings of some classical assumptions adopted in design practices have been
highlighted particularly those pertaining to the modelling of floor slabs in buildings. The study

277
showed that shear demands in the tower walls (above the podium interface level) would have
been underestimated if the rigid-diaphragm constraints were to be employed in the modelling of
floor slabs in a building. Accordingly, the recommendations set forth by the PEER/ATC 72-1
guidelines which instruct engineers against the use of this constraint in the podium storeys will
need to be extended to include the two-storeys above the interface level where the in-plane
(strutting) forces in the floor slabs are to be anticipated. Moreover, strutting in-plane slab forces
resulted in lower shear span ratios (M/V) in the interior tower wall above the podium level.
Consequently, the occurrence of shear-critical conditions in the tower walls attributed to
significant reductions in the deformation capacity of the building in projected earthquake events.
The limited-ductile response behaviour of the building was further demonstrated by way of
detailed nonlinear FE analyses performed using the program VecTor2 where these (shear-critical)
conditions in the interior tower walls were explicitly observed despite the wall being presumed to
be governed by the (more favourable) flexural response behaviour.

2. Develop analytical modelling tools for quantifying the effects of the podium in setback
buildings

Analytical modelling approaches have been introduced for quantifying the magnitudes of
internal strutting forces in the link elements (slabs and beams) connecting tower walls in setback
buildings. A stiffness-based approach has been developed where estimates of the strutting forces
have been evaluated proportionately to the differences in wall deformation (between the interior
and exterior tower walls) and the in-plane stiffness of the connecting elements (slabs and beams)
and the tower walls. The relative contributions of each of these parameters have been developed
into a 2 x 2 matrix where increments in the wall shear forces were evaluated.

Modelling of the restraint by the podium and the consequential strutting forces were further
simplified into a practical desktop tool for use in a design office. In the developed model, the
podium-tower setback building has been analysed analogously to a cantilever beam that is
propped (pinned) at the level of the interface and fixed at the base. Accordingly, shear reversals
manifested at the level of the interface were directly estimated by evaluating the internal (reaction)
forces at intermediate support (the hypothetical pin). The shear reaction at the ‘pin’ support was
refined by incorporating the flexibility of the podium structure in the lower storeys. The
magnitude of internal forces in the two-three storeys above the level of the podium (where these
forces are expected to occur in the links between the walls) has been expressed as proportionately
to the reactive (backstay) interface force. A verification study has been conducted to showcase
the applicability of the proposed methods (both the explicit stiffness-based and the rapid
approaches) for estimating the internal forces in the slabs and thus the additional shear forces in
the tower walls above the podium.

278
 A study has been conducted to investigate the effects of in-plane yielding of the backstay slab
(which is the slab at the interface level) on the potential loss of the force transfer between the
tower and podium (by the slab) in seismic conditions. It has been shown that the interferences by
the podium (on the tower walls) can be significantly reduced when considerations are made for
the potential yielding and cracking of the interfacial podium slab. Shear force redistributions in
the tower walls were shown to reduce as a result of reductions (or softening) in the in-plane
stiffness of the interface slab. Similar effects have been explored for buildings with settlement (or
expansion) joints are proportioned at the junction between the podium and the tower.

3. Develop analytical modelling tools for quantifying the effects of the podium in transfer
structures

Interferences by the podium in transfer structures are primarily represented by the out-of-plane
bending of the transfer plate. The flexibility of the plate has been identified as the primary
contributor to displacement incompatibilities between tower walls planted on a transfer plate. A
correlation was established between the difference in the transfer plate rotations evaluated at the
base of the tower walls and in-plane strains that were developed in the floor slabs connecting the
tower walls above the level of the plate. This correlation has been demonstrated to be constantly
linear, where the slope of the correlation (defined by the Flexibility Index parameter) has been
identified to be proportional to the relative stiffness of the transfer plate and the supported tower
wall. As such, in-plane strains and forces developed in the floor slabs between the walls can be
estimated once the values of the relative transfer plate rotation and the flexibility index are
determined.

An analytical model has been developed for estimating the relative rotations of the plate when
the building is subjected to earthquake ground excitations. In the developed model, the local
distortions of the transfer plate (at the base of the walls) were correlated to the resulting rotations
(drifts) of the tower structure evaluated at the centre of mass (at the mid-height of the tower).
Using this correlation, these rotations (of the plate) were estimated by the peak rotation demand
(PRD) parameter. A dynamic rotational coupling model has been derived to estimate the value of
the parameter PRD. By making use of this modelling technique, the multitudes of degrees of
freedom of the building were reduced to only two, one translational, and one rotational. The
parameter br was introduced to quantify the relative stiffness of the rotational and translation
degrees of freedom of the building. The value of PRD has been demonstrated to be characterised
by the displacement-controlled behaviour where the value (of the PRD) has been parametrised
into a simple closed-form expression defined by the parameter br and RSDmax of the ground
motion. Strictly speaking, br quantifies the extent of interferences by the transfer plate on shear
force demands on the planted tower walls. Higher values of br were attributed to high rise building
with a thick transfer plate. The interferences by the transfer plate were found to be more

279
pronounced in medium-rise buildings (characterised by lower values of br). Additional shear
forces in the tower walls were simply computed as the product of the Flexibility Index, the PRD
and the in-plane properties of the connecting floor slab. The robustness of the developed analytical
model in quantifying the complex interactions by the transfer plate has been verified by way of
detailed analyses on two case study buildings of different heights.

4. Assess the seismic performance and the response modification factors of podium-tower
buildings

Assessment of the seismic response modification factors (ductility and overstrength) has been
conducted to demonstrate the appropriateness of use of the code specified values in the seismic
design of a podium-tower building. A framework has been developed for quantifying these factors
taking into account the interferences by the podium on the response behaviour of walls above the
interface level. The limited-ductile response behaviour of the building had been confirmed by a
series of quasi-static and nonlinear time history analyses on sub-frame building models
representing different podium-tower building configurations. The results have shown that the
code specified force reduction factors (the product of the ductility and the overstrength factor) are
generally adequate for the seismic design of podium-tower buildings featuring setbacks. The
conservatism of code specified force reduction values was slightly reduced for buildings
incorporating a transfer plate. Despite this, ductility factors reported in this study for buildings
featuring a transfer plate were lower than recommended values in some design standards (e.g.
AS1170.4:2007). The poor deformation capacity of podium-tower buildings has been attributed
to the unfavourable interferences by the podium structure leading to limited-ductile mechanisms
forming in the storeys above the podium interface level. It has been shown that significant
concentrations of inelastic demands in the tower storeys (above the interface) can reduce the
global deformation capacity of the building when subjected to rare or very rare earthquake events.
The assessment involved the use of a large number of nonlinear time history analyses to evaluate
three distinct performance levels (damage limit states) for podium-tower buildings (IO, LS, and
CP). Results of the study demonstrated that the seismic performance of the building can vary
quite significantly depending on the extent of interferences by the podium structure on the
response behaviour of the building. The inter-storey drift limits (ISD) for the three performance
levels were lower than thee limits recommended by the ASCE/SEI 41-13.

For setback buildings, the positioning of the tower with respect to the podium structure
(centred vs. positioned at an offset) has shown to have significant effects on the seismic
performance behaviour of the building. Differences in the seismic performance (of the two
building configurations) were in the order of 20.0% -30.0% (in terms of the sustained ISD values
at the collapse prevention performance level). The poorer performance (lower ISD values) was
reported for buildings where the tower is positioned at an offset from the centre of the podium.

280
For buildings with a transfer plate, lower inter-storey drift limits have also been reported
particularly for the CP and LS performance levels compared to the drift limits listed in ASCE/SEI
41-13. The gradation of the seismic performance of the buildings investigated in this study placed
buildings with a transfer plate in the lower ranks (i.e. poor performance). The vulnerability
assessment (the developed fragility curves) further highlighted the described “ranking” of the
podium-tower building in terms of their respective seismic performance where higher exceedance
probabilities (i.e. fragilities) were computed for podium-tower buildings featuring a transfer plate.

8.2. Limitations and recommendations for future studies

The study investigated podium-tower buildings of two distinct configurations (i.e. featuring a
setback and a transfer plate). Only buildings with a single setback (i.e. a distinct podium interface
level) have been considered. Analytical models for quantifying the effects of setbacks on the shear
force intensities in the tower walls have also been derived assuming the shear reversal is
concentrated at a single level. It is possible, however, that the shear force reversal occurs at
various levels up the height of the building (e.g. when a building features setbacks at different
storeys).

The main thrust of this thesis was on quantifying the effects and magnitudes of the internal
strutting forces in the floor slabs between the tower walls. The capacity of the slab (or the beam)
in transferring these in-plane forces has not been explicitly examined. The accurate determination
of the in-plane strength and deformation of a floor slab element is particularly important for
modelling the backstay slab (at the interface level) where high transfer forces are expected to
develop. Nevertheless, the relatively high in-plane stiffness and strength of conventional
reinforced concrete floor slabs may well be adequate for transferring these forces even if they
have not been explicitly designed to resist such in-plane actions.

In Chapter 4, in-plane yielding mechanisms of the main backstay slab (slab at the interface
level) have been assumed based on some physical manifestations that can potentially occur at the
interfacial zone between the podium and the tower (e.g. cracking at construction joints or yielding
of dowel bars). Such assumptions may have been warranted given the lack of experimental
verifications reported in the literature (on the response behaviour of a wall-slab connection). This
is particularly crucial when construction joints are present at the interfacial zone between the
tower and the podium as a result of the sequential construction of the building (i.e. erecting the
tower walls and cores first and then ‘stitching’ the slabs at a later construction cycle). Further
studies are needed to investigate and quantify the load transfer mechanism at slab-wall
connections.

