Evaluation of the Seismic Design Methods for Steel Multi-Tiered Concentrically Braced

Frames

by

Pablo Antonio Cano

A thesis submitted in partial fulfillment of the requirements for the degree of

Master of Science

in
STRUCTURAL ENGINEERING

Department of Civil and Environmental Engineering

University of Alberta

© Pablo Antonio Cano, 2019

Abstract

Multi-tiered concentrically braced frames (MT-CBFs) are widely used in North America as the

lateral load-resisting system of tall single-storey buildings such as airplane hangars, recreational

facilities, shopping centres, and industrial buildings. MT-CBFs consist of multiple concentrically

braced panels along the height of the frame separated by horizontal struts. Multi-tier arrangements

are typically used when it is not practical nor economical to use a single bracing panel along the

height of the frame between the ground and roof levels. In multi-tiered braced frames, the length

of braces is reduced, which allows the selection of smaller brace sizes and easily satisfying code-

specified brace slenderness limits. The column buckling length in the in-plane direction is also

reduced due to the application of intermediate horizontal struts, which permits selection of a

smaller column section. When using shorter braces, result in smaller design forces on the adjacent

forced-controlled members including struts, beams, columns, and connections.

Past studies have shown that inelastic frame deformations tend to concentrate in one of the tiers

over the frame height, which induces large in-plane bending moments in braced frame columns

and high deformation demands in braces. This behaviour may lead to column buckling and/or

brace fracture. Design requirements have been included in the Canadian steel design standard

(CSA S16) and the U.S. Seismic Provisions to prevent such limit states. In the U.S., the Seismic

Provisions have included the design of multi-tiered ordinary and special concentrically braced

frames (MT-OCBFs and MT-SCBFs). However, there are no detailed numerical models or

experimental research to validate the design requirements.

This M.Sc. thesis research focuses on the evaluation of the seismic behaviour and design methods

for MT-CBFs. A two-tiered CBF prototype frame was first designed as a special concentrically

ii
braced frame using the 2010 and 2016 AISC Seismic Provisions. Then, a detailed numerical model

was developed and was analyzed using the cyclic pushover (static) analysis and the nonlinear

response history (dynamic) analysis. The global and local response of the prototype frames

together with the force demands in the columns were examined using the results obtained from the

numerical analyses. Special attention was paid to the stability condition of the column as well as

in-plane and out-of-plane moments induced in this member.

Results obtained for the prototype frame designed excluding the special seismic design provisions

confirmed column buckling and nonuniform distribution of the frame inelastic lateral deformations

in the tier where brace tensile yielding takes place first. A total of 13 column buckling cases were

observed using the dynamic analysis method among an ensemble of 40 ground motion records.

Moreover, excessive deformations, which is an indication of brace low-cycle fracture, were

observed in the yielding tier of this prototype frame. In contrast, the prototype braced frame that

was designed in accordance with the recent special seismic design provisions performed

satisfactorily. No column buckling occurred and the frame lateral response was stable. Braces in

both tiers yielded under most ground motion records and frame inelastic lateral deformations were

shared between both tiers. It was found that the column moment demands prescribed by the current

design provisions over estimates the moment demands obtained under a major earthquake event.

Additionally, expected storey drift was found to be higher than the code-specified design storey

drift, which resulted in large ductility demands in braced tiers, which poses concerns regarding the

adequacy of the current drift requirements. New brace force adjustment factors are proposed to

achieve more realistic brace nonlinear forces when computing column force demands and tier

drifts.

iii
Preface

This thesis is an original work by Pablo A. Cano. Parts of Chapters 3, 4, and 5 were presented at

the 2018 Annual Structural Stability Research Council (SSRC) conference held in Baltimore,

Maryland, U.S. and submitted to the 2019 American Society of Civil Engineers Structures

Congress to be held in Orlando, Florida, U.S. The work presented at these two conferences was

co-authored with Dr. Ali Imanpour. The conference paper published in the 2018 SSRC conference

proceeding associated with this research has been invited to be published in the American Institute

of Steel Construction (AISC) Engineering Journal, which will be submitted for publication.

iv
Dedication

To the victims of the 1999 Armenia earthquake

and

to my parents for encouraging me to strive to surpass myself each day.

v
Acknowledgements

First, I would like to express my most profound gratitude to my supervisor, Dr. Ali Imanpour. His

crucial guidance and continuous encouragement have led me to progress as an academic and as a

professional, but more important—as a person. Thank you for sharing your knowledge and passion

about seismic engineering. It is a privilege to continue my doctoral studies with you as my

supervisor. Many thanks to Dr. Robert G. Driver and Dr. Douglas Tomlinson for their time in

reviewing this thesis. I would also like to thank Dr. Robert Tremblay for his input in this research.

I am very thankful for the advice provided by Dr. Cristopher Stoakes in the development of the

numerical model used in this thesis. I am vey grateful with Yan Jiang for sharing the experimental

data of the bracing member and Dr. Ali Davaran for the advice provided.

Furthermore, I would like to express my genuine appreciation to the Steel Centre for providing

unique learning opportunities and encouragement to challenge traditional boundaries.

Additionally, I would like to thank my colleagues from the Steel Centre for their advice throughout

this project.

I would like to acknowledge the financial support from the Natural Sciences and Engineering

Research Council (NSERC), American Institute of Steel Construction (AISC), Canadian Institute

of Steel Construction (CISC), and the faculties of Engineering and Graduate Studies and Research

at the University of Alberta.

I would also like to thank my family, who have supported me throughout this journey and have

given me their perpetual love.

Last, but certainly not least, I would like to thank my beloved girlfriend, Paula, for believing in

me and always being there for me. Words do no exist to express how thankful I am for the

unremitting love and support that I have felt.

vi
Contents
Chapter 1 – Introduction ............................................................................................1

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

1.2 Research Problem ......................................................................................................... 3

1.3 Objectives ..................................................................................................................... 6

1.4 Methodology ................................................................................................................. 7

1.5 Thesis Organization ...................................................................................................... 8

Chapter 2 – Literature Review ...................................................................................9

2.1 General .......................................................................................................................... 9

2.2 Seismic Behaviour of Steel Braces ............................................................................... 9

2.3 Seismic Behaviour of Steel Concentrically Braced Frames ....................................... 17

2.4 Seismic Behaviour of Steel MT-CBFs ....................................................................... 21

2.5 Design of MT-CBFs ................................................................................................... 35

Chapter 3 – Design of Steel Multi-Tiered Concentrically Braced Frames ......42

3.1 General ........................................................................................................................ 42

3.2 Building and Frame Geometry.................................................................................... 42

3.3 Gravity and Seismic Loads ......................................................................................... 44

3.4 Brace Design ............................................................................................................... 45

3.5 Column Design ........................................................................................................... 46

3.6 Strut Design ................................................................................................................ 54

vii
 3.7 Drift Check.................................................................................................................. 56

3.8 Design Summary......................................................................................................... 58

Chapter 4 – Numerical Model of the Braced Frames ........................................60

4.1 General ........................................................................................................................ 60

4.2 Model Development.................................................................................................... 60

4.3 Analysis....................................................................................................................... 74

Chapter 5 – Seismic Response of Two-Tiered Steel Concentrically Braced Frames .81

5.1 General ........................................................................................................................ 81

5.2 Pushover Analysis ....................................................................................................... 81

5.3 Nonlinear Response History Analysis ........................................................................ 88

5.4 Summary ................................................................................................................... 109

Chapter 6 – Design Recommendations ...................................................................110

6.1 General ...................................................................................................................... 110

6.2 Strength Requirements .............................................................................................. 110

6.2 Drift Requirement ..................................................................................................... 115

6.3 Case Study ................................................................................................................ 118

Chapter 7 – Conclusions and Recommendations....................................................124

7.1 Summary ................................................................................................................... 124

7.2 Limitations ................................................................................................................ 126

viii
 7.3 Conclusions ............................................................................................................... 126

7.4 Recommendations for Further Research ................................................................... 129

Bibliography .........................................................................................................131

List of Tables

Table 3.1: Summary of design calculations for the prototype frame ............................................ 59

Table 4.1: Selected ground motions for dynamic analysis (Dehghani 2016) ............................... 76

Table 5.1: Tier drifts at a storey drift 2.0%................................................................................... 84

Table 5.2: Statisctics of the global response of the 2010 design .................................................. 91

Table 5.3: Statisctics of global response of the 2016 design ........................................................ 93

Table 5.4: Statisctics of column demands for 2010 design .......................................................... 99

Table 5.5: Statisctics of column demands for 2016 design ........................................................ 100

Table 6.1: Statistics of column axial force and biaxial moment demands plus brace forces at the

maximum interaction ratio .......................................................................................................... 111

Table 6.2: Statistics of brace forces at the maximum storey drift .............................................. 117

ix
List of Figures

Figure 1.1: a) four-tiered concentrically braced frame in an industrial building; b) two-tiered

concentrically braced frame in an industrial building. .................................................................... 2

Figure 1.2: MT-CBFs configurations a) X-shape; b) V-shape; c) diagonals; d) chevron; and e)

split-X ............................................................................................................................................. 2

Figure 1.3: Inelastic response of a) K-type CBF; and b) standard CBF ......................................... 5

Figure 2.1: Hysteretic response of steel bracing members ........................................................... 10

Figure 2.2: Influence of global slenderness ratio on the hysteretic response of steel braces (Ziemian

2010) ............................................................................................................................................. 11

Figure 2.3: Inelastic response of chevron braced frame when plastic hinge forming in the beam11

Figure 2.4: HSS brace out-of-plane buckling (Sabelli et al. 2013) ............................................... 13

Figure 2.5: Effect of loading velocity (v in mm/s) on the steel brace buckling capacity

(Kazemzadeh Azad et al. 2018) .................................................................................................... 14

Figure 2.6: Hysteretic response of a steel bracing member obtained from a) quasi-static loading,

and b) dynamic loading with v = 300mm/s (Kazemzadeh Azad et al. 2018) ............................... 14

Figure 2.7: HSS members under cyclic loading: a) Local buckling; and b) fracture of (Fell et al.

2009) ............................................................................................................................................. 15

Figure 2.8: Influence of the brace slenderness on the fracture life of HSS members (Tremblay

2001) ............................................................................................................................................. 16

Figure 2.9: rotation at fracture: test data versus values predicted (Tremblay et al. 2003) ............ 16

x
Figure 2.10 Brace fracture limit varying with a) brace width-to-thickness ratio (KL/r =52.6 and

E/Fy = 446 MPa); and b) brace slenderness ratio (b/t =11.3 and E/Fy = 446 MPa) (Hsiao et al. 2013)

....................................................................................................................................................... 17

Figure 2.11: Single-storey steel concentrically braced frame ....................................................... 18

Figure 2.12: Progression of seismic behaviour of single-storey CBF .......................................... 19

Figure 2.13: Linear hinge zone for out-of-plane buckling of steel braces (Sabelli et al. 2013) ... 20

Figure 2.14: Typical MT-CBF configurations (AISC 2016a) ...................................................... 22

Figure 2.15: Comparison between Multi-tiered CBFs and Multi-storey CBF ............................. 23

Figure 2.16: Drift in individual braced panels of a two-tiered CBF and column in-plane bending

demand recorded between the braced panels (Imanpour and Tremblay 2012) ............................ 24

Figure 2.17: Two-tiered concentrically braced frame geometry (Imanpour and Tremblay 2012) 24

Figure 2.18: Normal stress contour (ksi, where 1 MPa = 0.145 ksi) of the column at maximum Tier

1 drift (Imanpour et al. 2013) ........................................................................................................ 25

Figure 2.19: Frame deformed shape at the instant of column buckling (Imanpour et al. 2016a) . 27

Figure 2.20: Improved lateral response of a two-tiered steel CBF: a) Brace yielding and buckling

in Tier 1 and initiation of brace tensile yielding in Tier 2; and b) Proposed brace force scenario to

trigger yielding in the noncritical tier (critical tier: Tier 1; and noncritical tier: Tier 2) .............. 28

Figure 2.21:a) Column free-body diagram under unbalanced brace story shear; b) column shear

force diagram; and c) column bending moment diagram ............................................................. 29

Figure 2.22: Three-tiered steel braced frame with adjacent gravity columns............................... 30

Figure 2.23: W250x101 column part of a two-tiered CBF in the multi-directional hybrid testing

system (Imanpour et al. 2017) ...................................................................................................... 32

xi
Figure 2.24: Local buckling of Wide-flange steel columns as part of multi-storey CBFs (Newell

and Uang 2008) ............................................................................................................................. 32

Figure 2.25: Lateral instability of steel Wide-flange column as part of MRFs (Elkady and Lignos

2018) ............................................................................................................................................. 33

Figure 2.26: Failure modes observed in wide-flange steel MRF columns: a) symmetric flange

buckling; b) anti-symmetric local buckling; c) and coupled buckling (Ozkula et al. 2017) ........ 34

Figure 2.27: Out-of-plane notional load applied at the tier level .................................................. 36

Figure 2.28: Analysis cases prescribed by AISC Seismic Provisions for columns, beams, and

connections ................................................................................................................................... 38

Figure 2.29: Progression of brace buckling and yielding in MT-SCBFs (AISC 2016a) .............. 40

Figure 2.30: Forces arise from out-of-plane brace buckling (AISC 2016a) ................................. 41

Figure 3.1: Three-dimensional schematic of a single storey building with two-tiered CBFs (roof

trusses not shown), figure not to scale .......................................................................................... 43

Figure 3.2: Geometry of the prototype frame ............................................................................... 43

Figure 3.3: Design response spectrum .......................................................................................... 45

Figure 3.4: Brace loading scenarios: a) Analysis case A; and b) Analysis case B ....................... 48

Figure 3.5: Column buckling mode shapes: a) In-plane mode; and b) out-of-plane mode .......... 49

Figure 3.6: Brace loading scenario for Analysis case C ............................................................... 51

Figure 3.7: Analysis case C: a) frame deformed shape ; b) frame free-body diagram; c) column

free-body diagram; d) column shear force diagram; and e) column bending moment diagram under

∆𝑽𝒃𝒓 ............................................................................................................................................. 53

Figure 3.8: Strut-to-column connection a) 2010 design; and b) 2016 design ............................... 55

xii
Figure 3.9: Frame deformed shape at expected storey drift from analyses .................................. 58

Figure 3.10: Prototype frame designed in accordance with the a) 2010 AISC Seismic Provisions;

and b) 2016 AISC Seismic Provisions .......................................................................................... 59

Figure 4.1: Three-dimensional 4-node general-purpose shell element with reduced integration. 61

Figure 4.2: One-dimensional representation of the hardening in the nonlinear isotropic/kinematic

model (Adopted from Dassault Systèmes 2014) .......................................................................... 63

Figure 4.3: Finite element model of the isolated HSS 127×127×7.9 member ............................ 64

Figure 4.4: HSS brace mesh across the section wall with various divisions a) one shell element; b)

two shell elements; c) four shell elements; d) eight shell elements; e) and twelve shell elments 64

Figure 4.5: Axial force–axial deformation response of the HSS127×127×7.9 brace under

incremental cyclic loading (test data by Jiang 2012) .................................................................... 65

Figure 4.6: Lateral displacement at buckling using different mesh size elements ....................... 66

Figure 4.7: Simulation of flexural plastic hinging and local buckling of HSS127×127×7.9 δ =

4.5δy .............................................................................................................................................. 67

Figure 4.8: Finite element model of the steel two-tiered concentrically braced frame ................ 67

Figure 4.9: Constrain applied at the base of the frame ................................................................. 68

Figure 4.10: Boundary conditions assigned to the columns a) in the plane of the frame; and b) out

of the plane of the frame ............................................................................................................... 69

Figure 4.11: Two-tiered CBF with adjacent leaning column ....................................................... 70

Figure 4.12: In-plane and out-of-plane initial geometric imperfections: a) CBF elevations; and b)

CBF side view (deformations magnified) ..................................................................................... 71

Figure 4.13: Residual stress pattern assigned to Wide-flange sections ........................................ 73

xiii
Figure 4.14: Gravity analysis step ................................................................................................ 73

Figure 4.15: Pushover loading protocol ........................................................................................ 74

Figure 4.16: Inertia point masses assigned to the braced frame model ........................................ 77

Figure 4.17: HSS 89×89×6.4 brace hysteretic response under 1980 Irpinia, Italy (SCC1) ground

motion record using a) full density; and b) 0.1% density ............................................................. 78

Figure 4.18: Influence of column density on the out-of-plane moment under the 1980 Irpinia, Italy

(SCC1) ground motion record ...................................................................................................... 79

Figure 5.1: Normalized lateral load–lateral displacement response: (a) 2010 design; and (b)2016

design, (V is the design base shear as defined in Section 3.3)...................................................... 82

Figure 5.2: Frame deformed shape: (a) 2010 design: column buckling at story drift 2.0%; and (b)

2016 design final deformed shape at story drift 2.1% .................................................................. 83

Figure 5.3: Tier drifts: a) 2010 design; and b) 2016 design.......................................................... 84

Figure 5.4: Normalized brace axial force: (a) continuous braces of the 2010 design; (b)

discontinuous braces of the 2010 design; (c) continuous braces of the 2016 design; (d)

discontinuous braces of the 2016 design ...................................................................................... 85

Figure 5.5: Column in-plane bending moments for a) 2010 design; and b) 2016 design............. 87

Figure 5.6: Deformed shape with Von-Mises stress contour at the verge of RHS column buckling

at 2.0% storey drift ........................................................................................................................ 87

Figure 5.7: Column out-of-plane bending moments for a) the 2010 design; and b) the 2016 design

....................................................................................................................................................... 88

xiv
Figure 5.8: 2010 frame deformed shape under 1994 Northridge (SCC14) ground motion record: a)

onset of LHS column buckling at t = 4.70 s and story drift of 1.7%; b) LHS column buckling at t

= 5.38 s; and c) frame collapse at t = 5.49 s ................................................................................. 89

Figure 5.9: 2010 design a) maximum storey drift under 40 ground motion records; and b) storey

drifts at buckling under 13 collapsed ground motion records....................................................... 90

Figure 5.10: 2010 design response under 2007 Pisco, Peru (SCI7) ground motion record: a) axial

force–axial displacement response of LHS column; and b) base shear–storey drift response(frame

lateral shear force is normalized by the design base shear, V; and dots represent the instant of

column buckling) .......................................................................................................................... 92

Figure 5.11: 2010 design: LHS column buckling at 34.7 s and 1.6% storey drift under 2007 Pisco,

Peru (SCI7) ground motion record ............................................................................................... 94

Figure 5.12: 2016 design: maximum storey drifts under the 40 ground motion records.............. 94

Figure 5.13: 2010 design a) Tier 1 drift; and b) Tier 2 drift ......................................................... 95

Figure 5.14: 2016 design a) Tier 1 drift; and b) Tier 2 drift ......................................................... 96

Figure 5.15: Brace axial force response in Tier 1 and Tier 2 under 2007 Pisco, Peru (SCI7) ground

motion record: a) 2010 design; and b) 2016 design ...................................................................... 97

Figure 5.16: Local buckling of the first tier brace in 2010 design under 2011 Tohoku, Japan (SCI3)

ground motion record at 9.7 s 1.5% storey drift and 2.6% tier drift ............................................. 98

Figure 5.17: Column in-plane bending moment demand under the 2011 Southern Peru (SCI16)

ground acceleration record: a) 2010 design; and b) 2016 design ............................................... 101

xv
Figure 5.18: Column in-plane bending moments for the LHS column under a) the 1994 Northridge

(SCC14) ground motion record; and b) the 2001 Southern Peru (SCI16) ground motion record

..................................................................................................................................................... 103

Figure 5.19: Column in-plane bending moments for 2016 design ............................................. 104

Figure 5.20: Column out-of-plane bending moments for a) 2010 design; and b) 2016 design .. 104

Figure 5.21: Brace out-of-plane buckling (deformations magnified) ......................................... 105

Figure 5.22: Moments in the X- and Y-directions produced by plastic hinging of the brace gusset

plate ............................................................................................................................................. 106

Figure 5.23: History of column out-of-plane bending moment under 1980 Irpinia, Italy (SCC1)

ground motion record .................................................................................................................. 107

Figure 5.24: Column out-of-plane moment history under the 1980 Irpinia, Italy (SCC1) ground

motion record .............................................................................................................................. 107

Figure 5.25: Column out-of-plane bending moment for the 2016 design under the 2011 Tohoku,

Japan (SCI6) ground motion record ............................................................................................ 108

Figure 5.26: Column axial force for a) 2010 design; and b) 2016 design .................................. 108

Figure 6.1: Proposed adjusted brace resistances when yielding propagates between tiers: a) brace

tensile yielding has just initiated in Tier 2; and b) brace tensile yielding has just initiated in Tier 1

..................................................................................................................................................... 113

Figure 6.2: Slenderness effect on the difference between Cexp and C’exp ................................... 114

