Seismic Performance Assessment of Multi-Storey Buildings with Cold

Formed Steel Shear Wall Systems

by

Joel Martínez Martínez

A thesis presented to the

University of Waterloo

in fulfilment of the thesis

requirement for the degree of

Doctor of Philosophy

in

Civil Engineering

Waterloo, Ontario, Canada, 2007

© Joel Martínez Martínez 2007

I hereby declare that I am the sole author of this thesis. This is a true copy of
the thesis, including any required final revisions, as accepted by my
examiners.

I understand that my thesis may be made electronically available to the

public.

ii
 AKNOWLEDGEMENTS

I would like to express my sincere appreciation to my supervisors, Prof. Lei Xu

and Prof. D.E. Grierson, for their support, guidance, and many hours of supervision
required to complete this research. Their knowledge in the subjects involved in this
research played an important role in the completion of this research.

I would like to thank the National Council on Science and Technology of Mexico
(CONACYT) for providing financial support for my PhD studies.

Thanks are due to the Natural Sciences and Engineering Research Council of
Canada (NSERC) for providing financial support.

Thanks are due to Prof. Colin Rogers for providing experimental data that was of
enormous value to my research, and also for participating as the external-to-the-
university member of the examining committee.

Thanks are due to all the members of the examining committee, Prof. Jeff West,
Prof. Sanjeev Bedi, and especially to Prof. Reinhold Schuster who also taught me the
theory and applications of cold formed steel.

I am very grateful to my lovely wife Cinthia Escalera, for her love, support and
patience during these years. Also, I am very grateful to my parents and siblings for their
support during my studies. Especially, to my elder sister Ernestina and her husband
Omar.

Finally, I would like to express my appreciation to everyone who made these

years in Waterloo a wonderful experience.

iii
 To Cinthia and Joelito

iv
 ABSTRACT

Cold-Formed Steel (CFS) is a material used in the fabrication of structural and non-
structural elements for the construction of commercial and residential buildings. CFS
exhibits several advantages over other construction materials such as wood, concrete and
hot-rolled steel (structural steel). The outstanding advantages of CFS are its lower overall
cost and non-combustibility. The steel industry has promoted CFS in recent decades,
causing a notable increase in the usage of CFS in building construction. Yet, structural
steel elements are still more highly preferred, due to the complex analysis and design
procedures associated with CFS members. In addition, the seismic performance of CFS
buildings and their elements is not well known.

The primary objective of this study is to develop a method for the seismic assessment of
the lateral-load resistant shear wall panel elements of CFS buildings. The Performance-
Based Design (PBD) philosophy is adopted as the basis for conducting the seismic
assessment of low- and mid-rise CFS buildings, having from one to seven storeys.
Seismic standards have been developed to guide the design of buildings such that they do
not collapse when subjected to specified design earthquakes. PBD provides the designer
with options to choose the performance objectives to be satisfied by a building to achieve
a satisfactory design. A performance objective involves the combination of an earthquake
(i.e., seismic hazard) and a performance level (i.e., limit state) expected for the structure.
The building capacity related to each performance level is compared with the demand
imposed by the earthquake. If the earthquake demand is less than the building capacity,
the structure is appropriately designed.

The seismic performance of a CFS building is obtained using pushover analysis, a

nonlinear method of seismic analysis. This study proposes a Simplified Finite Element
Analysis (SFEA) method to carry out the nonlinear structural analysis. In this study,
lateral drifts associated with four performance levels are employed as acceptance criteria

v
for the PBD assessment of CFS buildings. The lateral drifts are determined from
experimental data.

In CFS buildings, one of the primary load-resistant elements is Shear Wall Panel (SWP).
The SWP is constructed with vertically spaced and aligned C-shape CFS studs. The ends
of the studs are screwed to the top and bottom tracks, and structural sheathing is installed
on one or both sides of the wall. For the analysis of CFS buildings, Conventional Finite
Element Analysis (CFEA) is typically adopted. However, CFEA is time consuming
because of the large number of shell and frame elements required to model the SWP
sheathing and studs. The SFEA proposed in this study consists of modeling each SWP in
the building with an equivalent shell element of the same dimensions; that is, a complete
SWP is modeled by a 16-node shell element. Thus, significantly fewer elements are
required to model a building for SFEA compared to that required for CFEA, saving both
time and resources. A model for the stiffness degradation of a SWP is developed as a
function of the lateral strength of the SWP. The model characterizes the nonlinear
behaviour of SWP under lateral loading, such that a realistic response of the building is
achieved by the pushover analysis.

The lateral strength of a SWP must be known before its seismic performance can be
assessed. In current practice, the lateral strength of a SWP is primarily determined by
experimental tests due to the lack of applicable analytical methods. In this investigation,
an analytical method is developed for determining the ultimate lateral strength of SWP,
and associated lateral displacement. The method takes into account the various factors
that affect the behaviour and the strength of SWP, such as material properties,
geometrical dimensions, and construction details.

To illustrate the effectiveness and practical application of the proposed methodology for
carrying out the PBD assessment of CFS buildings, several examples are presented. The
responses predicted by the SFEA are compared with responses determined
experimentally for isolated SWP. In addition, two building models are analyzed by
SFEA, and the results are compared with those found by SAP2000 (2006). Lastly, the

vi
PBD assessment of two buildings is conducted using SFEA and pushover analysis
accounting for the nonlinear behaviour of the SWP, to demonstrate the practicality of the
proposed technology.

vii
 TABLE OF CONTENTS

AKNOWLEDGEMENTS ..........................................................................................................................III
ABSTRACT ..................................................................................................................................................V
TABLE OF CONTENTS ........................................................................................................................ VIII
LIST OF TABLES........................................................................................................................................X
LIST OF FIGURES...................................................................................................................................XII
SYMBOLS ............................................................................................................................................... XIV
CHAPTER 1 INTRODUCTION.............................................................................................................1
1.1. GENERAL .................................................................................................................................1
1.2. OBJECTIVE OF STUDY .............................................................................................................6
1.3. SCOPE ......................................................................................................................................6
1.4. ASSUMPTIONS AND IDEALIZATIONS .......................................................................................7
1.5. THESIS OVERVIEW ..................................................................................................................9
CHAPTER 2 BACKGROUND AND LITERATURE REVIEW ON PERFORMANCE-BASED
DESIGN AND PUSHOVER ANALYSIS ..................................................................................................11
2.1. INTRODUCTION .....................................................................................................................11
2.2. PERFORMANCE-BASED DESIGN............................................................................................12
2.3. PUSHOVER ANALYSIS............................................................................................................16
2.4. NEW GENERATION PERFORMANCE-BASED DESIGN. ..........................................................19
2.5. SEISMIC ASSESSMENT OF COLD FORMED STEEL BUILDINGS. ............................................20
CHAPTER 3 ANALYTICAL METHOD FOR DETERMINING THE LATERAL STRENGTH
OF SHEAR WALL PANELS IN COLD FORMED STEEL BUILDINGS ...........................................21
3.1. INTRODUCTION .....................................................................................................................21
3.2. THE FAILURE MODES OF SWP ............................................................................................23
3.3. THE LATERAL STRENGTH OF SWP ASSOCIATED WITH SHEATHING FAILURE ..................23
3.3.1. LATERAL STRENGTH OF SHEATHING...................................................................................26
3.3.1.1. PROCEDURE FOR EVALUATING THE ULTIMATE STRENGTH REDUCTION FACTOR CU .............27
3.3.1.2. SIMPLIFIED EVALUATION OF THE ULTIMATE STRENGTH REDUCTION FACTOR CU .................30
3.4. THE LATERAL STRENGTH OF A SWP ASSOCIATED WITH FRAME FAILURE ......................32
3.5. ESTIMATION OF THE LATERAL DISPLACEMENT OF A SWP................................................33
3.6. ILLUSTRATIVE EXAMPLE......................................................................................................34
3.7. COMPARISON BETWEEN ANALYTICAL AND EXPERIMENTAL RESULTS ..............................38
CHAPTER 4 SIMPLIFIED FINITE ELEMENT ANALYSIS (SFEA) FOR CFS BUILDINGS ...48
4.1. INTRODUCTION .....................................................................................................................48
4.2. MODELING OF SWP WITH EQUIVALENT SHELL ELEMENTS ..............................................50
4.3. NONLINEAR FINITE ELEMENT FORMULATION ....................................................................54
4.4. ESTIMATING THE INTERNAL FORCES IN THE STUDS FROM THE EQUIVALENT SHELL.......63
4.5. COMPARISON OF RESULTS FOR SWP WITH SFEA AND CFEA...........................................65
4.6. STIFFNESS DEGRADATION MODEL FOR SWP......................................................................73
4.7. PREDICTION OF THE NONLINEAR RESPONSE OF SWP SUBJECTED TO LATERAL LOADING.
..............................................................................................................................................75
CHAPTER 5 PUSHOVER ANALYSIS FOR PERFORMANCE-BASED DESIGN ASSESSMENT
OF CFS BUILDINGS..................................................................................................................................83
5.1. INTRODUCTION .....................................................................................................................83
5.2. SPECTRUM-BASED PBD ASSESSMENT OF CFS BUILDINGS .................................................84

viii
 5.3. SPECTRUM-BASED PUSHOVER ANALYSIS FOR CFS BUILDINGS. ........................................88
5.4. PERFORMANCE ACCEPTANCE CRITERIA FOR CFS BUILDINGS ..........................................91
5.5. PROCEDURE OF PBD ASSESSMENT FOR CFS BUILDINGS ....................................................96
5.6. COMPUTER PROGRAM FOR THE PBD ASSESSMENT OF CFS BUILDINGS ..........................101
CHAPTER 6 EXAMPLES ..................................................................................................................103
6.1. INTRODUCTION ...................................................................................................................103
6.2. COMPARISON OF LINEAR SFEA AND CFEA .....................................................................103
6.2.1. EXAMPLE 1: ANALYSIS OF AN ISOLATED SWP USING SFEA AND CFEA ........................104
6.2.2. EXAMPLE 2: THREE-STOREY BUILDING ............................................................................107
6.2.3. EXAMPLE 3: FIVE-STOREY BUILDING ...............................................................................114
6.3. PERFORMANCE-BASED DESIGN ASSESSMENT OF CFS BUILDINGS...................................122
6.3.1. EXAMPLE 4: THREE STOREY BUILDING .............................................................................123
6.3.2. EXAMPLE 5. FIVE-STOREY BUILDING ................................................................................130
CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS ......................................................140
7.1. CONCLUSIONS .....................................................................................................................140
7.2. RECOMMENDATION FOR FUTURE WORK ..........................................................................143
REFERENCES ..........................................................................................................................................145
APPENDIX A SIMPLIFIED METHOD OF CALCULATING THE STRUCTURE PERIOD OF
VIBRATION..............................................................................................................................................151
APPENDIX B CALIBRATION OF THE ΑB AND ΑV COEFFICIENTS............................................157
APPENDIX C CALIBRATION OF THE Β EXPONENT.....................................................................159

ix
 LIST OF TABLES

TABLE 2.1. PERFORMANCE OBJECTIVES, SEAOC (1995) ..............................................................................15

TABLE 3.1 COMPARISON BETWEEN PREDICTED AND TEST (BRANSTON ET AL., 2006) RESULTS FOR LATERAL
STRENGTHS AND DISPLACEMENTS .........................................................................................................40
TABLE 3.2. COMPARISON BETWEEN PREDICTED AND TEST (BRANSTON ET AL., 2006) RESULTS FOR LATERAL
STRENGTHS AND DISPLACEMENTS .........................................................................................................41
TABLE 3.3. COMPARISON BETWEEN PREDICTED AND TEST (AISI, 2004) RESULTS FOR LATERAL STRENGTHS 42
TABLE 3.4. COMPARISON BETWEEN PREDICTED AND TEST (SERRETTE ET AL., 2002) RESULTS FOR LATERAL
STRENGTHS ...........................................................................................................................................43
TABLE 3.5. COMPARISON BETWEEN PREDICTED AND TEST (FULOP AND DUBINA, 2004) RESULTS FOR
LATERAL STRENGTHS AND DISPLACEMENTS .........................................................................................44
TABLE 3.6. COMPARISON BETWEEN PREDICTED AND TEST (SERRETTE ET AL., 1996) RESULTS FOR LATERAL
STRENGTHS AND DISPLACEMENTS .........................................................................................................45
TABLE 3.7. COMPARISON BETWEEN PREDICTED AND TEST (ER-5762, 2003) RESULTS FOR LATERAL
STRENGTHS AND DISPLACEMENTS .........................................................................................................46
TABLE 4.1 NUMBER OF SHELL AND FRAME ELEMENTS IN THE MODELING OF THE SWP..................................67
TABLE 4.2 PROPERTIES OF THE SHEATHING AND FRAMING STUDS EMPLOYED FOR DETERMINING THE
PROPERTIES OF THE EQUIVALENT SHELL ELEMENTS ..............................................................................68
TABLE 4.3 PROPERTIES OF THE EQUIVALENT SHELL ELEMENTS......................................................................68
TABLE 4.4 DISPLACEMENTS (MM) OF NODE 1 FOR THE SWP, PREDICTED WITH SFEA AND CFEA ................68
TABLE 4.5 MAXIMUM AXIAL FORCES IN THE STUDS OF THE SWP FROM CFEA AND SFEA ...........................69
TABLE 4.6 DISPLACEMENTS OF NODE 1 AND AXIAL FORCES OF THE SWP, PREDICTED WITH CFEA..............72
TABLE 5.1. PERFORMANCE OBJECTIVE SITE PARAMETERS, GONG (2003) ......................................................87
TABLE 5.2. VALUES OF FA AS A FUNCTION OF THE SITE CLASS AND MAPPED SHORT-PERIOD MAXIMUM
CONSIDERED EARTHQUAKE SPECTRAL ACCELERATION, FEMA 450 (2003) ..........................................88
TABLE 5.3. VALUES OF FV AS A FUNCTION OF THE SITE CLASS AND MAPPED 1 SECOND PERIOD MAXIMUM
CONSIDERED EARTHQUAKE SPECTRAL ACCELERATION, FEMA 450 (2003) ..........................................88
TABLE 5.4. DETERMINATION OF THE LIMIT DRIFT RATIOS FOR CFS SHEAR WALL PANELS .............................95
TABLE 6.1. QUANTITY OF ELEMENTS ...........................................................................................................106
TABLE 6.2. COMPARISON OF RESULTS FOR EXAMPLE 1................................................................................106
TABLE 6.3. PROPERTIES OF THE SHEATHING AND CFS FRAMING MATERIAL.................................................109
TABLE 6.4. CONSTITUTIVE PROPERTIES OF THE EQUIVALENT SHELL ELEMENTS ...........................................110
TABLE 6.5. DISPLACEMENT AND LATERAL DRIFT AT SELECTED JOINTS, MM ................................................111
TABLE 6.6. SWP LATERAL STRENGTH PR AND LATERAL FORCE PA, KN.......................................................112
TABLE 6.7. COMPRESSION FORCE IN THE MOST CRITICAL STUD OF EACH SWP, KN .....................................113
TABLE 6.8. TENSION FORCE IN THE MOST CRITICAL STUD OF EACH SWP, KN ..............................................114
TABLE 6.9. SHEATHING AND CFS FRAMING MATERIAL PROPERTIES ............................................................116
TABLE 6.10 CONSTITUTIVE PROPERTIES OF THE EQUIVALENT SHELL ELEMENTS ..........................................117
TABLE 6.11. DISPLACEMENT AND LATERAL DRIFT AT SELECTED JOINTS (MM).............................................118
TABLE 6.12. SWP LATERAL STRENGTH PR AND LATERAL FORCE PA ON THE, KN .........................................120
TABLE 6.13. COMPRESSION FORCE IN THE MOST CRITICAL STUD OF EACH SWP, KN ...................................121
TABLE 6.14. TENSION FORCE IN THE MOST CRITICAL STUD OF EACH SWP, KN ............................................122
TABLE 6.15. TARGET BASE SHEARS, KN.......................................................................................................125
TABLE 6.16. LATERAL LOADS IN THE BUILDING ...........................................................................................126
TABLE 6.17. SWP INTER-STOREY DRIFT, MM ...............................................................................................126
TABLE 6.18. SWP LATERAL STRENGTH PR AND LATERAL FORCE PA, KN......................................................128