In most analytical simulations presented in this study, the floor slabs have been modelled as
equivalent frame elements with a specified geometry. Effective widths of the floor slabs have

281
been determined based on some formulations available in the literature (summarised in Chapter
2) which were derived from experimental studies on column-slab assemblies. The adequacies of
use of these formulations to model floor slabs between walls have been verified (for some cases
in this study) by the use of FE numerical models of the building (e.g. by comparing to the width
of the column strip of the slab between the walls). It is preferable (and useful) to have a more
concise formulation for determining the effective width in wall-slab assemblies to assist engineers
and researchers in accurately modelling slabs in numerical simulations.

In the assessment study reported in Chapter 6, seismic damage of the transfer plate was limited
to the flexural deformations while no considerations were made for modelling punching shear
actions in the plate. Justifications for this modelling approach have been discussed in Chapter 6
(in Section 6.4.3). However, there is a need for conducting further verifications for the shear
response behaviour of transfer plates with a thickness of less than 1.5m (used in medium-rise
constructions). Furthermore, the definition of the CP performance level is, in some cases, overly
conservative. Notably, the definition of the limiting compressive strain value (-0.003) is lower
than what is generally adopted for assessment purposes. This is partly owed to the limited
investigations on structural walls with non-seismic detailing (e.g. without sufficient ties at the end
regions). More verification studies will need to be addressed (possibly through experiments)
before higher compressive strain values can be used for walls with similar detailing. The scope
of the study presented in Chapter 6 has been restricted to buildings with a height of 70m. Similar
studies will need to be conducted on taller buildings with different configurations. It should be
noted that for the case of the buildings featuring a transfer plate, the analytical model presented
in Chapter 5 suggests that the interferences by the transfer plate reduce as the thickness of the
plate increases (higher values of br). Accordingly, shear force anomalies in the tower walls are
less critical in taller buildings that incorporate a deeper plate. The assessment study involved the
use of a large number of ground motion records that characterise (on average) Australian seismic
conditions. Variations in seismic scenarios have not been considered in the context of the study
presented in Chapter 6 (i.e. different site classes or far-field and near-field ground motion
records). Such studies are important for quantifying the vulnerability of a building at a given site
or when the effects of the ground motion characteristics on the response behaviour of a building
are investigated.

282
 REFERENCES
Abdelbasset, Y. M., Sayed-Ahmed, E. Y., & Mourad, S. A. (2014). High-Rise Buildings with
Transfer Floors: Drift Calculations. Paper presented at the IABSE Symposium Report.
American Concrete Institute. (2014). ACI 318-14: Building Code Requirements for Structural
Concrete and Commentary: Farmington Hills.
Al Abadi, H., Lam, N., & Gad, E. (2006). A simple displacement-based model for predicting
seismically induced overturning. Journal of Earthquake Engineering, 10(06), 775-814.
Almeida, J. P., Tarquini, D., & Beyer, K. (2016). Modelling approaches for inelastic behaviour
of RC walls: Multi-level assessment and dependability of results. Archives of
Computational Methods in Engineering, 23(1), 69-100.
Alwaeli, W., Mwafy, A., Pilakoutas, K., & Guadagnini, M. (2017). A methodology for defining
seismic scenario‐structure‐based limit state criteria for rc high‐rise wall buildings using
net drift. Earthquake Engineering & Structural Dynamics, 46(8), 1325-1344.
Amirsardari, A., Goldsworthy, H. M., & Lumantarna, E. (2017). Seismic Site Response Analysis
Leading to Revised Design Response Spectra for Australia. Journal of Earthquake
Engineering, 21(6), 861-890.
American Society of Civil Engineers (ASCE/SEI). (2010). ASCE/SEI 7-10, Minimum design
loads for buildings and other structures. Reston, Virginia: American Society of Civil
Engineers Standard.
American Society of Civil Engineers (ASCE/SEI). (2013). ASCE 41-13: Seismic evaluation and
retrofit of existing buildings. Reston, Virginia: American Society of Civil Engineers
Standard Reston, Virginia.
Aslani, H., & Miranda, E. (2005). Probabilistic earthquake loss estimation and loss
disaggregation in buildings. Stanford University Stanford, CA.
Atkinson, G. M. (1993). Earthquake source spectra in eastern North America. Bulletin of the
Seismological Society of America, 83(6), 1778-1798.
Baker, J. W. (2015). Efficient analytical fragility function fitting using dynamic structural
analysis. Earthquake Spectra, 31(1), 579-599.
Bentz, E. C., & Collins, M. (2000). Response 2000 manual. Retrieved from:
http://www.ecf.utoronto.ca/~bentz/r2k.htm
Bevan-Pritchard, G., Man, E., & Anderson, D. (1983). Force distribution between core and sub-
grade structure of high-rise buildings subjected to lateral load induced forces. Paper
presented at the Proceeding of 4th Canadian conference on Earthquake Engineering,
Vancouver.
Beyer, K. (2005). Design and analysis of walls coupled by floor diaphragms. M. Sc. thesis. Rose
School, University of Pavia, Italy.
Beyer, K., Dazio, A., & Priestley, N. (2008). Inelastic wide-column models for U-shaped
reinforced concrete walls. Journal of Earthquake Engineering, 12(S1), 1-33.
Beyer, K., Dazio, A., & Priestley, N. (2008). Seismic design of torsionally eccentric buildings
with U-shaped RC walls. Research report ROSE – 2008/03. Pavia (Italy): IUSS Press.
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.
Beyer, K., Priestley, M., Calvi, G., & Pinho, R. (2005). Design and analysis of walls coupled by
floor diaphragms. M. Sc. thesis. Rose School, University of Pavia, Italy.

283
Beyer, K., Simonini, S., Constantin, R., & Rutenberg, A. (2014). Seismic shear distribution
among interconnected cantilever walls of different lengths. Earthquake Engineering &
Structural Dynamics, 43(10), 1423-1441.
Bradley, B. A., & Dhakal, R. P. (2008). Error estimation of closed‐form solution for annual rate
of structural collapse. Earthquake Engineering & Structural Dynamics, 37(15), 1721-
1737.
Bradley, B. A., Dhakal, R. P., Mander, J. B., & Li, L. (2008). Experimental multi‐level seismic
performance assessment of 3D RC frame designed for damage avoidance. Earthquake
Engineering & Structural Dynamics, 37(1), 1-20.
Brown, A., & Gibson, G. (2004). A multi-tiered earthquake hazard model for Australia.
Tectonophysics, 390(1-4), 25-43.
Brown, P. C., & Lowes, L. N. (2007). Fragility functions for modern reinforced-concrete beam-
column joints. Earthquake Spectra, 23(2), 263-289.
Bull, D. K. (2004). Understanding the complexities of designing diaphragms in buildings for
earthquakes. Bulletin of the New Zealand Society for Earthquake Engineering, 37(2), 70-
88.
Burbidge, D., Leonard, M., Robinson, D., & Gray, D. (2012). The 2012 Australian earthquake
hazard map. Geoscience Australia, Record, 71.
CEN. (2004). Eurocode-8: Design of structures for earthquake resistance Part 1: General rules,
seismic actions, and rules for buildings (Vol. 2, pp. 1998-1991: 2004). Brussels: CEN,
European Committee for Standardization.
Chintanapakdee, C., & Chopra, A. K. (2004a). Evaluation of modal pushover analysis using
vertically irregular frames. Paper presented at the Proceedings of the 13th World
conference on earthquake engineering, CD ROM. Vancouver.
Chintanapakdee, C., & Chopra, A. K. (2004b). Seismic response of vertically irregular frames:
response history and modal pushover analyses. Journal of Structural Engineering,
130(8), 1177-1185.
Chopra, A. K. (1995). Dynamics of structures (Vol. 3): Prentice Hall New Jersey.
Chopra, A. K., & Goel, R. K. (2001). A modal pushover analysis procedure to estimate seismic
demands for buildings: theory and preliminary evaluation. Civil and Environmental
Engineering, 55.
Clough, D. P. (1982). Considerations in the design and construction of precast concrete
diaphragms for earthquake loads.
Coleman, J., & Spacone, E. (2001). Localization issues in force-based frame elements. Journal
of Structural Engineering, 127(11), 1257-1265.
Collins, M. P. (1978). Towards a rational theory for RC members in shear. Journal of the
Structural Division, 104(4), 649-666.
Committee, S. V. (1995). Performance-based seismic engineering. Structural Engineers
Association of California, Sacramento, California.
Council), A. A. T. (1995). ATC-19,Structural Response Modification Factors. Retrieved from
Redwood City,California:
CSI, C. (2004). analysis reference manual for SAP2000, ETABS, and SAFE. Computers and
Structures, Inc., California, USA.
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.
Dovich, L. M., & Wight, J. K. (2005). Effective slab width model for seismic analysis of flat slab