Figure 6.3: a) Deformed shape of MT-CBF; b) deformed shape of compression column caused by

the notional out-of-plane load; c) column out-of-plane shear diagram; and d) column out-of-plane

moment diagram ......................................................................................................................... 115

xvi
Figure 6.4: Proposed adjusted brace force scenario to verify tier drifts ..................................... 118

Figure 6.5: Proposed analysis case C for the moment calcaulation including............................ 119

xvii
List of Symbols

A Cross-sectional area

Ag Gross cross-sectional area

B1 Multiplier for P-δ effects (AISC 360)

B Width of flange

Cd Deflection amplification factor (ASCE 7)

Cexp Brace expected strength in compression (AISC 341)

C’exp Brace expected post-buckling strength (AISC 341)

Cs Seismic response coefficient (ASCE 7)

D Depth of the section

Fa Acceleration-based site coefficient

Fcr Column inelastic buckling stress (AISC 360)

Fe Column elastic buckling stress (AISC 360)

Fcre Inelastic buckling stress for brace flexural buckling using expected yield strength

(AISC 360)

Fv Velocity-based site coefficient

Fy Yielding stress of steel

xviii
h Height of column or frame

hi Height of tier

Ie Importance factor of the structure (ASCE 7)

K Effective length factor of member

Kl Effective length

Kl/r Effective slenderness ratio or Brace slenderness limit

Mc Column seismic induced in-plane flexural bending moment

Mn Nominal flexural strength

Mp Nominal plastic flexural strength

Mv Higher mode effect factor

P Column axial force

Pn Column nominal compressive strength

Py Column nominal axial yield strength

R Response modification factor (ASCE 7)

r Radius of gyration

Rd Ductility related force modification factor (NBCC)

Ro Overstrength-related force modification factor (NBCC)

SDS Design spectral response acceleration parameter for short periods (ASCE 7)
xix
SD1 Design spectral response acceleration parameter at 1 s (ASCE 7)

SMS MCER spectral response acceleration parameter for short periods (ASCE 7)

SM1 MCER spectral response acceleration parameter at 1 s (ASCE 7)

Texp Brace expected strength in tension (AISC 341)

Ta Fundamental period of vibration of the structure

tf Thickness of the flange

tw Thickness of the web

V Design storey shear

Vc Column shear force

W Seismic weight

Z Plastic section modulus of the column

β Coefficient for bending in beam-columns

λ Slenderness parameter

φ Resistance factor

xx
Abbreviations

AISC American Institute of Steel Construction

ASCE American Society of Civil Engineers

BFCs Braced Frame Columns

BFs Braced frames

BRBF Buckling restraint braced frames

CBFs Concentrically braced frames

CSA Canadian Standards Association

EBFs Eccentrically braced frames

GCs Gravity columns

HSS Hollow Square Section

LD Limited ductility

LHS Left-hand-side

MD Moderately ductile

OCBF Ordinary concentrically braced frames

MCER Maximum considered earthquake

MPa Mega Pascal

xxi
MRFs Moment resisting frames

MT-BFs Multi-tiered braced frames

MT-BRBFs Multi-tiered buckling restraint braced frames

MT-CBFs Multi-tiered concentrically braced frames

MT-OCBFs Multi-tiered ordinary concentrically braced frames

MT-SCBFs Multi-tiered special concentrically braced frames

NBCC National Building Code of Canada

NLRH Nonlinear response history

OpenSees Open system for earthquake engineering simulation

RHS Right-hand-side; Rectangular Hollow Section

SCBF Special concentrically braced frames

SDOF Single-Degree-Of-Freedom

xxii
Chapter 1 – Introduction

1.1 Background

Multi-tiered concentrically braced frames (MT-CBFs) are widely used in North America as the

lateral load-resisting system of tall single-storey buildings such as airplane hangars, recreational

facilities, shopping centres, and industrial buildings (Figure 1.1). MT-CBFs consist of multiple

concentrically braced tiers along the height of the frame separated by horizontal strut members.

Various bracing configurations can be used in MT-CBFs including X-, V-, diagonal, chevron, and

split-X bracing (Figure 1.2). Bracing members of MT-CBFs can be designed as a

tension/compression (T/C) or tension-only (T/O) member. Multi-tier arrangements are typically

used when it is not practical or economical to use a single bracing panel along the height of the

frame between the ground and roof levels. By introducing multiple bracing tiers stacked on top of

each other, the length of braces is reduced, which allows smaller brace sizes to be used and easily

satisfy code-specified brace slenderness limits. The column buckling length in the in-plane

direction is also reduced due to the application of intermediate horizontal struts, which permits

selection of a smaller column section. From the seismic design perspective, the stringent limits on

width-to-thickness and global slenderness ratios can be easily satisfied when using shorter braces.

Moreover, when capacity design is imposed, reduced brace sizes result in smaller design forces on

the adjacent forced-controlled members including struts, beams, columns, and connections.

1
 a) b)

Figure 1.1: a) four-tiered concentrically braced frame in an industrial building; b) two-tiered

concentrically braced frame in an industrial building.

Figure 1.2: MT-CBFs configurations a) X-shape; b) V-shape; c) diagonals; d) chevron; and

e) split-X

2
MT-CBF columns are typically wide-flange (W-shape) sections oriented such that out-of-plane

bending moments act about the major axis of the section to resist the out-of-plane wind load along

the full height of the frame. No out-of-plane bracing exists along the height of the column, and the

column buckling length is taken equal to the full frame height in this direction. However, the

columns can be considered braced in the plane of the frame because of horizontal struts. It should

be noted that hollow structural sections (HSSs) are also used in multi-tiered braced frames as

columns in low and moderate seismicity regions of Canada.

1.2 Research Problem

Previous studies on the seismic performance of multi-tiered concentrically braced frames have

shown that frame lateral deformations under seismic loads are not uniformly distributed along the

height. Rather, the lateral deformation tends to concentrate in the tier where the brace tensile

yielding takes place first (Imanpour et al. 2012a; 2012b; and Imanpour and Tremblay 2012). Such

differential drift demands cause high in-plane flexural demands on the columns, which may lead

to flexural yielding, column buckling and even frame collapse. Additionally, past studies showed

that the tier where tensile yielding occurs first (i.e. critical tier) can be identified by comparing the

storey shear resistance obtained from probable brace resistance. The tier with the least shear

resistance will often yield first and prevent the increase in the lateral-load carrying capacity of the

frame, which in turn prevents the initiation of tensile yielding in other (non-critical) tiers. Non-

uniform yielding is also expected when identical tiers with the same storey shear resistance are

used. This is because of the variations in the brace geometry, initial out-of-straightness and

material properties between tiers that can affect the sequence of brace yielding and buckling.
3
Special seismic design provisions were introduced for the first time in the 2009 edition of the

Canadian steel design standard (CSA S16) (CSA 2009). The requirements were updated in 2014

edition of the S16 standard (CSA 2014) to reflect the findings of the numerical studies reported in

Imanpour et al. (2012a; 2012b) and Imanpour and Tremblay (2012). Based on these provisions, in

addition to considering the axial force, MT-CBF columns are required to resist in-plane and out-

of-plane flexural demands arising from the difference in tier-drifts and column initial

imperfections and brace out-of-plane buckling, respectively.

In the 2010 AISC Seismic Provisions (AISC 2010a), MT-CBFs were classified as K-type braced

frames, a braced-frame configuration in which the braces connect to a column at a location with

no out-of-plane support. K-type frames were not permitted for seismic applications due to the

unbalanced brace force induced on the columns between supports, which can result in large in-

plane bending moments. This response is illustrated in Figure 1.3a. However, designers were able

to mitigate this effect by including struts between columns (Figure 1.3b), and design MT-CBFs as

conventional multi-storey CBFs. The columns of such MT-CBFs were designed only for axial

compression force resulting from gravity loads plus brace axial forces due to seismic load effects

as required by 2010 AISC Seismic Provisions

The results obtained from past numerical studies in Canada and the U.S. (Imanpour et al. 2016a;

Imanpour and Tremblay 2016; Stoakes and Fahnestock 2016) have been adopted by the 2016

edition of the AISC Seismic Provisions (AISC 2016a). In this standard, new seismic design

requirements were introduced for the first time for multi-tiered ordinary concentrically braced

frames (MT-OCBFs), multi-tiered special concentrically frames (MT-SCBFs), and multi-tiered

buckling-restrained braced frames (MT-BRBFs). The 2016 AISC Seismic Provisions introduce a

4
set of design requirements for struts and columns of MT-SCBFs to protect the columns and force

yielding of braces in more than one braced panel. Similar to the Canadian standard, in addition to

considering the axial force, the MT-CBF columns are required to resist in-plane bending moment

demands arising from the difference in tier drifts and out-of-plane bending moment demands due

to column initial imperfections and brace out-of-plane buckling. Furthermore, columns are

required to be torsionally braced at each tier level to reduce the tendency of the column to twist at

the strut level.

a) b)

Figure 1.3: Inelastic response of a) K-type CBF; and b) standard CBF

Although significant improvements have been made over the past decade to develop seismic

design procedures for steel MT-CBFs, there is still a lack of background research into seismic

response and design of such frames. In particular, there is very limited detailed nonlinear numerical

analysis and no full-scale experimental test data to fully understand the instability observed in MT-

CBF columns designed excluding the special seismic provisions. Furthermore, the moments
5
induced in the columns as a result of the non-uniform yielding need to be characterized, and the

adequacy of the design requirements adopted by North American design standards (CSA 2014;

AISC 2016a) should be validated.

1.3 Objectives

The objective of this M.Sc. research is to evaluate the seismic performance of steel multi-tiered

concentrically braced frames with the focus on frames designed in accordance with the provisions

prescribed by the U.S. steel design standard (2016 AISC Seismic Provisions) for multi-tier special

concentrically braced frames (MT-SCBFs) using numerical methods. A set of special objectives

were defined as follows:

• To conduct a literature survey on the seismic response of steel CBFs and stability response

of steel wide-flange columns.

• To review the guidelines prescribed in the Canadian Steel Design Standard and the U.S.

Seismic Provisions.

• To examine the seismic behaviour of MT-SCBFs designed in accordance with 2010 and

2016 AISC Seismic Provisions using nonlinear static and dynamic analyses.

• To characterize the in-plane and out-of-plane bending moments induced in MT-SCBF

columns under seismic loads.

• To validate the seismic design requirements prescribed for the columns of MT-SCBFs in

2016 AISC Seismic Provisions including flexural bending moment demands and stiffness

requirements.

• To propose design recommendations to improve design practices for MT-SCBFs.

6
 • To produce nonlinear seismic analysis data for future research studies such as experimental

testing of MT-SCBFs.

Note that this research does not include the experimental testing of full-scale MT-SCBF frames;

however, this research lays the foundation for future experimental work.

1.4 Methodology

The objectives of this research were accomplished through five phases as follows:

• Phase 1: An extensive review of the existing literature was performed at the beginning and

throughout the course of the work to understand the seismic behaviour of steel

concentrically braced frames, multi-tiered braced frames, and steel wide-flange columns.

• Phase 2: A prototype frame consisting of a two-tiered SCBF was designed in accordance

with the 2010 and 2016 American Seismic Provisions to evaluate its global and local

response.

• Phase 3: A detailed numerical model using a finite element program was developed to

examine the behaviour of MT-SCBFs. Special considerations were made to simulate the

nonlinear response of the members including braces, columns, beams, and struts.

• Phase 4: A cyclic nonlinear static (pushover) analysis and a nonlinear response history

analysis were performed on the prototype frames of Phase 2.

• Phase 5: The results obtained from numerical analyses were examined to evaluate the

seismic response of the prototype frames, determine the seismic demands on the columns,

validate the current design requirements for MT-SCBFs and where possible, make

recommendations to enhance the design procedure.

7
1.5 Thesis Organization

This M.Sc. thesis is presented in seven chapters. Chapter 1 consists of the introduction and

background information. In Chapter 2, a survey of past studies on the seismic response of

conventional steel CBFs, MT-CBFs, steel wide-flange columns as well as a review of current

design requirements are presented. Chapter 3 discusses the design of a prototype frame in

accordance with the 2010 and 2016 AISC Seismic Provisions. The development of the numerical

model including the element type, material model, imperfections and analysis methods are then

outlined in Chapter 4. In Chapter 5, the results from the cyclic pushover analyses and the nonlinear

response history analyses are presented. In Chapter 6 recommendations for the design of MT-

CBFs are proposed. Finally, a summary of the work completed and the main conclusions are

outlined in Chapter 7; in addition, recommendations for future research are made in this chapter.

Appendix A contains the data corresponding to each ground motion for the nonlinear response

history analysis.

8
Chapter 2 – Literature Review

2.1 General

The objective of this Chapter is to review past studies on the seismic response of steel

concentrically braced frames (CBFs) plus the seismic design provisions currently used in Canada

and the U.S. to design steel multi-tiered concentrically braced frames (MT-CBFs). The nonlinear

cyclic behaviour of steel braces used in CBFs is described first. The seismic response of

conventional steel CBFs with the focus on MT-CBFs is then presented. Finally, the seismic design

provisions for steel MT-CBFs as prescribed by the Canadian steel design standard CSA S16 and

the AISC Seismic Provisions are presented.

2.2 Seismic Behaviour of Steel Braces

2.2.1 Brace Hysteretic Response

The bracing members of CBFs are an essential component of the seismic-force-resisting system

(SFRS). The bracing members have two main purposes 1) to transfer lateral loads from the top of

the frame to the foundation, and 2) to dissipate the energy introduced into the SFRS from an

earthquake. Several studies have examined the cyclic behaviour of steel bracing members (Jain et

al. 1980; Popov and Black 1981; Lee and Goel 1987; Bertero et al. 1989; Shaback 2001; and

Tremblay et al. 2003). The cyclic behaviour of steel braces is characterized by a non-symmetrical

hysteretic behaviour as illustrated in the axial force–axial deformation curve shown in Figure 2.1,

where P is the axial load and 𝛿 is the axial displacement. As shown, the compression capacity of

9
the brace is considerably lower than the tension capacity. The compression capacity degrades

significantly after its first buckling point (Cexp) is reached until reaching the expected post-buckling

load (C’exp). In comparison, once a brace reaches its tensile yielding load (AgFy), the brace can

uphold the high-tension load even after further axial displacement is applied.

Figure 2.1: Hysteretic response of steel bracing members

There are different factors that can influence the hysteretic behaviour of the bracing members,

such as slenderness, width-to-thickness, type of loading, and connection details. The slenderness

is the most influential parameter that dictates the overall behaviour of bracing members and its

energy dissipation capacity when subjected to cyclic loading as shown in Figure 2.2 (Tremblay

2002). Past experimental studies have also shown a strong correlation between slenderness and the

10
capacity of a member to dissipate energy. When the brace slenderness increases, the dissipation

capacity decreases.

Figure 2.2: Influence of global slenderness ratio on the hysteretic response of steel braces
(Ziemian 2010)

2.2.2 Bracing Configuration

The type of framing system can also affect the hysteretic behaviour of the bracing members. Plastic

hinge forming in the middle of the beam of chevron frames upon compression brace buckling

(Figure 2.3) can lead to more pronounced buckling and yielding in bracing members (Tremblay

2002). In X-braced frames, braces tend to undergo larger deformations over a shorter length in

contrast to single-bracing, resulting in higher plastic rotation of the braces (Tremblay et al. 2003).

Figure 2.3: Inelastic response of chevron braced frame when plastic hinge forming in the
beam

11
2.2.3 Brace Cross-section

Singly-symmetric, doubly symmetric, and built-up shapes can be used as bracing members of

CBFs. The type of cross-section has a significant effect on the hysteretic behaviour of steel bracing

members. Singly-symmetric braces such as T and C shapes have been found to be less efficient

due to their susceptibility to buckling in a torsional-flexure mode (Black et al. 1980). Similarly,

built-up shapes (e.g. double angles) are less effective in resisting compression load and dissipating

energy than doubly-symmetric shapes, due to early buckling of individual members (Astaneh-Asl

and Goel 1984; Astaneh-Asl et al. 1985). In regions of high seismicity, doubly-symmetric shapes

such as hollow structural sections (HSSs) are commonly used due to their effectiveness in resisting

compression loads (Lee and Goel 1987; Liu 1987; Tang and Goel 1987; and Foutch et al. 1986).

Although, past experimental studies have shown that square or rectangular HSSs exhibit a limited

inelastic deformation capacity under cyclic loading. This limitation is caused by the high strains

that are developed in the corners of the cross section upon local buckling of the member (Gugerli

1982; Lee and Goel 1987; Leowardi and Walpole 1996; and Liu 1987). Other doubly symmetric

shapes such as round HSS and W-sections have shown to be less prone to local buckling (Fell et

al. 2009).

Rectangular and square HSS members are a popular choice for bracing members in North America

largely because low-cycle fatigue fracture can be mitigated using appropriate slenderness and

width-to-thickness ratios (Tremblay 2002); furthermore, rectangular HSS members lend

themselves to convenient connections to gusset plates. Depending on the connections and

orientation of the member, the bracing member made of HSSs can buckle in-plane or out-of-plane.

A typical connection detail is to slot the connecting gusset plate to the end of the HSS section and
12
connect the two members using a fillet weld or bolt at the intercepting regions. This connection

generally results in the braces buckling out-of-plane (Figure 2.4) by developing a flexural plastic

hinge at each of the gusset plates and at the middle of the brace.

Figure 2.4: HSS brace out-of-plane buckling (Sabelli et al. 2013)

2.2.4 Dynamic Buckling

Buckling of steel braces under real-time dynamic loading have a significant effect on the hysteretic

response of braces (Kazemzadeh Azad et al. 2018). It was shown that the buckling resistance of a

brace under dynamic loading can exceed its static buckling capacity. This is because of higher

loading-velocities, as illustrated in Figure 2.5. This phenomenon can result in significant variations

of the hysteretic response of a member, as shown in Figures 2.6a and 2.6b.

13
 Figure 2.5: Effect of loading velocity (v in mm/s) on the steel brace buckling capacity
(Kazemzadeh Azad et al. 2018)

a) b)

Figure 2.6: Hysteretic response of a steel bracing member obtained from a) quasi-static
loading, and b) dynamic loading with v = 300mm/s (Kazemzadeh Azad et al. 2018)

14
2.2.5 Brace Fracture under Cyclic Loading

There have been several studies that have examined brace fracture under cyclic loading including

Goel 1987; Lee 1988; Hassan and Goel 1991; Archambault 1995; Tremblay 2002; Tremblay et al.

2003; Shaback and Brown 2003; Fell et al. 2009; and Haddad et al. 2011. Fracture under cyclic

loading typically occurs in the following tension cycle after the member has suffered severe local

buckling (Figure 2.7) under the previous compression cycle. When HSS members are used,

fracture initiates at the corners where plastic strains are the highest, then, it propagates across the

cross-section as shown in Figure 2.7 (Tremblay et al. 2003). It is concluded from these studies

that the most influential factors that affect the fractured life of a section are the slenderness, width-

to-thickness ratio, the yield stress, and section geometry. It was also found that slender members

tend to perform better than non-slender members as shown in Figure 2.8, where 𝜆 is the slenderness

parameter and ductility value is determined the ratio between the maximum displacement and the

displacement at yield. Tests by Tremblay et al. (2003) showed that higher width-to-thickness ratios

can result in fracture of the member at low levels of ductility because of severe local buckling

expected in such cross-sections. Fell et al. (2009) reported fracture initiation for HSS members in

the range from 2% to 3% of the storey drift.

Figure 2.7: HSS members under cyclic loading: a) Local buckling; and b) fracture of (Fell et
al. 2009)
15
Figure 2.8: Influence of the brace slenderness on the fracture life of HSS members (Tremblay
2001)

There are several models that have been proposed to predict the fracture life of HSS members (Lee

and Goel 1987; Tang and Goel 1987; Hassan and Goel 1991, Archambault 1995; Fell et al. 2009;

Hsiao et al. 2013). Tremblay (2002) proposed a relationship between total ductility reached at

fracture and the brace slenderness ratio where the total ductility is the sum of the peak ductility

reached in tension and the peak ductility reached in compression. This model estimates the fracture

life of diagonal bracing members well; however, it was found to be unconservative for X-bracing

configurations (Tremblay et al. 2003). Instead, Tremblay et al. (2003) proposed to use the rotation

experienced at the brace plastic hinging as an indicator of fracture life rather than a ductility-related

parameter. As shown in Figure 2.9, rotation values at fracture can be well estimated.

Figure 2.9: rotation at fracture: test data versus values predicted (Tremblay et al. 2003)
16
Hsiao et al. 2013 proposed a fracture model for nonlinear modelling of SCBFs with HSS members.