x
TABLE 6.19. SWP STIFFNESS DEGRADATION COEFFICIENT Λ........................................................................128
TABLE 6.20. COMPRESSION FORCE IN MOST CRITICAL STUD OF EACH SWP, KN ..........................................129
TABLE 6.21. TENSION FORCE IN MOST CRITICAL STUD OF EACH SWP, KN ...................................................130
TABLE 6.22. TARGET BASE SHEARS IN THE X DIRECTION ..............................................................................131
TABLE 6.23. LATERAL LOADS IN THE BUILDING IN THE X DIRECTION ...........................................................131
TABLE 6.24. SWP INTER-STOREY DRIFT IN THE X DIRECTION, MM................................................................132
TABLE 6.25. SWP STIFFNESS DEGRADATION COEFFICIENT Λ........................................................................133
TABLE 6.26. SWP LATERAL STRENGTH PR AND LATERAL FORCE PA, KN......................................................134
TABLE 6.27. COMPRESSION FORCE IN THE MOST CRITICAL STUD OF EACH SWP, KN ...................................134
TABLE 6.28 TENSION FORCE IN THE MOST CRITICAL STUD OF EACH SWP, KN .............................................135
TABLE 6.29. TARGET BASE SHEARS FOR ANALYSIS IN THE Y DIRECTION .......................................................136
TABLE 6.30. LATERAL LOADS IN THE BUILDING FOR ANALYSIS IN THE Y DIRECTION ....................................136
TABLE 6.31. SWP INTER-STOREY DRIFT, MM ...............................................................................................137
TABLE 6.32. SWP LATERAL STRENGTH PR, LATERAL FORCE PA, KN, AND STIFFNESS DEGRADATION
COEFFICIENT, Λ ...................................................................................................................................137
TABLE 6.33. COMPRESSION FORCE IN MOST CRITICAL STUD OF EACH SWP, KN ..........................................138

xi
 LIST OF FIGURES

FIGURE 3.1. FASTENER ARRANGEMENT NOTATION.........................................................................................27

FIGURE 3.2. PANEL ROTATION AND FORCE DISTRIBUTION ..............................................................................33
FIGURE 3.3. SHEAR WALL PANEL TESTED BY BRANSTON ET AL. (2006) .........................................................34
FIGURE 4.1 MODELING OF A CFS BUILDING USING FINITE ELEMENT ANALYSIS .............................................49
FIGURE 4.2. A) PANEL CROSS-SECTION, AND B) EQUIVALENT SHELL ..............................................................51
FIGURE 4.3. SIXTEEN-NODE SHELL ELEMENT .................................................................................................55
FIGURE 4.4. VECTORS AT NODE K ...................................................................................................................58
FIGURE 4.5. SWP MODEL FOR (A) SFEA AND (B) CFEA................................................................................67
FIGURE 4.6. DISPLACEMENT COMPARISONS AT NODE 1 WITH SFEA AND CFEA............................................69
FIGURE 4.7. MAXIMUM AXIAL FORCES IN THE STUDS OF THE SWP ................................................................70
FIGURE 4.8. MAXIMUM BENDING MOMENT IN STUDS OF THE SWP.................................................................70
FIGURE 4.9. MAXIMUM AXIAL FORCES IN THE STUDS OF THE SWP ................................................................72
FIGURE 4.10. CHARACTERIZATION OF THE SWP LOSS IN STRENGTH ..............................................................73
FIGURE 4.11. PREDICTED VS. EXPERIMENTAL (BRANSTON ET AL., 2004) CURVES OF OSB SWP (SC = 152MM)
.............................................................................................................................................................76
FIGURE 4.12. PREDICTED VS. EXPERIMENTAL (BRANSTON ET AL., 2004) CURVES OF OSB SWP (SC = 102MM)
.............................................................................................................................................................76
FIGURE 4.13. PREDICTED VS. EXPERIMENTAL (BRANSTON ET AL., 2004) CURVES OF OSB SWP (SC = 76 MM)
.............................................................................................................................................................77
FIGURE 4.14. PREDICTED VS. EXPERIMENTAL (BRANSTON ET AL., 2004) CURVES OF DFP SWP (SC = 152MM)
.............................................................................................................................................................77
FIGURE 4.15. PREDICTED VS. EXPERIMENTAL (BRANSTON ET AL., 2004) CURVES OF DFP SWP (SC = 102MM)
.............................................................................................................................................................78
FIGURE 4.16. PREDICTED VS. EXPERIMENTAL (BRANSTON ET AL., 2004) CURVES OF DFP SWP (SC = 76 MM)
.............................................................................................................................................................78
FIGURE 4.17. PREDICTED VS. EXPERIMENTAL (NAHBRC, 1997) CURVES OF A 40 FT OSB SWP (SC = 152MM)
.............................................................................................................................................................79
FIGURE 4.18. PREDICTION VS. EXPERIMENTAL (COLA-UCI, 2001) CURVES OF STR SWP (SC = 152MM)......80
FIGURE 4.19. PREDICTIED VS. EXPERIMENTAL (COLA-UCI, 2001) CURVES OF STR SWP (SC = 51MM) ........81
FIGURE 4.20. PREDICTION VS EXPERIMENTAL (COLA-UCI, 2001) CURVES OF OSB SWP (SC = 152MM) ......81
FIGURE 5.1. GENERAL RESPONSE SPECTRUM (FEMA 450, 2003)...................................................................85
FIGURE 5.2. SENSITIVITY TESTS OF THE RESPONSE CURVE FOR DIFFERENT SIZES OF LATERAL LOAD
INCREMENTS. ........................................................................................................................................90
FIGURE 5.3. COMPARISON OF THE BEHAVIOUR OF SWP WITH WOOD AND CFS FRAMING FOR DIFFERENT EDGE
SCREW SPACING: (A) 152MM AND (B) 51MM (SEAOC AND COLA-UCI, 2001)....................................93
FIGURE 5.4. NORMALIZED FORCE VERSUS DEFORMATION RATIO FOR WOOD ELEMENTS (FIG. 8-1, FEMA 273,
1997) ....................................................................................................................................................93
FIGURE 5.5. FLOWCHART FOR THE PBD ASSESSMENT OF CFS BUILDINGS ...................................................100
FIGURE 6.1. SWP MODEL FOR CFEA ...........................................................................................................105
FIGURE 6.2. THREE-STOREY CFS BUILDING .................................................................................................108
FIGURE 6.3. TYPICAL PLAN AND ELEMENT IDENTIFICATION OF THE THREE-STOREY CFS BUILDING ............108
FIGURE 6.4 BUILDING DEFORMATION DUE TO GRAVITY LOADS ....................................................................113
FIGURE 6.5. FIVE-STOREY CFS BUILDING ....................................................................................................115
FIGURE 6.6. TYPICAL PLAN AND ELEMENT IDENTIFICATION OF THE FIVE-STOREY CFS BUILDING ................116
FIGURE 6.7 A) FIFTH STOREY DISPLACEMENT IN THE Y DIRECTION, B) BUILDING INTER-STOREY DRIFT .......119

xii
FIGURE 6.8. INTER-STOREY DRIFTS AND DISPLACEMENTS OF THE SWP 1: LINEAR AND PUSHOVER ANALYSIS
...........................................................................................................................................................127
FIGURE 6.9. INTER-STOREY DRIFTS AND DISPLACEMENTS OF THE SWP 1: LINEAR AND PUSHOVER ANALYSIS
...........................................................................................................................................................133

xiii
 SYMBOLS

αB Sheathing’s bending stiffness reduction coefficient

αV Sheathing’s shear stiffness reduction coefficient
β Nonlinear exponent given as function of the screw edge spacing
∆ Shear wall panel lateral deformation, at its maximum strength
∆C Maximum deformation of a sheathing-to-framing connection
∆U Structure displacement increment vector
δy Distance from the elastic centre to the instantaneous centre of rotation
φ Rotation about the x local axis of the shell element
γ Rotation about the y local axis of the shell element
η Shear wall panel strength modification factor
ηs Frequency parameter, given as function of kp
κ Exponent depending on the period of the structure
λ Stiffness degradation coefficient
ρ Cross section shape factor
τ Cauchy stress matrix
τˆ Cauchy stress vector
νx, νy Sheathing’s Poisson ratios in the x and y directions, respectively
νxeq, νyeq Equivalent shell element’s Poisson ratios in the x and y directions,
respectively
A Profile’s cross section area
Ab Beam’s cross section area
Ac Column’s cross section area
AF Stud’s cross section area
AISI American Iron and Steel Institute
AS Sheathing’s cross section area
BL Linear strain-displacement matrix
BNL Nonlinear strain-displacement matrix

xiv
C Constitutive matrix
CFEA Conventional finite element analysis
CFS Cold-Formed Steel
CP Collapse Prevention
CSP Canadian Soft Plywood
Cu Ultimate strength factor
Cvx Vertical load distribution coefficient
c Distance from the stud’s centroid to sheathing’s mid-plane
D Shell’s bending rigidity
DFP Douglas Fir Plywood
Drift Inter-storey drift: displacement of the floor above minus the displacement
of the floor below
Dx, Dy Shell’s bending rigidities along the x and y directions respectively
d Normalized deformation
df Distance from the column’s centroid to floor plan’s centroid
di Distance from the screw i to the centre of rotation
dmax Distance between the centre of rotation and the farthest screw connection
ds Screw diameter
dxi, dyi x and y components of di
E Modulus of elasticity (Young modulus)
Eb Beam’s modulus of elasticity
Ec Column’s modulus of elasticity
EF Steel framing studs modulus of elasticity
ESx, ESy Sheathing’s modulii of elasticity in the x and y directions, respectively
Ex, Ey Reduced sheathing’s modulii of elasticity
Exeq, Eyeq Equivalent shell’s modulii of elasticity in the x and y directions,
respectively
e2 Unit vector in the direction of a shell’s local y axis
ey y component of the distance from the load to the elastic centre of rotation
F Vector of unbalanced internal forces
Fa Site coefficient

xv
Fgx Unbalanced force in the fastener group
Fu Material tensile strength
Fv Site coefficient
Fx Vector lateral forces in a building
Fy Material tensile yield strength
fb Building’s shear frequency
fs Building’s shear frequency
fg Frames’ bending frequency
Gb Column’s shear modulus of elasticity
GS Sheathing’s shear modulus of elasticity
GWB Gypsum Wall Board
Gxyeq Equivalent shell’s shear modulus of elasticity

g Gravity constant, 9.81 m/s2

H Building’s height
h Storey height
hx Height from the ground to the storey level x
hz Height from the ground to the storey level z
I Cross section moment of inertia (Second moment of area)
I3 3x3 identity matrix
Ib Beam’s cross section moment of inertia
Ic Column’s cross section moment of inertia
IF Steel stud’s cross section moment of inertia
IP Iterative procedure
IO Immediate occupancy
IS Sheathing’s cross section moment of inertia of
i Counter
J Polar moment of inertia of a fastener group
Jm Jacobian matrix
k Node number
Ka SWP stiffness at load increment q
Kb Frame global shear stiffness

xvi
Kc Columns’ shear stiffness
Ke Building’s effective shear stiffness
KF Steel framing studs’ lateral stiffness
Ki SWP initial stiffness (elastic)
KL Structure’s elastic stiffness matrix
KNL Structure’s nonlinear stiffness matrix
KS Sheathing’s lateral stiffness associated
Kt Steel stud’s effective length factor for torsion
Kx,y Steel stud’s effective length factors for x and y axes, respectively.
kL Shell element’s elastic stiffness matrix
kNL Shell element’s nonlinear stiffness matrix
kp Non-dimensional parameter
L Floor live load
Lr Roof live load
LDR Lateral drift ratio for SWP
LS Life Safety
l Shear wall panel length
M Moment strength of the fasteners group
Mo Moment produced by the unitary force about the elastic centre of rotation
Mp Moment produced by the unitary force about the instantaneous centre of
rotation
m Mass density per unit of building height
N Shell element shape functions
NEHRP National Earthquakes Hazards Reduction Program
Nx, Ny Shell axial rigidity along the x and y directions, respectively
n Number of nodes in a shell element
nC Total number of screws on the panel, used to attach the sheathing
nd Number of nodes per node line of the shell element
ne Number of shell elements
nF Number of studs
no Number of columns

xvii
np Number of shear wall panels
ns Number of screws along the top of the shear wall (i.e, l /sC)
nt Number of building storeys
OP Operational
OSB Oriented Strand Board
P Shell axial rigidity
Pa Lateral load on the SWP in consideration at load increment q
PBD Performance-Based Design
PF SWP steel framing lateral strength
Pfc Shear wall panel lateral strength due to failure of the end-stud
PL Performance level
Pn Nominal compressive strength of the end-stud
PO Performance objective
PR Shear wall panel lateral strength
PS Sheathing lateral strength
Px Lateral unitary force applied in the location of the actual force
q Load increment index for the nonlinear structural analysis
R Vector of external loads on the structure
Ri / Ru Normalized force on the screw connection i
r Shell element natural coordinate, along the Cartesian x coordinate
rc Stiffness combination factor
rf Reduction factor, given as function of the number of storeys
S1 Response acceleration parameter at one-second period
Sa Spectral acceleration
SFEA Simplified finite element analysis
SLI Strength loss index of shear wall panels
SM1 Design spectral response acceleration parameter at one second period
SMS Design short-period spectral response acceleration parameter
SP Simplified procedure
SS Short period response acceleration parameter
SWP Shear wall panel

xviii
s Shell element natural coordinate, along the Cartesian y coordinate
sC Edge screw spacing
T Fundamental period of the structure
Tϕ Transformation matrix for the shell element direction vectors
To Characteristic period
teq Equivalent shell thickness

tS Sheathing’s thickness
U Structure’s displacements vector
u Shell element’s displacement in the x direction
v Shell element’s displacement in the y direction
V1 Direction vector of the shell element local x axis
V2 Direction vector of the shell element local y axis
Va Building base shear at load increment q
Vb Base shear
Vb max Maximum target base shear
Vn Direction vector normal to the shell element surface
Vr Strength of a single sheathing-to-framing connection
W Structure’s seismic weight
Wg Gravitational loads
w Shell element’s displacement in the z direction
wSF Tributary width of a stud
wx Seismic weight corresponding to the storey level x
wz Seismic weight corresponding to the storey level z
x Cartesian coordinate in the x direction
xCi, yCi Coordinates of the i screw with respect the elastic centre of rotation
y Cartesian coordinate in the y direction
z Cartesian coordinate in the z direction

xix
 Chapter 1
Introduction

1.1. General
The construction of low-rise buildings using light-gauge materials such as Cold-Formed
Steel (CFS) is a common practice in North America, even in seismic areas. CFS has
numerous advantages over traditional construction materials such as wood, hot rolled
(structural) steel, and concrete. CFS framing is simpler to erect, reducing the construction
time and labour cost, and, consequently, the building overall cost. In addition, CFS is
durable and has a large strength-to-weight ratio. Homebuilders and consumers alike have
a general understanding of the benefits of CFS and are receptive to the use of steel
framing. This has led to a tremendous growth of the application of CFS in residential and
commercial construction that has long been predicted. Despite CFS virtues and its
growing application, however, opposition still exists to use CFS for structural elements in
the construction of mid-rise CFS buildings in seismic areas. The primary reasons are the
lack of information about the seismic performance of CFS buildings, and the complicated
structural analysis and design procedures associated with CFS members and systems.
Therefore, research and development, for the technical and practical aspects to eliminate
or substantially mitigate the impact of these barriers, is of primary importance.