284
 frames. ACI Structural Journal, 102(6), 868.
Elawady, A., Okail, H., Abdelrahman, A., & Sayed-Ahmed, E. (2014). Seismic Behaviour of
High-Rise Buildings with Transfer Floors. Electronic Journal of Structural Engineering,
14(2), 57-70.
Elnashai, A., & Mwafy, A. (2002). Overstrength and force reduction factors of multistorey
reinforced‐concrete buildings. The structural design of tall buildings, 11(5), 329-351.
Fajfar, P. (2000). A nonlinear analysis method for performance-based seismic design. Earthquake
Spectra, 16(3), 573-592.
Fanaie, N., & Shamlou, S. O. (2015). Response modification factor of mixed structures. Steel and
Composite Structures, 19(6), 1449-1466.
FEMA. (2000). FEMA 356, Prestandard and commentary for the seismic rehabilitation of
buildings. Washington, DC: FEMA.
FEMA. (2000). FEMA-356 Commentary for the Seismic Rehabilitation of Buildings. Federal
Emergency Management Agency, Washington, DC.
FEMA. (2005). FEMA 440, Improvement of nonlinear static seismic analysis procedures. FEMA-
440, Redwood City.
FEMA. (2009). FEMA P695. Quantification of Building Seismic Performance Factors Federal
Emergency Management Agency.
FEMA. (2009). FEMA P-750. NEHRP recommended seismic provisions for new buildings and
other structures. Washington, DC (USA): Building Seismic Safety Council.
Gardiner, D. R. (2011). Design recommendations and methods for reinforced concrete floor
diaphragms subjected to seismic forces.
Gérin, M., & Adebar, P. (2009). Simple rational model for reinforced concrete subjected to
seismic shear. Journal of Structural Engineering, 135(7), 753-761.
Greifenhagen, C. (2006). Seismic behavior of lightly reinforced concrete squat shear walls. Ph.D.
Thesis, École Polytechnique Fédérale De Lausanne, Switzerland.
Greifenhagen, C., & Lestuzzi, P. (2005). Static cyclic tests on lightly reinforced concrete shear
walls. Engineering Structures, 27(11), 1703-1712.
Grossman, J. S. (1997). Verification of proposed design methodologies for effective width of
slabs in slab-column frames. ACI Structural Journal, 94(2), 181-196.
Gulec, C. K., & Whittaker, A. S. (2009). Performance-based assessment and design of squat
reinforced concrete shear walls: Multidisciplinary Center for Earthquake Engineering
Research MCEER.
Gulec, C. K., Whittaker, A. S., & Stojadinovic, B. (2008). Shear strength of squat rectangular
reinforced concrete walls. ACI Structural Journal, 105(4).
Habibullah, A. (1997). ETABS-Three Dimensional Analysis of Building Systems, Users Manual.
Computers and Structures Inc., Berkeley, California.
Hashash, Y., Musgrove, M., Harmon, J., Groholski, D., Phillips, C., & Park, D. (2015).
DEEPSOIL 6.1, User manual. Board of Trustees of University of Illinois at Urbana-
Champaign, Urbana.
Henry, R. (2011). Self-centering precast concrete walls for buildings in regions with low to high
seismicity. ResearchSpace@ Auckland.
Hines, E. M., Restrepo, J. I., & Seible, F. (2004). Force-displacement characterization of well-
confined bridge piers. ACI Structural Journal, 101(4), 537-548.
Hoult, R. (2017). Seismic assessment of reinforced concrete walls in Australia. PhD Thesis,

285
 Department of Infrastructure Engineering. The University of Melbourne, Melbourne,
Australia.
Hoult, R., Goldsworthy, H., & Lumantarna, E. (2017). Plastic Hinge Length for Lightly
Reinforced Rectangular Concrete Walls. Journal of Earthquake Engineering, 1-32.
Hutchinson, G., Lam, N., & Wilson, J. L. (2003). Determination of earthquake loading and
seismic performance in intraplate regions. Progress in Structural Engineering and
Materials, 5(3), 181-194.
Hwang, S.-J., & Moehle, J. P. (2000). Models for laterally loaded slab-column frames. ACI
Structural Journal, 97(2).
Ibarra, L. F., & Krawinkler, H. (2005). Global collapse of frame structures under seismic
excitations: Pacific Earthquake Engineering Research Center Berkeley, CA.
Ibrahimbegovic, A., & Wilson, E. L. (1991). A unified formulation for triangular and quadrilateral
flat shell finite elements with six nodal degrees of freedom. International Journal for
Numerical Methods in Biomedical Engineering, 7(1), 1-9.
Jalayer, F. (2003). Direct probabilistic seismic anaysis: implementing non-linear dynamic
assessments. PhD Dissertation, Stanford University.
Ji, J., Elnashai, A. S., & Kuchma, D. A. (2007). An analytical framework for seismic fragility
analysis of RC high-rise buildings. Engineering Structures, 29(12), 3197-3209.
Kafle, B., Lam, N., Gad, E. F., & Wilson, J. (2011). Displacement controlled rocking behaviour
of rigid objects. Earthquake Engineering & Structural Dynamics, 40(15), 1653-1669.
Kayen, R. E., Carkin, B. A., Allen, T., Collins, C., McPherson, A., & Minasian, D. L. (2015).
Shear-wave velocity and site-amplification factors for 50 Australian sites determined by
the spectral analysis of surface waves method: U.S. Geological Survey Open-File Report
2014-1264. doi: http:/dx.doi.org/10.3133/ofr20141264
Kazaz, İ., Gülkan, P., & Yakut, A. (2012). Deformation limits for structural walls with confined
boundaries. Earthquake Spectra, 28(3), 1019-1046.
Kelly, T. (2004). Nonlinear analysis of reinforced concrete shear wall structures. Bulletin of the
New Zealand Society for Earthquake Engineering, 37(4), 156-180.
Khalifa, S. (2015). Assessment of Multi-Story Building Seismic Design Factors with Structural
Irregularity, MSc Thesis, United Arab Emirates University, UAE.
Kim, J., & Choi, H. (2005). Response modification factors of chevron-braced frames.
Engineering Structures, 27(2), 285-300.
Kowalsky, M. J., & Priestley, M. N. (2000). Improved analytical model for shear strength of
circular reinforced concrete columns in seismic regions. ACI Structural Journal, 97(3).
Krolicki, J., Maffei, J., & Calvi, G. (2011). Shear strength of reinforced concrete walls subjected
to cyclic loading. Journal of Earthquake Engineering, 15(S1), 30-71.
Kuang, J. S., & Li, S. (2001). Interaction‐based design tables for transfer beams supporting in‐
plane loaded shear walls. The Structural Design of Tall and Special Buildings, 10(2), 121-
133.
Lam, N., Wilson, J., & Hutchinson, G. (1996). Building ductility demand: Interplate versus
intraplate earthquakes. Earthquake Engineering & Structural Dynamics, 25(9), 965-985.
Lam, N., Wilson, J., & Hutchinson, G. (1998). The ductility reduction factor in the seismic design
of buildings. Earthquake Engineering & Structural Dynamics, 27(7), 749-769.
Lam, N. (1999). Program “GENQKE” User’s Guide‐Program for generating synthetic earthquake
accelerograms based on stochastic simulations of seismological models. Department of
Civil and Environmental Engineering, The University of Melbourne, Australia.

286
Lam, N., Wilson, J., Chandler, A., & Hutchinson, G. (2000a). Response spectral relationships for
rock sites derived from the component attenuation model. Earthquake Engineering &
Structural Dynamics, 29(10), 1457-1489.
Lam, N., Wilson, J., Chandler, A., & Hutchinson, G. (2000b). Response spectrum modelling for
rock sites in low and moderate seismicity regions combining velocity, displacement and
acceleration predictions. Earthquake Engineering & Structural Dynamics, 29(10), 1491-
1525.
Lam, N. , Tsang, H.H., Lumantarna, E., & Wilson, J. L. (2016). Minimum loading requirements
for areas of low seismicity. Earthquakes and Structures, 11(4), 539.
Lam, N., Lumantarna, E., & Wilson, J. (2016). Simplified elastic design checks for torsionally
balanced and unbalanced low-medium rise buildings in lower seismicity regions.
Earthquake and Structures, 741-777. doi:10.12989/eas.2016.11.5.741
LATBSDC. (2005). An Alternative Procedure for Seismic Analysis and Design of Tall Buildings
Located in the LA Region. LA Tall Buildings Structural Design Council.
Lee, H.-S., & Ko, D.-W. (2004). Seismic response of high-rise RC bearing-wall structures with
irregularities at bottom stories. Paper presented at the Proceedings of the 13th World
Conference on Earthquake Engineering, Vancouver, Canada.
Lee, H.-S., & Ko, D.-W. (2007). Seismic response characteristics of high-rise RC wall buildings
having different irregularities in lower stories. Engineering Structures, 29(11), 3149-
3167.
Li, B., Pan, Z., & Xiang, W. (2015). Experimental Evaluation of Seismic Performance of Squat
RC Structural Walls with Limited Ductility Reinforcing Details. Journal of Earthquake
Engineering, 19(2), 313-331.
Li, C. S. (2005). Response of transfer plate when subjected to earthquake. (PhD Thesis), The
Hong Kong Polytechnic University.
Li, C. S., Lam, S., Chen, A., & Wong, Y. (2008). Seismic Performance of a Transfer Plate
Structure. Journal of Structural Engineering, 134(11), 1705-1716.
Li, C. S., Lam, S., Zhang, M., & Wong, Y. (2006). Shaking table test of a 1: 20 scale high-rise
building with a transfer plate system. Journal of Structural Engineering, 132(11), 1732-
1744.
Looi, D. T. W., Su, R. K. L., Cheng, B., & Tsang, H. H. (2017). Effects of axial load on seismic
performance of reinforced concrete walls with short shear span. Engineering Structures,
151, 312-326. doi:https://doi.org/10.1016/j.engstruct.2017.08.030
Lu, X., Su, N., & Zhou, Y. (2013). Nonlinear time history analysis of a super‐tall building with
setbacks in elevation. The Structural Design of Tall and Special Buildings, 22(7), 593-
614.
Luco, N., & Cornell, C. A. (1998). Effects of random connection fractures on the demands and
reliability for a 3-story pre-Northridge SMRF structure. Paper presented at the
Proceedings of the 6th US national conference on earthquake engineering.
Lumantarna, E., Lam, N., & Wilson, J. (2013). Displacement-controlled behavior of
asymmetrical single-story building models. Journal of Earthquake Engineering, 17(6),
902-917.
Lumantarna, E., Lam, N., Wilson, J., & Griffith, M. (2010). Inelastic displacement demand of
strength-degraded structures. Journal of Earthquake Engineering, 14(4), 487-511.
Lumantarna, E., Wilson, J. L., & Lam, N. (2012). Bi-linear displacement response spectrum
model for engineering applications in low and moderate seismicity regions. Soil
Dynamics and Earthquake Engineering, 43, 85-96.