This model accounts for the width-to-thickness ratio of the cross-section, overall slenderness ratio

of the brace, and yield strength of the bracing members. It was found that the maximum strain can

be used as the best variable to predict brace fracture as shown in Figure 2.10. It is confirmed that the

maximum storey drift before fracture occurs for square HSS braces satisfies a drift of approximately

2%.

Figure 2.10 Brace fracture limit varying with a) brace width-to-thickness ratio (KL/r =52.6
and E/Fy = 446 MPa); and b) brace slenderness ratio (b/t =11.3 and E/Fy = 446 MPa) (Hsiao
et al. 2013)

2.3 Seismic Behaviour of Steel Concentrically Braced Frames

Steel concentrically braced frames (CBFs) consist of vertical, horizontal, and diagonal members

that resist lateral wind and seismic loads through vertical truss action. An example of such braced

frames is shown in Figure 2.11. Similar to other seismic force resisting systems, the main objective

of CBFs is to prevent structural collapse and maintain the integrity of gravity load-carrying system

under major earthquake events. In steel CBFs, bracing members are designed as sacrificial

elements (seismic fuses) to yield and buckle under earthquake loads and safely dissipate seismic
17
energy. The other elements of the structure including beams, columns, roof diaphragm, footing,

anchor rods and connections are intended to remain elastic (Filiatrault et al. 2013). This design

methodology is referred to as capacity design where weaker elements are engineered in the

structure to respond in the nonlinear range of the material while the rest of the structure are

designed stronger than those elements and must remain elastic under major earthquake events.

Figure 2.11: Single-storey steel concentrically braced frame

Steel concentrically braced frames are categorized in seismic design provisiosns by their ductility,

which is defined as the system capability to undergo inelastic deformation and dissipate seismic-

input energy. The ductily-related force modification factor Rd in Canada and the response

modification factor R in the U.S. are used to represent this aspect of the structure in seismic design.

In Canada, steel concentrically braced frames are categorized into two ductility levels: limited-

ductility (LD) and moderately-ductile (MD) systems. In the U.S. there are two ductility levels for
18
concentrically braced frames: ordinary concentrically braced frames (OCBFs) and special

concentrically braced frames (SCBFs).

The expected seismic behaviour of CBFs and the concept of capacity design is described using a

single-storey chevron braced frame shown in Figure 2.12a. Under a lateral seismic load, one of the

braces will act in tension, and the other one responds in compression (Figure 2.12b). By increasing

the lateral load (Figure 2.12c), the compression member reaches its maximum compression

capacity and subsequently buckles. Beyond this point, the compression brace force will begin to

degrade. As the lateral load increases, the tension brace achieves its maximum tension capacity,

and beyond this point, tensile yielding is initiated in the tension brace. The reduction in the

compression brace force continues until the force reaches the brace post-buckling strength (Figure

2.12d).

Figure 2.12: Progression of seismic behaviour of single-storey CBF

19
To ensure that the expected forces are achieved in the bracing members of CBFs, columns, beams

and adjacent connections must resist the forces arising from the yielding and buckling of the

bracing members.

Brace connections should be designed and detailed so that braces can develop a stable cyclic

response under reversal loads expected in a major seismic event. Brace gusset plates should be

designed to allow the development of plastic rotations when the brace buckles in-plane or out-of-

plane. For out-of-plane buckling, the connection detail with a fold line on the gusset plate shown

in Figure 2.13 is recommended (AISC 2016a). The clear spacing at the end of the brace is to be

taken equal to 2 times the thickness of the gusset plate as recommended by Astaneh-Asl et al.

(1986).

Figure 2.13: Linear hinge zone for out-of-plane buckling of steel braces (Sabelli et al. 2013)

20
2.4 Seismic Behaviour of Steel MT-CBFs
Multi-tiered concentrically braced frames (MT-CBFs) are commonly used in tall single-storey

buildings, such as airport hangers, industrial warehouses, and recreation centres or tall-storey-

buildings such as hotels, residential, and office buildings. MT-CBFs are made of multiple panels

of CBFs stacked on top of each other along the height of a storey.

Different configurations of MT-CBFs exist as illustrated in Figure 2.14. Although MT-CBFs

appear to be similar to multi-storey concentrically braced frame (MS-CBFs), there are two major

distinctions between the two frames: 1) MT-CBFs lack floor diaphragms between the ground and

roof levels (Figure 2.15), which creates a single-degree-of-freedom (SDOF) system, and 2) MT-

CBF columns lack out-of-plane bracing along the height of the frame, which poses concerns on

the out-of-plane stability of such members (Figure 2.15).

21
Figure 2.14: Typical MT-CBF configurations (AISC 2016a)

22
 Figure 2.15: Comparison between Multi-tiered CBFs and Multi-storey CBF

2.4.1 Expected Seismic Behaviour

The seismic behaviour of MT-CBFs was the focus of several research programs in recent years.

The numerical seismic response evaluation using the OpenSees platform (McKenna and Fenves

2004) showed that the lateral frame deformations are not uniformly distributed along the frame

height, but rather concentrated in one of the tiers, which produces in-plane bending moments on

the columns as shown in Figure 2.16 for a two-tiered frame studied shown in Figure 2.17

(Imanpour and Tremblay 2012; Imanpour et al. 2012b).

23
Figure 2.16: Drift in individual braced panels of a two-tiered CBF and column in-plane
bending demand recorded between the braced panels (Imanpour and Tremblay 2012)

Figure 2.17: Two-tiered concentrically braced frame geometry (Imanpour and Tremblay
2012)

24
Imanpour et al. (2013) examined the seismic stability response of columns in a four-tier braced

frame designed in accordance with the 2010 AISC Seismic Provisions (AISC 2010a). The frame

was analyzed using the nonlinear response history method in OpenSees (McKenna and Fenves

2004). The study also examined the response of an isolated columns part of the four-tiered CBF

using the Abaqus program (Dassault Systèms 2012) under the displacements obtained from the

nonlinear response time history analysis in OpenSees. It was found that the frame designed

according to the 2010 AISC Seismic Provisions is prone to a concentration of inelastic drift in one

of the bracing tiers. A significant in-plane flexural demand was observed in the columns of the

frame (Figure 2.18). The study found that flexural yielding of the column could compromise the

stability condition of the column. Additionally, this study identified excessive ductility demands

on the bracing members of the tier where large inelastic deformations are concentrated.

Figure 2.18: Normal stress contour (ksi, where 1 MPa = 0.145 ksi) of the column at maximum
Tier 1 drift (Imanpour et al. 2013)
25
Stoakes and Fahnestock (2014) examined the seismic stability of the steel MT-CBF columns using

a three-dimensional finite element model. Columns were subjected to realistic axial loads and

torsional rotation due to the buckling of braces. This study was able to reassure the positive effects

of providing a torsional restraint on the columns at the tier levels; however, large tier drifts were

observed in the analyses, which could lead to brace brittle fracture due to low cycle fatigue.

Imanpour et al. (2016a) conducted a numerical parametric study to analyze the behaviour of MT-

CBFs designed in accordance with the 2010 AISC Seismic Provisions. In this study, the number

of tiers ranged between two and six, and the height ratio between tiers was also varied.

Furthermore, this study investigated the effects of column base fixity and low-ductile steel MT-

CBFs. The seismic behaviour of three braced frames was also analyzed using a three-dimensional

finite element model. The results of these three analyses were used to validate a computationally

efficient approach fibre-based numerical model to perform the parametric study. This study found

a concentration of inelastic lateral deformations that led to high in-plane flexural demands on the

columns. In several cases, column buckling was observed as shown in Figure 2.19. The buckling

mode took place about the weak-axis of the column section and changed to torsional-flexural

buckling as a result of the out-of-plane displacement mainly at the mid-height of the column as the

column has no lateral out-of-plane support along the frame height. The study showed that the MT-

CBFs designed in accordance with the 2010 AISC Seismic Provisions can exhibit an unsatisfactory

seismic response. This study also found that the use of higher seismic forces in design did not have

a significant benefit to the flexural demand produced on the columns nor prevented column

buckling. Nonetheless, using a fixed condition at the base of the column improved the seismic

response of the column and prevented column instability.

26
 Figure 2.19: Frame deformed shape at the instant of column buckling (Imanpour et al.
2016a)

Imanpour et al. (2016b) proposed new design requirements to improve the seismic response of

MT-CBFs to prevent column instability and brace fracture by propagating brace tensile yielding

between the braced panels. Based on the proposed design method, an in-plane flexural bending

moment should be included in the column design. To determine the column in-plane moment, the

tier where brace yielding takes place first should be identified. The critical tier can be identified

by comparing the storey shear resistance at brace yielding and buckling. The tier with the lowest

storey shear resistance is identified as the critical tier. Then, the column in-plane moment demand

can be estimated using the brace force scenario shown in Figure 2.20 where the tension brace in

the critical tier (Tier 1) reaches its expected tensile resistance Texp , brace force in the compression

member of the critical tier reaches the post-buckling compression resistance C’exp, while in the

adjacent tier, the tension brace is assumed to reach its expected tensile resistance Texp and the

compression brace force in the noncritical tier is set equal to the buckling resistance Cexp. If

27
multiple tiers have the same storey shear resistance, multiple analyses where the critical tier

location varies by slightly reducing (e.g. 20%) the brace yield strength. This analysis results in an

unbalanced brace storey shear force that should be resisted by the columns:

∆𝑉𝑏𝑟 = 𝑉𝑏𝑟 1 − 𝑉𝑏𝑟 2

where,

′ ′
𝑉𝑏𝑟 1 = 𝑐𝑜𝑠𝜃1 (𝑇𝑒𝑥𝑝 + 𝐶𝑒𝑥𝑝 ) and 𝑉𝑏𝑟 2 = 𝑐𝑜𝑠𝜃2 (𝑇𝑒𝑥𝑝 + 𝐶𝑒𝑥𝑝 ).

The design in-plane bending moment of the column can be calculated based on the shear force

diagram under ∆𝑉𝑏𝑟 as shown in Figure 2.21. The required strength of the column should be

verified using the interaction equation specified in the AISC Specification (AISC 2010a) under

axial compression force and weak-axis bending moment demands.

Figure 2.20: Improved lateral response of a two-tiered steel CBF: a) Brace yielding and
buckling in Tier 1 and initiation of brace tensile yielding in Tier 2; and b) Proposed brace
force scenario to trigger yielding in the noncritical tier (critical tier: Tier 1; and noncritical
tier: Tier 2)

28
 Figure 2.21:a) Column free-body diagram under unbalanced brace story shear; b) column
shear force diagram; and c) column bending moment diagram

The second requirement that Imanpour et al. (2016b) proposed was to limit the tier drift to avoid

brace fracture due to excessive deformation demands in bracing tiers. This was achieved by

providing the column with sufficient in-plane flexural stiffness. This study proposed a tier drift

limit of 2% when the storey drift is equal to an expected storey drift, ∆𝑒𝑥𝑝 , of 2.0∆, where ∆ is the

design storey drift predicted in accordance with the applicable building code (ASCE 2010). A

prototype frame designed in accordance with the proposed requirements was analyzed using

OpenSees and it was shown that the proposed requirements can effectively prevent undesirable

limit states observed in the frames designed excluding such provisions.

In another study, the seismic behaviour of MT-CBFs was examined by involving the contribution

from the adjacent gravity columns in the lateral-load carrying capacity as shown in Figure 2.22. It

29
was found that the involvement of the gravity columns can improve the seismic response of MT-

CBFs (Imanpour et al. 2016c).

Figure 2.22: Three-tiered steel braced frame with adjacent gravity columns

Two analysis methods were proposed by Imanpour and Tremblay (2016a) to predict the seismic

induced in-plane bending moment of the columns and tier drifts in tall MT-CBFs with three or

more tiers. The first method is the extension of the method initially proposed for two-tiered CBFs

and can be used when yielding propagates from the bottom tier to the top one or from top tier to

the bottom one. However, the second method involves a more general stiffness-based analysis that

can be used for various yielding scenarios.

30
2.4.2 Seismic Behaviour of Wide-Flange Steel Columns

Limited research has been conducted to understand the seismic behaviour of steel wide-flange

columns in MT-CBFs. Stoakes and Fahnestock (2012; 2016) evaluated the seismic behaviour of

isolated columns using numerical analyses. Various limit states were identified, which includes

strong axis flexural buckling, weak-axis flexural buckling, flexural-torsional buckling due to

biaxial moment demands, and flange and web local buckling. It was found that the governing limit

state depends on the force distribution on the columns. Furthermore, it was shown that weak-axis

flexural yielding of a column can significantly degrade the strong-axis buckling strength of a

column. Sections with large slenderness ratios are prone to strength degradation when weak-axis

flexural yielding has occurred.

More recently, Imanpour et al. (2017) experimentally tested wide-flange columns, which was part

of a two-tiered CBFs, using a pseudo-dynamic hybrid simulation method. The column testing

machine is shown in Figure 2.23. A W250x101 column was physically tested using the input

produced by the finite element analysis under a ground motion record. The study successfully

simulated column instability observed in the past numerical simulations.

Newell and Uang (2008) experimentally studied the cyclic behaviour of wide-flange steel columns

in multi-storey CBFs. Nine full-scale wide-flange specimens were tested under various axial force

demands in combination with applied story drifts to examine the column stability response under

cyclic loading. The local buckling was the dominant buckling mode as shown in Figure 2.24. No

global buckling was observed in any of these tests.

31
Figure 2.23: W250x101 column part of a two-tiered CBF in the multi-directional hybrid
testing system (Imanpour et al. 2017)

Figure 2.24: Local buckling of Wide-flange steel columns as part of multi-storey CBFs
(Newell and Uang 2008)

32
Elkady and Lignos (2018) performed an extensive experimental study to evaluate the seismic

stability of steel wide-flange columns in moment-resisting frames (MRF) under multi-axis cyclic

loading. Local flange and web buckling at the column base were observed. Severe local buckling

led to large axial shortening and, in some cases, out-of-plane buckling of columns as shown in

Figure 2.25. Moreover, column twist was observed at large storey drifts.

Figure 2.25: Lateral instability of steel Wide-flange column as part of MRFs (Elkady and
Lignos 2018)

33
Ozkula et al. (2017) experimentally tested more than 10 column specimens under various loading

conditions to study the seismic response of steel MRF wide-flange columns. Three instability

modes were observed: 1) symmetric flange buckling; 2) anti-symmetric local buckling and 3)

coupled buckling. An example of each of the buckling modes is presented in Figure 2.26, where

the left image is at the west side of the column, the middle image shows the overall column, and

the right image is at the east side of the column. A criterion to identify the governing buckling

mode and consequently the hysteretic response of the wide-flange steel MRF column subjected to

cyclic loading was proposed.

Figure 2.26: Failure modes observed in wide-flange steel MRF columns: a) symmetric flange
buckling; b) anti-symmetric local buckling; c) and coupled buckling (Ozkula et al. 2017)
34
2.5 Design of MT-CBFs

2.5.1 Canadian Design Provisions

2.5.1.1 2009 Canadian steel design standard (S16)

Design requirements and guidelines for the design of columns in MT-CBFs were originally

introduced in Canada as part of the 2009 Design of Steel Structures standard, S16-09 (CSA 2009).

These requirements were given under Section 27.6.6 titled ‘Columns with Braces Intersecting

Between Diaphragms.’ The requirements introduced were applicable only to limited-ductility

(LD) MT-CBFs, as these systems are expected to experience limited inelastic behaviour (RdRo =

2.0 × 1.3 = 2.6) where Rd is the ductility-related force modification factor, and Ro is the

overstrength-related modification factor. The 2009 Design of Steel Structures required a horizontal

strut to be placed between the columns at the tier levels to transfer the unbalance force, which is

created once the compression braces buckle and the tension forces increase to reach yielding. In

addition to the gravity loads, the columns were required to resist the forces induced from the

yielding and buckling of the braces at a roof displacement equal to the design storey drift, RdRo∆𝑒 ,

assuming that yielding has occurred in the tension brace located at any tier of the braced frame

(where ∆𝑒 is the elastic storey drift under the base shear of the corresponding building code).

Furthermore, columns in MT-CBFs were required to resist the effects of an out-of-plane transverse

load acting on the column at each tier level. Each of these loads was equivalent to 10% of the load

in the compression member meeting at the intersecting point, as illustrated in Figure 2.27.

35
 Figure 2.27: Out-of-plane notional load applied at the tier level

2.5.1.2 2014 Canadian steel design standard (S16)

In the latest edition of the Canadian standard, CSA S16-14 (CSA 2014), the provisions to design

MT-CBFs have been modified to permitted moderately-ductile (Type MD) (RdRo = 3.0 × 1.3 =

3.9) in addition to limited-ductility (Type LD) frames. Type MD frames are allowed to be designed

up to three tiers high, and Type LD frames have been extended to five tiers. The design

requirements introduced in S16-09 still apply in the latest standard; however, the out-of-plane

notional load has been reduced from 10% to 2% of the compression force acting in the column

below the brace-to-column connection to reflect the finding of Imanpour and Tremblay (2012). A

strut is still required between columns at each tier level in the latest edition; however, the standard

suggests using the flexural capacity of the strut to constrain the axial rotation of the columns at

36
each connection to reduce the effective length of the column against flexural-torsional buckling.

Lastly, the 2014 Design of Steel Structures standard advises avoiding tier-drift in excess of 1.5%-

2.0%, as this can lead to premature brace fracture caused by ultra-low cycle fatigue.

Imanpour and Tremblay (2016) investigated the seismic response of MT-CBFs designed in

accordance with the 2014 Canadian standard (CSA S16-14). A set of prototype frames were

analyzed using the nonlinear response history analysis method in OpenSees. The results showed

that frames complying to the CSA S16 requirements can develop brace tension yielding in a single

tier without excessive tier drifts. The in-plane flexural bending demand can be predicted well using

the current CSA S16 requirements. The study suggested that brace tensile yielding can occur in

more than one tier and the corresponding force demands should be considered in the column

design. This study found that the in-plane flexural demand on the columns is sensitive to the storey

drift assumed in the design; therefore, more representative storey drift estimates should be used in

design to prevent undervaluing the in-plane moments on the columns. The authors also proposed

that further studies should investigate the validity of the out-of-plane notional load imposed on the

design of the MT-CBF columns in the current standard, which is evaluated in Chapter 5 of this

M.Sc. thesis.

2.5.2 Design Provisions in the U.S.

2.5.2.1 2010 U.S. Seismic Provisions for Structural Steel Buildings (AISC 341)

The 2010 American Institute of Steel Construction (AISC), Seismic Provisions for Structural Steel

Buildings (AISC 2010a), did not include design provisions for MT-CBFs. In the absence of special

design provisions, MT-SCBFs were designed based on the provisions prescribed for conventional

multi-storey braced frames. The required strength of the bracing members was calculated based
37
on the seismic base shear as determined in the applicable building code. The bracing members

were also sized to meet desired energy dissipation capacities and to increase the fracture life by

limiting the slenderness ratio (kL/r ≤ 200) and satisfying the requirements for width-to-thickness

ratios as prescribed in section D1.1 of the Seismic Provisions. The design of the other members of

the frame, such as columns, struts, and connections, was conducted using the ultimate strengths of

the bracing members using capacity design. In the 2010 Seismic Provisions, two analyses were

prescribed for SCBFs to determine the forces in the columns, struts, and connections (Figure 2.28).

Analysis Case A (Figure 2.8a) represents the brace force scenario where braces reach their

expected tensile strength (Texp) and expected compressive strength (Cexp). Case B (Figure 2.8b)

represents the brace force scenario where the braces achieved their expected tensile strength (Texp)

and expected post-buckling strength (C’exp).

a) b)

Figure 2.28: Analysis cases prescribed by AISC Seismic Provisions for columns, beams, and
connections

38
2.5.2.2 2016 U.S. Seismic Provisions for Structural Steel Buildings (AISC 341)

Requirements to design MT-CBFs were introduced for the first time in the 2016 Seismic

Provisions (AISC 2016a). Special seismic provisions were prescribed for multi-tiered ordinary and

special CBFs, as well as multi-tiered buckling-restrained braced frames. Under the current

provisions, multi-tiered special concentrically braced frames (MT-SCBFs) are required to be

analyzed under three different analysis cases as specified in Chapter F of the 2016 AISC Seismic

Provisions. The first two analyses are the same prescribed for CBFs (Figure 2.28). The third

analysis case represents the progressive yielding and buckling of braces in MT-SCBFs, which

corresponds to the initiation of tensile yielding in the weakest tiers and propagation to the strongest

(Figure 2.29). For this analysis, it is assumed that the compression brace in the critical tier (lower

tier in this case) has reached its post-buckling strength, C’exp, and the compression brace in the

adjacent tier has reached its expected buckling strength, Cexp. Concurrently, the tension braces in

the critical tier and the adjacent tier are assumed to be at their expected tension strength, Texp. This

analysis case is shown in Figure 2.20b for a two-tiered frame. The unbalance storey shear force

can be determined by analyzing the frame under applied brace loads. Then, the column in-plane

bending demand can be computed for design purposes. This estimate is considered conservative

since it is possible that the brace in the adjacent non-critical tier has already experienced several

loading cycles, which leads to a decrease in buckling strength at the time when the tension brace

yields in the critical tier (Tremblay 2002; Imanpour and Tremblay 2014a).