Typically, the structural system of CFS buildings is different than that of structural steel
buildings. In CFS buildings, the primary structural elements are Shear Wall Panels
(SWP), load-bearing wall panels, and floor and roof panels; while in structural steel
buildings, the primary structural elements are columns, beams, and cross-bracings. The
functions of SWP, in addition to maintaining the stability and integrity of the structural
system, are to resist in-plane lateral and gravity loads, and out-of-plane wind loads if the
SWP are used as exterior walls. In general, SWP in CFS framing are constructed with

1
vertically spaced and aligned C-shape CFS studs. The ends of the studs are screwed to the
top and bottom tracks. The structural sheathing can be installed on one or both sides of
the wall. Load-bearing wall panels are built similarly to SWP, except that no structural
sheathing is attached to the framing. Consequently, load-bearing wall panels lack the
lateral strength and, therefore, are only capable to resist gravity loads. Although the floor
and roof panels are built in the same fashion as the SWP, a concrete slab poured on the
CFS deck can be used instead of wood sheathing. The primary loads on the floor and roof
panels are gravity loads, and in-plane lateral loads associated with wind, or earthquakes
when the panels serve as horizontal diaphragms.

The seismic assessment of CFS buildings is not addressed by the National Building Code
of Canada (NBCC, 2005), nor by the North American Specification for the Design of
Cold-Formed Structural Members (S136-01). Therefore, engineers are not provided with
guidelines for the seismic assessment of CFS buildings. The seismic design of CFS
buildings is carried out in accordance with seismic design standards. Traditionally, the
principal objective of seismic design standards is to prevent the collapse of structures
subjected to design earthquakes, disregarding the associated economical losses. In the last
decade, a new philosophy of seismic design known as Performance-Based Design (PBD)
was introduced by the Federal Emergency and Management Agency (FEMA 273, 1997)
for the seismic assessment and rehabilitation of existing structures. FEMA combines
multiple building performance levels with seismic hazards, represented by design
earthquakes, to “assure” the expected behaviour of the structure. For a building subjected
to a design earthquake, a performance level refers to a limit damage state. A limit damage
state is represented by displacements, inter-storey drifts, base shears, loads, stresses,
accelerations, or by other limit states (Ghobarah, 2001). Using PBD assessment
procedures, designers and building owners are aware of the damage in a building that
satisfies the specified performance levels for the design earthquake. Therefore, PBD
assessment can provide a better understanding of the behaviour of CFS buildings,
resulting in safer and more economical designs.

2
The PBD assessment of buildings is carried out by different analysis methods, such as
linear static, linear dynamic, nonlinear static, and nonlinear dynamic analyses (FEMA
273, 1997). As such, the accuracy of the results and the complexity of the structural
analysis of a building vary, depending on the chosen method. In this study, the nonlinear
static analysis method, also known as pushover analysis, is employed due to its simplicity
and accuracy for multi-storey buildings with a predominant fundamental mode of
vibration in their response (Krawinkler and Seneviratna, 1998). Essentially, pushover
analysis consists of transforming the building under consideration into an equivalent
single degree of freedom system. Then, the seismic loads are applied on the building in
small increments and structural analysis is carried out after each load increment, and the
results are accumulated over the loading history.

Pushover analysis for PBD assessment has been successfully implemented for buildings
constructed of different materials such as concrete, steel, and wood. However, PBD
assessment has not yet been implemented in CFS buildings. To do this, it is first
necessary to characterize force-displacement response curves, develop acceptance criteria
associated with the performance levels, and establish a procedure for the analysis and
design of the lateral-load resistant elements.

In this study, the acceptance criteria (limit damage state) for CFS buildings are
established as a function of the lateral drift and lateral strength of the SWP. Although
FEMA 273 (1997) provides limit drift ratios as acceptance criteria for different
performance levels and types of structural systems, FEMA does not provide such limit
drift ratios for CFS SWP. The limit drift ratios for SWP are determined from
experimental data. The acceptance criteria include account for the lateral strength of
SWP, which must be checked to assure that the SWP do not fail prior to satisfying the
specified performance objectives.

Since SWP exhibit a nonlinear behaviour when subjected to lateral loads, as

demonstrated in experimental investigations (Gad et al., 1999; Branston et al., 2006;
Serrette et al., 2002), a stiffness degradation model to characterize the nonlinear response

3
of CFS buildings is needed for the pushover analysis. Only then can the response of CFS
buildings be accurately predicted. The stiffness degradation model represents the loss of
the lateral strength of the SWP as the lateral loads applied on the building are increased.
Several models characterizing the nonlinear behaviour of SWP are found in the literature,
including those developed by Branston et al. (2006), and by Fulop and Dubina (2004).
These models are determined from experimental testing on SWP. This study proposes a
general model, defined by a function of the lateral strength of the SWP and the spacing of
the screws at the edge of the panel.

Due to the nature of the structural elements used for CFS buildings, the structural analysis
of such buildings should be carried out by Conventional Finite Element Analysis
(CFEA). However, CFEA is time-consuming and cumbersome for a mid-rise CFS
building because it requires a large number of elements to generate the structural model.
In addition, if conventional nonlinear pushover analysis is employed for the PBD
assessment of a CFS building, the computation time is significant. Therefore, the
development of a simplified analysis method which is also practical to use is of primary
importance.

Only a few methods to simplify the analysis of SWP are published in the literature. Fulop
and Dubina (2004) have proposed a simplified method that involves replacing the
sheathing by equivalent cross-bracing. The method provides accurate results for isolated
SWP under seismic loads, but not for SWP subjected to gravity or out-of-plane loads. In
this study, the simplification is carried even further by recognizing that the individual
modeling of the studs and sheathing plates in a building is not needed. Instead, the SWP
are transformed into flat shell elements with equivalent properties for modeling complete
panels. The equivalent shell elements account for the properties of the studs and
sheathing, and are modeled by using a sixteen-node shell element for each panel. Then, a
nonlinear finite element method is employed to perform the structural analysis. This
procedure is called Simplified Finite Element Analysis (SFEA), for which the number of
elements required to model a mid-rise CFS building is significantly less than that for

4
CFEA. As a result, less, time and resources are required to generate the model and
conduct the analysis of a mid-rise CFS building.

In the PBD assessment of a CFS building, the lateral strength and stiffness of the SWP
must be computed and checked with applicable acceptance criteria. Due to the
complexity of the interaction among wall components, evaluating the lateral strength and
stiffness of a SWP has challenged structural engineers. The lateral strength of a SWP
cannot be determined by the strength of the sheathing alone, as the interaction among the
sheathing, studs and fasteners affect considerably both the behaviour and lateral strength
of SWP. Typically, due to the lack of analytical methods, the lateral strength of a SWP
must be determined by experimental testing.

The American Iron and Steel Institute (AISI, 2004) has published design tables for the
lateral strength of SWP having different characteristics, such as sheathing materials,
screw diameter and spacing, and stud thickness. However, the tables are derived from
experimental testing, so that their application is limited to the types of SWP tested. In this
study, a method for determining the lateral strength of SWP is developed, which is
versatile because SWP with different configurations and material properties can be
analyzed. The principal characteristics and properties of SWP are accounted for, such as
thickness and material properties of the sheathing, cross-section and material properties
of the studs, spacing of the studs, and diameter and spacing of the screws. The proposed
method is an extension of that developed by Brant (1982), which is currently used in steel
standards such as CSA S16-01 (2003) to calculate the moment strength of steel bolted
connections with eccentric loads. Brant’s method is employed because of the similarity of
SWP with steel moment connections, in that both are composed of arrays of fasteners
subjected to eccentric loads.

5
1.2. Objective of Study
The primary objective of this study is to develop a methodology for conducting the PBD
assessment of mid-rise CFS buildings. The accomplishment of this objective involves
achieving a number of particular goals that are described in the following:

• Development of a method for determining the lateral strength of SWP

• Development of an equivalent shell formulation to transform CFS panels into
shell elements with equivalent properties
• Establishment of a formulation for SFEA of CFS buildings
• Validation of the proposed SFEA
• Development of a stiffness degradation model for SWP
• Establishment of an appropriate pushover analysis procedure
• Establishment of acceptance criteria related to specified performance levels for
SWP
• Presentation of example applications of PBD assessment of CFS buildings using
the proposed methodology

1.3. Scope
The method for determining the lateral strength of SWP accounts for the failure of the
sheathing-to-framing connections, and the failure of the end-studs in compression due to
in-plane loads. The method can be used to calculate the lateral deformation of isolated
SWP. However, the failure of a SWP due to the possible local buckling of sheathing
material is not considered in the strength evaluation.

The panels in a CFS building are transformed into flat shell elements, with equivalent
properties that account for the axial and bending stiffness of the SWP. It is assumed that
the studs are uniformly distributed along the length of the wall, and that a continuous
bond exists between the studs and sheathing.

6
The SFEA employs sixteen-node shell elements to model the CFS panels. One or more
sixteen-node shell elements are used to model each panel. An updated Lagrangian
formulation is adopted for the nonlinear finite element analysis, accounting for the
geometric and material nonlinearities. The geometric nonlinearities are accounted for in
the nonlinear analysis procedure by updating the nodal coordinates and member forces at
each load increment. The material nonlinearities are accounted by using a stiffness
degradation model for the SWP.

The stiffness degradation model characterizes the lateral nonlinear behaviour of a SWP.
This is accomplished by reducing SWP lateral stiffness as the lateral loads on the
building increase. The model is a function of the lateral strength of the SWP, and the
screw spacing at the edge of the SWP. It is noted that the model does not account for
axial or bending stiffness degradation of a SWP.

A single-mode pushover analysis is applicable for low- and mid-rise buildings whose
seismic response is governed by the first mode of vibration. Thus, high-rise buildings
may not be appropriately analyzed with this method as account for the influence of higher
vibration modes may required for such buildings. Spectrum-based (also known as force-
controlled) pushover analysis is employed. The lateral loads are applied on the building
in a pre-defined pattern in load increments of 1% of the total loading applied on the
building.

The acceptance criteria for CFS buildings are established as a function of the SWP inter-
storey drifts and lateral strength. This study relates the inter-storey drifts to the four
performance levels defined by FEMA 273 (1997).

1.4. Assumptions and Idealizations

The following assumptions and idealizations are adopted to determine the lateral strength
of SWP:

7
 • Two types of failure are considered in the prediction of SWP lateral strength: 1)
failure of the sheathing-to-framing connections, and 2) failure of the end-studs
under axial forces.
• The proposed procedure does not account for the effect of local or shear buckling
in the sheathing; thus, it is not applicable for SWP with thin sheathing such as
CFS sheets or soft plywood.
• It is assumed that the hold-down anchors have been property designed to resist the
uplift forces in the SWP. Thus, hold-down anchor failure is not accounted for in
the determination of the SWP lateral strength.
• Lateral deformation due to hold-down anchor rotation is not accounted for in the
prediction of the SWP lateral drifts, nor in the finite element analysis.
• It is assumed that the strength of the sheathing-to-framing connection is the same
in all directions.
• The lateral deformation of a SWP is computed at the load level corresponding to
the lateral strength of the SWP.

The following assumptions and idealizations are adopted for determining the properties
of the equivalent shell elements:
• The studs are spaced uniformly along the length of the panel.
• A continuous bond exists between the sheathing and the studs.
• The gross cross-section properties of the CFS studs are used in the calculations.

The following assumptions and idealizations are adopted for the SFEA:
• The normal stresses transverse to the equivalent panel are disregarded for plane
stress analysis.
• When a SWP fails under lateral loads, the vertical stiffness contributed by the
studs is accounted for in the analysis.
• The sixteen-node shell element used to model CFS panels has five degrees of
freedom; i.e., three translations along the x, y and z axes, and two rotations about
the x and y axes in the plane of the shell.
• An updated Lagrangian formulation is adopted for carrying out the SFEA.

8
The following assumptions and idealizations are adopted for the pushover analysis:
• The pushover analysis procedure is only used for low-rise and mid-rise CFS
buildings.
• The seismic loads are distributed over the height of the building using the
distribution pattern specified by FEMA 273 (1997).
• The degradation of the stud-track and stud-joist connections is not taken into
consideration for the pushover analysis.

1.5. Thesis Overview

Chapter 1 briefly describes the problem to be solved by this study, and the proposed
methodology to solve it. The assumptions, idealizations and scope of the methodology for
carrying out the PBD assessment of CFS buildings are described.

Chapter 2 presents a literature review of PBD and pushover analysis for seismic
assessment. Their features, background, development and current trends are summarized.

Chapter 3 develops the methodology for determining the lateral strength of SWP. The
chapter describes the SWP modes of failure that are accounted for in the proposed
methodology. An iterative and simplified method for computing the lateral strength of
shear wall panels is introduced. A hand-worked example application of the method is
presented. Finally, the strengths and deformations of isolated SWP are compared using
predicted and experimental results.

Chapter 4 develops the methodology for carrying out the simplified finite element
analysis (SFEA) of CFS buildings, including: the formulation for transforming the panels
into equivalent shell elements, the stiffness degradation model for the SWP, the equations
for estimating the axial force in studs, and the nonlinear finite element formulation for the
sixteen-node shell elements. The results of linear SFEA of isolated SWP are compared
with the results computed using SAP2000 (2006). Finally, the experimental results for

9
isolated SWP obtained by Branston et al. (2006) are compared with the numerical
predictions obtained using the SFEA proposed by this study.

Chapter 5 develops the methodology for carrying out the PBD assessment of CFS
buildings using spectrum-based pushover analysis. The acceptance criteria for SWP are
derived from experimental data. A step-by-step procedure for the PBD assessment of
CFS buildings is presented along with a corresponding flowchart. Finally, a computer
program created for carrying out the PBD assessment of CFS buildings is described.

Chapter 6 presents two simplified building models analyzed using linear SFEA and
CFEA. The results from the two analyses are compared. Pushover analysis is carried out
for the two building models to demonstrate the practical application of PBD assessment
for CFS buildings. Finally, the results from the linear analyses are compared to those
from the pushover analyses to evaluate the differences between the various approaches.

Chapter 7 discusses the findings and conclusions of this study. Several recommendations
are given for future studies regarding the PBD assessment of CFS buildings.

Appendix A provides a simplified method for computing the fundamental period of

vibration for buildings comprised of frameworks and shear walls.

10
 Chapter 2
Background and Literature Review on Performance-
Based Design and Pushover Analysis

2.1. Introduction
The seismic assessment of existing structures and the seismic design of new structures
have taken a step forward since the publication of the National Earthquake Hazards
Reduction Program (NEHRP) guidelines for the seismic rehabilitation of buildings
(FEMA 273, 1997). These guidelines introduced the Performance-Based Design (PBD)
philosophy, which provided designers and building owners more options to design
buildings than seismic codes. For instance, with PBD, a building can be designed to
remain operable, or to undergo limited damage after an earthquake. Also, FEMA’s
guidelines have introduced nonlinear static analysis, better known as pushover analysis,
as an alternative to dynamic analysis, which is a convenient tool for seismic analysis. An
earthquake is simulated by applying lateral forces on the structure, in small increments, in
a predefined pattern. The pattern of lateral forces remains constant throughout the
analysis, which is determined assuming that the response of the building is controlled by
a single mode of vibration, usually the fundamental mode (Krawinkler and Senervitna,
1998). As such, this method of analysis is appropriate for low and mid-rise structures,
including CFS buildings, the objective of study in this research.