287
Luo, Y., & Durrani, A. (1995a). Equivalent Beam Model for Flat-Slab Buildings--Part II: Exterior
Connections. Structural Journal, 92(2), 250-257.
Luo, Y., & Durrani, A. (1995b). Equivalent beam model for flat-slab buildings: part I: interior
connections. Structural Journal, 92(1), 115-124.
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. (2006). RC wall shear–flexure interaction: Analytical and experimental responses.
PhD Dissertation.
McCue, K., Gibson, G., Love, D., & Cranswick, E. (2008). Engineering for earthquakes in
Australia. Australian Journal of Structural Engineering, 8(1), 1-12.
Melek, M., Darama, H., Gogus, A., & Kang, T. (2012). Effects of Modeling of RC Flat Slabs on
Nonlinear Response of High Rise Building Systems. Paper presented at the 15th World
Conference in Earthquake Engineering.
Menegon, S., Wilson, J., Lam, N., & Gad, E. (2017). RC walls in Australia: reconnaissance survey
of industry and literature review of experimental testing. Australian Journal of Structural
Engineering, 18(1), 24-40.
Menegotto, M., & Pinto, E. (1973). Method of Analysis for Cyclically Loaded Reinforced
Concrete Plane Frames Including Changes in Geometry and Non-elastic Behavior of
Elements Under Combined Normal Force and Bending. Paper presented at the IABSE
Symposium on the Resistance and Ultimate Deformability of Structures Acted on by
Well-Defined Repeated Loads, Lisbon.
Mergos, P., & Beyer, K. (2014). Modelling shear–flexure interaction in equivalent frame models
of slender reinforced concrete walls. The Structural Design of Tall and Special Buildings,
23(15), 1171-1189.
Mitchell, D., & Collins, M. P. (1974). Diagonal compression field theory-a rational model for
structural concrete in pure torsion. Paper presented at the ACI Journal Proceedings.
Moehle, J. P., & Diebold, J. W. (1985). Lateral load response of flat-plate frame. Journal of
Structural Engineering, 111(10), 2149-2164.
Moehle, J. P., Hooper, J. D., Kelly, D. J., & Meyer, R. (2010). Seismic design of cast-in-place
concrete diaphragms, chords, and collectors: A guide for practicing engineers. NEHRP
seismic design technical brief(3).
Mwafy, A. (2011). Assessment of seismic design response factors of concrete wall buildings.
Earthquake Engineering and Engineering Vibration, 10(1), 115-127.
Mwafy, A. (2013). Use of overstrength and inelastic response in seismic design. Proceedings of
the Institution of Civil Engineers-Structures and Buildings, 166(6), 282-297.
Mwafy, A., & Elkholy, S. (2017). Performance assessment and prioritization of mitigation
approaches for pre-seismic code structures. Advances in Structural Engineering, 20(6),
917-939.
Mwafy, A., & Elnashai, A. (2001). Static pushover versus dynamic collapse analysis of RC
buildings. Engineering Structures, 23(5), 407-424.
Mwafy, A., & Elnashai, A. (2002). Calibration of force reduction factors of RC buildings. Journal
of Earthquake Engineering, 6(02), 239-273.
Mwafy, A., & Khalifa, S. (2017). Effect of vertical structural irregularity on seismic design of tall
buildings. The Structural Design of Tall and Special Buildings.
Nakaki, S. D. (2000). Design guidelines for precast and cast-in-place concrete diaphragms:
Earthquake Engineering Research Institute.

288
Newmark, N. M., & Hall, W. J. (1982). Earthquake spectra and design. Earth System Dynamics.
Orakcal, K., Massone, L. M., & Wallace, J. W. (2009). Shear strength of lightly reinforced wall
piers and spandrels. ACI Structural Journal, 106(4), 455.
Orakcal, K., & Wallace, J. W. (2004). Nonlinear modeling and analysis of slender reinforced
concrete walls. ACI Structural Journal(5), 688-698.
Otani, S. (1981). Hysteresis models of reinforced concrete for earthquake response analysis.
Journal of Faculty of Engineering, 36(2), 407-441.
Ozcebe, G., & Saatcioglu, M. (1989). Hysteretic shear model for reinforced concrete members.
Journal of Structural Engineering, 115(1), 132-148.
Park, R., & Paulay, T. (1975). Reinforced Concrete Structures: Wiley.
Paulay, T., & Priestley, J. N. (1992). Seismic Design of Reinforced Concrete and Masonry
Buildings: Wiley.
Pecknold, D. A. (1975). Slab effective width for eqivalent frame analysis. Paper presented at the
Journal Proceedings.
PEER. (2013). Pacific earthquake engineering research (PEER) Center Ground Motion Database.
Retrieved from http://peer.berkeley.edu/peer_ground_motion_database/
PEER/ATC72-1. (2010). Modeling and acceptance criteria for seismic design and analysis of tall
buildings. Redwood City, CA: Applied Technology Council in cooperation with the
Pacific Earthquake Engineering Research Center.
Pinho, R., & Antoniou, S. (2005). A displacement-based adaptive pushover algorithm for
assessment of vertically irregular frames. Paper presented at the Proceedings of the
Fourth European Workshop on the Seismic Behaviour of Irregular and Complex
Structures.
Pinho, R., Bhatt, C., Antoniou, S., & Bento, R. (2008). Modelling of the horizontal slab of a 3D
irregular building for nonlinear static assessment. Paper presented at the The 14th World
Conference on Earthquake Engineering, Beijing, China.
Porter, K., Kennedy, R., & Bachman, R. (2007). Creating fragility functions for performance-
based earthquake engineering. Earthquake Spectra, 23(2), 471-489.
Priestley, J. N., Calvi, G. M., & Kowalsky, M. J. (2007). Displacement-based Seismic Design of
Structures: IUSS Press.
Priestley, M. (1997). Displacement-based seismic assessment of reinforced concrete buildings.
Journal of Earthquake Engineering, 1(01), 157-192.
Priestley, M., & Grant, D. (2005). Viscous damping in seismic design and analysis. Journal of
Earthquake Engineering, 9(spec02), 229-255.
Priestley, M. N. (1993). Myths and fallacies in earthquake engineering-conflicts between design
and reality. Bulletin of the New Zealand National Society for Earthquake Engineering,
26(3), 329-341.
Priestley, M. N., Verma, R., & Xiao, Y. (1994). Seismic shear strength of reinforced concrete
columns. Journal of Structural Engineering, 120(8), 2310-2329.
Pugh, J. S., Lowes, L. N., & Lehman, D. E. (2015). Nonlinear line-element modeling of flexural
reinforced concrete walls. Engineering Structures, 104, 174-192.
Qian, C.-g., & Wang, W. (2006). Effect of the thickness of transfer slab on seismic behavior of
tall building structure. Optimization of Capital Construction, 27(4), 98-100.
Rad, B., & Adebar, P. (2009). Seismic Design of High-Rise Concrete Walls: Reverse Shear due
to Diaphragms below Flexural Hinge. Journal of Structural Engineering, 135(8), 916-

289
 924. doi:10.1061/(ASCE)0733-9445(2009)135:8(916)
Robertson, I. N. (1997). Analysis of flat slab structures subjected to combined lateral and gravity
loads. Structural Journal, 94(6), 723-729.
Rutenberg, A. (2004). The seismic shear of ductile cantilever wall systems in multistorey
structures. Earthquake Engineering & Structural Dynamics, 33(7), 881-896.
Rutenberg, A., & Nsieri, E. (2006). The seismic shear demand in ductile cantilever wall systems
and the EC8 provisions. BULLETIN OF EARTHQUAKE ENGINEERING, 4(1), 1-21.
Salonikios, T. N., Kappos, A. J., Tegos, I. A., & Penelis, G. G. (1999). Cyclic load behavior of
low-slenderness reinforced concrete walls: Design basis and test results. ACI Structural
Journal, 96(4).
Sedgh, R., Dhakal, R., & Carr, A. (2015). State of the Art: Challenges in analytical modelling of
multi-storey shear wall buildings. Paper presented at the New Zealand Society for
Earthquake Engineering Annual Conference (NZSEE2015), Rotorua, New Zealand, pO-
15.
Seismosoft. (2012). SeismoArtif. Pavia, Italy. Retrived from www.seismosoft.com.
SeismoSoft. (2016). A computer program for static and dynamic nonlinear analysis of framed
structures. Retrived from www.seismosoft.com.
Shahi, R., Lam, N., Gad, E., Wilson, J., & Watson, K. (2017). Seismic performance behavior of
cold-formed steel wall panels by quasi-static tests and incremental dynamic analyses.
Journal of Earthquake Engineering, 21(3), 411-438.
Shahrooz, B. M., & Moehle, J. P. (1987). Experimental study of seismic response of RC setback
buildings: Earthquake Engineering Research Center, College of Engineering, University
of California Springfield, VA.
Shahrooz, B. M., & Moehle, J. P. (1990a). Evaluation of seismic performance of reinforced
concrete frames. Journal of Structural Engineering, 116(5), 1403-1422.
Shahrooz, B. M., & Moehle, J. P. (1990b). Seismic response and design of setback buildings.
Journal of Structural Engineering, 116(5), 1423-1439.
Shin, M., Kang, T. H. K., & Grossman, J. S. (2010). Practical modelling of high-rise dual systems
with reinforced concrete slab-column frames. The Structural Design of Tall and Special
Buildings, 19(7), 728-749. doi:10.1002/tal.509
Sivaselvan, M. V., & Reinhorn, A. M. (2000). Hysteretic models for deteriorating inelastic
structures. Journal of Engineering Mechanics, 126(6), 633-640.
Spacone, E., Filippou, F. C., & Taucer, F. F. (1996a). Fibre beam-column model for non-linear
analysis of R/C frames: Part I. Formulation. Earthquake engineering and structural
dynamics, 25(7), 711-726.
Spacone, E., Filippou, F., & Taucer, F. F. (1996b). Fibre beam-column model for non-linear
analysis of R/C frames: part II. Applications. Earthquake engineering and structural
dynamics, 25(7), 727-742.
Spence, S. M., & Kareem, A. (2013). Tall buildings and damping: a concept-based data-driven
model. Journal of Structural Engineering, 140(5), 04014005.
Stafford-Smith, B. (1986). Deficiencies in the wide column analogy for shearwall core analysis.
Concrete International, 8(4), 58-61.
Standards Australia. (1993). AS1170.4 Minimum design loads on structures- Part 4: Earthquake
Loads. Sydney, NSW.
Standards Australia. (2002). AS/NZ 1170.0:2002, Structural design actions - Part 0: General
principals. (2002). Sydney, NSW 2001, Australia.