39
 Figure 2.29: Progression of brace buckling and yielding in MT-SCBFs (AISC 2016a)

An out-of-plane bending moment can also be induced on the columns of MT-SCBFs due to initial

geometric imperfections in columns, out-of-plane buckling of braces, and plastic hinge forming in

the gusset plate (Figure 2.30). To account for such demands, the 2016 AISC Seismic Provisions

requires an out-of-plane horizontal notional load be applied on the column at the strut level. The

notional load is equal to 0.006 times the vertical component of the compression brace that meets

the column at the tier level. In addition, the columns must be designed to resist the out-of-plane

moment that the braces produce upon buckling, but less than the maximum bending resistance of

the brace connections.

40
 Figure 2.30: Forces arise from out-of-plane brace buckling (AISC 2016a)

Additionally, the 2016 Seismic Provisions require MT-SCBF columns to be torsionally braced at

the strut-to-column connections. Stoakes and Fahnestock (2012; 2016) showed that providing

rotational bracing, along the height of the column at the strut-to-column connections, can improve

the strong-axis buckling strength in the presence of in-plane flexural yielding, particularly when

the location of weak-axis flexural moment matches the location of the strong-axis flexural moment

(e.g. two-tier braced frame with identical tier heights). The 2016 AISC Seismic Provisions also

require that a strut is placed between two tiers to prevent the unsatisfactory K-brace frame

response. Finally, the provisions have established a maximum tier drift ratio of 2% to prevent

excessive brace deformations that can cause brace fracture (Tremblay et al. 2003).

41
Chapter 3 – Design of Steel Multi-Tiered Concentrically
Braced Frames
3.1 General

The purpose of this chapter is to present the design of a prototype frame using the 2010 and the

2016 AISC Seismic Provisions (AISC 2010a, 2016a). First, the building and frame geometry are

introduced, followed by frame loading. Finally, the detailed design calculations for the prototype

frame members including braces, columns, and strut are presented.

3.2 Building and Frame Geometry

A single-storey steel building located in Seattle, Washington, U.S., was selected for the case study.

The building has plan dimensions of 35 m x 189 m, and a height of 9.0 m. In each principal

direction, the building has four concentrically braced frames (two per each exterior walls) as shown

in Figure 3.1. The frame height is divided into two tiers with X-bracing configuration. As

illustrated in Figure 3.2, the bottom tier, Tier 1, is 4.7 m tall, and the top tier, Tier 2, is 4.3 m tall.

Non-uniform tier heights were selected to intentionally reduce the storey shear resistance of one

of the tiers (Tier 1) such that under the lateral load its braces yield first. This tier is referred to as

the ‘critical tier’ or the ‘weakest tier.’ This non-uniformity of storey shear resistance may occur in

an actual building with identical tier heights as a result of various material properties, brace end

conditions or initial geometric imperfections (Schmidt and Barlett 2002). The braced frame was

designed as a Special Concentrated Braced Frame (SCBF) system. The braces were sized to carry

the lateral seismic load in tension and compression.

42
Figure 3.1: Three-dimensional schematic of a single storey building with two-tiered CBFs
(roof trusses not shown), figure not to scale

Figure 3.2: Geometry of the prototype frame

43
The height of each column is 9.0 m. The columns are made of W-shapes and are oriented such that

the out-of-plane bending occurs about the strong-axis of the section as seen above in Figure 3.2.

A 7.0 m long intermediate horizontal strut is placed between tiers to prevent K-braced frame

response (AISC 2010a; 2016a) and ensure the seismic load is properly transferred to the base of

the structure through truss-action once the braces respond in inelastic range.

3.3 Gravity and Seismic Loads

The design loads for the selected building were determined in accordance with the ASCE 7-16

standard (ASCE 2016). A Risk Category II was chosen, and it was assumed that the building is

located on a Site Class C with a Seismic Design Category D. The gravity loads were calculated

using the roof dead load Droof = 1.0 kPa, the exterior wall dead load Dwall = 0.5 kPa based on a light

cladding structure resting on the ground, and a live load L = 0.96 kPa as prescribed by ASCE 7.

The tributary area considered per column was calculated on the basis that steel roof trusses support

the roof system between the exterior columns of the building. The resulting gravity factored load

at the top of each column was then calculated to be 227 kN.

The seismic load parameters include a response modification factor R = 6.0, overstrength factor

Ωo = 2, and a deflection amplification factor Cd = 5.0. The mapped risk-targeted Maximum

Considered Earthquake (MCER) ground motion response parameters, SS = 1.362g and S1 = 0.458g

for short and 1.0 s periods, respectively, were used to obtain the design spectral response

acceleration parameter SDS = 0.908g and SD1 = 0.458g. Design Response Spectrum is shown in

Figure 3.3. The empirical fundamental period was calculated using Ct = 0.0488 and x = 0.75, and

44
is equal to Ta = 0.25 s (red dash line in Figure 3.3). Using these values, the seismic design

coefficient Cs = 0.151 was obtained. The seismic weight of the building W is equal to 7624 kN,

based on the roof and the exterior wall dead loads. The equivalent lateral force procedure was used

to calculate the frame seismic base shear V, which is the product of the seismic coefficient and the

seismic weight tributary of the frame. This force was amplified to account for accidental torsion,

resulting in a seismic design base shear equal to 316 kN per frame.

Figure 3.3: Design response spectrum

3.4 Brace Design

The braces in both tiers were designed to resist the seismic load effects in tension and compression.

The brace design force in compression is equal to Pr,b = 200 kN, which includes the seismic-

induced axial force PE,b = 191 kN plus the gravity induced axial compression force PG,b = 9 kN.

The braces are designed using square Hollow Structural Section (HSS) members. Such members

45
are more efficient than singly-symmetric sections as they have an identical radius of gyration about

both principal axes of the section (Black et al. 1980). The braces are made of ASTM A1085 Grade

A steel (ASTM 2015a) with a yield stress Fy = 345 MPa and an expected yield stress RyFy = 431

MPa as specified in the AISC Seismic Provisions (AISC 2016a). Braces were designed such that

they buckle out of the plane of the frame. An effective length of 0.45 times the total length of the

brace, which is measured between the brace working points, was used in design to account for the

lateral bracing provided by the brace acting in tension and the length of the end connections that

do not contribute to brace buckling (Wakabayashi et al. 1974; Nakashima and Wakabayashi 1992;

and El-Tayem and Goel 1985; 1986; Sabelli et al. 1999). The brace axial compression resistance

in Tier 1 was calculated to equal to Pc,b1 = 215 kN using the AISC Specification for Structural

Steel Buildings (AISC 2010b). Although the brace lengths are slightly different between tiers, an

identical HSS 89×89×6.4 section was selected for both tiers as typically done in practice to

maintain a similar connection size. The selected section complies with the width-to-thickness ratio

limit b/t < 14 for highly ductile members where b is the effective width of the member and t is the

thickness, and with the slenderness limit of Lc/r ≤ 200, where Lc is the effective length and r is the

radius of gyration.

3.5 Column Design

3.5.1 Design in Accordance with the 2010 AISC 341 Seismic Provisions

The columns were first designed in accordance with the 2010 AISC Seismic Provisions. The frame

with the selected columns is referred to as the 2010 design. The columns were designed to resist

46
the gravity loads PG,c = 227 kN plus the maximum axial load induced by the summation of the

vertical forces due to the brace expected resistances in tension and compression. For the later, two

analysis cases, A and B, are prescribed by the 2010 AISC Seismic Provisions as shown in Figures

3.4a and 3.4b with their respective loading. Analysis Case A represents a brace load scenario where

braces reach their expected tensile strength (Texp) and expected compresison strength (Cexp). Case

B represents a brace load scenario where braces reach their expected tensile strength (Texp) and

expected post-buckling strength (C’exp). The maximum axial compression force, PE,c = 1103 kN,

was obtained under the first analysis case. The columns are made of ASTM A992 Grade 50 steel

(ASTM 2015b) with yield stress Fy = 345 MPa. The effective length of the column in the strong-

axis Kx = 0.84h, weak-axis Ky = 0.80h1, and torsion Kz = 1.0h1 were used in design where h is the

total frame height, and h1 is the height of Tier 1 as shown in Figure 3.2. The effective length factors

were derived using the structural analysis software (S-Frame 2017) where an individual column

was modelled using frame elements, and the corresponding gravity and seismic loads were applied

at the top of the column and at the strut level. An elastic Eigen buckling analysis, also referred to

as Linear Perturbation Frequency analysis, was then performed (Figure 3.5). The eigenvalues were

used to calculate the effective length of the column about its strong and weak axes. An effective

length smaller than unity was used to account for the distributed axial load applied on the MT-

CBF column segments (Dalal 1969). A W410×67 section was finally selected for the columns of

the 2010 design to achieve the most efficient cross-section. The column axial resistance was

obtained from the AISC Specification Equation E3-1 is equal to Pn = 1391 kN. The web and flange

width-to-thickness ratios were verified to comply with a moderately ductile W-shape in

47
accordance with the 2010 Seismic Provisions, such that b/t ≤ 9.1 and h/t ≤ 49.3 were satisfied, for

the flange and the web of the section, respectively.

Figure 3.4: Brace loading scenarios: a) Analysis case A; and b) Analysis case B

48
 Figure 3.5: Column buckling mode shapes: a) In-plane mode; and b) out-of-plane mode

3.5.2 Design in Accordance with the AISC 341-16 Seismic Provisions

The columns of the prototype frame (Figure 3.2) were redesigned according to the AISC 341-16

Seismic Provisions, which is referred to as the 2016 design. The required strength of columns was

determined considering the maximum forces obtained from three analyses cases A, B, and C, as

illustrated in Figures 3.4a, 3.4b and Figure 3.6. Analysis case C represents the progressive buckling

and yielding of the braces from the critical tier to the non-critical tier, where the bracing members

reach their expected tensile strength (Texp) and expected compression strength (Cexp) in the non-

critical tier and the expected post-buckling strength (C’exp) in the critical tier. This analysis case

was introduced in the 2016 AISC Seismic Provisions to address the unsatisfactory limit states

observed in the past numerical analyses (Imanpour et al. 2016a; 2016b). Variations in brace

material properties, brace slenderness ratio, brace end conditions and initial out-of-straightness can

trigger brace buckling and subsequently yielding in one of the braced tiers. It was shown that such

response can impose large in-plane flexural bending moment demands on the braced frame
49
columns, which may lead to column instability in the presence of axial compression force demands

if not considered in design. As shown in Analysis case C, the columns should also be verified

under the forces arising from the progressive yielding of bracing members along the height of the

frame. For the two-tiered CBF studied here, brace tensile yielding is expected to initiate in the first

tier, which has a lower expected storey shear resistance (Tier 1) and propagates to Tier 2 with a

higher expected storey shear resistance. Under the brace loading scenario corresponding to

analysis case C (Figure 3.6), the braced frame column experiences the most critical seismic force

demands that includes 1) axial compression force due to the brace expected resistances in tension

and compression plus the axial compression force induced by gravity loads; and 2) in-plane

bending moment caused by uneven yielding of braces in two adjacent tiers and is obtained from

the differences between the expected storey shear resistances of adjacent tiers as described in

Section 2.5.2.2. The 2016 AISC Seismic Provisions also requires an additional out-of-plane

bending moment demand arising from the brace out-of-plane buckling, flexural plastic hinging of

brace connections, column initial out-of-straightness. Similar to the 2010 design, the maximum

axial compression force is induced in the first tier segment of the right-hand-side column and is

equal to Pc,c = 1330 kN due to brace expected forces PE,c = 1103 kN and the gravity load PG,c =

227 kN.

50
 Figure 3.6: Brace loading scenario for Analysis case C

The first step in the calculation of the column in-plane bending moment is to identify the critical

tier. The critical (or weakest) tier is the tier with the least expected storey shear resistance. The

shear resistance is obtained from the summation of the horizontal components of the brace

resistances in tension and compression Vexp = (Texp + Cexp) cosθ, where θ is the angle between the

brace and the horizontal plane. In cases where tier heights and members are the same, tiers will

have different shear resistance due to inherent differences as a result of connection details, material

variability (Schmidt and Barlett 2002), or geometric imperfections. In that case, multiple analyses

should be performed by switching the critical tier. For the frame of Figure 3.2, expected storey

shear resistance in Tier 1 is Vexp,1 = 970 kN, which is smaller than that of Tier 2 Vexp,2 = 1014 kN.

The column in-plane bending demand Mry is then calculated using unbalanced brace storey shear

force ΔVbr using Equation 3.1 as follows:

∆𝑉𝑏𝑟 ℎ1 ℎ2
𝑀𝑟𝑦 = (3.1)
2 ℎ

51
where h1 and h2 are the height of Tiers 1 and 2, respectively, as shown in Figure 3.2. ΔVbr is

computed as follows:

′
∆𝑉𝑏𝑟 = (𝑇𝑒𝑥𝑝 + 𝐶𝑒𝑥𝑝 )2 𝑐𝑜𝑠𝜃2 − (𝑇𝑒𝑥𝑝 + 𝐶𝑒𝑥𝑝 )1 𝑐𝑜𝑠𝜃1 (3.2)

For Analysis case C, as shown in Figure 3.6, ΔVbr = 227 kN and the corresponding in-plane

bending moment on the columns is Mry = 254 kN-m. Figure 3.7 illustrates how the in-plane flexure

demand is derived from Analysis case C.

Column out-of-plane bending moment demand is obtained from two components, the moment

induced by applying an out-of-plane horizontal notional load at the strut level that is 0.006 times

the vertical load contributed by the compression brace (amplified by multiplier B1 = 1.16 to account

for the P-δ effect) and the respective component of the moment caused by the buckling of braces

in the out-of-plane direction (Figure 2.30), where the moment induced by the buckling of the

braces is equal to 1.1RyMp/𝛼𝑠 where Ry is the ratio of expected yield stress to the specified

minimum yield stress, Mp is the plastic bending moment of the minimum between the compression

brace and the brace connection, and 𝛼𝑠 is the LRFD force level adjustment factor and is taken

equal to 1.0. For the prototype frame, the total out-of-plane bending moment demand is equal to

5.7 kN-m.

52
a) b) c) d) e)

Figure 3.7: Analysis case C: a) frame deformed shape ; b) frame free-body diagram; c) column free-body diagram; d) column
shear force diagram; and e) column bending moment diagram under ∆𝑽𝒃𝒓

53
A W310×143 column was selected to carry the gravity, and seismic-induced forces described here.

This section was selected as it results in the most efficient cross-section for the column that

satisfied the strength requirements in accordance with the interaction equation (Equation 3.3)

specified in Chapter H (H1-1a) of the 2016 AISC Specification.

𝑃𝑟 8 𝑀 𝑀𝑟𝑦
+ ( 𝑟𝑥 + ) ≤ 1.0 (3.3)
𝑃𝑐 9 𝑀 𝑀
𝑐𝑥 𝑐𝑦

The flange and web width-to-thickness ratios were verified using the 2016 AISC Seismic

Provisions. The column is classified as a moderately ductile member. The width-to-thickness ratios

b/t ≤ 8.6 and h/t ≤ 52.4 were satisfied for the flange and the web of the section, respectively.

Comparing the final column sections obtained between the two designs, it is observed that in the

2016 design the presence of combined axial compression force and bi-axial bending moment

demands in two-tiered CBF columns leads to a more severe loading scenario.

3.6 Strut Design

An intermediate horizontal strut was placed between columns at each tier to resist the unbalance

load that is developed after brace buckling and yielding. For both 2010 and 2016 designs, the

maximum axial design force for the strut is obtained from Analysis case B when the tension braces

in both tiers reach Texp and compression brace forces are equal to C’exp as shown in Figure 3.4b.

Hence, the strut design force is equal to Pr,s = 663 kN.

For the 2010 and 2016 designs, the strut was designed assuming that the strut is connected to a

stub-section using a set of splice plates (Figure 3.8), which act as a simple connection and prevent

the development of in-plane moment due to partial rigidity of the gusset plate connection on the

54
strut member (Carter et al. 2016). Thus, the in-plane moment acting on the strut was calculated

based on the self-weight of the member only for both designs. In addition to the in-plane bending

moment, the 2016 Seismic Provisions requires that the strut be designed for an additional out-of-

plane bending moment induced by the corresponding component of the minimum value between

the plastic moment induced by the brace or the connection, which is caused by brace buckling out-

of-plane. This moment was calculated as 4.6 kN-m.

The struts for both designs use a W-shape conforming to ASTM A992 Grade 50 steel with yield

stress Fy, = 345 MPa. A W250×67 strut was selected in both designs to carry the design loads. For

the 2010 design, the strut was oriented such that the web is in the plane of the frame as shown in

Figure 3.8a; however, the strut web was placed in the horizontal plane for the 2016 design as

shown in Figure 3.8b to provide torsional bracing to the column at the strut-to-column connection

through the stiffness and strength provided by major axis bending of the strut (Imanpour et al.

2016b; Stoakes and Fahnestock 2014).

a) b)

Figure 3.8: Strut-to-column connection a) 2010 design; and b) 2016 design

55
3.7 Drift Check

As prescribed by ASCE 7, the design storey drift must be limited to 2.5% for an SCBFs with risk

category II. This limit was verified for both 2010 and 2016 designs. The elastic drift, ∆𝑒 , can be

calculated manually using structural analysis principles or using a structural analysis program

where the design seismic base shear is applied at the top of the frame. Thus, the design storey drift,

Cd∆𝑒 , for the 2010 design is 0.60% and for the 2016 design is 0.55%. In both cases, the storey drift

limit is satisfied.

The 2016 AISC Seismic Provisions require that in addition to limiting the storey drift to 2.5%, the

tier drift be limited to 2% to prevent premature failure of the bracing members (Tremblay et al.

2003; Fell et al. 2009; and Roeder et al. 2011). The 2016 Seismic Provisions require each tier in

an MT-SCBF to be subjected to the drift limitation of the applicable building code (ASCE 7), but

the drift shall not exceed 2% of the tier height. This requirement was satisfied by verifying the tier

drifts at the maximum anticipated storey drift 2.0𝐶𝑑 ∆𝑒 , as obtained from the nonlinear response

history (NLRH) analyses performed on multi-tiered concentrically braced frames designed in

accordance to the 2010 AISC Seismic Provisions (Imanpour et al. 2016b). For the prototype frame

studied here, 2.0𝐶𝑑 ∆𝑒 is equal to 1.1%, which corresponds to a lateral displacement of 99.5 mm.

To calculate the tier displacement in the critical tier (Tier 1), it is assumed that the tier drift is

composed of two components (Figure 3.9): 1) the displacement associated with linear variation

over the length of the frame, 𝛿𝐹,1, and 2) the displacement associated with column bending (shear

distortion) caused by the unbalanced brace storey shear, ∆𝑉′𝑏𝑟 as obtained at the maximum

anticipated storey drift 𝛿𝐶,1 . Equation 3.4 was used to calculate the tier displacement in Tier 1, 𝛿1 :

56
 ∆𝑉′𝑏𝑟 ℎ2 ℎ2
𝛿1 = 𝛿𝐹,1 + ( 1 2
) (3𝐸𝐼 ) (3.4)
2 𝑐ℎ

As shown in Figure 3.9, the first component of the drift in Tier 1 is calculated using similar

ℎ
triangles when the storey drift of 2.0𝐶𝑑 𝛿𝑒 is achieved, which results in 𝛿𝐹,1 = 2.0𝐶𝑑 𝛿𝑒 ( ℎ1 ) = 52

mm. For the deflection due to column shear distortion, 𝛿𝐶,1 , an unbalanced brace storey force ∆𝑉′𝑏𝑟

is calculated when the braces in both tiers have reached their expected post-buckling capacity,

C’exp, as expected for well-proportioned frames when they reach the maximum anticipated storey

drift (Imanpour et al. 2016a).

Thus,

′
∆𝑉′𝑏𝑟 = (𝑇𝑒𝑥𝑝 + 𝐶′𝑒𝑥𝑝 )2 𝑐𝑜𝑠𝜃2 − (𝑇𝑒𝑥𝑝 + 𝐶𝑒𝑥𝑝 )1 𝑐𝑜𝑠𝜃1

For the prototype frame,

∆𝑉′𝑏𝑟 = 32 kN.

Substituting ∆𝑉′𝑏𝑟 , 𝛿𝑒,1 , the tier heights, h1 and h2, the frame height h, and column flexural

stiffness in the plane of the frame EIy,c = 2.26𝑥1013 N∙ 𝑚𝑚2 into Equation 3.4, the displacement

in Tier 1, 𝛿1 = 63 𝑚𝑚, is obtained, which corresponds to a tier drift of 1.33%—satisfying the 2%

limit prescribed by the 2016 AISC Seismic Provisions.