Discussed in the first part of this chapter, is the background of PBD, mainly based on
FEMA (1997, 2003), SEAOC (1995), and ATC (2003, 2005) documents. Summarized in
this chapter is the background of pushover analysis, including the primary considerations
and assumptions in the application of the method. Also described is the so-called second-
generation PBD. The guidelines of this latter, new, philosophy have not been completed,
but they are expected to be completed by 2010 (ATC, 2005). At this time, a draft with
25% of the guidelines has been published by ATC (2005). An overview of the second-

11
generation PBD objectives and characteristics is given. Described in the last part of the
chapter is the current situation for the seismic assessment of CFS buildings.

2.2. Performance-Based Design

After several major seismic events that occurred in the recent past, such as the 1989
Loma Prieta and the 1994 Northridge earthquakes, the structural engineering community
and building owners began to question the effectiveness of current building codes to
protect property (Gong, 2003). Seismic codes at that time were prescriptive and primarily
concerned with life safety, their primary objective being to protect building occupants
and pedestrians from life threatening situations. Thus, buildings that did not collapse
during an earthquake and did not generate large amounts of falling debris, were
considered to fully comply with seismic code requirements. This said, however, there are
many buildings for which it is desirable that they be functional after an earthquake event,
such as hospitals, shelters, and important facilities (ATC 58-2, 2003).

In recognition of the different performance demands possible for different building types,
in 1993 the Federal Emergency Management Agency (FEMA) provided funding to
various organizations to develop the NEHRP guidelines; namely FEMA 273 (1997) and
FEMA 274 (1997), for Seismic Rehabilitation of Buildings (ATC 58-2, 2003). The
organizations in charge of developing the guidelines were the Applied Technology
Council (ATC), American Society of Civil Engineers (ASCE), and the Building Seismic
Safety Council (BSSC). These guidelines laid down the foundation for the PBD
philosophy, which were primarily created for seismic assessment and rehabilitation of
existing structures. Later in 1994, FEMA also awarded the Structural Engineers
Association of California (SEAOC) a project to develop a framework for the PBD of new
buildings, extending the concepts of FEMA 273. The project was known as VISION
2000 (ATC 58-2, 2003).

Currently, the PBD philosophy is widely accepted and used for assessing the performance
of existing and new buildings, subjected to seismic loads. PBD assessment provides a

12
good understanding of a structure’s behaviour, and allows building owners to have a
better idea of a building’s damages at different levels of earthquake intensity. The PBD
philosophy can be defined as multi-level design that not only has explicit concern for the
performance of a building at the ultimate-strength limit states, but also at intermediate
and serviceability limit states (Hasan et al., 2002). In this philosophy, the design criteria
are expressed in terms of the specified performance objectives that are chosen depending
on the performance expected for the structure. A performance objective involves the
combination of the structure’s expected performance level with a seismic hazard (Bertero
and Bertero, 2002). That is, a performance objective dictates the intensity of the seismic
hazard that the building will be subjected to, and the limit damage the building should
experience.

A performance level is a discrete damage state, selected from among a number of damage
possibilities (Gong, 2003). FEMA 273 (1997) describes three performance levels for
structural components and four for non-structural components, which are combined to
generate four performance levels for the assembled building. For the latter, the most
common and representative performance levels in the design and rehabilitation of
buildings are Operational (OP), Immediate Occupancy (IO), Life Safety (LS), and
Collapse Prevention (CP).

At the OP level, the building is expected to be suitable for normal use and occupancy
after an earthquake, and the risk to life safety in the building is extremely low, but some
non-essential services may not function (FEMA 273, 1997). Buildings at the IO
performance level are safe to reoccupy after an earthquake. However, non-structural
systems may not function due to either lack of electrical power or damage of the
equipment. Although the building may require some reparations before re-occupancy,
minimal or no damage to structural elements is expected and only minimal damage to
non-structural components is expected (FEMA 273, 1997). Buildings at the LS level
undergo extensive damage to structural and non-structural components, and reparations
must be done before re-occupancy. Although reparations may be costly, risk to life safety
is low in buildings meeting this performance level (FEMA 273, 1997). Buildings in the

13
CP level have reached a state of impeding partial or total collapse, and they may have
suffered a significant loss of strength and stiffness with some permanent lateral
deformation. Yet, the major components of the gravity load carrying system should
continue carrying the gravity load demands (Gong, 2003). This building may be
dangerous to life safety due to the failure of non-structural components. Most buildings at
this performance level are considered complete economical losses.

A seismic hazard at a given site is represented by ground motions and its associated
probability of occurrence (Bertero and Bertero, 2002). FEMA 273 (1997) identifies four
seismic hazard levels with different mean return periods rounded to 2500, 500, 225 and
75 years, respectively. These seismic hazard levels are usually represented by their
probability of exceedance in a 50 year period (i.e. 2%/50, 10%/50, 20%/50, and 50%/50
for severe to light ground motion intensities, respectively).

Performance objectives can be generated by combining the aforementioned performance

levels and earthquake hazards. Typically, multiple performance objectives are selected
for the design of new structures and the rehabilitation of existing structures. In the design
of new structures, the performance objectives can be established in accordance with the
importance of the building and the budget available to build it. The performance
objectives can range from minimum code requirements (e.g., the OP performance level
for a 50%/50 year seismic hazard, and the LS performance level for a 10%/50 year
seismic hazard) to high performance requirements (e.g., the OP performance level for a
2%/50 year seismic hazard) (Bertero and Bertero, 2002). The described high performance
objective, poses high demands on buildings, since the building must remain operational
for the largest seismic hazard. FEMA 273 (1997) proposes three different performance
objectives for the rehabilitation of structures: basic, enhanced and limited safety. These
are based on the performance and damage expected for the structure rather than the type
of structure. On the contrary, the performance objectives established by SEAOC (1995),
for new structures, are established depending on the importance of the structure.

14
Shown in Table 2.1 are different performance objectives for the design of new structures.
Each cell represents a performance objective which is the result of combining a
performance level with a seismic hazard. Also shown in the table are the different multi-
performance objectives that should be satisfied by a structure in accordance with its
importance, such as ordinary building, essential building or hazardous facility. For
instance, a hazardous facility should meet the OP and IO performance levels for 10%/50
and 2%/50 earthquake hazards, respectively. Although different performance objectives
can be created combining other performance levels with earthquake hazards, the
performance objectives shown in Table 2.1 are suitable for most buildings.

Table 2.1. Performance Objectives, SEAOC (1995)

Once the multiple performance objectives are selected for designing a structure, they
need to be transformed into damage targets and acceptance criteria to facilitate the
analysis. The procedure to transform a performance objective into a damage target
involves consideration of factors such as the building location and natural period of
vibration, among others (this is discussed for CFS buildings in Chapter 5). Typically, the
damage targets are established as a function of the lateral displacement and base shear for
each earthquake hazard. The acceptance criteria represent the capacity of the structural
elements in the building at the different performance levels, given as a function of plastic
hinge rotation capacity, shear distortion capacity, and inter-storey drift capacity (FEMA
274, 1997). Once the damage targets are established, the analysis of the building is then

15
carried out to determine the displacement demand of the building and compare it to the
limit drifts.

FEMA 273 (1997) has established four structural analysis procedures that can be used
with the PBD assessment: linear static analysis, linear dynamic analysis, nonlinear static
analysis, and nonlinear dynamic analysis. The advantages and disadvantages of each
approach should be assessed to determine which procedure is appropriate for each
particular building to be analyzed. Although the linear static analysis procedure is the
simplest applicable method, its capabilities are very limited. It gives wrong answers when
used for analyzing buildings that have inelastic responses, such as buildings with highly
irregular structural systems or with energy dissipating devices (FEMA 273, 1997). On the
other hand, the dynamic analysis procedures (linear and nonlinear) can be used for most
structures and generate good results. However, the main disadvantage of dynamic
methods is that their application is complicated, and the designer needs to have a deep
knowledge of the mathematics of the method (FEMA 273, 1997).

The nonlinear static analysis procedure, better known as pushover analysis, is simple to
apply and often yields good results for structures with a predominant fundamental period
of vibration. However, pushover analysis should not be used for analyzing structures for
which higher-mode vibration effects are significant, such as structures of irregular plan,
structures with irregular distribution of their mass along their height, and structures with
seismic isolation devices (FEMA 273, 1997).

2.3. Pushover Analysis.

Pushover analysis consists in subjecting a building’s model to constant gravity loads and
monotonically increasing lateral forces or displacements, until either the specified
damage target is exceeded or the building collapses. This method accounts for inelastic
behaviour and provides a reasonable approximation of the response of the building when
subjected to the design earthquake (FEMA 273, 1997). The inelastic response of a
building is the consequence of geometric and material nonlinear behaviour of its

16
structural elements. Geometric nonlinearities are obtained from the analysis procedure by
updating the nodal coordinates of the building’s model at each load increment.
Depending on the chosen formulation for the structural analysis, geometric stiffness
matrices can be used to consider second-order geometric effects. Material nonlinearities
are accounted for in the nonlinear behaviour of the lateral-load resistant elements in the
building.

Pushover analysis is employed to evaluate the demand of a building subjected to design

earthquakes associated with the specified performance objectives. Then, the building’s
demand is compared to the building’s capacity. Here, the designer can determine if the
building is appropriately designed to satisfy the acceptance criteria associated with the
performance levels.

Single-mode pushover analysis is based on the assumption that a structure can be

modeled as a single degree of freedom system. The building response is typically
assumed to be dominated by only the first mode shape, which is assumed to remain
constant throughout the analysis. Although these assumptions are somewhat simplistic,
studies have indicated that the response of multi-degree of freedom systems that have one
dominant mode shape is modeled quite well (Krawinkler and Seneviratna, 1998). A
structure that has more than 75% of its mass participating in its first vibration mode is
considered to have a predominant mode shape (FEMA 273, 1997). If other mode shapes
also have a high participation in the structure response, multi-mode (modal) pushover
analysis should be employed (Chopra and Goel, 2002).

When assessing the behaviour of a building using pushover analysis, the designer first
must transform each performance objective into a corresponding damage target. The most
common damage targets are displacements and forces. For example, suppose a damage
target is represented by a displacement that is the maximum roof-level displacement
likely to be experienced by the structure during the design earthquake. The target
displacement is compared to the displacement measured at a roof-level control node, as
the lateral loads or displacements are incrementally applied on the building. If the

17
displacement at the control node exceeds the target displacement of any of the
performance levels, the analysis is terminated, because the performance objective has
been exceeded. Otherwise, the acceptance criteria for the structural elements are checked.
Usually, the control node is typically located at the centre of mass at the roof level of the
building (FEMA 273, 1997). This analysis approach is known as the displacement-based
pushover analysis.

In the so-called spectrum-based approach, the base shear of the building is selected as the
damage target: i.e., the maximum base shear most likely to be experienced by the
building subjected to the design earthquake. In this approach, the incremental lateral
loads are applied in the same fashion as those for the displacement-based approach, and
the base shear of the building is evaluated after each load increment. When the base shear
in the building is equal or greater than the target base shear for any of the specified
performance objectives, the analysis is “paused” and the acceptance criteria of the
structural elements are checked. Thus, it can be determined if the structural elements
satisfy the specified performance levels. Although both methods have pros and cons, the
displacement-based method has been shown to be more convenient for existing
structures, whereas the spectral method has proven to be better for new structures (Gong,
2003).

FEMA 273 (1997) provides limit drift ratios for several types of structures, including
steel moment frames, concrete walls, masonry infill walls, and wood stud walls (see
Section 5.4). The drift ratios are associated with the four performance levels discussed
previously, and are provided as a means to estimate the lateral drift of buildings
depending on the type of structure. It is common to use the drift ratios to compute target
displacements for the displacement-based PBD assessment of buildings (Hasan et al.,
2002; Gong, 2003). However, pushover analysis becomes complex, when the drift ratios
for the structure analyzed are not provided; for example, for CFS shear wall panels.
Although CFS buildings are built in a similar fashion as wood buildings, the drift ratios
for wood-framed buildings cannot be used to carry out the PBD of CFS buildings, since

18
there is no proof of equal behaviour. Therefore, it is necessary and appropriate to
determine the drift ratios for CFS buildings from experimental test results.

2.4. New Generation Performance-Based Design.

The PBD guidelines published in the mid-nineties, such as FEMA 273 (1997) and
SEAOC (1995), are referred to by ATC (2005) as first-generation PBD (PBD-1). In 2001,
FEMA provided funding to ATC to develop second-generation PBD guidelines (PBD-2)
for new and existing buildings, in accordance with the FEMA 445 program plan (ATC,
2005). The Pacific Earthquake Research Center (PEER) is in charge of developing the
theoretical basis for these new guidelines.

One of the objectives of PBD-2 is to provide guidelines for the seismic assessment of
buildings considering their unique design and construction characteristics. Another
objective is to present the guidelines in terms of the risk associated with earthquake-
induced losses (ATC, 2005). Using earthquake-induced losses, building owners have a
better sense of the cost of repairing a building after an earthquake. These risks relate to
casualties, economic losses, and downtime of the building. Casualties refer to the risk of
loss of life and serious injury to occupants as a consequence of earthquake-induced
damage. Economic losses include the cost of repairing or replacing damaged elements or
components after an earthquake. Downtime refers to the time that buildings are closed
due to repairs, inspections, cleaning, and re-occupation preparation. Establishing the
performance objectives in terms of these three types of risks, not only the expected
behaviour of the building is considered, but also the associated losses can be translated
into an economic loss.

According to the PBD-2 philosophy, three types of procedures can be followed to

perform the seismic assessment of buildings. The first is an intensity-based assessment, in
which the risk is estimated for a specific seismic hazard, defined by a response spectrum.
The second is a scenario-based assessment that estimates the risk from a specific
earthquake scenario, as defined by earthquake magnitude and distance from the site. The

19
third is a time-based assessment which measures the risk over a period of time, where all
the earthquakes that may occur in that period and their associated probability are
considered. To date, 25% of the guidelines for the seismic performance assessment of
buildings have been completed; they are expected to be fully completed around 2010
(ATC, 2005).

2.5. Seismic Assessment of Cold Formed Steel Buildings.

Research on PBD has been carried out for the design of new structures, such as structural
steel frames (Gong, 2003; Liew and Chen, 2004; Lee and Foutch, 2002), and concrete
structures (Zou and Chan, 2005). Also, investigations of PBD assessment for lightweight
wood-frame buildings have been conducted (Filiatrault and Folz, 2002), though not as
extensively as for conventional steel or concrete buildings. Conversely, this study reveals
that there is neither a corresponding design standard available nor any reported research
results for the seismic assessment of cold-formed steel buildings. Presently, therefore, a
designer has only one option for the seismic design of a CFS building; that is, carrying
out the seismic design in accordance with the seismic provisions contained in the
applicable seismic codes. However, the designers are unable to assess the seismic
behaviour of CFS buildings because there are no guidelines in the literature for this
purpose.

Since PBD has advantages over seismic design standards and it has been successfully
employed for the seismic assessment of steel and concrete structures, in this study PBD is
used for carrying out the seismic assessment of CFS buildings. The procedure for
conducting the PBD assessment of CFS buildings is adapted from PBD guidelines, such
as FEMA 273 (1997), and FEMA 450 (2003). As described in Section 2.2, the PBD
assessment of a building can be carried out in a number of ways. Thus, an appropriate
analysis method and damage targets should be established in accordance with CFS
buildings structural systems.