290
Standards Australia. (2007). AS 1170.4 Structural Design Actions, Part 4: Earthquake actions in
Australia. Sydney, NSW.
Standards Australia. (2009). AS 3600-2009: Concrete Structures.
Su, R.K.L. (2008). Seismic behaviour of buildings with transfer structures in low-to-moderate
seismicity regions. Department of Civil Engineering, The University of Hong Kong, Hong
Kong, China.
Su, R.K.L., Chandler, A., Li, J., & Lam, N. (2002). Seismic assessment of transfer plate high rise
buildings. Structural Engineering and Mechanics, 14(3), 287-306.
Su, R.K.L., & Cheng, M. (2009). Earthquake‐induced shear concentration in shear walls above
transfer structures. The Structural Design of Tall and Special Buildings, 18(6), 657-671.
Su, R.K.L., Tsang, H., & Lam, N. (2011). Seismic design of buildings for Hong Kong conditions:
Hong Kong Institution of Engineers.
Su, R.K.L., & Wong, S. (2007). Seismic behaviour of slender reinforced concrete shear walls
under high axial load ratio. Engineering Structures, 29(8), 1957-1965.
Su, R.K.L., Chandler, A. M., Li, J. H., & Lam, N. (2002). Seismic assessment of transfer plate
high rise buildings. Structural Engineering and Mechanics, 14(3), 287.
Sullivan, T. J., Calvi, G. M., & Priestley, M. N. (2006). Seismic design of frame-wall structures:
IUSS Press.
Tang, T. O., & Su, R.K.L. (2014). Gravity-induced shear force in reinforced concrete walls above
transfer structures.
Tang, T. O., & Su, R.K.L. (2015). Gravity-induced shear force in reinforced concrete walls above
transfer structures. Proceedings of the Institution of Civil Engineers: Structures and
Buildings.
Taucer, F., Spacone, E., & Filippou, F. C. (1991). A fiber beam-column element for seismic
response analysis of reinforced concrete structures (Vol. 91): Earthquake Engineering
Research Center, College of Engineering, University of California Berkekey, California.
TBI. (2017). Guidelines for Performance-Based Seismic Design of Tall Buildings. Redwood City,
CA: Tall Buildings Initative, Pacific Earthquake Engineering Research Center.
Tsang, H.H. (2011). Should we design buildings for lower-probability earthquake motion?
Natural hazards, 58(3), 853-857.
Tsang, H.H., Wilson, J., Lam, N., & Su, R.K.L. (2017). A design spectrum model for flexible soil
sites in regions of low-to-moderate seismicity. Soil Dynamics and Earthquake
Engineering, 92, 36-45.
UBC, (1997). ‘Uniform building code. Paper presented at the Int. Conf. Building Officials,
Whittier, Califronia.
Vamvatsikos, D., & Allin Cornell, C. (2006). Direct estimation of the seismic demand and
capacity of oscillators with multi‐linear static pushovers through IDA. Earthquake
Engineering & Structural Dynamics, 35(9), 1097-1117.
Vamvatsikos, D., & Cornell, C. A. (2002). Incremental dynamic analysis. Earthquake
Engineering & Structural Dynamics, 31(3), 491-514.
Vecchio, F. J., & Collins, M. P. (1986). The modified compression-field theory for reinforced
concrete elements subjected to shear. Paper presented at the ACI Journal Proceedings.
Walker, G. (2010). Comparison of the impacts of Cyclone Tracy and the Newcastle Earthquake
on the Australian building and insurance industries. Australian Journal of Structural
Engineering, 11(3), 283-293.

291
Wang, S., & Wei, L. (2002). Studies on dynamic characteristics and seismic responses of the tall
buildings with different level transfer stories. Building Structures, 32(8), p54-58.
Waugh, J. D., & Sritharan, S. (2010). Lessons learned from seismic analysis of a seven-story
concrete test building. Journal of Earthquake Engineering, 14(3), 448-469.
Whittaker, A., Hart, G., & Rojahn, C. (1999). Seismic response modification factors. Journal of
Structural Engineering, 125(4), 438-444.
Wibowo, A., Wilson, J. L., Lam, N. T., & Gad, E. F. (2013). Seismic performance of lightly
reinforced structural walls for design purposes. Magazine of Concrete Research, 65(13),
809-828.
Wilson, J. L., Lam, N. T., & Pham, L. (2008). Development of the new Australian earthquake
loading standard. EJSE Special Issue: Earthquake Engineering in the low and moderate
seismic regions of Southeast Asia and Australia, 8, 25-31.
Wong, P., Vecchio, F., & Trommels, H. (2014). VecTor2—software for nonlinear analysis of
two-dimensional reinforced concrete membrane structures.
Wong, R. (2013). Construction of Transfer Plate - from various case studies.
Wood, S. L. (1990). Shear strength of low-rise reinforced concrete walls. ACI Structural Journal,
87(1).
Wood, S. L. (1992). Seismic response of R/C frames with irregular profiles. Journal of Structural
Engineering, 118(2), 545-566.
Wu, M., Qian, J., Fang, X., & Yan, W. (2007). Experimental and analytical studies on tall
buildings with a high‐level transfer story. The Structural Design of Tall and Special
Buildings, 16(3), 301-319.
Xu, P., Wang, C., Hao, R., & Xiao, C. (2000). The effects on seismic behavior of the transfer
story level on the column-supported shear wall structures. Building Structures, 30(1), 38-
42.
Yacoubian, M., Lam, N., & Lumantarna, E. (2015). Performance of limited ductility reinforced
concrete walls in low to moderate seismicity regions. Paper presented at the 10th Pacific
Conference on Earthquake Engineering; Building an Earthquake-Resilient Pacific,
Sydney, Australia (6-8 November).
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. L. (2016a). Effects of podium interferences
on shear forces distributions in RC walls supporting tall buildings. Paper presented at the
Australasian Structural Engineering Conference, Brisbane, Australia.
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. (2016b). Seismic performance of high-
rise buildings featuring a transfer plate taking into account displacement-controlled
behaviour. Paper presented at the Australian Earthquake Engineering Society, Melbourne
(25-27 November).
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. L. (2017a). Effects of podium interference
on shear force distributions in tower walls supporting tall buildings. Engineering
Structures, 148, 639-659.
doi:https://doi.org/10.1016/j.engstruct.2017.06.075
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. (2017b). Analytical Modelling of Podium
Interference on Tower Walls in Buildings. Australian Journal of Structural Engineering.
doi: 10.1080/13287982.2017.1396870
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. (2017c). Simplified design checks for
buildings featuring transfer structures in regions of lower seismicity. Paper presented at
the 2017 World Congress on Advances in Structural Engineering and Mechanics
(ASEM17), Ilsan (Seoul), Korea (28th of August-1st of September).

292
Yacoubian, M., Lam, N., Lumantarna, E., & Wilson, J. (2017d). Seismic performance assessment
of structural walls supporting towers in tall buildings. Paper presented at the Australian
Earthquake Engineering Society conference, Canberra, Australia (24-26 November).
Yacoubian, M., Lam, N., & Lumantarna, E. (2018). Simplified Design Checks for Buildings with a
Transfer Structure in Regions of Low Seismicity, in: Design of Buildings and Structures in
Low to Moderate Seismicity Regions (Chapter 8, pp. 113-140): Chinese National
Engineering Research Centre for Steel Construction (Hong Kong Branch) at the Hong
Kong Polytechnic University.
Yang, T., Hurtado, G., & Moehle, J. (2010). Seismic behavior and modeling of flat-plate gravity
framing in tall buildings. Paper presented at the Proceedings of the 9th US National and
10th Canadian Conference on Earthquake Engineering.
Zekioglu, A., Willford, M., Jin, L., & Melek, M. (2007). Case study using the Los Angeles tall
buildings structural design council guidelines: 40‐storey concrete core wall building. The
Structural Design of Tall and Special Buildings, 16(5), 583-597.
Zhao, J., & Sritharan, S. (2007). Modeling of strain penetration effects in fiber-based analysis of
reinforced concrete structures. ACI Structural Journal, 104(2), 133.
Zhitao, Z. J. W. G. L. (2000). Dynamic Properties and Response of the Thick Slab of Transfer
Plate Models with Dual Rectangular Shape in Tall Building [J]. Building Structure, 6,
010.