57
 Figure 3.9: Frame deformed shape at expected storey drift from analyses

3.8 Design Summary

Table 3.1 gives a summary of the selected members for the two-tiered CBFs designed in

accordance with the 2010 and 2016 AISC Seismic Provisions. The 2010 and 2016 frames are

presented in Figure 3.10. The two differences between the two designs are the section of the

columns and the orientation of the strut beam.

58
 Table 3.1: Summary of design calculations for the prototype frame

Figure 3.10: Prototype frame designed in accordance with the a) 2010 AISC Seismic
Provisions; and b) 2016 AISC Seismic Provisions

59
Chapter 4 – Numerical Model of the Braced Frames

4.1 General

The numerical models of the prototype frames designed in Chapter 3 were constructed and

analyzed using the Abaqus finite element program (Dassault Systèmes 2014) to examine the

seismic behaviour of the frames, and evaluate the seismic demands induced in the frame members.

The Abaqus program was chosen to conduct this examination because of its ability to efficiently

simulate complex nonlinearities including material, geometry, interactions between different

members, and loading. Furthermore, the program can appropriately simulate initial geometric out-

of-straightness and residual stresses while capturing the local response of the frame such as local

buckling, twist, and connection behaviour. This chapter presents modelling assumptions and

analysis techniques used for nonlinear pushover (static) and nonlinear response history (dynamic)

analyses.

4.2 Model Development

4.2.1 Element

The three-dimensional deformable quadrilateral (4-node), stress/displacement shell elements with

reduced integration and a large-strain formulation shell element (S4R) were used to simulate the

braces, columns, strut, connections, and roof beam of the prototype frames. The geometry and

integration point of the selected element are shown in Figure 4.1. Shell elements are more

computationally-efficient than other three-dimensional (solid) elements because of the lesser

number of integration points; however, they provide slightly less accurate measurements

60
particularly across the thickness of the elements. Nonetheless, this type of element is capable of

estimating displacements and strains well (Dassault Systèms 2014) and consequently the forces

and stresses of a given discretized part. Therefore, due to the large size of the model produced,

three-dimensional shell elements were considered as the best option based on a mixed criterion

between computational efficiency and accuracy.

Figure 4.1: Three-dimensional 4-node general-purpose shell element with reduced integration

4.2.2 Material Properties and Plasticity Model

The elastic behaviour was simulated using Young’s modulus equal to E = 200,000 MPa and the

Poisson’s ratio 𝑣 = 0.3. Under seismic loading, it is expected that the braced frame members

experience severe yielding and undergo large plastic deformations. Hence, a material model that

could accurately simulate the cyclic inelastic behaviour of the steel material while considering the

kinematic and cyclic hardening was selected.

Kinematic/isotropic plastic material model was chosen from the the Abaqus material library to

simulate the inelastic response of the material. This allows to use the kinematic formulation to

track the yield surface shift, and the isotropic formulations to define the uniform expansion of the

yield surface. The parameters used to describe the combined hardening response of the steel

61
material in the Abaqus model were obtained from the calibration performed by Suzuki and Lignos

(2015). Although these parameters were only calibrated for ASTM A992 steel, they were used

here to describe the material properties of the other steel grades due to limited access to cyclic

coupon test data. Within the Abaqus model, there are three parameters that define the

kinematic/isotropic material model: yield stress at zero strain (Fy), initial kinematic hardening (C1),

and the rate at which C1 decreases (𝛾) with increasing plastic deformation, 𝜀 𝑝𝑙 . Figure 4.2 shows

how the back stress, 𝛼, is defined using C1 and 𝛾, and combined with the Fy to determine the

envelop of the yield surface. The kinematic component of the material model was defined as C1 =

3378 MPa and 𝛾 = 20. A nominal yield stress Fy = 345 MPa was used for the beam, strut and

columns made of ASTM A992 steel. The yield stress equal to 300 MPa was assigned to connection

plates conforming to CSA G40.21 300W. For HSS braces selected from ASTM A1085 steel, the

expected yield stress RyFy = 431 MPa where Ry is the ratio between the measure and nominal yield

stresses (Ry = 1.25 as per the 2016 AISC Seismic Provisions) was used.

Cyclic hardening of steel was defined using the maximum change in the size of the yield surface

𝑄∞ and the rate at which the yield surface changes with plastic deformation b, as shown in Figure

4.2. The cyclic hardening parameters equal to 𝑄∞ = 90 MPa and b = 12 were used.

62
 Figure 4.2: One-dimensional representation of the hardening in the nonlinear
isotropic/kinematic model (Adopted from Dassault Systèmes 2014)

4.2.3 Mesh analysis and calibration

An independent model was constructed using a 4.5 m long isolated HSS127×127×7.9 member to

determine an appropriate size of the shell elements and ensure that brace flexural plastic hinging

and local buckling can be reproduced using the selected brace elements upon global buckling

(Figure 4.3). Then, the analysis results were compared against experimental data conducted by

Jiang (2012). This calibration was beneficial not only to determine an appropriate size of mesh but

also to validate the use of the parameters obtained from Suzuki and Lignos (2015) used to define

the plastic material model for the brace.

63
 Figure 4.3: Finite element model of the isolated HSS 127×127×7.9 member

Five mesh sizes were studied using the isolated brace model as shown in Fig. 4.4: a single element,

two elements, four elements, eight elements, and twelve elements across the width of the section.

Figure 4.4: HSS brace mesh across the section wall with various divisions a) one shell
element; b) two shell elements; c) four shell elements; d) eight shell elements; e) and twelve
shell elments

The hysteretic response from the finite element analysis was compared against the test data, as

shown in Figure 4.5. As shown, a good agreement was obtained between the experimental results

and numerical prediction. A slightly higher compressive capacity by the numerical model can be

64
attributed to the fact that residual stresses were neglected in the numerical model and since the

global slenderness ration (KL/r = 93) of the selected brace falls under inelastic buckling range it is

expected that residual stresses reduce the buckling load.

Figure 4.5: Axial force–axial deformation response of the HSS127×127×7.9 brace under
incremental cyclic loading (test data by Jiang 2012)

Then, the lateral displacement of the member at buckling was obtained to study the computational

efficiency and convergence rate of the finite element model. It can be observed in Figure 4.6 that

the lateral displacement at buckling tends to converge when using more than four elements across

the width of the section. Additionally, a higher number of elements allows for better prediction of

local buckling.

65
 Figure 4.6: Lateral displacement at buckling using different mesh size elements

A flexural plastic hinge and a local buckling were properly captured in the analysis when using

four elements across the width of the section, as shown in Figure 4.7.

As shown in Figure 4.8, a finer mesh density (approximately 25 mm square mesh) was used around

the connections to better capture the interaction between members and resulting complex stress

state at connection plates and parts of the members adjacent to the connections. The remaining

parts of the columns, strut, and beam were assigned coarser elements (25-30 mm in length) as

shown in Figure 4.8, based on the convergence analysis performed on four W-sections by Stoakes

and Fahnestock (2016).

66
Figure 4.7: Simulation of flexural plastic hinging and local buckling of HSS127×127×7.9 δ =
4.5δy

Figure 4.8: Finite element model of the steel two-tiered concentrically braced frame

67
4.2.4 Boundary Conditions

The base of the columns and bottom-edge of the base-gusset-plates were constrained to a reference

point at the centre of the column as shown in Figure 4.9. The translational degrees-of-freedom of

this reference point were fixed in all three principal directions. Also, a torsional constraint was

defined at the reference point. The reference point was free to rotate in and out of the plane of the

frame (UR3 and UR1) to simulate a pinned base condition. Similarly, at the top of each column,

the web and flanges were constrained to a reference point at the middle of the column web. These

reference points at the top of the columns were restrained from out-of-plane movement and

torsional rotation. Similar to the base of the frame, the reference points at the top were free to rotate

in and out of the plane of the frame to simulate a pinned-roller condition as shown in Figure 4.10.

Figure 4.9: Constrain applied at the base of the frame

4.2.5 Geometric nonlinearity

4.2.5.1 P-Δ Simulation

Nonlinear analyses are load path dependent, and the results depend on the combined gravity and

lateral load effects. Therefore, the vertical gravity loads acting on the entire structure, not only on

68
the braced frame, should be simulated in the numerical model in order to capture destabilizing P-

Δ effects. As shown in Figure 4.11 a leaning column with gravity loads tributary to the gravity-

load-resisting system was included in the model to represent the proportion effects of the

corresponding gravity columns, which relies on the seismic-force-resisting system for lateral

stability. The leaning column was linked to the braced frame at the roof level to represent in-plane

rigidity of the roof diaphragm. Note that although the performance of multi-tiered braced frames

can be improved by involving gravity columns in the lateral load-carrying capacity as shown by

(Imanpour et al. 2016c), the effect of gravity columns in providing the lateral stiffness was not

considered in this numerical analysis to portray the worst scenario possible.

Figure 4.10: Boundary conditions assigned to the columns a) in the plane of the frame; and b)
out of the plane of the frame

69
 Figure 4.11: Two-tiered CBF with adjacent leaning column

A three-dimensional deformable wire element was used to simulate the leaning column. The

benefit of using this element is the low computational effort required when performing nonlinear

dynamic analyses. Within the model, the leaning column was pinned at the base, and the top was

constrained to the top of the braced frame in the plane of the frame using an equation-constraint.

The leaning column was also torsionally fixed at both ends, and its out-of-plane movement was

constrained. P-Δ tributary to the braced frame was simulated by applying the braced frame

tributary gravity loads on top of its columns.

4.2.5.2 P-δ Simulation

To simulate the effects of axial forces on the braces and columns, initial geometric imperfections,

specifically out-of-straightness, were explicitly considered in the numerical model. Initial

70
geometric imperfections corresponding to the first buckling mode of the bracing members and

columns, which were obtained from an Eigen buckling analysis, were assigned to these members.

The amplitude of the initial imperfections was taken equal to 1/1000 times the unbraced length of

the member in the direction of buckling as allowed in construction (AISC 2016c). For the columns,

the total height of the frame was considered as the unbraced length in the out-of-plane direction,

and the corresponding tier heights were considered as the unbraced length in the in-plane direction

as shown in Figure 4.12. For the braces, 0.45 times the length of each brace was considered as the

unbraced length (Wakabayashi et al. 1974; Nakashima and Wakabayashi 1992; and El-Tayem and

Goel 1985; 1986; Sabelli et al. 1999) to account for the lateral support provided by the other

bracing member of the tier and the restraints imposed by brace end connections. The directions of

the imperfections were set to initiate the in-plane bending of the columns towards the negative X-

direction in the bottom tier and towards the positive X-direction in the top tier (Figure 4.12a). The

direction of the out-of-plane imperfections applied on the braces and columns were set to produce

the maximum out-of-plane bending moment on the column upon brace buckling and yielding as

shown in Figure 4.12b.

a) b)

Figure 4.12: In-plane and out-of-plane initial geometric imperfections: a) CBF elevations;
and b) CBF side view (deformations magnified)
71
4.2.6 Residual Stress

Residual stresses, which are produced due to the differential cooling of hot-rolled elements during

the manufacturing process, were considered in the model. The results of the residual stress

measurement for wide-flange hot-rolled shapes show that compression stresses tend to develop at

the tips of the flanges, while tension-stresses are produced near the intersections of the web and

flanges (Ziemian 2010). Such residual stresses can result in non-uniform yielding in wide-flange

sections and cause premature yielding of parts of the cross-section and reduce the stiffness of the

section (Ziemian 2010). This effect can be severe when wide-flange member bends in the weak-

axis direction since the tips of the flanges have built-in compressive stresses that overlap with the

compressive stresses produced by the flexural bending moment. Residual stresses were

incorporated into beams and columns of the finite element model based on the pattern proposed

by Galambos and Ketter (1958) as illustrated in Figure 4.13. The residual stresses developed in

the HSS braces due to the cold-forming process were neglected in the model because they are

deemed to have a limited influence on the brace capacities (Izvernari 2007).

4.2.7 Gravity Loading

Seismic evaluation of the CBF prototypes was carried out using two analysis steps. In the first

step, the gravity load was applied at the top of the leaning and braced frame columns using the

static/general procedure (Dassault Systèmes 2014). A vertical downwards load of 227 kN was

applied on the braced frame columns, and a 3 MN load was applied at the top end of the leaning

column, as shown in Figure 4.14. The load at the top of the leaning column is based on the dead

and live load acting on one-quarter of the tributary area of the building plus the corresponding

weight associated with the wall as described in Section 3.3.

72
Figure 4.13: Residual stress pattern assigned to Wide-flange sections

Figure 4.14: Gravity analysis step

73
4.3 Analysis

4.3.1 Nonlinear Static (Pushover) Analysis

Once the gravity load was applied, a cyclic horizontal displacement history shown in Figure 4.15

was applied at the roof level of the frame to simulate the seismic load effects. In Figure 4.15, Δby

and Δbm are the ratios of brace yield deformation and the design storey drift, respectively. A

static/general step in the Abaqus program was employed using the Full Newton solution technique

to perform the pushover analysis. The horizontal displacement applied has 14 cycles based on the

loading protocol proposed by Appendix K of the 2016 Seismic Provisions for experimental testing

of buckling restraint braces (BRBs). The loading protocol includes two cycles at 1.0Δby,

0.5Δbm,1.0Δbm, 1.5Δbm, 2.0Δbm, 3.0Δbm, and 4.0Δbm. The last 4 cycles were modified by applying

higher displacement demands to ensure that the 2% tier drift limit, as permitted by 2016 AISC

Seismic Provisions, was captured during the pushover analysis.

Figure 4.15: Pushover loading protocol

74
4.3.2 Nonlinear Response History (Dynamic) Analysis

A nonlinear response history (dynamic) analysis was used to examine the seismic response of the

selected prototype braced frames. The application of such analysis is essential when evaluating the

response parameters including internal forces, reactions, displacements, rotations as it leads to a

realistic estimation of such parameters and provides reliable information about the seismic-induced

demands used in structural design (ASCE 2017).

4.3.2.1 Ground Motion Records

Ground motion records used as input for the dynamic analysis were applied in the plane of the

frame to the base of the frame and leaning column in the horizontal X-direction of the braced

frame. The set of ground motions used comprises 40 historical ground motions (Table 4.1)

obtained from Dehghani (2016). The ensemble contains 21 records representing interplate

earthquakes (70-300 km deep), 14 records representing crustal earthquakes (0-300 km), and 5

records representing in-slab earthquakes (300-700 km). The horizontal component of selected

records was selected and scaled using the method proposed by Dehghani and Tremblay (2016) to

match, on average, the code-prescribed MCER response spectra as given in ASCE 7 at the

fundamental period of the braced frame.

4.3.2.2 Analysis method

The ‘Dynamic, implicit’ procedure was selected to conduct the analysis. This analysis uses the

Hilbert-Hughes-Taylor implicit time integration method with parameter 𝛼 = −0.05 to solve the

differential equations, where the Full Newton technique is implemented to solve the nonlinear

dynamic equilibrium (Dassault Systèmes 2014). The initial step was set to 0.01 and the minimum

step increment was set to 2E-15.

75
Table 4.1: Selected ground motions for dynamic analysis (Dehghani 2016)

76
4.3.2.3 Inertia Masses

Inertial forces developed at the roof level were reproduced using two point-masses at the top end

of the braced frame column as shown in Figure 4.16. The masses represent the weight equal to

one-eighth of the total building seismic weight, which equals to 97.13 tonnes (as calculated in

section 3.3 of Chapter 3) carried by the braced frame.

Figure 4.16: Inertia point masses assigned to the braced frame model

The mass corresponding to the self-weight of the braces was modelled using two approaches: 1)

mass corresponding to the self-weight was assigned to the brace components; and 2) only 0.1% of

the mass corresponding to the self-weight is considered. The results of the dynamic analysis

obtained from the first modelling approach showed that the maximum compression capacity of the

brace did not appropriately match the expected buckling capacity, Cexp. It was found that dynamic

77
buckling can produce higher buckling load as a result of overshoot effects, which can also delay

brace buckling as confirmed by Kazemzadeh Azad et al. (2017). Although this effect may be

representative of what a brace might experience during an earthquake, the variation of the dynamic

characteristics of ground motions including natural period, input energy, frequency content and

duration can substantially change the overshoot effects resulting in a significant variation in the

response parameters used to evaluate the frames herein. The overshoot effect was diminished by

reducing the density of the braces from its standard value to 0.1% of the regular density. Figure

4.17 shows an example of the hysteretic response of a continuous brace in the 2016 design under

a ground motion record with 100% and 0.1% mass assigned to the frame components.

Figure 4.17: HSS 89×89×6.4 brace hysteretic response under 1980 Irpinia, Italy (SCC1)
ground motion record using a) full density; and b) 0.1% density

Similarly, the results of NLRH analyses showed that higher mode of vibrations associated with

the individual column member with a long-unbraced length could significantly affect the out-of-

plane moment demands induced in the columns. The long-unbraced length of the column in the

out-of-plane direction can amplify the local dynamic response of the member (e.g. out-of-plane

bending moment) under dynamic loading. As discussed earlier, two modelling approaches were

78
examined to evaluate the effects of the local vibration modes. Figure 4.18 compares the response

of the 2016 frame under the 1980 Irpinia, Italy (SCC1) ground motion record when 1) the column

mass is fully considered, and 2) when the columns are modelled using only 0.1% of their full

density. As shown, the effects of higher vibration modes were minimized when using the smaller

density.

Figure 4.18: Influence of column density on the out-of-plane moment under the 1980 Irpinia,
Italy (SCC1) ground motion record

4.3.2.3 Damping

Rayleigh’s damping was used to determine the classical damping matrix as defined in the equation

of motion to generate the internal damping forces, which cause the attenuation of the frame

movement under dynamic loading. Rayleigh’s damping is described as c = 𝛼m + 𝛽k, where m and

k are mass and stiffness matrices, respectively. α and β are the mass- and stiffness-proportional

damping coefficients, respectively (Chopra 2011). To obtain these coefficients, an Eigen-buckling

analysis was performed to calculate the frame natural frequencies. Subsequently, the first two

natural frequencies, 𝜔𝑖 =13.4 rad/s and 𝜔𝑗 = 55.04 rad/s, were used in combination with the

critical damping ratio 𝜁 = 2%, in Equations 4.1 and 4.2 to determine the damping coefficients 𝛼 =

0.43 and 𝛽 =0.00058. These coefficients were finally introduced in the finite element model.
79
 2𝜔𝑖 𝜔𝑗
𝛼=𝜁 (4.1)
𝜔𝑖 +𝜔𝑗

2
𝛽=𝜁 (4.2)
𝜔𝑖 +𝜔𝑗

80
Chapter 5 – Seismic Response of Two-Tiered Steel
Concentrically Braced Frames
5.1 General

The seismic response of the prototype two-tiered concentrically braced frames was evaluated using

the pushover and nonlinear response history (NLRH) analyses. The global response of the frames

including the lateral force, storey drift and tier drifts is first presented. Then, the seismic-induced

demands including the column and brace forces are discussed.

5.2 Pushover Analysis

Pushover analysis is a static analysis with a nonlinear material definition where the roof

displacement is incrementally increased up to a target displacement under a given displacement

pattern. This analysis was used to simulate the lateral response of the frame and its components.

The displacement scheme applied was based on the protocol proposed by Appendix K of the 2016

Seismic Provision (AISC 2016a) for the experimental testing of buckling restraint braces. A full

description of the displacement scheme is presented in Section 4.3.1.

The static loading scheme was applied to the frame designed using the 2010 AISC Seismic

Provisions, hereafter referred to as 2010 design, and the frame designed in accordance with the

2016 Seismic Provisions, hereafter referred to as the 2016 design.

5.2.1 Frame Global Response

Figure 5.1 shows the lateral response of 2010 and 2016 designs. For the 2010 design, a stable

response with an increased lateral load was observed up to a storey drift equal to 2.0%, at which

column buckling occurred and analysis was halted because of a convergence issue. As shown in

81
Figure 5.2a column buckling was observed in the right-hand-side (RHS) column. In contrast, the

2016 design was able to complete the cyclic pushover analysis without column buckling or frame

instability. A stable lateral load–lateral displacement response was obtained as shown in Figure

5.1b. The 2016 frame reached a maximum lateral displacement applied corresponding to a storey

drift of 2.1%. Figure 5.2 shows the frame deformed shape at 2.1% storey drift.