20
 Chapter 3
Analytical method for determining the lateral strength of
shear wall panels in cold formed steel buildings

3.1. Introduction
SWP are the primary structural component of CFS buildings, determining the SWP
lateral strength is crucial prior to carrying out PBD assessment of a CFS building. The
reason is that the lateral strength of the SWP is employed to design or review the design
of the SWP. In addition, the lateral strength of SWP is employed in this study for
evaluating the stiffness degradation of SWP, which characterizes their nonlinear
behaviour during the pushover analysis (the stiffness degradation model and pushover
analysis procedure are described in Sections 4.6 and 5.3, respectively).

In recent years, extensive experimental investigations have been carried out to understand
the behaviour, and to determine the lateral strength, of SWP. Serrette et al. (1996, 2002)
have conducted series of tests on SWP in CFS framing with different sheathing materials,
such as Gypsum Wall Board (GWB), Oriented Strand Board (OSB), plywood, and thin
CFS plates. The outcome of the tests has contributed to the development of the Standard
for Cold-Formed Steel Framing-Lateral Design, published by the American Iron and
Steel Institute (AISI, 2004). More recently, Branston et al. (2006) have extensively tested
SWP using American and Canadian wood-based sheathing materials, such as Canadian
Softwood Plywood (CSP), Douglas Fir Plywood (DFP), and OSB. The objective of
Rogers’s testing program is to develop guidelines for the seismic design of CFS SWP for
use with the National Building Code of Canada (Branston et al., 2006). Also, Fulop and
Dubina (2004, 2004b) have conducted series of tests on SWP, to generate a hysteretic
model and a simplified modeling method for the SWP.

21
In current design practice, structural engineers obtain the lateral strength of SWP
primarily from the published values in design standards and guidelines, such as AISI
(2004, 1998). The nominal lateral strengths of SWP, presented in a tabulated form in the
standards, are convenient to use. They are determined on the basis of experimental tests,
which provide degree of confidence to practitioners. However, the application of design
tables is limited by the number of tests carried out on SWP with different material
characteristics and construction details, such as the type and thickness of the sheathing
material, the size and configuration of the stud, and the size and spacing of the fastener.
The values given in design tables for certain SWP cannot be applied or extrapolated to
other SWP with different configurations or construction details. Although research has
been carried out by using numerical simulations to determine the lateral strength of SWP,
the approach is time consuming and may not be suitable for daily engineering practice.
The development of a practical and reliable analytical method for determining the lateral
strength of SWP is of primary importance to practitioners concerned with CFS framing
construction.

Presented in this Chapter is an analytical method for determining the ultimate lateral
strength of SWP, and their lateral displacement. Both, the lateral strength and lateral
displacement, are predicted for isolated SWP. The proposed method takes into account a
broad range of factors that affect the behaviour and strength of SWP, namely: material
property, thickness and geometry of sheathing and studs, spacing of studs, and
construction details such as size and spacing of the sheathing-to-stud fasteners. A hand-
worked example is presented to demonstrate the practical and efficient nature of the
proposed method. To validate the effectiveness of the proposed method, its predictions
of the ultimate lateral strengths of different SWP are compared to test results of recent
experimental investigations (AISI, 2004; Fulop and Dubina, 2004; Branston et al., 2006;
Serrette et al., 1996, 2002). The comparison of predicted versus experimental results is
presented in the last section of this chapter.

22
3.2. The Failure Modes of SWP
The failure of a SWP subjected to in-plane lateral loading at the ultimate strength state,
occurs when the panel has no further strength to resist lateral forces. According to tests
that have been carried out, such as by Gad et al. (1999), Branston et al. (2006), and
Serrette et al. (2002), the predominant failure mode of SWP is associated with sheathing
failure. It is observed that failure is often initiated at the sheathing-to-framing connections
for common sheathing materials such as plywood, OSB, and GWB. The failure of the
sheathing is evident due to rupture of the sheathing-to-framing connections and, as
observed in some tests, the complete separation of the sheathing from the frame.
However, when the thickness of the CFS studs is relatively small (e.g., thickness ≤
0.84mm [33 mils]), the failure of the SWP can be initiated by the local buckling of the
steel studs, even though the studs are braced by the sheathing. The failure of the steel
studs can also occur when the sheathing is installed on both sides of the framework so as
to double the sheathing thickness and thereby enhance the panel lateral strength.
However, the resulting forces in the end studs are amplified, which can result in stud
failure in compression, prior to failure of the sheathing. This mode of failure is explained
in detail in Section 3.4.

Both failure modes discussed in the foregoing are accounted for in the proposed method
for evaluating the lateral strength of SWP. The failure of the sheathing is identified by the
failure associated with the sheathing-to-framing connections; the failure of steel studs is
detected, when the axial force in the end post of the SWP exceeds the compressive
strength of the end steel stud. The proposed method cannot be applied to SWP with light-
gauge steel sheathing, since local buckling of the sheathing is not accounted for in this
study.

3.3. The Lateral Strength of SWP Associated with Sheathing Failure

The lateral strength of a SWP, which is contributed to by the assembly of sheathing and
steel framing studs, can be expressed as,

PR = PS + PF (3. 1)

23
where PS is the lateral strength associated with the sheathing. When the sheathing is
provided on both sides of a shear wall panel, the lateral strength is given by,
2
PS = ∑ PS ,i (3. 2)
i =1

where PS,i (i = 1, 2) are the lateral strengths of the sheathing installed on sides 1 and 2 of
the panel, respectively (see Section 3.3.1). The lateral strength PF associated with the
steel framing studs is determined as,

PF = K F ∆ (3. 3)

where KF is the lateral stiffness associated with the framing studs, and ∆ is the lateral
deflection of the sheathing at impending failure at the ultimate lateral load level.
Compared to the sheathing, the framing studs contribute little to the ultimate lateral
strength of the SWP because their lateral stiffness is insignificant. Experimental results,
such those published by the AISI (2004), provide evidence of the framing contribution to
the SWP lateral strength. The test results show that the lateral strength of the SWP
increase as the steel stud thickness increases (see Table 3.3). Therefore, for simplicity, KF
is taken as the elastic lateral stiffness of the framing studs, i.e.,

3E F I F
KF = ∑
studs h3
(3. 4)

where EF and IF are the Young’s modulus and the moment of inertia of the framing studs,
respectively, and h is the height of the panel.

Considering the compatibility of lateral deformation between sheathing and framing studs
prior to the failure of the panel, the relationship between the sheathing strength and the
lateral deformation of the panel is,

PS
∆= (3. 5)
KS

Substituting Eq. (3.5) into Eq. (3.3) yields,

24
 KF
PF = PS (3. 6)
KS

Then substituting Eq. (3.6) into Eq.(3.1), the lateral strength of the SWP is given by,

⎛ K ⎞
PR = ⎜⎜1 + F ⎟⎟ PS (3. 7)
⎝ KS ⎠

where PS is the lateral strength of sheathing defined in Eq. (3.2), and KS is the effective
sheathing stiffness that can be calculated as,

GS AS 3E I
KS = α V + S3 S α B (3. 8)
1.2h h

where ES and GS are the Young’s and shear modulus of the sheathing, respectively; h is
the height of the SWP; αV and αB are sheathing lateral stiffness reduction coefficients for
shear and bending deformation, respectively; and AS and IS are the cross-section area and
moment of inertia of the sheathing, defined as,

tS l 3
AS = t S l , IS = (3. 9 a,b)
12

in which tS is the sheathing thickness, and l is the length of the SWP.

In addition to the material and cross-section properties of the sheathing, the lateral
strength of the SWP is also highly affected by the characteristics and arrangement of the
sheathing-to-framing connections. At its imminent state of failure, the lateral stiffness of
the SWP is substantially less than its initial elastic stiffness. The degradation of the lateral
stiffness is primarily contributed to by the failure of the sheathing-to-framing
connections, as evidenced by experimental tests (Branston et al., 2006; Serrette et al.,
2002). The sheathing stiffness reduction coefficients αV and αB, are introduced to account
for failure effects of the connections that affect the sheathing lateral stiffness. Both
reduction coefficients are calibrated from the test results reported by Branston et al.
(2006), as described in Appendix B, and are evaluated as follows:
1.8
⎛ Cu ⎞ ⎛ 6in ⎞
αV = ⎜⎜ ⎟⎟ ⋅ ⎜⎜ ⎟⎟ (3. 10)
⎝ 3.3 ⋅ nC ⎠ ⎝ C ⎠s

25
 1.3⋅nC
2
⎛ 6 ⎞ ⎛ 6in ⎞ Cu
α B = ⎜⎜ ⎟⎟ ⋅ ⎜⎜ ⎟⎟ (3. 11)
⎝ Cu ⎠ ⎝ s C ⎠

where sC is the edge screw spacing in inches; nC is the total number of screws used to
fasten the sheathing to the steel framing; Cu is the ultimate strength factor, representing
the number of effective screws of the SWP at its imminent state of failure (determined in
Sections 3.3.1.1 and 3.3.1.2).

3.3.1. Lateral Strength of Sheathing

Consider the analogy between a SWP and an eccentrically loaded bolted steel connection.
For both, the loads are applied eccentrically, and the strength reduction is primarily the
result of the failure of connections initiated at locations which are far from the centre of
rotation. In this study, an inelastic method for evaluating the strength of eccentrically
loaded bolted connections (Brandt, 1982) is employed to evaluate the ultimate lateral
strength of the sheathing. Brandt’s method involves an iterative process to locate the
instantaneous centre of inelastic rotation of a bolt group. As an extension of Brandt’s
method, the ultimate lateral strength of sheathing is evaluated as,

PS ,i = CuVr η ; (i = 1, 2) (3. 12)

where Vr is the strength of a single sheathing-to-framing connection, as determined by the

minimum value of the bearing resistance of the sheathing material, the shear resistance of
the fastener, and the bearing resistance of the steel stud. The parameter Cu is the ultimate
strength reduction factor related to the eccentrically applied load. The determination of
Cu involves locating the so-called instantaneous centre of rotation of the fastener group
(this is accomplished by using the iterative procedure in Section 3.3.1.1, or the simplified
procedure in Section 3.3.1.2). The parameter η is a modification factor that accounts for
the variation of the height-to-length ratio of the SWP, and is given by,

h
η = 8 .0 − − 1.45 ≥ 0 (3. 13)
l

26
 Figure 3.1. Fastener arrangement notation

3.3.1.1. Procedure for Evaluating the Ultimate Strength Reduction Factor

Cu
Brandt’s method of determining the ultimate strength reduction factor for eccentrically
loaded bolted steel connections, involved an iterative process of locating the
instantaneous centre of inelastic rotation of the bolt group. The process consists of
locating the elastic centre of rotation (i.e., the centroid of the SWP), and then the distance
between the elastic centre and the instantaneous centre is computed. By locating the
instantaneous centre, the moments caused by the external loads and internal force in the
screws are computed. Then, Cu is evaluated in addition to the unbalanced forces that
cause the moments. Finally, the process is repeated until the equilibrium condition is
reached to obtain Cu. In the following, Brandt’s method is modified and extended to
determine the ultimate strength reduction factor of sheathing.

For the eccentrically loaded fastener group shown in Figure 3.1, the x, y components of
the distance from fastener i to the elastic centre of the fastener group are,
(0 )
d x i = xC i ; d y i = yC i (3. 14 a,b)

27
where xCi, yCi are the coordinates of the fasteners with respect to the elastic centre. In Eq.
(3.14b) the superscript in parentheses denotes the iteration number. As the applied force
Px is in the x direction, it is noted that the component dxi is invariant throughout the
iterative process.

The polar moment of inertia of the fastener group, with respect to the elastic centre is,

( )
nC
J = ∑ xC2 i + yC2 i (3. 15)
i =1

Let Px be a unit force applied on the location of the actual lateral force, which typically is
on the top of the panel. The moment associated with Px is,

M 0 = Px e (y0) (3. 16)

where ey(0) is the initial eccentricity of Px with respect to the elastic centre, and the
superscript denotes the iteration number.

The iterative process of determining the instantaneous centre of rotation and the ultimate
strength reduction factor Cu is as follows:

Step 1. Set the index of iteration j = 0, the tolerance of convergence ε = 0.01, the initial
values of the unbalanced force Fgx( 0 ) = Px, and the strength reduction factor Cu( 0) = nC;

Compute the polar moment of inertia of the fastener group, J, and the moment, M0, based
on Eqs. (3.15) and (3.16), respectively.

Step 2. Calculate the distance between the instantaneous centre of rotation and the elastic
centre, and update the eccentricity of the applied unit force with respect to the
instantaneous centre of rotation,

δ y( j ) = (Fgx( j ) nC )( J M 0 ) (3. 17)

( j +1) ( j)
ey = ey + δ (yj ) (3. 18)

Evaluate the moment associated with the applied unitary force,

28
 M (pj ) = Px e (y j +1) (3. 19)

Step 3. Update the distance between each fastener and the instantaneous centre of
rotation, and compute the normalized deformation for all fasteners,

( j +1) ( j)
d yi = d y i + δ (yj ) (i = 1, 2, 3,…nC) (3. 20)

d i( j +1) = (d ) + (d ( ) )
xi
2
yi
j 2
(i = 1, 2, 3,…nC) (3. 21)

(
∆(Cij ) = d i( j +1) d max
( j)
∆ C max ) (i = 1, 2, 3,…nC) (3. 22)

where: dmax is the distance between the instantaneous centre of rotation and the
furthest fastener; ∆max is the maximum deformation of the sheathing-to-framing
connection (taken herein as 10 mm (0.39 inch) for wood based sheathing material
in accordance with tests carried out by Okasha (2004)).

Step 4. Calculate the normalized force in all the fasteners and the resulting moment with
respect to the instantaneous centre of rotation,

(R i
( j)
(
Ru ) = 1 − e −10 ∆Ci )
( j ) 0.55
(i = 1, 2, 3,…nC) (3. 23)

( )( ) d (
nC
M ( j ) = ∑ Ri Ru j +1)
j
i (3. 24)
i =1

Step 5. Compute the ultimate strength reduction factor,

( j +1)
M ( j)
Cu = (3. 25)
M (pj )

Cu( j +1) − Cu( j )

If ≤ ε, then the equilibrium condition is satisfied, i.e.,
Cu( j +1)

Cu = Cu( j +1) (3. 26)

and the iterative process is terminated. Otherwise, go to Step 6.

29
Step 6. Compute the normalized ultimate fastener force,

( j) M (p j )
Ru = (3. 27)
M ( j)

and the normalized x components of the fastener forces,

( j)
(
Rx i = − d y i
( j +1)
)
d i( j +1) (Ri Ru ) Ru( j ) (i = 1, 2, 3,…nC)
( j)
(3. 28)

The unbalanced force is then,

nC
( j)
Fgx( j ) = Px + ∑ Rx i (3. 29)
i =1

Return to Step 2.

As stated by Brandt (1982), and confirmed by this study, the iterative procedure for
evaluating the ultimate strength reduction factor Cu converges after only a few iterations.

3.3.1.2. Simplified Evaluation of the Ultimate Strength Reduction Factor

Cu
In design practice, it is desirable to avoid iterative processes even though they are simple
to use. This study reveals that the ultimate strength reduction factor, Cu, converged at the
second iteration for all of the cases presented in Section 3.7. Moreover, it is found that the
difference between the values of Cu, obtained in the first and second iteration is often
quite small. Thus, it is possible to further simplify the foregoing procedure by eliminating
the iterative process without significantly affecting the accuracy of the ultimate strength
reduction factor, Cu.