293
294
APPENDICES

295
 Appendix A- Summary of the ground motion records
Table A-1 Details of the input ground motions used in the study

Record Earthquake Magnitude Distance Faulting Sa(T=2.0),

Earthquake Event (year) Source PGA, [g's] PGV, [mm/s] Sd, [mm]
Number description/Bin (Mw) (R) (km) mechanism [g's]
1 - Stochastic GENQKE 6.5 20* Reverse 0.10 115.78 67.42 0.06
2 - Stochastic GENQKE 6.5 20* Reverse 0.15 177.84 61.47 0.05
3 - Stochastic GENQKE 6.5 20* Reverse 0.11 118.04 83.05 0.04
4 - Stochastic GENQKE 6.5 20* Reverse 0.11 115.23 61.09 0.03
5 - Stochastic GENQKE 6.5 20* Reverse 0.13 95.95 80.23 0.04
6 - Stochastic GENQKE 6.5 20* Reverse 0.10 109.74 75.64 0.04
7 - Stochastic GENQKE 6.5 20* Reverse 0.12 96.54 62.02 0.05
8 - Stochastic GENQKE 6.5 20* Reverse 0.13 132.89 80.37 0.03
9 - Stochastic GENQKE 6.5 20* Reverse 0.11 108.56 70.65 0.04
10 - Stochastic GENQKE 6 28* Reverse 0.19 173.44 63.72 0.03
11 - Stochastic GENQKE 6 28* Reverse 0.20 137.22 63.73 0.02
12 - Stochastic GENQKE 6 28* Reverse 0.16 118.93 59.28 0.03
13 - Stochastic GENQKE 6 28* Reverse 0.13 133.93 68.38 0.03
14 - Stochastic GENQKE 7 90* Reverse 0.08 105.35 78.31 0.03
15 - Stochastic GENQKE 7 90* Reverse 0.13 94.00 74.18 0.03
16 - Stochastic GENQKE 7 90* Reverse 0.11 102.82 90.26 0.03
17 - Stochastic GENQKE 7 90* Reverse 0.12 130.76 92.84 0.04
18 - Stochastic GENQKE 7.5 30* Reverse 0.25 192.21 118.59 0.07
19 - Stochastic GENQKE 7.5 30* Reverse 0.31 218.24 82.19 0.06
20 - Stochastic GENQKE 7.5 30* Reverse 0.25 227.87 131.38 0.10
21 - Stochastic GENQKE 7.5 30* Reverse 0.28 242.89 106.81 0.06

Table A-1 continued on next page

296
Record Earthquake Magnitude Distance Faulting
Earthquake Event (year) Source PGA, [g's] PGV, [mm/s] Sd, [mm] Sa(T=2.0), [g's]
Number description/Bin (Mw) (R) (km) mechanism
22 - Stochastic GENQKE 7.5 30* Reverse 0.31 207.31 109.14 0.08
23 - Stochastic GENQKE 7.5 30* Reverse 0.26 265.32 106.39 0.06
24 - Stochastic GENQKE 7.5 35* Reverse 0.21 159.23 99.31 0.06
25 - Stochastic GENQKE 7.5 35* Reverse 0.20 190.12 110.98 0.08
26 - Stochastic GENQKE 7.5 35* Reverse 0.23 205.55 116.28 0.08
27 - Stochastic GENQKE 7.5 35* Reverse 0.27 286.74 135.66 0.07
28 - Stochastic GENQKE 7.5 35* Reverse 0.22 179.76 130.38 0.09
29 - Artificial SeismoArtif See Note 1 Reverse 0.24 198.02 122.49 0.11
30 - Artificial SeismoArtif See Note 1 Reverse 0.28 221.75 115.42 0.10
31 - Artificial SeismoArtif See Note 1 Reverse 0.25 274.46 129.29 0.09
32 - Artificial SeismoArtif See Note 1 Reverse 0.26 264.82 118.65 0.10
33 - Artificial SeismoArtif See Note 1 Reverse 0.29 251.77 112.48 0.11
34 - Artificial SeismoArtif See Note 1 Reverse 0.26 196.78 120.99 0.09
35 - Artificial SeismoArtif See Note 1 Reverse 0.24 172.18 132.66 0.10
Reverse
36 N. Palm Springs (1986) Hemet Fire Station PEER 6.06 34.48* 0.11 114.71 116.40 0.09
Oblique
Fremont - Emerson Reverse
37 Loma Prieta (1989) PEER 6.93 39.66* 0.11 136.94 112.12 0.07
Court Oblique
38 Chi-Chi Taiwan (1999) CHY025 PEER 6.3 39.07* Reverse 0.09 110.32 121.26 0.04
39 Friuli Italy (1976) - ISESD 6.5 21.1† Reverse 0.32 325.87 110.11 0.05
LA - Griffith Park PEER
40 Northridge (1994) 6.69 21.2* Reverse 0.29 265.40 119.82 0.04
Observatory

Table A-1 continued on next page

297
Record Earthquake Earthquake Magnitude Distance Faulting
Source PGA, [g's] PGV, [mm/s] Sd, [mm] Sa(T=2.0), [g's]
Number Event (year) description/Bin (Mw) (R) (km) mechanism

41 - Artificial SeismoArtif See Note 2 Reverse 0.184 184.53 77.46 0.08

42 - Artificial SeismoArtif See Note 2 Reverse 0.26 147.81 76.01 0.07

43 - Artificial SeismoArtif See Note 2 Reverse 0.25 217.40 80.36 0.06

44 - Artificial SeismoArtif See Note 2 Reverse 0.21 200.53 76.33 0.07

45 - Artificial SeismoArtif See Note 2 Reverse 0.21 148.61 83.49 0.07

46 - Artificial SeismoArtif See Note 2 Reverse 0.23 159.31 85.38 0.07
47 - Artificial SeismoArtif See Note 2 Reverse 0.65 176.53 95.80 0.10
48 - Artificial SeismoArtif See Note 3 Reverse 0.13 105.56 40.28 0.04
49 - Artificial SeismoArtif See Note 3 Reverse 0.13 97.05 42.84 0.04
50 - Artificial SeismoArtif See Note 3 Reverse 0.15 73.89 44.18 0.04
51 - Artificial SeismoArtif See Note 3 Reverse 0.13 157.01 43.97 0.04
52 - Artificial SeismoArtif See Note 3 Reverse 0.11 97.85 42.36 0.04
53 - Artificial SeismoArtif See Note 3 Reverse 0.13 81.28 39.83 0.04
54 - Artificial SeismoArtif See Note 3 Reverse 0.12 84.51 38.64 0.04
* Hypocentral distance
† Epicentral distance
Note 1 Code-Compliant Suite of records based on the response spectrum of the Australian Standard 1170.4 for site class De- SeismoArtif
Note 2 Code-Compliant Suite of records based on the response spectrum of the Australian Standard 1170.4 for site class Ce- SeismoArtif
Note 3 Code-Compliant Suite of records based on the response spectrum of the Australian Standard 1170.4 for site class Ae- SeismoArtif
Records hatched in grey are used in Chapter 5

298
Figure A-1 Illustration of the procedure for obtaining ground motion accelerograms evaluated on different sub-soil conditions

299
(a) Record nos. 29-35; Artificial records simulated to match the AS 1170.4:2007 spectrum (site class De)

Figure A-2 continued on next page

300
(b) Record nos. 41-47; Artificial records simulated to match the AS 1170.4:2007 spectrum (site class Ce)

Figure A-2 continued on next page

301
 (c) Record nos. 48-54; Artificial records simulated to match the AS 1170.4:2007 spectrum (site class Ae)

Figure A-2 Acceleration (Sa(T)) and displacement (Sd(T)) response spectra of artificial records used in the study (generated using SeismoArtif)

302
Appendix B- Details of the 2D and 3D case study building models

(a) Floor plans (b) 3D render

Figure B-1 Floor plans and elevations of Case study building 2 (Chapter 3)

303
 (a) Podium floor plan

(b) Tower floor plan (c) 3D render

Figure B-2 Floor plans and elevations of Case study building 3 (Chapter 3/4)

304
 (a) Floor Plans (b) 3D render
Figure B-3 Floor plans and elevations of Case study building 4 (Chapter 3/5)

305
(a) Reinforcement layout (elevation) of the buildings SB-1 and SB-2 (Chapter 6)

(b) Reinforcement layout (elevation) of the buildings SB-1 and SB-2 (Chapter 6)
Figure B-4 continued on next page

306
(c) Section/ reinforcement details of tower walls in buildings SB-1,SB-2, TS-1 &
TS-2

(d) Section/ reinforcement details of podium columns in buildings TS-1 and TS-2

Figure B-4 Reinforcement detailing of building models: TS-1, TS-2, SB-1 & SB-2
(Chapter 6)

307
Table B-1 Floor slab loads applied on the floor slabs of buildings SB-1, SB-2, TS-1 & TS-2
(Chapter 6)

Podium Floors Tower floors

Additional dead loads,

2.5 1.8
kPa

Imposed loads,
5.6 4.0
kPa

General Note: for 2D analyses, the gravity loads are modelled as line loads on the floor slabs
along the length of the span. The conversion of the above listed loads (in kPa) to line loads
(kN/m) is achieved by multiplying the value of the pressure by 6 and 8 for the tower and podium
storeys respectively.