Figure 5.1: Normalized lateral load–lateral displacement response: (a) 2010 design; and
(b)2016 design, (V is the design base shear as defined in Section 3.3)

The tier drift is plotted against the storey drift in Figure 5.3a and 5.3b for the 2010 and 2016

designs, respectively. For both designs, the tier drifts in both tiers were nearly the same in the

initial elastic cycles. However, under larger storey drifts, in the 2010 design, lateral frame

deformation was concentered in Tier 1 as brace tensile yielding occurred first in this tier; however,

very limited lateral deformations were concentrated in Tier 2 and the braces in this tier remained

relatively elastic. As a result of this response, nonlinear lateral deformations of the frame were

unevenly distributed along the height of the frame as shown in Figure 5.3a. By further increasing

the roof displacement, column instability occurred in the first-tier segment of the RHS column.

82
 a) b)

Figure 5.2: Frame deformed shape: (a) 2010 design: column buckling at story drift 2.0%; and
(b) 2016 design final deformed shape at story drift 2.1%

The lateral response of the 2016 design was different compared to its 2010 counterpart. Beyond

0.5% storey drift, brace tensile yielding was observed first in Tier 1 and then in Tier 2 (noncritical

tier). Although the storey shear in Tier 1 reduced as tension brace yielded and compression brace

buckled at 0.5% storey drift, the columns compensated for the storey shear difference between the

tiers, which led to brace tensile yielding to take place in Tier 2 at 1.0% storey drift. As a result of

this response, frame lateral deformations were more uniformly distributed between the tiers. Table

5.1 compares the tier drifts at of the two designs at 2.0%—the maximum storey drift applied to the

2010 design—which highlights the improvement of the seismic behaviour of the 2016 design.

83
 Figure 5.3: Tier drifts: a) 2010 design; and b) 2016 design

Table 5.1: Tier drifts at a storey drift 2.0%

2010 Design 2016 Design

Tier 2 (Δ2) 0.5% 1.3%
Tier 1 (Δ1) 3.5% 2.6%

5.2.2 Brace Behaviour

The brace axial forces (normalized by the maximum expected tensile strength, ARyFy) are plotted

against the tier drift in Figures 5.4 for continuous and discontinuous braces of both tiers. The

buckling and tensile resistances of the braces were appropriately predicted when compared to the

values calculated using the Specification (ASIC 2016b). For the 2010 design, both continuous and

discontinuous braces in Tier 2 (Figure 5.4a and 5.4b) remain nearly elastic. No tension yielding

occurred in Tier 2 and the compression brace force slightly degraded as a result of cyclic loading.

However, severe inelastic deformations were induced in Tier 1 due to the buckling and yielding of

the braces in this tier as expected. In contrast to the 2010 design, the braces in both tiers of the

84
2016 design contributed to the inelastic response of the frame and underwent yielding and buckling

as shown in Figures 5.4c and 5.4d.

Figure 5.4: Normalized brace axial force: (a) continuous braces of the 2010 design; (b)
discontinuous braces of the 2010 design; (c) continuous braces of the 2016 design; (d)
discontinuous braces of the 2016 design

5.2.3 Column Behaviour

5.2.3.1 In-Plane Response

In-plane bending moment demand induced in the columns caused by the progressive yielding and

buckling of braced panels was used to examine the in-plane response of the prototype frames. The

bending moment was measured just below the brace-to-column connection at the strut level where

the maximum in-plane moment occurs. The moments were plotted for the 2010 and 2016 designs

85
in Figures 5.5a and 5.5b, respectively. The moments were normalized by the plastic moment of

the corresponding section about its minor axis Mpy. Note that under positive storey drifts, the right

column was axially loaded in compression; similarly, under negative storey drifts the left column

was axially loaded in compression. The maximum normalized moment demand in the compression

column of the 2010 and 2016 designs were 0.34 and 0.33, respectively. The larger maximum

absolute value of the in-plane moment in the columns of the 2016 design, in comparison to that of

the 2010 design (126 vs 28 kN-m), is caused by the higher stiffness, which attracts higher moments

to compensate for the unbalanced brace storey shear force between the braced panels. Moreover,

in the 2010 design, when the columns are in tension, the in-plane moment demand increases

linearly; whereas when the columns are in compression, the moment rapidly decreases after the

attainment of the maximum in-plane moment that corresponds to the maximum unbalanced brace

storey shear force. The column moment degrades beyond the maximum value mainly because of

the loss of strength in the compression column caused by the yielding of the member. The first

plastic hinge formed within the first-tier segment of the compression column (Figure 5.6), which

led to the reversal of the column bending moment sign as shown Figure 5.5a. However, the in-

plane moment of the 2016 design columns increased proportionally with the lateral displacement

until it reached a plateau (Figure 5.5b). Although the moment arising from the unbalance brace

storey shear force decreased, in-plane bending moment resulting from the P-𝛿 effect increased

with enlarged storey drift, which prevented a net decrease of the in-plane moment.

86
Figure 5.5: Column in-plane bending moments for a) 2010 design; and b) 2016 design

Figure 5.6: Deformed shape with Von-Mises stress contour at the verge of RHS column
buckling at 2.0% storey drift

87
5.2.3.2 Out-of-Plane Response

Column out-of-plane bending moments were used to evaluate the response of the frame in the out-

of-plane direction. Figures 5.7a and 5.7b show the column moments for 2010 and 2016 designs,

respectively. The moments were measured on the column at strut-to-column connection and were

normalized by the corresponding plastic section moment about the section strong axis Mpx. The

columns of the 2010 and 2016 designs experienced a maximum out-of-plane demand of 23 kN-m

(0.05Mpx) and 25 kN-m (0.031Mpx), respectively. It was found that the maximum out-of-plane

moments do not coincide with the maximum story drifts, which may be attributed to the fact that

compression brace forces reduce at higher storey drifts, resulting in lower out-of-plane bending

moments imposed on the columns.

Figure 5.7: Column out-of-plane bending moments for a) the 2010 design; and b) the 2016
design

5.3 Nonlinear Response History Analysis

Nonlinear response history (NLRH) analysis was used to evaluate the dynamic response of the

prototype braced frames under earthquake loads. NLRH analysis is a powerful method in

88
earthquake engineering that considers in detail the dynamic effects of seismic loading. This

analysis method is performed by applying an acceleration record to the base of the structure

(Filiatrault et al. 2013). The ground motion input used for this study contains 40 ground motions

which were scaled by Dehghani (2016) to match, on average, the code-prescribed MCER response

spectra at the fundamental period of the braced frame. Section 4.3.2 provides further information

on the selected ground motions.

5.3.1 Frame Global Response

Column instability was observed for the 2010 design under 13 ground motion records. Column

buckling triggered inelastic instability and led to frame collapse in all 13 cases. Figure 5.8 shows

an example of the frame collapse under the 1994 Northridge (SCC14) ground motion record.

Column buckling occurred under the combination of large in-plane bending moment and axial

compression force demands. The in-plane flexural buckling, with a limited twist, was observed

first; the instability mode then changed to out-of-plane buckling due to the lack of out-of-plane

support, resulting in a flexural-torsional buckling mode.

a) b) c)

Figure 5.8: 2010 frame deformed shape under 1994 Northridge (SCC14) ground motion
record: a) onset of LHS column buckling at t = 4.70 s and story drift of 1.7%; b) LHS column
buckling at t = 5.38 s; and c) frame collapse at t = 5.49 s
89
The statistics of the NLRH analysis results were used to evaluate the seismic response of the

prototype frames further. The maximum values of the frame storey drift and tier drifts were

recorded under each ground motion. Maximum storey drifts are given in Table 5.2 and plotted in

Figure 5.9a for the 2010 design. The ground motions in the horizontal axis of the plot are arranged

in the order presented in Table 4.1. The squares highlighted in red are the cases where column

instability occurred; additionally, the expected storey drift, Cd∆e, as per ASCE 7 is shown. Three

indicators were used to identify column buckling in the NLRH analysis:

• stiffness reduction in the axial force–axial displacement response of the column (Figure

5.10a);

• stiffness reduction in the lateral shear force–storey drift response of the frame (Figure

5.10b); and

• curvature reversal due to plastic hinge forming at two locations within the first-tier segment

(Figure 5.11).

Figure 5.9: 2010 design a) maximum storey drift under 40 ground motion records; and b)
storey drifts at buckling under 13 collapsed ground motion records

90
Table 5.2: Statisctics of the global response of the 2010 design

91
 Figure 5.10: 2010 design response under 2007 Pisco, Peru (SCI7) ground motion record: a)
axial force–axial displacement response of LHS column; and b) base shear–storey drift
response(frame lateral shear force is normalized by the design base shear, V; and dots
represent the instant of column buckling)

The median value of the storey drift obtained from NLRH analyses is 1.1%. Note that the median,

as well as the 84th percentile values, were computed based on the ground motion records where

the frame did not collapse. For the 13 ground motion records where column buckling was

observed, the storey drift at the onset of buckling is plotted in Figure 5.9b. As shown, the storey

drifts range between 1.6% and 2.2% with an average of 1.8%. It was found that the LHS column

buckled upon reaching lower storey drifts in comparison to the RHS column due to the direction

of initial geometric imperfections (see Figure 4.12), which favoured the in-plane buckling of the

LHS column.

The results obtained from the NLRH analysis of the 2016 design indicated that no column buckling

nor inelastic instability occured under any of the 40 ground motion records. Table 5.3 present a

summary of the frame displacement response for the 2016 designs under the selected ground

motion records. Figure 5.12 shows the maximum storey drift under the 40 ground motion records.

The median storey drift is 1.4%, and the maximum storey drift of 2.6% occurred under the 1979

92
Montenegro, Yugo earthquake (SCC13). Note that the median storey drift for the 2016 frame

appears higher that that the 2010 design; however, this is because the collapsed cases are not

included in the calculation of median for the 2010 design.

Table 5.3: Statisctics of global response of the 2016 design

93
 Figure 5.11: 2010 design: LHS column buckling at 34.7 s and 1.6% storey drift under 2007
Pisco, Peru (SCI7) ground motion record

Figure 5.12: 2016 design: maximum storey drifts under the 40 ground motion records

The maximum tier drifts recorded in Tiers 1 and 2 are shown for the 2010 and 2016 designs in

Figures 5.13 and 5.14, respectively. As shown in Figure 5.13 for the 2010 design, a larger tier

drift was observed in Tier 1 while consistent elastic lateral deformations were obtained in Tier 2
94
without initiation of brace yielding under any of the ground motion records. The Tier 2 drifts were

near a 0.5% tier drift, which corresponds to the lateral semi elastic deformation of the tier. The

median tier drifts for the 2010 design were 1.7% and 0.5% for the first and second tier,

respectively. The maximum tier drift observed before frame collapse was 3.7%, which was

measured under 2011 Tohoku, Japan record (SCI4).

Figure 5.13: 2010 design a) Tier 1 drift; and b) Tier 2 drift

Tier drift results are plotted in Figure 5.14 for the 2016 design. As shown, frame nonlinear

deformations are distributed between the tiers, although it is more pronounced in Tier 1 where

brace yielding is initiated first. The results suggest that an improved seismic response is expected

in the 2016 frame with larger columns compared to the 2010 design. The median tier drifts for the

2016 design were 2.0% and 0.7% for the first and second tier, respectively. The maximum value

of the tier drift was observed under the 1979 Montenegro, Yugo record and was equal to 3.3%.

95
 Figure 5.14: 2016 design a) Tier 1 drift; and b) Tier 2 drift

5.3.2 Brace Behaviour

Tension yielding and buckling in the braces of the 2010 design was observed under all the ground

motion records. However, braces in the second tier of this frame did not yield in tension and

suffered a modest degradation of their buckling strength. An example of the brace axial force–tier

drift response is presented in Figure 5.15a. As shown, a larger tier drift was developed for the brace

in Tier 1, while in Tier 2 limited nonlinear response was obtained in the compression region.

Braces in the first and second tier of the 2016 design buckled and yielded under the ground motions

with a storey drift greater than 1.0%, which confirms the more uniform lateral response expected

in the 2016 design. An example of such a response is shown in Figure 5.15b.

96
 Figure 5.15: Brace axial force response in Tier 1 and Tier 2 under 2007 Pisco, Peru (SCI7)
ground motion record: a) 2010 design; and b) 2016 design

Brace local buckling occurred in braces of Tier 1 of the 2010 design in several occasions including

the ground motion records under which the frame collapsed (Table 5.2) and in three other ground

motions (2011 Tohoku, Japan, SCI3; 2011 Tohoku, Japan, SCI6; and 2007 Pisco, Peru, SCI20)

under which the frame did not collapse (Figure 5.16). Although brace fracture was not explicitly

simulated in the finite element model, the occurrence of the brace local buckling together with

large tier drifts observed (Figure 5.13a) suggests that brace fracture could be a limit state for the

2010 design. The results of past experimental study showed the drift where HSS brace fractures

range between 2.0%–3.0% for various slenderness ratios (Tremblay et al. 2003; Yang and Mahin

2005; Uriz and Mahin 2008; Fell et al. 2009; Roeder et al. 2011; Hsiao et al. 2013). Despite the

large tier drifts induced in Tier 1 of the 2016 design, local buckling was not observed under any

of the ground motions.

97
 Figure 5.16: Local buckling of the first tier brace in 2010 design under 2011 Tohoku, Japan
(SCI3) ground motion record at 9.7 s 1.5% storey drift and 2.6% tier drift

5.3.3 Column Behaviour

The statistics of the maximum column demands obtained from the NLRH analyses for the 2010

and 2016 designs are given in Tables 5.4 and 5.5, respectively. In each table, the bending moment

demands and axial forces are presented. For the 2010 design, the axial forces were compared

against the design value, and for the 2016 design, both the moment and axial force components

were compared against the designed values.

98
Table 5.4: Statisctics of column demands for 2010 design

99
Table 5.5: Statisctics of column demands for 2016 design

100
5.3.3.1 In-Plane Response

Column moment demands were recorded for each ground motion just below the brace-to-column

connection. The in-plane bending moments were normalized by the corresponding plastic moment

Mpy and plotted against the storey drift for the 2010 and 2016 designs under 40 ground motions

records. Appendix A provides the details of the measured response parameters including column

moment demands, tier and storey drift under the ground motion records analyzed.

Figure 5.17 shows the in-plane bending moment of the column for the 2010 and 2016 designs

under the 2011 Southern Peru ground motion record. Relatively insignificant moments were

induced in the columns before yielding of the tension brace in the critical tier, which corresponds

to 0.5% storey drift. Once the tension brace in Tier 1 yielded, the member elongated while the

noncritical tier remained elastic. As a result of differential deformation between Tiers 1 and 2, an

in-plane moment was imposed on the column.

Figure 5.17: Column in-plane bending moment demand under the 2011 Southern Peru
(SCI16) ground acceleration record: a) 2010 design; and b) 2016 design

The in-plane moment results for the 2010 design shown in Figure 5.17a indicates that the column

loaded in compression reaches its maximum moment value and subsequently decreases as lateral
101
displacement was increased because the flexural stiffness of the column reduces upon large lateral

displacements. The in-plane moment demand in the compression column reduced suddenly and

changed sign when a plastic hinge (second plastic hinge) formed at the strut level, which led to

column buckling as shown in Figure 5.11. Opposite to the compression column, the in-plane

demand of the tension column increased nearly linearly as the frame was moved laterally.

In the 2010 design, the maximum flexural demand of the compression column before buckling

was not consistent throughout all the ground motions. For the ground motions with multiple

nonlinear cycles, the in-plane bending moment of the compression column reached higher values

before column buckling. However, when the column buckled within the first cycle of the ground

motion, a lower in-plane demand was observed in the columns. Figure 5.18 compares two column

buckling scenarios under the 1994 Northridge and the 2001 Southern Peru ground motion records.

The former is presented in Figure 5.18a, and Figure 5.18b shows the latter. Under 1994 Northridge

ground motion record, the maximum moment before column buckling was 0.11Mpy; however, a

higher moment of 0.17Mpy was recorded in the LHS column under 2001 Southern Peru ground

motion. This behaviour is due to the degradation of compression force in the bracing members

after multiple cycles, which results in lower compression forces in the columns and in turn, allows

for higher in-plane moments to develop prior to column buckling.

102
 Figure 5.18: Column in-plane bending moments for the LHS column under a) the 1994
Northridge (SCC14) ground motion record; and b) the 2001 Southern Peru (SCI16) ground
motion record

In-plane flexural demands on the columns of the 2016 design obtained from the NLRH analysis

was consistent with the results obtained from the pushover analysis in Section 5.2. The in-plane

moment increases steeply as the unbalanced brace storey shear force develops due to the inelastic

brace response (Figure 5.17b). This increase was not pronounced when the frame reached storey

drifts greater than 1.0% as the moment caused the unbalance braces storey shear force decreased.

The maximum in-plane flexural demands observed under the 40 ground motion records are

provided in Figure 5.19. The red-dashed line represents the design in-plane moment Mry in

accordance with the 2016 AISC Seismic Provisions. The maximum in-plane moment measured on

a braced frame column was 0.41Mry, and the median was 0.30Mry.

103
 Figure 5.19: Column in-plane bending moments for 2016 design

5.3.3.2 Out-of-Plane Response

The out-of-plane response of the frame was examined using the bending moment demands induced

in the columns of the prototype frames obtained from the NLRH analyses. The out-of-plane

moments under to the selected ground motion records are shown in Figures 5.20a and 5.20b for

the 2010 and 2016 designs, respectively. The median moment was 0.05Mpx and 0.044Mpx for the

2010 and 2016 designs, respectively, where Mpx is the strong-axis plastic moment of the

corresponding column. The out-of-plane moments observed for the 2016 frame exceeded the

design out-of-plane moment as per the 2016 AISC Seismic Provisions, Mrx = 0.007Mpx, in all 40

cases.

Figure 5.20: Column out-of-plane bending moments for a) 2010 design; and b) 2016 design
104
The out-of-plane moment was investigated further by differentiating the contributing components

including (tension and compression) brace forces, strut forces, gusset plate plastic moments, and

P-δ effects. Once a brace buckles out-of-plane, a portion the compression force was directed out

of the plane (Figure 5.21), which produced a moment on the column. The strut force produced

insignificant out-of-plane forces due to misalignment of the frame. Also, a portion of the moment

was developed by the axial load of the column acting on an imperfect column, known as P-𝛿

effects. Additionally, as the braces buckled, the connections yielded and produced a moment,

which had components in the out-of-plane direction of the column and the strut member as

illustrated in Figure 5.22. These components were plotted for the 2016 design under 1980 Irpinia,

Italy (SCC1) ground motion record in Figure 5.23. As shown, the out-of-plane moment induced

by the out-of-plane component of the brace forces and the P-δ effects are the key contributers to

the out-of-plane bending moment of the compression column.

Figure 5.21: Brace out-of-plane buckling (deformations magnified)

105
 Figure 5.22: Moments in the X- and Y-directions produced by plastic hinging of the brace
gusset plate

The contribution from the braces on the out-of-plane moment of the columns was not only limited

to the buckled compression braces. It was found that tension braces also contribute to the out-of-

plane bending moment imposed on the column. This is because residual plastic deformations

developed upon brace out-of-plane buckling results in an elongated brace in the subsequent loading

cycle where the brace in tension, which in turn develops out-of-plane forces on the column prior

to the development of full tension capacity of the member. However, such out-of-plane

deformations in the tension brace were considerably smaller than those in the compression brace,

which led to lower out-of-plane bending moments in the columns.

106
 Figure 5.23: History of column out-of-plane bending moment under 1980 Irpinia, Italy
(SCC1) ground motion record

The summation of the moments arising from the four individual components was compared to the

measured out-of-plane moment on the column in Figure 5.24. A strong correlation was observed

between the summation of identified components and the measured demand on the column.

Figure 5.24: Column out-of-plane moment history under the 1980 Irpinia, Italy (SCC1)
ground motion record

Since the most significant contribution to the out-of-plane bending moment on the columns is the

force produced form braces buckling out-of-plane, which represents a correlation between the

compression brace force and the column out-of-plane moments. As the compression brace force

degrades in large tier drifts, smaller out-of-plane bending moments are imposed in the columns.

107
Figure 5.25 shows the out-of-plane moment measured in the columns of the 2016 design under the

2011 Tohoku, Japan (SCI6) ground motion record. As shown, the out-of-plane moment decreases

significantly when the frame is pushed to the storey drifts greater than 1% in both columns.

Figure 5.25: Column out-of-plane bending moment for the 2016 design under the 2011
Tohoku, Japan (SCI6) ground motion record

5.3.3.2 Axial Force

The maximum axial forces exerted on the columns of the 2010 and 2016 designs are plotted for

all 40 ground motions in Figures 5.26a and 5.26b. In both designs, the axial force on the columns

is well estimated in comparison to the design value. The median values for the maximum axial

load 2010 and 2016 designs are 1.04 and 0.99, respectively.