It is noted that in the iterative procedure, the deformations of the fasteners are linearly
proportional to their distances from the instantaneous centre (Eq. 3.22). Having the
deformations of the fasteners computed, the normalized forces in the fasteners are
obtained from Eq. (3.23). According to Eq. (3.23), fasteners with larger deformation yield
larger normalized forces. Instead of evaluating the normalized force for each fastener, the

30
simplified procedure adopted a normalized force of 0.93 (i.e., R/Ru = 0.93) for all the
fasteners regardless of their distance from the instantaneous center. A value of 0.93 is
calibrated from the results of the iterative procedure; that is, all the SWP presented in
Branston et al. (2006) are analyzed using the iterative procedure and the average
normalized force for all the SWP is 0.93. Since the purpose of iterating is to determine
the forces in the fasteners, adopting a fixed value for the normalized forces eliminates the
iterative procedure. For the iterative procedure, Fgx in Eq. (3.17) is replaced by Px. The
simplified procedure of evaluating the ultimate strength reduction factor Cu is as follows:

Step 1. Compute the polar moment of inertia of the fastener group J and the moment, M0,
based on Eqs. (3.15) and (3.16), respectively.

Step 2. Calculate the distance between the instantaneous centre of rotation and the elastic
centre, and evaluate the eccentricity of the applied unitary force with respect to
the instantaneous centre of rotation,

δ y = (Px nC )( J M o ) (3. 30)

ey = ey + δy
(0)
(3. 31)

Evaluate the moment associated with the applied unit force,

M p = Px e y (3. 32)

Step 3. Compute the distance between each fastener and the instantaneous centre of
rotation,

d yi = yC i + δ y (3. 33)

di = (xC i )2 + (d y i )2 (3. 34)

Calculate the moment associated with the fasteners,

nC
M = 0.93∑ d i (3. 35)
i =1

31
Step 4. Compute the ultimate strength reduction factor,

M
Cu = (3. 36)
Mp

It is clear that the simplified procedure for evaluating the ultimate strength reduction
factor Cu involves significantly less computational effort than that of the iterative
procedure. The accuracy of the simplified procedure is demonstrated in Section 3.6. The
maximum difference between the predicted lateral strengths of a SWP, found by using
the iterative and simplified procedures, is only about 3% for the comparisons in Section
3.7.

3.4. The Lateral Strength of a SWP Associated with Frame Failure

In resisting the applied lateral load, the end framing studs of a SWP undergo either
tension or compression, as they work to prevent over-turning of the panel, as shown in
Figure 3.2. Typically the axial load in the studs are maximum at the end and zero at the
centre of the SWP, thus the middle studs carry less load than the end studs. If a SWP is
also subjected to gravity load, one end stud will experience larger compression force than
the other end stud because the forces caused there by the lateral and gravity loads are
additive, while they are subtractive at the other end. Generally, the failure of the steel
framing for a SWP is associated with the failure of an end stud in compression, as
observed in experimental tests (Roger et al., 2004; Serrette et al., 2002). As such, the
lateral strength Pfc, of a SWP corresponds to the compressive strength of an end stud, and
can be found from the equilibrium of the forces in Figure 3.2 to be,

l
Pfc = Pn (3. 37)
h

where l and h are the length and height of the SWP, respectively, and Pn is the nominal
compressive strength of the end stud which, for this study, is evaluated in accordance
with Chapter D of the North American Specification for the Design of Cold-formed Steel
Structural Members, S136 (2001). As recommended by Telue and Mahendran (2001)
through their experimental investigation, the effective length factors associated with the
end stud are taken as Kex=0.75, and Key=Ket=0.10 for SWP with sheathing installed on

32
its two sides, and Kex=0.75, Key=0.10, and Ket=0.20 for SWP with sheathing installed on
one side. Where Ke = h / KL, where K is the effective length factor (K=1 is assumed for
the studs), and L is the unbraced length of the SWP. In the case where the lateral strength
of a SWP, given by Eq. (3.37), is less than the strength value PR computed by Eq. (3.1),
the lateral strength of the SWP is taken as,

PR = Pfc (3. 38)

Figure 3.2. Panel rotation and force distribution

3.5. Estimation of the Lateral Displacement of a SWP

Having calculated the lateral strength of a SWP, the lateral displacement of the panel at
the ultimate lateral load level can be conveniently evaluated as

PR
∆= (3. 39)
KF + KS

where PR is the lateral strength of the SWP, and KF and KS are the lateral stiffnesses
defined by Eqs. (3.4) and (3.8) for the steel framing and sheathing, respectively.

33
3.6. Illustrative Example
A step-by-step example for evaluating the lateral strength and deformation of the SWP in
Figure 3.3 is presented in the following. For purposes of manual evaluation, the
simplified procedure, developed in Section 3.3.1.2, is employed to demonstrate the
practical and efficient nature of the proposed method. The material properties and
construction details of the SWP are adopted from an experimental study, conducted by
Branston et al. (2006), and the here results are compared with those of the same study.

Figure 3.3. Shear wall panel tested by Branston et al. (2006)

For the SWP in Figure 3.3, the height and length of the panel are h = 2438 mm and l=
1219 mm, respectively, and the sheathing is present on only one side of the frame. The
characteristics and material properties of the steel framing studs are:
Yield and tensile strength: Fy-steel = 230 MPa and Fu-steel = 344 MPa
Young’s modulus: EF = 203000 MPa
End steel stud shape: Two C-shape 362S162-44mils sections connected back-to-back
End steel stud thickness: tF = 1.12 mm
Moment of inertia of the double end-stud: IF = 1.816×105 mm4
Moment of inertia of the intermediate stud: IF = 5.124×104 mm4

34
The characteristics and material properties of the Douglas Fir Plywood sheathing are
CANPLY (2003):
The thickness of the sheathing: tS = 12.5 mm
The bearing strength of the sheathing: Fu-sheathing = 4.5 MPa
Young’s and shear modulus of the sheathing: ES = 10445 MPa and GS = 825 MPa
The characteristics of the screw fasteners are:
No. 8 screw fasteners and screw diameter: dC = 4.064 mm
Screw spacing on the sheathing perimeter: sC = 152mm
Total number of screws on the sheathing: nC = 50
The evaluation of the lateral strength of the SWP, corresponding to failure of the
sheathing, is carried out as follows:
Step 1. Compute the polar moment of inertia of the fastener group, J, and the moment,
M0, using Eqs. (3.15) and (3.16), respectively,

J = ∑ (xC2 i + yC2 i ) = 5.137 × 10 7 mm 2

nC

i =1

M o = Px e (y0 ) = 1 × 1219 = 1219 mm

Step 2. Compute the distance between the instantaneous centre of rotation and the elastic
centre, δy, and the eccentricity of the applied unit force Px. Both shown in Figure
3.1 with respect to the instantaneous centre of rotation, ey, using Eqs. (3.30) and
(3.31), respectively,
Px J 1 × 5.137 × 10 7
δy = = = 842.82 mm
nC M o 50 × 1219

e y = e (y0) + δ y = 1219 + 842.82 = 2061.82 mm

Then compute the moment associated with the applied unit force using Eq. (3.32),
M p = Px e y = 1 × 2061.82 = 2061.82 mm

Step 3. Compute the distance between each fastener and the instantaneous centre of
rotation using Eqs. (3.33) and (3.34). Compute the moment associated with the
fasteners using Eq. (3.35),

35
 nC 50
M = 0.93∑ d i = 0.93∑ d i = 53992 mm
i =1 i =1

Step 4. Compute the strength reduction factor using Eq. (3.36),

M 53992
Cu = = = 26.186
M p 2061.82

Step 5. Compute the lateral strength of a single sheathing-to-framing connection Vr to

determine the lateral strength of the sheathing, as expressed in Eq. (3.12). The
parameter Vr is the minimum strength of a sheathing-to-framing connection,
accounting for: the bearing strength of the sheathing, bearing strength of the
framing, shear strength of the screws, and screw pullout.
Bearing strength of the sheathing and framing (S136, 2001):
Br −sheathing = 3.0 ⋅ t S ⋅ d C ⋅ Fu −sheathing = 3.0 × 12.5 × 4.064 × 4.5 = 685.8 N

Br −steel = 3.0 ⋅ t F ⋅ d C ⋅ Fu = 3 × 1.12 × 4.064 × 344 = 4687 N

Shear and pullout strength of the screws, based on Steel Stud Manufacturers
Association (SSMA, 2001), Vr − Sscrew = 3256 N

Vr − Pscrew = 1255 N

Vr = min{Br − sheathing , Br − steel ,Vr − Sscrew ,Vr − Pcrew } = 685.8 N

Step 6. Compute the lateral strength of the sheathing using Eqs. (3.13) and (3.12),

h 2438
η = 8− − 1.45 = 8 − − 1.45 = 1.0
l 1219
PS = CuVrη = 26.186 × 685.8 × 1.0 = 17958 N
Step 7. Compute the sheathing and framing lateral stiffness:
Sheathing stiffness reduction factors using Eqs. (3.10) and (3.11),
1.8
⎛ Cu ⎞ ⎛ 6in ⎞ ⎛ 26.186 ⎞1.8 ⎛ 6in ⎞
αV = ⎜⎜ ⎟ ⋅⎜ ⎟= ⋅
⎟ ⎜ s ⎟ ⎜⎝ 3.3 ⋅ 50 ⎟⎠ ⎜⎝ 6in ⎟⎠
= 0.036
⎝ 3.3 ⋅ n f ⎠ ⎝ C ⎠
1.3⋅n f 1.3⋅50
2
⎛ 6 ⎞ ⎛ 6in ⎞
2
Cu
⎛ 6 ⎞ ⎛ 6in ⎞ 26.186
α B = ⎜⎜ ⎟⎟ ⋅ ⎜⎜ ⎟⎟ =⎜ ⎟ ⋅⎜ ⎟ = 0.053
⎝ Cu ⎠ ⎝ C ⎠
s ⎝ 26.186 ⎠ ⎝ 6in ⎠
then, compute the cross section area, the moment of inertia, and the lateral
stiffness of the sheathing based on Eqs. (3.9 a,b) and (3.8), respectively.

36
 AS = t S l = 12.5 ×1219 = 15238 mm 2

tS l 3 12.5 × 12193
IS = = = 1.887 × 109 mm 4
12 12
GS AS 3E I 825 × 15238 3 × 10445 × 1.887 × 10 9
KS = αV + S3 S α B = 0.036 + 0.053
1 .2 h h 1.2 × 2438 24383
= 370.95 N mm
Upon noting from Figure 3.3 that there are two end studs with double C- sections
connected back-to-back and one intermediate stud with single C-shape section,
compute the frame stiffness using Eq. (3.4),

3E F I F ⎛ 3 × 203000 × 1.816 × 105 ⎞ ⎛ 3 × 203000 × 5.124 × 10 4 ⎞

KF = ∑ h3 = 2⎜
⎜ 2438 3
⎟+⎜
⎟ ⎜ 2438 3
⎟
⎟
studs ⎝ ⎠ ⎝ ⎠
= 17.42 N mm

Step 8. Finally, compute the lateral strength of the SWP corresponding to sheathing
failure using Eq. (3.7),
⎛ Kf ⎞ ⎛ 17.42 ⎞
PR = ⎜⎜1 + ⎟⎟ PS = ⎜1 + ⎟17958 = 18801 N
⎝ KS ⎠ ⎝ 370.95 ⎠

To further evaluate the lateral strength of the SWP corresponding to frame failure, the
compressive strength of the steel end studs must be determined in advance. The end studs
are formed with two C-shape 362S162-44 steel studs connected back-to-back. Based on
the recommendation of Telue and Mahendran (2001), the effective length factors of the
studs are taken as Kx = 0.75, Ly = 0.10, and Lt = 0.20. The unrestrained length of the studs
for the x and y directions and torsion are Lx = Ly = Lt = 2438mm, respectively. Based on
CSA S136-01 (2001), the nominal compressive strength of the end studs is: Pn = 71.166
kN. Using Eq. (3.37), the lateral strength of the SWP corresponding to frame failure is
then computed as,
l 1219
Pfc = Pn = 71166 = 35583 N
h 2438
As Pfc = 35583 > 18801 N (see Step 8), sheathing failure defines the lateral strength of
the SWP; i.e., PR = 18801 N.

37
The lateral displacement associated with sheathing failure computed using Eq. (3.39) as,
PR 18801
∆= = = 48.4mm
K F + K S 17.42 + 375.90
Compared to the computed results found in the foregoing for the SWP, the experimental
results reported by Branston et al. (2006) for the lateral strength by unit of length and
displacement of the same panel are 16000 N/m and 54.8 mm, respectively; therefore, the
calculated-to- test ratios for the lateral strength and displacement of the SWP are:
18801N 48.4
Strength ratio = = 0.96 , Displacement ratio = = 0.88
16000 N / m × 1.219m 54.8

3.7. Comparison between Analytical and Experimental Results

Experimental results reported from AISI (2004), Fulop and Dubina (2004), Branston et
al. (2006), and Serrette et al. (1996, 2002) are used to validate the accuracy of the
proposed analytical method of evaluating the ultimate lateral strength of a SWP. The
experimental results for strength and displacement, obtained by all the researches
referenced in this section, are at the ultimate limit state. Since not all properties are
reported in the foregoing literature, the material properties this study adopted in the
evaluation may not be identical to those of the tested materials. The geometric gross
properties of the steel studs are computed based on the cross-section dimensions reported
in the literature. Unless otherwise specified in individual cases, the yield strength and
Young’s modulus values used in the calculations for steel are 230 MPa and 203000 MPa,
respectively. Unless otherwise specified, the following material properties for sheathing
material are employed to evaluate the ultimate lateral strength of the SWP: the shear
modulus of elasticity for OSB, DFP and CSP = 925MPa, 825MPa, and 497MPa (Okasha,
2004), respectively; the modulus of elasticity for OSB = 9917MPa (OSB, 1995), while
that for DFP and CSP = 10445MPa and 7376MPa (CANPLY, 2003), respectively; the
elastic and shear moduli for GWB = 1290MPa (GA-235, 2001) and 561MPa,
respectively. The connection strength corresponding to the bearing strength of GWB is
Br=228 N (51.25 lb), which is the lowest experimental value tabulated by Lee (1999).

38
For all of the comparisons made in the following, the SWP lateral strengths are predicted
using both the Iterative Procedure (IP) and the Simplified Procedure (SP) so that the
differences between their predictions can be identified. However, the strengths computed
using the simplified procedure are alone used to calculate the ratios of the predicted-to-
test values since it is the recommended procedure. Furthermore, the lateral displacements
from the experimental tests are compared to the lateral displacements predicted by the SP
as illustrated in the example in Section 3.6.

Shown in Table 3.1 and Table 3.2 are the comparisons between predicted and test results
for lateral strengths and lateral displacements. The experimental results listed on Tables
3.1 and 3.2 were obtained from monotonic testing. Experimental results for cyclic testing
are not listed in these tables because they were not available at the time of preparation of
this chapter. Prof. Rogers provided the author experimental results derived from the first
part of his testing schedule, which was for monotonic testing. The three different
sheathing materials investigated were OSB, DFP and CSP with thicknesses of 11 mm,
12.5 mm, and 12.5 mm, respectively. The C-shape cold-formed steel studs were 92S41-
1.12 mm (362S162-44mils), spaced 610 mm (24 in) on centre. Double C-shape back-to-
back studs were placed at the ends of the panel and the end studs were attached with
S/HD10 hold-downs to the floor. The sheathing was attached on one side of the panel
using No. 8 (diameter = 4.06 mm) screws at every 305 mm (12 in) in the field of the
panel (i.e., location other than the edge). Three edge screw spacings of 76 mm (3 in), 102
mm (4 in) and 152 mm (6 in) were investigated for SWP lengths of 1219 mm (4 ft) and
2438 mm (8 ft), as tabulated in Tables 3.1 and 3.2, respectively. Two edge screw
spacings of 102 mm (4 in) and 152 mm (6 in) were investigated for SWP length of 609
mm (2 ft), as listed in Table 3.2.