308
 Appendix C- Calculation of the podium and tower lateral stiffnesses
A summary of the procedure for calculating the translational stiffness of the tower (𝐾𝑇 ) and
the podium (𝐾𝑃 ) structures is described. The translational stiffness parameters 𝐾𝑇 and 𝐾𝑃 for the
various case study buildings are computed by evaluating the effective displacement and base shear
of the building when subject to lateral loads. The storey masses ( 𝑚𝑖 ) and the applied lateral loads
( 𝐹𝑖 ) have been computed and distributed by considering the self-weight of the structural
members, the dead load and 30% of the live loads applied at each storey level as recommended
by the Australian Standard (AS 1170.4:2007).

Table C-1 Calculations for determining the lateral stiffness of the podium and the tower
structures (case study building Section 4.4)
Lateral stiffness of the podium (KP)

Level mi [kg] zi [m] 𝛿i [mm] Fi [kN] mi 𝛿i mi 𝛿i2

4 6226305 12 1.65 20177.06 17341835.83 48301407.03

3 6773368 9 1.22 16462.41 13958872.43 28767096.01

2 6777571 6 0.74 10981.75 8490086.81 10635310.04

1 6763031 3 0.29 5479.10 3288562.95 1599082.82

GF 571997 0 0 0 0 0

∑ 53100.32 4.31E+07 8.93E+07

δeff = ∑ mi δ2i /
2.07
∑ mi 𝛿i [mm]

Base shear
53100.32
Vb [kN]

𝑘P [kN/m] = BS/δeff 2.56E+07

Lateral stiffness of the tower (KT)

Level mi [kg] zi [m] 𝛿i [mm] Fi [kN] mi 𝛿i mi 𝛿i2

10 1893595.55 30 7.15 5505.82 13541050.67 96831687.94

9 2062386.94 27 6.34 5338.98 13083536.74 83000396.39

8 2062386.94 24 5.50 4688.81 11351235.78 62476420.56

7 2062386.94 21 4.64 4046.95 9564364.66 44354950.83

6 2062386.94 18 3.76 3414.44 7754460.24 29156339.40

309
5 2062386.94 15 2.89 2792.69 5969412.43 17277982.15

4 2526754.50 12 2.08 2662.01 5243502.30 10881277.31

3 2548349.78 9 1.34 1954.59 3403428.83 4545423.04

2 2548349.78 6 0.70 1250.02 1793186.88 1261804.48

1 2551254.99 3 0.23 582.72 587793.27 135423.91

0 197185.46 0 0.00 0.00 0.00 0.00

∑ 35625.38 72291971.8 349921706

δeff =
∑ mi δ2i /
4.84
∑ mi 𝛿i
[mm]

Base
32237.04
shear Vb [kN]

k T [kN/m]
6.66E+06
= BS/δeff

310
 Table C-2 Calculations for determining the lateral stiffness of the podium and the tower
structures (case study building Section 5.7.1)
Lateral stiffness of the podium (KP)

Level mi [kg] zi [m] 𝛿i [mm] Fi [kN] mi 𝛿i mi 𝛿i2

4 7083894 16.8 8.696 19218.90 97839256 867149330.3

3 2019757 12.6 7.03 4109.76 14035291 97531237.0

2 2019757 8.4 4.054 2739.84 8371893 34701494.5

1 2019757 4.2 1.487 1369.92 3037714 4568722.5

GF 0 0 0 0 0

∑ 27438.42 1.23E+08 1003950784

δeff = ∑ mi δ2i / ∑ mi 𝛿i
8.14
[mm]

Base shear Vb [kN] 27438

K P [kN/m] = BS/δeff 3369410

Lateral stiffness of the tower (KT)

Level mi [kg] zi [m] 𝛿i [mm] Fi [kN] mi 𝛿i mi 𝛿i2

34 371514 102 242.92 200.23 90248341 21923139862

33 439004 99 236.65 224.30 1.04E+08 24586030012

32 439003 96 230.49 212.52 1.01E+08 23321953215

31 439003 93 224.14 201.02 98396251 22054095449

30 439003 90 217.56 189.79 95510829 20779609165

29 439003 87 210.75 178.84 92520570 19498838487

28 439003 84 203.69 168.17 89422047 18214672236

27 439003 81 196.39 157.78 86215261 16931695198

26 439003 78 188.85 147.68 82903638 15655946317

25 439003 75 181.07 137.87 79491289 14393659465

24 439003 72 173.09 128.35 75986436 13152378561

23 439003 69 164.91 119.12 72397303 11939246548

22 439003 66 156.57 110.19 68733482 10761401400

311
21 439003 63 148.07 101.56 65005250 9625626715

20 439003 60 139.46 93.24 61222887 8538071010

19 439003 57 130.75 85.22 57398727 7504757495

18 439003 54 121.97 77.51 53544417 6530711977

17 439004 51 113.15 70.12 49671695 5620168457

16 439003 48 104.31 63.05 45792629 4776648592

15 439003 45 95.49 56.31 41921187 4003126420

14 439003 42 86.73 49.89 38073044 3301926224

13 439003 39 78.05 43.81 34263958 2674282137

12 439003 36 69.50 38.08 30509004 2120255588

11 439003 33 61.11 32.69 26828738 1639580080

10 439003 30 52.94 27.66 23241660 1230456996

9 439003 27 45.04 23.00 19771068 890414892.4

8 439003 24 37.45 18.71 16441629 615774738.8

7 439003 21 30.26 14.80 13284179 401977348.1

6 439003 18 23.54 11.30 10332291 243178649.7

5 439003 15 17.37 8.21 7627080 132510021.9

4 439003 12 11.89 5.55 5217879 62018345.41

3 439003 9 7.21 3.35 3164988 22817932.16

2 439003 6 3.51 1.65 1541039 5409526.729

1 439003 3 1.00 0.49 439905.2 440808.9628

GF 67488.9 0 0.00 0.00 0 0

3002 1.74E+09 2.93E+11

δeff = ∑ mi δ2i / ∑ mi 𝛿i
168.27
[mm]

Base shear Vb [kN] 3002.04

312
K T [kN/m] = BS/δeff 17841

313
 Table C-3 Calculations for determining the lateral stiffness of the podium and the tower
structures (case study building Section 5.7.2)
Lateral stiffness of the podium (KP)

Level mi [kg] zi [m] 𝛿i [mm] Fi [kN] mi 𝛿i mi 𝛿i2

11.00 5767549.06 36.50 22 12268 124202562 2674667557

10.00 3857243.26 33.30 20 7465 75363557 1472467587

9.00 3889716.53 30.10 18 6784 68720325 1214094420

8.00 3889716.53 26.90 16 6042 60854994 952082296

7.00 3889716.53 23.70 14 5303 52547106 709871346

6.00 3889716.53 20.50 11 4567 43958505 496784326

5.00 3889716.53 17.30 9 3835 35287212 320122894

4.00 3889716.53 14.10 7 3107 26772220 184268374

3.00 3889716.53 10.90 5 2383 18693079 89834624

2.00 3889716.53 7.70 3 1666 11385599 33326814

1.00 4008653.43 4.50 1 987 5360932 7169388

GF 0.00 0.00 0 0 0 0

∑ 54407.40 523146090 8154689626

δeff = ∑ mi δ2i /
15.59
∑ mi 𝛿i [mm]

Base shear Vb [kN] 54407.40

K P [kN/m] = BS/δeff 3490386

Lateral stiffness of the tower (KT)

Level mi [kg] zi [m] 𝛿i [mm] Fi [kN] mi 𝛿i mi 𝛿i2

23 584688.31 36.00 21.74 1370.47 12712528.27 276400899.74

22 694277.14 33.00 19.54 1488.65 13568039.72 265155931.78

21 694277.14 30.00 17.39 1350.27 12073145.14 209946180.40

20 694277.14 27.00 15.19 1212.21 10547441.78 160236484.44

19 694277.14 24.00 12.97 1074.52 9006770.65 116843710.99

314
18 694277.14 21.00 10.77 937.24 7474104.93 80461016.60

17 694277.14 18.00 8.61 800.42 5977794.20 51469393.80

16 694277.14 15.00 6.56 664.14 4551015.80 29832099.68

15 694277.14 12.00 4.65 528.51 3231572.81 15041634.24

14 694277.14 9.00 2.97 393.69 2060986.10 6118109.72

13 694277.14 6.00 1.57 259.95 1087085.11 1702135.88

12 694314.44 3.00 0.52 127.87 361108.87 187810.61

TF 0.00 0.00 0.00 0.00 0.00 0.00

10207.96 82651593.39 1213395407.86

δeff = ∑ mi δ2i /
168.27
∑ mi 𝛿i [mm]

Base shear Vb [kN] 10207.96

K T [kN/m] = BS/δeff 695325

315
 Appendix D- Summary of the key parameters used in the analytical model of
Chapter 5

Table D-1 Detailed calculation summary of the case study building in Section 5.7.1

Parameter Calculation Value

𝐊𝐓 , kN/m The procedure is summarised in the Appendix C (Table C-2) 17841

𝐊𝐏 , kN/m The procedure is summarised in the Appendix C (Table C-2) 3369410

−1
𝐊𝐱 , kN/m 1 1
( + ) 17747
(Eq. 5-7) 17841 3369410

𝐫, m 1022 + 282
√ 30.53
(Eq. 5-14) 12

𝐊𝛉𝐓𝐏 kNm/rad
28 1.5
31.62 × 106 × 2.72 × 22.35 × ( ) 7.41 × 108
(Eq. 5-11) 102

𝐊𝛉 , kNm/rad For infinitely stiff podium columns and fixed at the base 𝐊𝛉 =
7.41 × 108
(Eq. 5-10) 𝐊𝛉𝐓𝐏