Figure 5.26: Column axial force for a) 2010 design; and b) 2016 design
108
5.4 Summary

Nonlinear static and dynamic analyses were carried out to evaluate the response of the prototype

two-tiered braced frames designed in accordance with the 2010 and 2016 AISC Seismic

Provisions. The results show that column stability is compromised in the 2010 design. The results

of NLRH confirmed column buckling and subsequent frame instability under 13 ground motion

records. Column stiffness was found insufficient in the 2010 design to propagate brace tensile

yielding between the tiers, which led to the concentration of drift in the critical tier.

The analysis results indicate that the frame designed in accordance with the 2016 AISC Seismic

Provisions remain stable under the applied seismic loads. No column buckling nor frame instability

occurred under the ground motion records. The columns possess sufficient strength and stiffness

to prevent column buckling and trigger brace tensile yielding in both braced panels. However, the

in-plane moment demand was found to be over-estimated and the out-of-plane moment demand

be under-estimated. Moreover, the median storey drift was 2.5 times larger than the design storey

drift. Consequently, the tier drift was also larger than the expected value.

109
Chapter 6 – Design Recommendations

6.1 General

In this chapter, improved design recommendations are proposed based on the results obtained in

Chapter 5 to achieve an efficient design for MT-SCBF design. The results obtained from the NLRH

analyses of the two-tiered concentrically braced frame designed in accordance with the 2016 AISC

Seismic Provisions were used to make recommendations on the column force demands and frame

drift requirements.

6.2 Strength Requirements

Improved in-plane and out-of-plane bending moment requirements are proposed for columns of

MT-CBFs. The column interaction ratio was measured under the axial compression force Pr/Pn

and biaxial bending moment demands, weak-axis bending moment Mry/Mpy and strong-axis

bending moment, Mrx/Mpx, at the instant of a ground motion when the interaction ratio resulted in

the highest value from the combination of axial compression force and flexural bending moments.

Note that only ground motions where the brace tensile yielding occurs in both braced tiers were

considered. Table 6.1 presents the results of the column interaction ratios, the ratio between the

measured axial force and moment demands and the corresponding design values, and brace axial

forces at the instant when the interaction ratio results in the highest value. Interaction ratios

represent the column in-plane buckling limit state as per AISC 360 in Chapter H. Design values

are obtained from AISC 341-16 as given in Chapter 3 Section 3.5.2.

110
Table 6.1: Statistics of column axial force and biaxial moment demands plus brace forces at the maximum interaction ratio

111
6.2.1 Column Design Moments: In-plane

In-plane design moment for the MT-CBF columns is estimated based on the progressive yielding

and buckling of the braces in two adjacent tiers, which is derived by the unbalanced brace storey

shear force. Brace force adjustment factors are proposed to account for the inelastic cyclic response

of the bracing members that are in compression when the columns reach their maximum force

demands. Brace force adjustment factors for strength, 𝛼 and 𝛼′, are proposed based on the results

obtained from the NLRH analyses to estimate a realistic brace axial compression force for multi-

tiered concentrically braced frames. The adjustment factor 𝛼 represents the brace force for the tier

where yielding has just been triggered, and 𝛼′ represents the brace force adjustment factor for the

tier where brace tensile yielding has been developed already. The 𝛼 factor accounts for the limited

force degradation of the compression brace in the noncritical tier where yielding has just initiated,

and 𝛼′ accounts for the fact that compression brace force in the critical tier has not reached its post-

buckling strength when the combined demand in the column is maximum. Figure 6.1 shows the

brace force scenario proposed to calculate the column in-plane bending moment demands using

updated compression brace forces. Two yielding scenarios are considered: yielding first takes

place in Tier 1 and yielding first occurs in Tier 2.

112
 Figure 6.1: Proposed adjusted brace resistances when yielding propagates between tiers: a)
brace tensile yielding has just initiated in Tier 2; and b) brace tensile yielding has just initiated
in Tier 1

The column in-plane bending moment should be computed by incorporating 𝛼 and 𝛼′ in the

equation (Equation 3.2 in Chapter 3) represeting the unbalance brace storey shear force, ∆𝑉𝑏𝑟 :

′
∆𝑉𝑏𝑟 = (𝑇𝑒𝑥𝑝 + 𝛼𝐶𝑒𝑥𝑝 )𝑚 𝑐𝑜𝑠𝜃𝑚 − (𝑇𝑒𝑥𝑝 + 𝛼′𝐶𝑒𝑥𝑝 )𝑛 𝑐𝑜𝑠𝜃𝑛 (6.1)

where m and n correspond to the tier where brace tensile yielding is just initiated and the tier where

brace tensile yielding has been developed, respectively.

Based on the results of the NLRH analyses, it is proposed to set the force adjustment factors as

𝛼 = 0.80 and 𝛼′ = 1.10 based on the 84th percentile values from Table 6.1. The 84th percentile

values are suggested to account for the possibility of using stockier braces, which could lead to a

higher unbalanced brace storey shear force (Figure 6.2). Using the adjustmen brace force factors

and the brace strengths, ∆𝑉𝑏𝑟 can then be calculated to obtain the column in-plane moment demand

Mry as described in Chapter 3.

113
 0.6

0.5

(Cexp ̶ C'exp) / Texp 0.4

0.3

0.2

0.1

0
50 100 150 200
kL/r

Figure 6.2: Slenderness effect on the difference between Cexp and C’exp

6.1.2 Column Design Moments: Out-of-plane

The out-of-plane bending moment induced in the columns of the prototype frame was measured

as a fraction of the plastic section modulus about its strong axis Mpx. This moment, however, can

be represented as a horizontal notional load acting at the strut to column connection similar to CSA

S16 procedure, which produces the same bending moment demand on the column as shown in

Figure 6.3. The 84th percentile value of the out-of-plane moment demand obtained from the NLRH

analyses is 0.02 Mpx at the instant when the column interaction ratio results in the most extreme

value, which is greater than the design value of 0.007 Mpx. However, to further protect the column

from potential out-of-plane buckling (Chapter 5, Figure 5.8), the 84th percentile value of the

absolute maximum out-of-plane moment is proposed to be used as the design column out-of-plane

moment demand. This value, as shown in Table 5.5, is equal to 0.05Mpx, which corresponds to a

horizontal notional load equal to 0.1 times the vertical component of the compression brace acting

at each tier level as per the 2016 AISC Seismic Provisions, or 0.014 times the load in the

114
compression force acting in the column below the brace-to-column connection as per the CSA

S16. Although individual components of the out-of-plane moment (Figure 5.23) including brace

forces, P-𝛿 effects, and out-of-plane buckling and yielding of the brace connection contribute to

the total out-of-plane moment on the column, it is proposed here to use a single notional load that

represents the contribution from all sources as described in Section 5.3.3.2 to facilitate the design

process.

Figure 6.3: a) Deformed shape of MT-CBF; b) deformed shape of compression column

caused by the notional out-of-plane load; c) column out-of-plane shear diagram; and d)
column out-of-plane moment diagram

6.2 Drift Requirement

A new drift requirement is proposed for two-tiered concentrically braced frame based on the results

obtained from the NLRH analyses. Total storey drift and individual tier drift should be verified to

ensure the frame satisfies the stiffness requirement and brace deformation demands do not exceed

the demands corresponding to brace premature low-cycle fatigue fracture. Tier drift is obtained

115
from the summation of the overall frame drift and distortion due to column bending when the

frame roof displacement reaches the displacement corresponding to the design storey drift. To

obtain drift due to column distortion, one can use the unbalance brace storey shear force similar to

the in-plane bending moment calculations in Section 6.2.1. The NLRH analysis results suggest

that the maximum storey drift experienced under the selected ground motions is significantly

higher than the design value prescribed by ASCE 7-16. This finding from this research agrees with

the findings by Imanpour et al. (2016a) and Imanpour and Tremblay (2016b).

Table 6.2 presents the storey drift values and brace forces at the maximum storey drift for the

analysis where yielding is triggered in both tiers and corresponding brace forces. The results of the

NLRH analyses for all the selected ground motion records are given in Table 5.3. Design storey

drift prescribed by ASCE 7 for the frame designed using the 2016 AISC Seismic Provisions is

Cd∆e = 0.55%, which is 2.5 times lower than the median values obtained from the NLRH analyses

as shown in Table 5.3. Such large expected storey drift can result in large ductility demands in

braced tiers, which poses concerns regarding the adequacy of the current drift requirements. An

amplified design storey drift corresponding to 2.5 times the design storey drift is proposed for two-

tiered concentrically braced frames when verifying the storey and tier drifts based on the median

value from Table 5.3.

New brace force adjustment factors are proposed to verify the tier drift limitation that include 𝛽

and 𝛽′ for the noncritical tier and the critical tier, respectively. This brace force scenario is

illustrated in Figure 6.4. According to the results obtained from the NLRH analyses, the brace

force adjustment factors for drift 𝛽 = 0.60 and 𝛽′ = 1.15 are proposed based on the median

(statistically neutral) brace force values as given in Table 6.2, where the 𝛽 factor accounts for the

116
strength degradation of strength the compression brace in the noncritical tier, and 𝛽′ amplifies the

strength of the compression brace in the critical tier.

Table 6.2: Statistics of brace forces at the maximum storey drift

117
 Figure 6.4: Proposed adjusted brace force scenario to verify tier drifts

6.3 Case Study

A case study is presented here to illustrate how the proposed strength and drift requirements would

be applied to the prototype frame designed in Chapter 3.

6.3.1 Column In-plane Moment

Maximum column in-plane moment is calculated using the maximum expected brace strengths in

the analysis representative of the progressive yielding and buckling of the braces plus the brace

force adjusted factor as shown in Figure 6.5.

118
 Figure 6.5: Proposed analysis case C for the moment calcaulation including

The unbalance brace storey shear force ∆𝑉𝑏𝑟 is calculated as follows:

′
∆𝑉𝑏𝑟 = (𝑇𝑒𝑥𝑝 + 𝛼𝐶𝑒𝑥𝑝 )𝑚 𝑐𝑜𝑠𝜃𝑚 − (𝑇𝑒𝑥𝑝 + 𝛼′𝐶𝑒𝑥𝑝 )𝑛 𝑐𝑜𝑠𝜃𝑛 (6.1)

∆𝑉𝑏𝑟 = (859 + 0.8(323))2 𝑐𝑜𝑠(31.6𝑜 )2 − (859 + 1.1(92))1 𝑐𝑜𝑠(33.9𝑜 )1

∆𝑉𝑏𝑟 = 155 𝑘𝑁

substituting ∆𝑉𝑏𝑟 into Equation 3.1 , Mry can be calculated for each column:

∆𝑉𝑏𝑟 ℎ1 ℎ2
𝑀𝑟𝑦 = (3.1)
2 ℎ

155 (4.7)(4.3)
𝑀𝑟𝑦 =
2 9

𝑀𝑟𝑦 = 174 𝑘𝑁– 𝑚 (originally calculated as 254 𝑘𝑁– 𝑚)

119
6.3.2 Column Out-of-plane Moment

The proposed method to calculate the out-of-plane moment consist on calculating the notional load

and then determining the corresponding the out-of-plane moment. The notional load is first

calculated using Equation 6.2.

𝑁 = 0.1𝐶𝑒𝑥𝑝 2 sin(𝜃2 ) (6.2)

𝑁 = 0.1 (323 𝑘𝑁) sin(33.9𝑜 ) = 18 𝑘𝑁

The moment induced in the column due to the notional load is equal to:

ℎ1 ℎ2 (4.7)(4.3)
𝑀𝑟𝑥 = 𝑁 = 18 𝑘𝑁 = 40 𝑘𝑁– 𝑚 (6.3)
ℎ 9

To compare the design value to the one prescribed by the CSA S16, the notional load is calculated

using the CSA S16 procedure:

𝑁 = 0.02𝑃𝑟 = 0.02 (1331 𝑘𝑁) = 26.6 𝑘𝑁 (6.4)

ℎ1 ℎ2 (4.7)(4.3)
𝑀𝑟𝑥 = 𝑁 = 26.6 𝑘𝑁 = 60 𝑘𝑁– 𝑚 (50% larger than the value in Equation 6.3)
ℎ 9

The column strength is verified for a W310×129 section using the interaction equation H1-1a of

the AISC Specification as follows:

𝑃𝑟 8 𝑀 𝑀𝑟𝑦
+ 9 (𝑀𝑟𝑥 + 𝑀 ) ≤ 1.0 (6.5)
𝑃𝑐 𝑐𝑥 𝑐𝑦

8
0.32 + 9 (0.06 + 0.57) = 0.88 < 1.0 OK

The selected W310×129 section satisfies the strength requirement.

120
6.3.3 Tier Drift Limit Verification

The tier drift in the critical tier is calculated for the column using Equation 6.6, where ∆𝐹,1 is the

drift in Tier 1 corresponding to the linear drift of the frame, and ∆𝐶,1 is the drift in Tier 1 caused

by the column distortion under the unbalanced brace storey shear force:

For W310×129

∆1 = ∆𝐹,1 + ∆𝐶,1 ≤ 0.02 (6.6)

where ∆𝐹,1 = 2.5𝐶𝑑 ∆𝑒 (6.7)

2.5(5)(10.1 𝑚𝑚)
∆𝐹,1 = = 0.014
9000 𝑚𝑚

and

∆𝑉𝑏𝑟 ℎ1 ℎ22
∆𝐶,1 = (6.8)
2 3𝐸𝐼𝑦 ℎ

Thus,

′
∆𝑉𝑏𝑟 = ∆𝑉𝑏𝑟 = (𝑇𝑒𝑥𝑝 + 𝛽𝐶𝑒𝑥𝑝 )𝑚 𝑐𝑜𝑠𝜃𝑚 − (𝑇𝑒𝑥𝑝 + 𝛽′𝐶𝑒𝑥𝑝 )𝑛 𝑐𝑜𝑠𝜃𝑛 (6.9)

∆𝑉𝑏𝑟 = (859 + 0.60(323))2 𝑐𝑜𝑠(31.6𝑜 )2 − (859 + 1.15(92))1 𝑐𝑜𝑠(33.9𝑜 )1

∆𝑉𝑏𝑟 = 95.9 𝑘𝑁

Substituting the ∆𝑐,1 value in Equation 6.8:

95.9𝑘𝑁 (4700)(4300 𝑚𝑚)2

∆𝐶,1 =
2 3(200,000𝑀𝑃𝑎)(100𝑥106 𝑚𝑚4 )(9000)

∆𝐶,1 = 0.0077

The tier drift is the calcualted by summing the values from equations 6.7 and 6.8 as follows:

∴ ∆1 = 0.014 + 0.0077 = 0.022 > 0.02 NOT OK

121
The selected W310X129 does not satisfy the proposed drift requirement.

A larger W310×143 section is verified to meet the the tier drift using Equation 6.6, where ∆𝐹,1

and ∆𝐶,1 are calculated using Equations 6.7 and 6.8, respectivly.

2.5(5)(9.95 𝑚𝑚)
∆𝐹,1 = = 0.014 (6.7)
9000 𝑚𝑚

∆𝑉𝑏𝑟 ℎ1 ℎ22
∆𝐶,1 = (6.8)
2 3𝐸𝐼𝑦 ℎ

Thus, the tier drift is equal to:

95.900 𝑘𝑁 (4700)(4300 𝑚𝑚)2

∆𝐶,1 =
2 3(200,000𝑀𝑃𝑎)(113𝑥106 𝑚𝑚4 )(9000)
∆𝐶,1 = 0.0064
∴ ∆1 = 0.014 + 0.0064 = 0.02 = 0.02 OK

The W310×143 section satisfies the proposed tier drift check. Since, this new section has been

selected based on the drift check only, its strength should be verified using Equation. 6.5 :

8
0.29 + 9 (0.05 + 0.50) = 0.79 < 1.0 OK

The selected W310×143 section satisfies the strength requirement.

In this case study, the column section did not change from the original desing in Chapter 3 by

implemnting the recommendations proposed in this chapter; howerver, the proposed

recommendations represent a relistic estimation of the force and drift demands on MT-CBFs under

seismic loads. The key difference between the design presented here and the one shown in Chapter

3 is that the column design was originally governed by the strength and not stiffness limit; in

constrast, the design of the column when implementing the proposed design recommendations is

122
dominated by the required stiffness to meet the tier drift limit. Although, if the tier-heights were

modified, such that h1 = 4.6 m and h2 = 4.4 m to reduced the stiffness demand caused by the

unbalanced braced storey shear force, the column size could be decreased from W310×143 to a

W310×129 and result in a more efficient design by implementing the design recomedations.

The proposed design recommnedations to determine the bending moment demads can also be

considered in the design of frames with lower ductility levels such as moderatly ductile frames in

Canada.

For MT-CBFs with more than two tiers, the design recommendations can be applied in the cases

where brace tensile yielding propagates progressively from the bottom tier or the top tier following

the method proposed by Imanpour et al. (2016a).

123
Chapter 7 – Conclusions and Recommendations

7.1 Summary

Steel multi-tiered concentrically braced frames (MT-CBFs) are commonly used in North America

to resist lateral loads in tall single-storey buildings. Multi-tiered configurations involve multiple

bracing panels stacked along the height of the storey that divide the height of a storey into several

bracing panels. The primary advantage of using this configuration is to avoid using a single bracing

panel between the ground and roof levels, which is not practical in most of cases. This

configuration reduces the buckling length of the columns in the plane of the frame. Furthermore,

the length of the bracing members is reduced, which results in smaller brace sizes that can easily

satisfy the stringent width-to-thickness ratio and slenderness limits in high seismic areas. If the

capacity design is required, lower capacity-induced forces are imposed on the adjacent members

and connections of the braced frame.

Past studies confirmed unfavourable limit states under seismic load effects when the columns are

designed under the axial load only. Moreover, under lateral seismic load brace tensile yielding is

not distributed along the height of the frame and rather takes place in one of the tiers. As a result

of this response, large in-plane flexural bending moment is induced in the column, which in some

cases leads to plastic hinging and subsequent buckling of columns, and in some extreme cases

frame collapse. Additionally, nonuniform brace yielding develops large inelastic deformations in

one of the tiers, which may cause brace fracture. Improved design guidelines were introduced in

the Canadian Steel Design Standard (CSA S16) and the AISC Seismic Provisions (AISC 341-16)

to address these concerns and protect the columns of multi-tiered braced frames. The special

124
seismic provisions in Canada are limited to limit ductility CBFs (Type LD) and moderately ductile

(Type MD) CBFs. The U.S. seismic provisions address ordinary CBFs, special CBFs, and

Buckling Restrained Brace Frames (BRBFs) with multi-tiered configurations. The recent

provisions require designers to design columns under the combined effects of the axial force, the

in-plane bending moment due to progressive yielding of the braced panels, and out-of-plane

bending moment due to brace out-of-plane buckling and column initial imperfections. Tier drift

must also be verified to ensure that columns have enough flexural stiffness to prevent large tier

drift that can cause premature brace fracture. Finally, the provisions require that the multi-tiered

braced frames have intermediate struts to avoid unsatisfactory K-braced frame response.

Although significant improvement has been achieved in the design methodology of MT-CBFs,

there is very limited detail numerical models and no experimental test data available to understand

the stability response of the columns in such frames, verify the column moment demands, validate

and improve the current design guidelines. The objective of this M.Sc. thesis is to evaluate the

seismic behaviour of and the design methods for multi-tired special concentrically braced frames.

A survey of the existing literature was conducted. Furthermore, a prototype frame consisting of a

two-tiered special concentrically braced frame was designed using the 2010 and 2016 AISC

Seismic Provisions. The seismic behaviour of the prototype frames was then evaluated in detail

with the Abaqus finite element software (Dassault Systèmes 2014) using two nonlinear analysis

methods: static (pushover) and nonlinear response history (dynamic) analyses. The global response

of the selected prototype braced frames was examined and the force demands induced in the

columns including the in-plane and out-of-plane bending moment demands were investigated. The

results obtained from NLRH analyses were finally used to propose seismic design

recommendations to improve the design of multi-tiered concentrically braced frames.

125
7.2 Limitations

This study only examined a single braced frame geometry and applied the AISC Seismic

Provisions to design the braced frame. The geometrical properties including the frame height,

number of tiers, tier height ratio was not examined in this study. Although extensive efforts were

made to develop the detailed finite element model, certain assumptions were used in cases were

enough information was not available. Specifically, the hardening properties used for the definition

of the inelastic material behaviour for the bracing members are similar to those used for wide-

flange sections. Furthermore, the base condition of the braced frame was assumed to be pinned to

create a severe inelastic instability condition by reducing the frame redundancy.

7.3 Conclusions

The main findings of this M.Sc. research project are summarized as follows:

Numerical model of the braced frame:

• Detailed three-dimensional finite element model was developed, which is capable of

predicting the seismic response of MT-CBFs and provides a good understanding of the

column stability condition.

• Numerical analysis results produced in this study offers valuable input data for future

experimental studies, in particular, the displacement history obtained from the NLRH

analyses can be used to performed seismic testing of the full-scale two-tiered braced frame.