From Table 3.1 and Table 3.2, it can be seen that the maximum difference between the
test results and predicted results for the lateral strength of the SWP is 10%, whereas the
maximum difference between the test and predicted results for the SWP lateral
displacements is 21%. It is noteworthy that the predicted SWP lateral strengths are
governed by sheathing failure, except for the panels with DFP sheathing and screw

39
spacing of 76 mm listed in Table 3.1. The SWP lateral strengths predicted by both the
iterative and simplified procedure indicate that stud failure governs, which is confirmed
by the test results.

Table 3.1 Comparison between predicted and test (Branston et al., 2006) results for lateral strengths
and displacements

Steel stud 92S41-1.12mm. Screw size: No. 8. Field screw spacing: 305mm
Wall length and height: 1219mm, 2438mm. Sheathing length and height: 1219mm, 2438mm
Lateral
Lateral strength, kN/m
Sheathing Edge displacement, mm
Test ID 2
material screw Predicted
(Branston et Ratio
and spacing Test Sheathing Ratio
al., 2006) Stud Test Test SP
thickness mm Branston failure Test
failure Pred3 Pred3
IP1 SP1
OSB 21-A,B,C 152 13.20 12.26 12.24 1.08 41.1 42.1 0.98
11 mm 23-A,B,C 102 19.30 18.06 18.01 1.07 39.5 46.2 0.85
One side 25-A,B,C 76 23.50 23.83 23.76 0.99 40.7 49.3 0.83
DFP 11-A,B,C 152 16.00 15.45 15.42 1.04 54.8 48.4 1.13
12.5mm 5-A,B,C,D 101 23.80 22.78 22.72 29.19 1.05 60.6 53.6 1.13
One side 13-A,B,C 76 29.70 30.10 30.00 1.02 58.2 57.5 1.01
CSP 7-A,B,C 152 12.70 12.49 12.46 1.02 50.7 57.8 0.88
12.5mm 1-A,B,C 102 16.60 18.32 18.27 0.91 60.6 64.5 0.94
One side 9-A,B,C 76 25.10 24.13 24.05 1.04 61.0 69.7 0.88
Average 1.02 0.96
Standard deviation 0.05 0.11
Coefficient of variation 0.05 0.11
1
IP = Iterative procedure, SP = Simplified Procedure
2
Predicted SWP strengths are evaluated based on sheathing and stud failures
3
The “Pred” is the smaller one of the predicted strengths, based on the sheathing failure, using SP,
and stud failure

40
Table 3.2. Comparison between predicted and test (Branston et al., 2006) results for lateral strengths
and displacements

Steel stud 92S41-1.12mm. Screw size: No. 8, field screw spacing: 305mm. Sheathing on one side
Lateral
Lateral strength, kN/m
Sheathing Edge displacement, mm
Test ID
material screw Predicted
(Branston et Ratio
and spacing Test Sheathing Ratio
al., 2006) Stud Test SP Test
thickness mm failure Test
failure Pred
IP SP Pred

Wall length and height: 609mm, 2438mm. Sheathing length and height: 609mm, 2438mm
OSB 19-A,B,C 152 12.50 10.95 11.09 1.13 78.4 56.0 1.40
11 mm 27-A,B,C 102 18.40 15.48 15.68 1.17 78.0 56.3 1.39
29.19
CSP 15-A,B,C 152 12.20 11.63 11.50 1.06 103.3 80.3 1.29
12.5mm 17-A,B,C 102 18.00 16.17 16.10 1.12 107.0 81.2 1.32
Wall length and height: 2438mm, 2438mm. Sheathing length and height: (2) 1219mm, 2438mm
29-A,B,C 152 13.60 13.35 13.22 1.03 50.5 50.9 0.99
CSP
31ABCDEF 102 20.50 18.96 18.71 29.19 1.10 55.6 55.6 1.00
12.5mm
33-A,B,C 76 26.30 24.55 24.19 1.09 64.1 60.0 1.07
Average 1.10 1.21
Standard deviation 0.05 0.18
Coefficient of variation 0.05 0.15

Table 3.3 presents a comparison between the predicted lateral strength of SWP with OSB
sheathing to that published in the Standard for Cold-Formed Steel Framing-Lateral
Design, (AISI, 2004). Table 3.3 lists the lateral strengths provided by the AISI (2004) for
seismic design. The length and height of the SWP are 1219 mm (4 ft) and 2438 mm (8
ft), respectively. The steel stud size is 89S41 (350S162), and the four different stud
thicknesses that are described in the AISI (2004) standard are 0.838 mm (33 mils), 1.092
mm (43 mils), 1.372 mm (54 mils), and 1.727 mm (68 mils). Double studs are used for
the end studs, and information of the hold-downs is not provided. OSB sheathing was
attached on one side of the panel using No. 8 or No. 10 screws at every 305 mm (12 in) in
the field. The four edge screw spacings used are 51 mm (2 in), 76 mm (3 in), 102 mm (4
in), and 152 mm (6 in). Table 3.3 indicates that the lateral strengths of all SWP are

41
governed by sheathing failure, and that there is a good correlation between predicted and
test results for the SWP lateral strength. However, Table 3.3 shows a larger standard
deviation than Tables 3.1 and 3.2, which may result from the difference in the material
properties, assumed for the OSB in the predictions and the properties used in the testing.

Table 3.3. Comparison between predicted and test (AISI, 2004) results for lateral strengths

OSB sheathing: 11.1mm on one side. Stud designation: 89S41mm

(350S162). Wall length and height: 1219mm, 2438mm. Sheathing
length and height: 1219mm, 2438mm
Stud Edge Lateral strength, kN/m
Screw thick. screw Predicted
AISI Ratio
size mm spacing Sheathing failure Stud
(2004) AISI
(mils) mm IP SP failure Pred
0.838 152.4 10.22 12.15 12.12 0.84
21.17
(33) 101.6 13.35 17.93 17.88 0.75
152.4 12.04 12.28 12.25 0.98
1.092 101.6 18.02 18.07 18.02 1.00
No. 8 29.88
(43) 76.2 22.55 23.85 23.77 0.95
(4.06
50.8 30.06 35.37 35.26 0.85
mm)
152.4 13.72 12.41 12.38 1.11
1.372 101.6 20.58 18.22 18.17 1.13
39.59
(54) 76.2 25.69 24.00 23.93 1.07
50.8 34.30 35.54 35.43 0.97
No.1 152.4 17.98 14.91 14.88 1.21
0 1.727 101.6 26.97 21.82 21.76 1.24
52.30
(4.83 (68) 76.2 33.71 28.70 28.61 1.18
mm) 50.8 44.95 42.41 42.27 1.06
Average 1.02
Standard deviation 0.15
Coefficient of variation 0.14

42
Shown in Table 3.4 are the comparisons between predicted and test results for lateral
strengths and displacements, conducted by Serrette et al. (2002). The SWP dimensions
were 2438 mm (8 ft) by 2438 mm (8 ft). OSB sheathing was fastened on one side or both
sides of the panel using No. 8 or No. 10 screws. The screw spacing was 51 mm (2 in) on
the edge and 305 mm (12 in) in the field of the sheathing. The framing steel studs
investigated in the two tests were 350S162 (89S41) with thicknesses of 1.37mm (54 mils)
and 1.73mm (68 mils), and yield strength of 407MPa (59 ksi) and 386MPa (56 ksi),
respectively. The studs were spaced at 610 mm (24 in) on centre, and double studs were
placed at the ends of the SWP; information of the hold-downs is not provided. The
ultimate test lateral strengths shown in Table 3.4 are the average values from two
specimens, tested under reversed cyclic loading protocol (SEAOC, 1997). For the two
cases, where OSB sheathing is applied on only one side of the panel, the SWP lateral
strength is governed by sheathing failure; when the OSB sheathing is installed on both
sides of the panel, the strength is governed by stud failure.

Table 3.4. Comparison between predicted and test (Serrette et al., 2002) results for lateral strengths

SWP description Lateral strength, kN/m

Walls length and height: Screw Stud Predicted
2438mm, 2438mm. Test ID Size thick. Sheathing Ratio
Sheathing length and No. mm Test failure Stud
Test
height: (2) 1219mm, (mils) failure Pred
IP SP
2438mm
OSB sheathing on one side: 1.37
1-2 8 34.38 35.68 34.98 57.83 1.02
11mm. Steel stud: 89S41 (54)
Screw spacing (mm) Edge: 1.73
3-4 10 44.96 42.44 41.61 74.21 0.93
51; Field: 305 (68)
OSB sheathing on both 6-7 1.37
8 60.96 70.91 69.51 60.49 0.99
sides: 11mm. Steel stud: (54)
89S41. Screw spacing (mm) 8-9 1.73
10 76.53 84.24 82.58 77.39 1.08
Edge: 51; Field: 305 (68)
Average 1.00
Standard deviation 0.06
Coefficient of variation 0.06

43
Shown in Table 3.5 are the comparisons between predicted and test results for lateral
strength and displacement, conducted by Fulop and Dubina (2004). In this case OSB
sheathing was presented on one side of the panel, and the steel studs were spaced 610 mm
(24 in) on centre. Information regarding the type of hold-downs employed to attach the
SWP to the floor is not provided. Differing from the foregoing experimental

investigations, the dimensions of the SWP panel were 3600 mm (≈12 ft) in length and

2440 mm (8 ft) in height. The lateral strength listed in Table 3.5 for the test was obtained
from only one specimen.

Table 3.5. Comparison between predicted and test (Fulop and Dubina, 2004) results for lateral
strengths and displacements

Lateral
SWP Lateral strength, kN/m
displacement, mm
description
Screw Test Predicted
OSB sheathing: 10mm Dia. Fulop Sheathing Ratio Ratio
Steel stud: 150S41-1.5mm mm and failure Stud Test Test SP
Test
Screw spacing (mm) Dubina failure Pred Pred
IP SP
Edge: 102; Field: 305 (2004)
Wall length and height:
3600mm, 2440mm.
Sheathing length and 4.2 21.88 20.16 19.53 47.66 1.12 42.85 44.16 0.97
height: (3) 1200mm,
2440mm

Shown in Table 3.6 are the comparisons between predicted and test results for lateral
strengths and displacements, conducted by Serrette et al. (1996). The SWP length and
height were 2438 mm (8 ft) and 2438 mm (8 ft), respectively. The GWB sheathing was
fastened on both sides of the panel with No. 6 screws. The screw spacing was 152 mm (6
in) on the edge and 305 mm (12 in) in the field of the panel. The steel studs were 150S32
(600S125) with thickness of 0.88 mm (33 mils), and information of the hold-downs
employed is not provided. The studs were spaced 610 mm (24 in) on centre, and the

44
sheathing was attached to the studs using No. 6 screws. The lateral strength shown in
Table 3.6 for the test was the average value for four specimens.

Table 3.6. Comparison between predicted and test (Serrette et al., 1996) results for lateral strengths
and displacements

Lateral
SWP Lateral strength, kN/m
displacement, mm
description
Screw Predicted
Ratio1 Ratio
GWB sheathing: 12.7mm No. Sheathing
Test Stud Test Test SP Test
Steel stud: 150S32-0.88mm failure
failure Pred Pred
Screw spacing (mm) IP SP
Edge: 152; Field: 305
Wall length and height:
2438mm, 2438mm.
6 10.93 11.15 10.97 23.46 1.00 40.00 37.1 0.93
Sheathing length and
height: (2) 1219mm,
2438mm
1
The “Pred” is the smaller one of the predicted strengths based on sheathing (SP) and stud failures

Recently, CFS steel laminates have been employed as sheathing for SWP. Although CFS
laminates are thin plates that are prone to local buckling, they can be used efficiently
when bonded to thicker plates of other materials, such as GWB. For example, there is a
marketplace double layer board called Sureboard (2006), which is composed of a 0.69
mm (27 mils) thick CFS steel sheathing bonded to a 12.7 mm (0.5 in) thick GWB.

Table 3.7 presents a comparison of experimentally determined SWP lateral strengths of

Sureboard, obtained by the ICC Evaluation Service (ER-5762, 2003), with corresponding
strengths predicted by this study. The comparisons are presented for steel studs of three
different thicknesses, i.e., 33 mils, 43 mils, and 54 mils. In all comparisons the sheathing
is attached to the framing with No. 6 screws with different spacing at the edge of the
panel (i.e., 152 mm, 102 mm, 76 mm, and 51 mm). Information regarding the type of
hold-downs employed to attach the SWP to the floor is not provided. Unlike all of the
previous comparisons where failure of a sheathing-to-framing connection was governed
by the bearing capacity of the sheathing material, Sureboard failure is governed by screw

45
pullout. This was determined comparing the strengths of a sheathing-to-framing
connection for failure of the sheathing, stud, and screw as described in the example of
Section 3.6.
A considerable discrepancy can be observed in Table 3.7 among of several of the
predictions made by this study and those reported, by ER-5762 (2003). A factor affecting
the precision of the prediction calculations is the value adopted for the strength of the
sheathing-to-framing connection, Vr. The same value was used for all predictions, even
though studs with three different thicknesses are used. The strength of the sheathing-to-
framing connections is governed by the thickness of the sheathing steel, which is smaller
than the thickness of the studs. Even so, the thickness of the studs can affect the strength
of the connections. Therefore, accurate information, regarding the strength of sheathing-
to-framing connections for Sureboard, is needed to yield accurate estimations of the
lateral strength of SWP with Sureboard.

Table 3.7. Comparison between predicted and test (ER-5762, 2003) results for lateral strengths and
displacements

SWP Lateral
Lateral strength, kN/m
description Edge displacement, mm
screw Predicted
Sureboard Series 200, Ratio Ratio
spacing Sheathing
Steel thick. 0.69 mm Test Stud Test Test SP Test
mm failure
GWB 12.7 mm (1/2 in) failure Pred Pred
IP SP
152 15.83 14.56 14.53 1.09 13.97 16.97 0.82
Stud thick. 0.84 mm 102 22.55 21.62 21.56 1.05 17.78 17.53 1.01
36.44
(33mils) @ 610 mm oc. 76 25.25 28.68 28.59 0.88 17.78 17.72 1.00
51 27.95 42.79 42.65 0.77 17.78 18.16 0.98
152 20.50 14.62 14.58 1.41 20.83 17.02 1.22
Stud thick. 1.09 mm 102 28.09 21.68 21.62 1.30 24.64 17.53 1.41
55.70
(43mils) @ 610 mm oc. 76 31.30 28.74 28.65 1.09 24.64 17.78 1.39
51 34.44 42.86 42.72 0.81 21.08 18.16 1.16
Stud thick. 1.37 mm 76 42.25 28.8 28.76 1.47 25.65 17.78 1.44
73.31
(43mils) @ 610 mm oc. 51 50.49 42.92 42.78 1.18 31.5 18.16 1.73
Average
1.10 1.22

46
 Standard deviation
0.24 0.28
Coefficient of variation
0.22 0.23

47
 Chapter 4
Simplified Finite Element Analysis (SFEA) for CFS
Buildings

4.1. Introduction
Typically, CFS buildings are built with structural SWP, load-bearing and non-load
bearing wall panels, floor panels, and roof panels SWP are built with CFS studs and
tracks, covered with structural sheathing attached with screws on one or two sides, as
shown in Figure 3.3. Wall panels are built similarly, but without structural sheathing. The
floor panels can be built in three ways: flat slabs, flat slabs supported by CFS joists, or
panels supported by CFS joists. The slabs can be constructed of concrete, and the panels
can be constructed of CFS, structural steel or wood. Roof panels are built in the same
fashion as the floor panels, though steel and wood are more commonly used than
concrete.