7.41 × 108
𝐛𝐫 √Kθ /Kx /r [√ ] / 30.53 6.69
17747

Records No: 29-35 107

(99346) [−0.2 ln(6.69) + 0.6] * 0.00024
RSDmax = 107mm
𝐏𝐑𝐃
Records No: 41-47 67
(99346) [−0.2 ln(6.69) + 0.6] * 0.00015
(Eq. 5-23(a)-(b)) RSDmax = 67mm

Records No: 48-54 38

(99346) [−0.2 ln(6.69) + 0.6] * 8.41× 10−5
RSDmax = 38 mm
𝛂𝐫 (Eq. 5- 1
( × 2.73 × 5.02)
12
2) √ 6.42
1
( × 2.03 × 0.3)
12

𝐅𝐈 From Figure 5-10 0.40

𝐅𝐒𝐓𝐑𝐔𝐓 , 𝑘𝑁 (Eq. Records No: 29-35

0.4 × 0.00024 × 5.88 × 106 † 557
5-24) RSDmax = 107mm

316
 Records No: 41-47
0.4 × 0.00015 × 5.88 × 106 †
349
RSDmax = 67mm

Records No: 48-54

0.4 × 8.41 × 10−5 × 5.88 × 106 † 198
RSDmax = 38 mm

Notes:

*
The value of ℎ𝑥 used in the calculation of PRD has been obtained directly from the FE analysis using ETABS
program package. In the absence of a detailed numerical model, the recommended value of hx = 0.7 hb may be used
for preliminary (conservative) estimate of the value of PRD (Eq. 5-23a-5-23b).
†
The value of 𝐸𝐶 𝐴𝑒𝑓𝑓 is calculated based on the in-plane sectional properties of the connecting slab with concrete
grade (characteristic compressive strength) of 𝑓𝑐′ = 24𝑀𝑃𝑎: (5000 × √24 × (2.4 × 0.2) × 0.5 × 103 ). The term
(2.4 x 0.2) is the product of the width of the column strip as obtained from the FE analysis (using the program
package ETABS) and the gross thickness of the slab. The factor (0.5) is recommended by TBI (2017) guidelines for
computing the in-plane stiffness of reinforced concrete slabs subjected to transfer forces (Table 4-3 of TBI 2017).

317
 Table D-2 Detailed calculation summary of the case study building in Section 5.7.2

Parameter Calculation Value

𝐊𝐓 , kN/m The procedure is summarised in the Appendix C (Table C-3) 695325

𝐊𝐏 , kN/m The procedure is summarised in the Appendix C (Table C-3) 3490386

𝐊𝐱 , kN/m 1 1 −1
( + ) 579818
(Eq. 5-7) 695325 3490386

𝐫, m 362 + 38.62
√ 15.24
(Eq. 5-14) 12

𝐊𝛉𝐓𝐏
38.6 1.5
kNm/rad 5000 × √32 × 103 × (0.62 ) × 20.5 × ( ) 2.32 × 108
36
(Eq. 5-11)

𝐊𝛉 , kN-
38.6 1.5
m/rad 5000 × √32 × 103 × (0.62 ) × 20.5 × ( ) 2.32 × 108
36
(Eq. 5-10)

2.32 × 108
𝐛𝐫 √Kθ /Kx /r [√ ] / 15.24 1.31
579818

Records No: 29-35 107

( ) [−0.2 ln(1.31) + 0.6] 0.00115
RSDmax = 107mm 50750
𝐏𝐑𝐃
Records No: 41-47 67
(Eq. 5.23a- ( ) [−0.2 ln(1.31) + 0.6] 0.00072
RSDmax = 67mm 50750
b)
Records No: 48-54 38
( ) [−0.2 ln(1.31) + 0.6] 0.00041
RSDmax = 38 mm 50750

𝛂𝐫 1
(12 0.63 6.8)
√ 1.31
1
(Eq. 5-2) (12 1.53 0.25)

FI From Figure 5-10 0.4

Records No: 29-35 0.4 × 0.00115 × 5000 × √20 × (1.1 × 0.2) × 0.5
𝐅𝐒𝐓𝐑𝐔𝐓 , 𝑘𝑁 1132
RSDmax = 107mm × 103

318
Records No: 41-47 0.4 × 0.00072 × 5000 × √20 × (1.1 × 0.2) × 0.5
709
RSDmax = 67mm × 103

Records No: 48-54 0.4 × 0.00041 × 5000 × √20 × (1.1 × 0.2) × 0.5
402
RSDmax = 38 mm × 103

319
 Appendix E- Strutting force generation above the TFL
In this section, results from the time history analyses performed on the FE building models of
Section 5.7 are outlined.

320
 (a) Wall 1
Figure E-1 continued on next page

321
 (b) Wall 2
Figure E-1 𝑭𝑺𝑻𝑹𝑼𝑻 vs ∆𝜽𝑻𝑷 time histories of walls 1 & 2 (high-rise building) analysed using records 48-54 (Section 5.7.1)

322
 (a) Record nos. 29-35

Figure E-2 continued on next page

323
 (b) Record nos. 41-47

Figure E-2 continued on next page

324
(c) Record nos. 48-54

325
Figure E-2 𝑭𝑺𝑻𝑹𝑼𝑻 vs ∆𝜽𝑻𝑷 time histories for Wall 3 (Medium-rise building) analysed using different intensity ground motion records

326
Appendix F- Summary of RC wall specimens considered in the study

327
Table F-1 Summary of the geometric, reinforcement, material details of the wall specimens used in the calibration of numerical models (Chapter 3 &
6)
𝑓𝑐′ Shear span ALR
𝑀/𝑉𝑙𝑤 Reinforcement Details
Specimen Name Reference Dimension
N 𝜌𝑣 𝜌𝑡 𝜌𝑏
(𝑙𝑤 × 𝑡𝑤 ) (MPa) (M/V) (m) ( )
fc′ Ag (%) (%) (%)
M3 (Greifenhagen & Lestuzzi, 2005) 900 x 80 20.10 0.69 0.77 0.05 0.3 0.3 -
M4 (Greifenhagen & Lestuzzi, 2005) 900 x 80 24.40 0.69 0.77 0.095 0.3 0.3 -
MSW3 (Salonikios et al., 1999) 1200 x 100 24.6 1.920 1.6 0.07 0.277 0.277 1.3

LSW3 (Salonikios et al., 1999) 1200 x 100 24 1.320 1.1 0.07 0.277 0.277 1.3
W1 (Su & Wong, 2007) 400x80 50.2 1.640 4.0 0.25 1.96 0.54 -
W2 (Su & Wong, 2007) 400x80 41.8 1.640 4.0 0.5 1.96 0.54 -
W3 (Su & Wong, 2007) 400x80 42.9 1.640 4.0 0.5 1.96 1.08 -
C-3-N_ALR01 (Looi et al., 2017) 800x80 29.1 0.950 1.19 0.10 2 1.4 -
C-3-N_ALR02 (Looi et al., 2017) 800x80 26.4 0.950 1.19 0.2 2 1.4 -
C-3-N_ALR03 (Looi et al., 2017) 800x80 27.6 0.950 1.19 0.30 2.0 1.4 -
WS-T1-S1 (Massone, 2006) 1520 x 152 25.5 0.76 0.5 0 0.428 0.278 3.12
WP-T5-N10-S2 (Massone, 2006) 1370 x 152 31.4 1.00 0.44 0.10 0.227 0.278 1.33
WSH4 (Dazio et al., 2009) 2000 x150 40.9 4.56 2.28 0.057 1.54 0.25 1.54
WSH5 (Dazio et al., 2009) 2000 x150 38.3 4.56 2.28 0.13 0.67 0.25 0.67
WSH6 (Beyer et al., 2008) 2000 x150 45.6 4.52 2.26 0.11 1.54 0.25 1.54

LW2 (Li et al., 2015) 2000 x 120* 41.6 1.13 0.05 0.50 0.50 1.40

* Wall with a boundary element of dimension: 150 x 300 (Barbell wall) used in the Vector2 numerical model calibration (Section 3.6.1)

328
 Appendix G- Summary of the results used in constructing the fragility
Function
The results from the numerous nonlinear time history analyses described in Section 6.7-6.8
are assembled in Figures E-1 – E-2. The white bars represent the total number of analyses
performed on each of the four building models (SB-1, SB-2, TS-1and TS-2). The coloured hatching
represent the number of analyses where a specified performance level (limit state) had been
observed in the response behaviour of the building (green for IO, blue for LS and red for CP
performance levels).

(a) PGA as IM

Figure G-1 is continued on next page

329
 (b) PGV as IM

(c) RSDmax as IM
Figure G-1 Number of nonlinear time history analyses performed on building SB-1 and
the fraction of analyses where a performance level had been observed in the building

330
 (a) PGA as IM

(b) PGV as IM
Figure G-2 is continued on next page

331
 (c) RSDmax as IM

Figure G-2 Number of nonlinear time history analyses performed on building SB-2 and
the fraction of analyses where a performance level had been observed in the building

332
 (a) PGA as IM

(b) PGV as IM

Figure G-3 is continued on next page

333
 (c) RSDmax as IM

Figure G-3 Number of nonlinear time history analyses performed on building TS-1 and
the fraction of analyses where a performance level had been observed in the building

334
 (a) PGA as IM

(b) PGV as IM

Figure G-4 is continued on next page

335
 (c) RSDmax as IM
Figure G-4 Number of nonlinear time history analyses performed on building TS-2 and
the fraction of analyses where a performance level had been observed in the building

336