Two-tiered SCBF designed in accordance with the 2010 AISC Seismic Provisions:

• Non-uniform distribution of inelastic frame deformations was observed due to brace tensile

yielding in one of the tiers only.

126
• Non-uniform distribution of lateral displacement induced large in-plane flexural demands

on the columns.

• Excessive tier lateral deformations occurred in the tier where brace yielded initiated. Tier

drifts ranged between 0.5% and 3.6%, where larger tier deformations combined with local

buckling observed can lead to brace premature fracture.

• Strength and stiffness of the columns were not sufficient to propagate yielding along the

height of the frame.

• Column buckling occurred in the cyclic pushover analysis because of the combined high

axial compression force and in-plane flexural demand at 2.0% storey drifts.

• High strength degradation was observed in the lateral force–lateral displacement response

obtained from the pushover analysis, which confirmed column instability.

• Pushover analysis was able to predict well the seismic behaviour of the frame plus force

and deformation demands in the members compared to the NLRH analysis.

• Column buckling was observed under 13 ground motions record out of 40 records studied

using the NLRH analysis method. In-plane buckling dominated the instability of the

column that changed to bi-axial buckling as a result of the large out-of-plane displacement

and lack of out-of-plane lateral bracing. Column buckling occurred in the storey drift range

from 1.6% to 2.2%. Variablity of storey drifts is attributed to the properties of the ground

motion record and in part to the direction of initial geometric imperfections assigned to the

column.

127
 • Brace local buckling was observed under several ground motions records. Local buckling

was observed in tier drifts ranging from 2.3% to 3.6%.

Two-tiered SCBF designed in accordance with the 2016 AISC Seismic Provisions:

• Frame exhibited a more uniform lateral deformation response under lateral seismic load.

Brace tensile yielding was triggered in both braced tiers along the height of the frame in

the pushover analysis and under most of the ground motion records.

• Progressive yielding of braces resulted in in-plane moment demands on the columns.

• Strength and stiffness of the columns were sufficient to propagate yielding along the height

of the frame.

• A stable and satisfactory response was obtained using both pushover and dynamic analysis

methods. No column instability nor frame collapse was observed.

• Maximum axial force induced in the columns from NLRH analyses agreed with the design

value prescribed by 2016 AISC Seismic Provisions.

• Results obtained form NLRH analyses found that in-plane bending moment induced in the

columns due to progressive yielding of braces is over-estimated by the 2016 AISC Seismic

Provisions. The median value of the moments is 46% of the design demand.

• Results obtained form NLRH analyses found that out-of-plane bending moment induced

in the columns due to brace out-of-plane buckling, P-𝛿 effects, plastic hinging of the brace

connection, and strut forces is under-estimated by the 2016 AISC Seismic Provisions. The

median value of the moments is 163% of the design demand.

• Pushover analysis results compare well with the results obtained from the NLRH analysis.

128
Proposed design recommendations for two-tiered CBFs:

• Brace force adjustment factors were proposed to amplify the compression brace resistance

in the tier which yields first by 𝛼 ′ = 1.10 and de-amplify the compressive brace resistance

in the other tier by 𝛼 = 0.80 when calculating the unbalance brace storey shear force on

the column, which then is used to obtained the column in-plane bending moment.

• An out-of-plane horizontal notional load applied at the tier level on the column was

proposed to be equal to 0.1 times the vertical component of the compression brace force

meeting the column at the same level.

• An amplified design storey drift of 2.5 times the code specified design storey drift is

proposed for MT-SCBFs. Brace force adjustment factors were proposed to amplify the

compression brace resistance in the tier which yield first by 𝛽′ = 1.15 and to de-amplify

the compressive brace resistance in the other tier by 𝛽 = 0.60. Such adjusted brace forces

should be used to calculate the unbalance brace storey shear force on the column when

verifying the tier drift under the proposed design storey drift.

• For MT-CBFs with three or more tiers, the proposed design recommendations can be

applied in the cases where brace tensile yielding propagates progressively from the bottom

tier or the top tier following the method proposed by Imanpour et al. (2016a).

7.4 Recommendations for Further Research

This research has contributed to understanding the seismic stability response of MT-CBFs using

the detailed finite element simulation. This research has also made recommendations to improve

the current seismic design provisions implicit in North American design standards. However,

129
further investigation will be beneficial in understanding the complex stability response of such

frames and further improve the design guidelines. Proposals for future studies are as follows:

• An experimental evaluation of full-scale two-tiered SCBF specimens is needed to verify

column buckling, validate the numerical models used in this study and further validate the

current design guidelines.

• A parametric numerical study should be performed to examine the effect of frame

geometries such as frame height, tier height ratios, the number of tiers, bracing

configurations such as chevron, diagonal and split-X, braced frame system such as

eccentrically braced frames and buckling-restraint braced frames.

• The effects of the column base connection on the seismic behaviour of MT-CBFs should

be examined.

• Further research is recommended into the effects of dynamic brace response such as the

brace overshoot phenomenon at buckling.

• Given the concerns raised regarding the dynamic stability of MT-SCBFs designed

excluding the special seismic design provisions, research studies to assess the performance

of, identify deficiencies of and propose retrofit strategies for existing MT-SCBFs are

recommended.

130
Bibliography

AISC. (2010a). ANSI/AISC 341-10, Seismic Provisions for Structural Steel Buildings, American
Institute of Steel Construction, Chicago, IL.

AISC. (2010b). ANSI/AISC 360-10, Specifications for Structural Steel Buildings, American
Institute of Steel Construction, Chicago, IL.

AISC. (2016a). ANSI/AISC 341-16, Seismic Provisions for Structural Steel Buildings, American
Institute of Steel Construction, Chicago, IL.

AISC. (2016b). ANSI/AISC 360-16, Specifications for Structural Steel Buildings, American
Institute of Steel Construction, Chicago, IL.

AISC. (2016c). Code of Standard Practice for Steel Buildings and Bridges. American Institute of
Steel Construction, Chicago, IL.

Archambault, M.H. (1995). “Étude du comportement séismique des contreventements ductiles en

X avec profiles tubulaires en acier.” EPM/GCS-1995-09, Department of Civil Engineering,
École Polytechnique, Montréal, QC.

ASCE. (2016). SEI/ASCE 7-16, Minimum Design Loads for Buildings and Other Structures,
American Society of Civil Engineers.

ASCE. (2017). SEI/ASCE 7-41, Seismic Evaluation and Retrofit of Existing Buildings, American
Society of Civil Engineers.

Astaneh-Asl, A., Goel, S.C. (1984). “Cyclic in-plane buckling of double-angle bracing.” J Struct
Eng., ASCE;109:2036–55.

Astaneh-Asl, A., Goel, S.C., Hanson, R.D. (1985) “Cyclic out-of-plane buckling of double-angle
bracing.” J Struct Eng., ASCE;111:1135–53.

131
Astaneh-Asl, A., Goel, S.C., and Hanson, R.D (1986). “Earthquake design of double-angle
bracing,” AISC Engineering Journal, Vol. 23, No. 4, 4th Quarter, 1986, pp. 133-147.

ASTM. (2015a). A1085/A1085M-15 Standard Specification for Cold-Formed Welded Carbon

Steel Hollow Structural Sections (HSS). ASTM International, West Conshohocken, PA.

ASTM. (2015b). A992/A992M-11 Standard Specification for Structural Steel Shapes. ASTM
International, West Conshohocken, PA.

CSA. (2009, 2014). CSA S16-09, CSA S16-14, Design of Steel Structures, Canadian Standards
Association, Mississauga, ON.

Bertero, V. V., Uang, C.-M., Llopiz, C. R., and Igarashi, K. (1989). “Earthquake simulator testing
of concentric braced dual system.” J. Struct. Eng., 115(8), 1877–1894.

Black, R. Gary, W. A. Wenger, and E. P. Popov. (1980). “Inelastic Buckling of Steel Struts Under
Cyclic Load Reversals.” Berkeley, Calif: Earthquake Engineering Research Center,
University of California.

Carter, C.J., Muir, L.S., Dowswell, B. (2016) “Establishing and developing the weak-axis strength
of plates subjected to applied loads.” 8th International Workshop Connections in Steel
Structures, Boston, MA, May 24-26, 2016.

Chopra, A.K. (2011). Dynamics of Structures: Theory and Applications to Earthquake

Engineering, 4th Edition, Prentice Hall, Englewood Cliffs, New Jersey, 2011.

Dalal, S.T. (1969). “Some non-conventional cases of column design.” Eng. J. AISC, 6(1), 28–39.

Dassault Systèmes. (21012;2014). "Abaqus Analysis User’s Guide" Version 6.14, Dassault Systèmes
Simulia Corp, Providence, RI.

Dehghani, M. (2016). Seismic design and qualification of all-steel buckling-restrained braced frames
for Canadian applications. (Order No. 10806518, Ecole Polytechnique, Montreal (Canada)).

132
 ProQuest Dissertations and Theses, 440. Retrieved from
http://login.ezproxy.library.ualberta.ca/login?url=https://search.proquest.com/docview/20241
62852?accountid=14474

Dehghani M., and Tremblay, R. (2016) “Robust Period-Independent Ground Motion Selection and
Scaling for Effective Seismic Design and Assessment,” Journal of Earthquake Engineering,
20:2, 185-218, DOI: 10.1080/13632469.2015.1051635

Elkady, A., & Lignos, D. G. (2018). “Full-scale testing of deep wide-flange steel columns under
multiaxis cyclic loading: Loading sequence, boundary effects, and lateral stability bracing
force demands.” Journal of Structural Engineering, 144(2), 04017189.

El-Tayem, A. A., and Goel, S. C (1985). “Cyclic behavior of angle X-bracing with welded
connections.” Research Rep. No. UMCE 85-4, Univ. of Michigan, Ann Arbor, Mich.

El-Tayem, A. A., and Goel, S. C. (1986). “Effective length factor for the design of X-bracing
systems.” AISC Eng. J., 24(1), 41–45.

Fell, B.V., Kanvinde, A.M., Deierlein, G.G., and Myers, A.T. (2009). “Experimental Investigation
of Inelastic Cyclic Buckling and Fracture of Steel Braces.” J. Struct. Eng., ASCE, 135(1),
19–32.

Foutch, D. A., Goel, S. C., and Roeder, C. W. (1987). ‘‘Seismic testing of full-scale steel
building—Part I.’’ J. Struct. Eng., 113(11), 2111–2129.

Galambos, T.V., and Ketter, R.L. (1958). Columns Under Combined Bending and Thrust, Fritz
Engineering Laboratory Report 205A.21, Bethlehem, Pennsylvania.

Gugerli, H. (1982). ‘‘Inelastic cyclic behavior of steel bracing members.’’ PhD thesis, Univ. of
Michigan, Ann Arbor, Mich.

Haddad, M., Brown, T., & Shrive, N. (2011). “Experimental cyclic loading of concentric HSS
braces.” Canadian Journal of Civil Engineering, 38(1), 110-123.
133
Hassan, O. F., and Goel, S. C. (1991). ‘‘Modeling of bracing members and seismic behavior of
concentrically braced steel structures.’’ Research Rep. No. UMCE 91-1, Univ. of
Michigan, Ann Arbor, Mich.

Hsiao, P-C., Lehman, D.E., and Roeder, C.W. (2013). “A model to simulate special concentrically
braced frames beyond brace fracture”, Earthquake Engineering and Structural Dynamics,
42, 183-200.

Imanpour, A., Tremblay, R., and Davaran, A. (2012a). “Seismic performance of steel
concentrically braced frames with bracing members intersecting columns between floors.”
Proc.,7th STESSA Conference, Santiago, Chile, 447–453.

Imanpour, A., Tremblay, R., and Davaran, A. (2012b). “Seismic Evaluation of Multi-Panel Steel
Concentrically Braced Frames.” 15WCEE, Lisbon, Portugal, Paper No. 2996.

Imanpour, A., and Tremblay, R. (2012). “Analytical Assessment of Stability of Unbraced Column
in Two Panel Concentrically Braced Frames.” 3rd International Structural Specialty
Conference, Edmonton, Alberta, Canada, Paper No. 1218.

Imanpour, A., Stoakes, C., Tremblay, R., Fahnestock, L., and Davaran, A. (2013). “Seismic
Stability Response of Columns in Multi-Tiered Braced Steel Frames for Industrial
Applications.” ASCE Structures Congress, Pittsburgh, PA, 2650–2661.

Imanpour, A., and Tremblay, R. (2014a). “Seismic performance evaluation and design of
multitiered steel concentrically braced frames.” Proc., 10th National Earthquake
Engineering Conf., Anchorage, AK, Paper No. 1347.

Imanpour, A., Tremblay, R., Davaran, A., Stoakes, C., and Fahnestock, L. (2016a). “Seismic
performance assessment of multi-tiered steel concentrically braced frames designed in
accordance with 2010 AISC seismic provisions.” Journal of Structural Engineering.
ASCE.

134
Imanpour, A., Tremblay, Fahnestock, L., and A., Stoakes, C. (2016b). “Analysis and Design of
Two-Tiered Steel Braced Frames under In-Plane Seismic Demand.” Journal of Structural
Engineering. ASCE.

Imanpour, A., Tremblay, R., and Auger, K. (2016c). “Seismic design and performance of
multitiered steel braced frames including the contribution from gravity column under in-
plane seismic demand.” Advanced in Engineering Software.

Imanpour, A. and Tremblay, R. (2016a). “Analysis Methods for the Design of Special
Concentrically Braced Frames with Three or More Tiers for In-Plane Seismic Demand.”
Journal of Structural Engineering. ASCE.

Imanpour, A., & Tremblay, R. (2016b). “Seismic design and response of steel multi-tiered
concentrically braced frames in Canada.” Canadian Journal of Civil Engineering, 43(10),
908-919.

Imanpour, A., Tremblay, R., Leclerc, M., & Siguier, R. (2018). “Development of a Hybrid
Simulation Computational Model for Steel Braced Frames.” In Key Engineering Materials
(Vol. 763, pp. 609-618). Trans Tech Publications.

Izvernari, C. (2007). The Seismic Behavior of Steel Braces with Large Sections, Doctoral
Dissertation, Department of Civil, Geological, and Mining Engineering, Ecole
Polytechnique, Montreal, Quebec, Canada.

Jain, A. K., Goel, S. C., and Hanson, R. D. (1980). ‘‘Hysteretic cycles of axially loaded steel
members.’’ J. Struct. Div. ASCE, 106(8), 1777–1795.

Jiang, Y. (2013). Numerical and experimental seismic assessment and retrofit of steel tension-only
double angle braced frames designed before the implementation of detailing provisions for
ductile seismic response. (Order No. 1528147, Ecole Polytechnique, Montreal
(Canada)). ProQuest Dissertations and Theses, 253. Retrieved from

135
 http://login.ezproxy.library.ualberta.ca/login?url=https://search.proquest.com/docview/15
56108729?accountid=14474

Kazemzadeh Azad, S., Topkaya, C., and Bybordiani, M. (2018) “Dynamic buckling of braces in
concentrically braced frames”. Earthquake Engng Struct Dyn.;47: 613–633.
https://doi.org/10.1002/eqe.2982

Lee, S., and Goel, S. C. (1987). ‘‘Seismic behavior of hollow and concrete-filled square tubular
bracing members.’’ Research Rep. No. UMCE 87-11, Univ. of Michigan, Ann Arbor,
Mich.

Lee, S. 1988. Seismic behaviour of hollow and concrete-filled square tubular bracing members.
Ph.D. thesis, Department of Civil Engineering, The University of Michigan, Ann Arbor,
Michigan.

Liu, Z. (1987). ‘‘Investigation of concrete-filled steel tubes under cyclic bending and buckling.’’
PhD thesis, Univ. of Michigan, Ann Arbor, Mich.

McKenna, F., and Fenves, G.L. (2004). “Open System for Earthquake Engineering Simulation
(OpenSees).” Pacific Earthquake Engineering Research Center (PEER), University of
California, Berkeley, CA. http://opensees.berkeley.edu/.

Nakashima, M., and Wakabayashi, M. (1992). “Analysis and design of steel braces and braced
frames in building structures.” Stability and ductility of steel structures under cyclic
loading, Y. Fukumoto and G. C. Lee, eds., CRC Press, Boca Raton, Fla., 309–321.

Ozkula, G., Harris, J., & Uang, C. M. (2017). “Classifying Cyclic Buckling Modes of Steel Wide-
Flange Columns under Cyclic Loading.” In Structures Congress 2017 (pp. 155-167).

Popov, E.P., and Black, R.G. (1981). “Steel struts under severe cyclic loadings.” J. Struct. Eng.,
107(9), 1857–1881.

136
Roeder, C.W., Lehman, D.E., Clark, K, Powell, J., Yoo, J-H, Tsai, K-C, Lin, C-H, and Wei, C-Y
(2011). “Influence of Gusset Plate Connection and Braces on the Seismic Performance of
X-Braced Frames.” Earthquake Engineering and Structural Dynamics, 40(4), 355-374.

Sabelli, R., Roeder, C.W., and Hajjar, J.F. (2013). “Seismic design of steel special concentrically
braced frame systems: A guide for practicing engineers,” NEHRP Seismic Design
Technical Brief No. 8, produced by the NEHRP Consultants Joint Venture, a partnership
of the Applied Technology Council and the Consortium of Universities for Research in
Earthquake Engineering, for the National Institute of Standards and Technology,
Gaithersburg, MD, NIST GCR 13-917-24.

S-Frame Software. (2017). Structural analysis software.

Shaback, J. B. (2001). ‘‘Behaviour of square HSS braces with end connections under reversed
cyclic axial loading.’’ Master’s thesis, Univ. of Calgary, Calgary, Alta., Canada.

Schmidt, B.J., and Bartlett, F.M. (2002). “Review of resistance factor for steel: data collection,”
Can. J. Civ. Eng. 29: 109–118.

Stoakes, C.D., and Fahnestock, L.A. (2012). “Influence of weak-axis flexural yielding on strong
axis buckling strength of wide flange columns.” Proc., Annual Stability Conference, SSRC,
Grapevine, Texas, April 17–20.

Stoakes, C. D., and Fahnestock, L. A. (2014). “Three-dimensional Finite Element Simulation of

the Seismic Behavior of Multitier Concentrically Braced Frames” ASCE Structures
Congress, Boston, MA, 3-5 April.

Stoakes, C. D., and Fahnestock, L. A. (2016). “Strong-axis stability of wide flange steel columns
in the presence of weak-axis flexure.” Journal of Structural Engineering., 10.1061/
(ASCE)ST.1943-541X.00001448, 04016004.

137
Suzuki Y, Lignos DG (2015). “Large scale collapse experiments of wide flange steel beam-
columns.” Proceedings of the 8th International Conference on Behavior of Steel Structures
in Seismic Areas (STESSA), Shanghai, China, July 1-3, 2015.

Tang, X., and Goel, S. C. (1987). ‘‘Seismic analysis and design considerations of braced steel
structures.’’ Research Rep. No. UMCE 87-4, Univ. of Michigan, Ann Arbor, Mich.

Tremblay, R. (2001). “Seismic Behavior and Design of Concentrically Braced Steel Frames.” Eng.
J. AISC, 38(3), 148–166.

Tremblay, R. (2002). “Inelastic seismic response of steel bracing members.” J. Constr. Steel Res.,
58(5-8), 665–701.

Tremblay, R. (2003). “Achieving a stable inelastic seismic response for multi-story concentrically
braced steel frames.” Eng. J. AISC, 40(2), 111–129.

Tremblay, R., Archambault, M.H., and Filiatrault, A. (2003). “Seismic response of concentrically
braced steel frames made with rectangular hollow bracing members.” J. Struct. Eng.,
ASCE, 129(12), 1626–1636.

Uriz, P., Filippou, F.C., Mahin, S.A. (2008). “Model for cyclic inelastic buckling for steel
member.” J. Struct. Eng., ASCE, 134(4), 619–28.

Newell, J. D., & Uang, C. M. (2008). “Cyclic behavior of steel wide-flange columns subjected to
large drift.” Journal of structural engineering, 134(8), 1334-1342.

Walpole, W. R. (1996). “Behaviour of cold-formed steel RHS members under cyclic loading.”
Research Rep. No. 96-4, Univ. of Canterbury, Christchurch, New Zealand.

Wakabayashi, M., Matsui, C., Minami, K., and Mitani, I. (1974). “Inelastic behavior of full-scale
steel frames with and without bracings.” Bull. Disaster Prevention Research Institute,
Kyoto Univ., 24(216), 1–23.

138
Yang, F., and Mahin, S.A. (2008). “Limiting Net Section Fracture in Slotted Tube Braces.” Steel
Tips Series, Structural Steel Education Council, Moraga, CA.

Ziemian, R. D. (Ed.). (2010). Guide to stability design criteria for metal structures. John Wiley &
Sons.

139
APPENDIX A

Nonlinear response history analyses

140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220