In engineering practice, typically the analysis of CFS buildings is conducted using

Conventional Finite Element Analysis (CFEA). In this approach, the shear wall, wall,
floor and roof panels are modeled using frame elements for the studs, tracks, and joists,
and shell elements for the sheathing and slabs. Shown in Figure 4.1, is a typical model of
a three-storey CFS building. Although CFEA is effective for analyzing CFS buildings,
generating and analyzing the model is time consuming due to the large number of
elements. Reviewing the results from CFEA is also time consuming, because all the studs
in the building need to be reviewed in order to identify the studs with larger axial forces.

Presented in this Chapter is the proposed methodology to carry out a SFEA of CFS
buildings. The objective of the development of the methodology is to provide a simple
analysis procedure having acceptable accuracy for the PBD assessment of CFS buildings.

48
Specifically, the SFEA method is devised to reduce the number of elements in the model,
thus the model can be generated and analyzed in less time. Also, reviewing the results
becomes simpler, because the output can be set up to print out the lateral force in each
SWP along with the forces in the most critical studs. Although the advantages may not be
significant for a single linear analysis, for the pushover analysis of CFS buildings the
advantages become significant, because usually the analysis must be carried out many
times, depending on the number of applied loading increments.

Basically, the SFEA of CFS buildings consists of two aspects: 1) transforming the SWP,
wall panels, floor panels, and roof panels into flat shell elements with equivalent material
and geometric properties; 2) modeling each panel with an equivalent (e.g., sixteen-node)
shell element in the finite element analysis. Shown in Figure 4.1 is the model of the
building that is analyzed using CFEA, in which the SWP are modeled with shell and
frame elements. Also depicted in Figure 4.1 are the equivalent shell elements, represented
by the large dark rectangles that are used for the SFEA model. Thus, the relative
difference in size of the elements that would be used for CFEA and SFEA for this
building is evident.

Figure 4.1 Modeling of a CFS building using finite element analysis

49
In Section 4.2 the method for transforming the panels into shell elements of equivalent
material and geometric properties is described. A nonlinear finite element formulation is
presented in Section 4.3. After the SFEA of a CFS building is conducted, the internal
forces in the studs are estimated using the procedure in Section 4.4. Next, nine SWP are
analyzed by SFEA and CFEA, and the results are compared and discussed in Section 4.5.
To account for the SWP material nonlinearities in the pushover analysis, a stiffness
degradation model is developed in Section 4.6. Finally, in Section 4.7 several SWP are
analyzed by SFEA, and the results are compared to those from experimental testing.

4.2. Modeling of SWP with Equivalent Shell Elements

Panels are generally complex to analyze because they are geometric orthotropic elements
having different cross section geometry in their orthogonal directions. To simplify the
analysis, however, this study considers panels as natural orthotropic elements having the
same geometry but different material properties in their orthogonal directions. Typically,
common finite element formulations allow natural orthotropic elements but do not accept
geometric orthotropic elements. Figure 4.2 shows the cross section of a panel and its
equivalent shell cross section, where: sF is the spacing between studs, sC is the distance
between screws on the edge of the panel, and c1 and c2 are the distances from the centre
of the studs to the mid-plane of the sheathing on side 1 and 2, respectively. The
parameters EF, AF, and IF denote the modulus of elasticity, cross sectional area, and
moment of inertia about the strong bending axis of the studs, respectively, while l is the
length of the panel and equivalent shell element, and tS1 and tS2 are thicknesses of the
sheathing on side 1 and side 2, respectively.

Troitsky (1976) developed equations to determine the equivalent rigidity of orthotropic

ribbed plates in bending. In this study the Troitsky equations are extended to account for
the axial rigidity of the panel in the direction along the longitudinal axis of the studs (i.e.,
y direction). The equivalent thickness and modulus of elasticity in the y direction of a
shell element are determined by equating and solving the axial and flexural rigidities of a
panel and an orthotropic element. Afterwards, the equivalent modulus of elasticity in the

50
x direction is determined by equating and solving the bending rigidity of a panel and a
orthotropic shell .

Figure 4.2. a) Panel cross-section, and b) equivalent shell

The general equations for the axial and flexural rigidities of an orthotropic plate or shell
are given by,
Ei ⋅ Ai Ei ⋅ I i
T= , D= (4. 1 a,b)
1 − ν xν y 1 − ν xν y

where T, and D, are the axial and flexural rigidities of an orthotropic flat element; Ai, Ii,
and νi are the cross sectional area, moment of inertia of the orthotropic element, and
Poisson ratio, respectively. The index i denotes the direction (e.g., x or y) in which the
properties are measured. Substituting the properties of a panel into Eqs. (4.1 a,b) and
dividing through by the length l, the axial and flexural rigidities per unit of length are
expressed as follows.

(q ) E y(q1 ) ⋅ t S(q1) E y(q2) ⋅ t S(q2) EF AF nF

Ty = + + (4. 2)
1 − ν x1ν y1 1 − ν x 2ν y 2 l

E y( q1 ) ⎛ t S31 ⎞ E y( q2) ⎛ t S3 2 ⎞ E I n
Dy( q ) = ⎜ + t S 1 ⋅ c12 ⎟ + ⎜ + t S 2 ⋅ c22 ⎟ + F F F (4. 3)
1 − ν x1ν y1 ⎝ 12 ⎠ 1 − ν x1ν y1 ⎝ 12 ⎠ l

E y( q1 ) =λ ( q )ESy1 E y( q2) =λ ( q )ESy 2 (4. 4 a,b)

51
where: ESy is the sheathing modulus of elasticity in the y direction; ts is the thickness; λ is
the stiffness degradation coefficient, computed from Eq. (4.52); subscripts 1 and 2
represent the side of the panel in which the properties are measured; νx and νy are the
sheathing Poisson ratios in the x and y directions; nF is the number of studs in the panel;
and q is a load increment index associated with the nonlinear analysis.

Substituting the properties of the equivalent shell into Eqs. (4.1 a,b), and dividing through
by l, the axial and flexural rigidities per unit length are expressed as follows.
(q ) (q ) (q ) 3 (q )
(q ) E yeq ⋅ teq (q ) E yeq ⋅ teq
Tyeq = =
12(1 −ν xeqν yeq )
, Dyeq (4. 5 a,b)
1 − ν xeqν yeq

where: Eyeq and teq are the modulus of elasticity and thickness of the equivalent shell,
respectively; νxeq and νyeq are the equivalent shell Poisson ratios in the x and y directions,
respectively.

By equating Ty(q ) and Tyeq

(q )
, the following equation is obtained,
(q ) (q )
E yeq ⋅ t yeq E (q ) ⋅ t E (q ) ⋅ t E An
= y1 S 1 + y 2 S 2 + F F F (4. 6)
1 −ν xeqν yeq 1 −ν x1ν y1 1 −ν x 2ν y 2 l

By equating Dy(q ) and Dyeq

(q )
, the following equation is obtained,
(q ) 3 (q )
E yeq ⋅ teq E y( q1 ) ⎛ t S31 ⎞ E y( q2) ⎛ t S3 2 ⎞ E I n
= ⎜ + t S 1 ⋅ c12 ⎟ + ⎜ + t S 2 ⋅ c22 ⎟ + F F F
12(1 −ν xeqν yeq ) 1 −ν x1ν y1 ⎝ 12
(4. 7)
⎠ 1 −ν x1ν y1 ⎝ 12 ⎠ l

By solving Eqs. (4.6) and (4.7), the equivalent modulus of elasticity in the y direction and
thickness of the equivalent shell element, at load increment q, are expressed as

(q )
E yeq =
(α ( ) + α ) (1 −ν ν )
1
q
2
2
xeq yeq

l ⋅ (1 −ν x 2ν y 2 )(1 −ν ν )[(α ( ) + α )(α ( ) + α ( ) + α ( ) )]

q q q q 1/ 2 (4. 8)
x1 y1 1 2 3 4 5

teq( q ) =
[(α ( ) + α )(α ( ) + α ( ) + α ( ) )]
1
q
2 3
q
4
q
5
q 1/ 2

(α ( ) + α ) q
(4. 9)
1 2

where

52
 α1( q ) = EF ⋅ AF ⋅ nF + l ⋅ [q E y1 ⋅ tS 1 (1 − ν x 2ν y 2 )+ q E y 2 ⋅ tS 2 (1 − ν x1ν y1 )]

α 2 = − EF ⋅ AF ⋅ nF ⋅ [ν x 2ν y 2 (1 − ν x1ν y1 ) + ν x1ν y1 ]

α 3( q ) = 12 ⋅ l⋅q E y1 ⋅ tS 1 ⋅ c12 (1 − ν x 2ν y 2 ) − 12 ⋅ EF ⋅ I F ⋅ nF (ν x 2ν y 2 + ν x1ν y )

α 4( q ) = l ⋅q E y 2 ⋅ tS 2 ⋅ (tS 2 ⋅ (1 − ν x1ν y ) + 12 ⋅ c12 ) − l⋅q E y1 ⋅ tS 1 ⋅ν x 2ν y 2

3 3

α 5( q ) = l ⋅q E y1 ⋅ tS 1 + 12 ⋅ EF ⋅ I F ⋅ nF (1 + ν x 2ν y 2ν x1ν y ) − 12 ⋅ l⋅q E y 2 ⋅ tS 2 ⋅ c22 ⋅ν x1ν y1

3

It is noted that all the right terms in Eq. (4.9) are contained in Eq. (4.8), thus the terms in
Eq. (4.8) are replaced by teq( q ) , such that,

(q )
=
(α ( ) + α )(1 − ν
q
ν yeq )
teq ⋅ l ⋅ (1 − ν x 2ν y 2 )(1 − ν x1ν y1 )
1 2 xeq
E yeq (q ) (4. 10)

The flexural rigidity of the panel in the x direction at load increment q is given by,
Ex(1q ) ⎛ tS31 ⎞ Ex( q2 ) ⎛ tS3 2 ⎞
Dx( q ) = ⎜⎜ + tS 1 ⋅ c12 ⎟⎟ + ⎜⎜ + tS 2 ⋅ c22 ⎟⎟ (4. 11)
1 −ν x1ν y1 ⎝ 12 ⎠ 1 −ν x 2ν y 2 ⎝ 12 ⎠
E x(1q ) = λ( q ) ESx( q1) E x( q2 ) = λ( q ) ESx( q2) (4. 12 a,b)
,
where ESx is the sheathing Young’s modulus in the x direction. The flexural rigidity of the
equivalent shell in the x direction is given by,
(q ) 3 (q )
(q ) Exeq ⋅ teq
=
12(1 − ν xeqν yeq )
Dxeq (4. 13)

By equating and solving Dx(q ) and Dxeq

(q )
, the modulus of elasticity of the equivalent shell

element in the x direction is given by,

⎛ Ex(1q ) ⎛ t S31
⎜ 2⎞ E2( q ) ⎛ t S3 2 ⎞ ⎞ 12(1 −ν xν y )
(q )
E xeq =⎜ ⎜ + t S 1 ⋅ c1 ⎟ + ⎜ + t S 2 ⋅ c22 ⎟ ⎟⎟ ⋅ (q ) (4. 14)
⎝ 1 −ν x1ν y1 ⎝ 12 ⎠ 1 −ν x 2ν y 2 ⎝ 12 ⎠⎠ teq3

and the shear rigidity of the panel in its plane (i.e., x-y plane) is given by,

53
 Dxy(q ) = Gxy(q1) ⋅ l ⋅ t S 1 + Gxy(q2) ⋅ l ⋅ t S 2 (4. 15)

G xy(q ) = λ(q ) ⋅ GS (4. 16)

where GS denotes the shear modulus of elasticity of the sheathing. The shear rigidity of
the equivalent shell is given by,
(q )
Dxyeq (q )
= Gxyeq (l ⋅ teq(q ) ) (4. 17)

By equating the equivalent shell and the panel rigidity, Dxy(q ) and Dxyeq
(q )
, and solving for

the equivalent shear modulus of elasticity,

(q ) Gxy( q1) ⋅ t S 1 + Gxy( q2) ⋅ t S 2

Gxyeq = (4. 18)
teq( q )

For panels with different types of sheathing on the two sides, the Poisson ratios for the
equivalent shell in the x and y directions are given by,
ν x 2ν x1 (ESx1 ⋅ t S 1 + ESx 2 ⋅ t S 2 ) ν ν (E ⋅ t + ESy 2 ⋅ t S 2 )
ν xeq = , ν yeq = y 2 y1 Sy1 S1 (4. 19 a,b)
ν x 2 ⋅ ESx1 ⋅ t S1 + ν x1 ⋅ ESx 2 ⋅ t S 2 ν y 2 ⋅ ESy1 ⋅ t S 1 + ν y1 ⋅ ESy 2 ⋅ t S 2

For panels with sheathing on only one side, or the same type of sheathing on both sides,
the Poisson ratios for the equivalent shell are given by,
ν xeq = ν x1 ν yeq = ν y1 (4. 20 a,b)

These equivalent Poisson ratio equations were obtained by assuming that the
deformations in the x and y directions of the panel and equivalent shell are the same.
Furthermore, it is assumed that compatibility of deformations exists between the
sheathing on either side of the panel.

4.3. Nonlinear Finite Element Formulation

The shell element used in this study to model the panels in a CFS building is an
isoparametric element, which typically has between four and sixteen nodes. The sixteen-
node shell element, in Figure 4.3, is selected to model the CFS equivalent shell elements.

54
The reason for this is that this element has the ability to provide more accurate results
than elements of fewer nodes, when a typical CFS panel is modeled with one shell
element. Each node of the sixteen-node shell element has five degrees of freedom (i.e.,
three translations along the x, y, and z axes, and two rotations about the x and y axes),
which is adequate to simulate CFS panels used in construction practice.

An updated Lagrangian formulation is adopted to develop the element stiffness matrix of

the sixteen-node shell element, where the values of all parameters, static and kinematic,
are computed in accordance with the latest deformed configuration of the element. The
procedure to achieve the elastic and geometric stiffness matrices for the sixteen-node
shell element is presented in the following (Bathe and Bolourchi, 1982).

Figure 4.3. Sixteen-node shell element

The shape functions of a shell element are obtained by means of the following Lagrange
interpolation equation:

Nk =
(x1 − α )(x2 − α )L (xnd − α ) for nd ≠ j
(x1 )( )(
− x j x2 − x j xnd − x j ) (4. 21)

where Nk is a cubic shape function for node k, where x is the natural coordinate of the
node k measured in the direction α (i.e., s or r, as shown in Figure 4.3), and nd is the
number of nodes in the direction α (i.e., nd=4 for the sixteen-node shell element in the s
and r directions). The natural coordinates s or r, represent the “normalized” dimensions
of the shell element, since the length of each side of the shell element is always 2. For
instance, the coordinates of node 1 in the r and s directions are -1 and 1, respectively; the

55
coordinates of node16 in the r and s directions are 1/3 and -1/3, respectively. The index j
represents the node under consideration (j=1,2,3 or 4). From Eq. (4.21), the following
expressions for the cubic shape functions for each node of the shell element are,

N1 = −
1
(1 + s )(1 + 3s )(− 1 + 3s )(1 + 3r )(− 1 + 3r )(− 1 + r ) (4. 22 a)
256

N2 =
1
(1 + s )(1 + 3s )(− 1 + 3s )(1 + 3r )(− 1 + 3r )(1 + r ) (4. 22 b)
256

N3 =
1
(1 + 3s )(− 1 + 3s )(− 1 + s )(1 + 3r )(− 1 + 3r )(− 1 + r ) (4. 22 c)
256
−1
N4 = (1 + 3s )(− 1 + 3s )(− 1 + s )(1 + 3r )(− 1 + 3r )(1 + r ) (4. 22 d)
256

N5 =
9
(− 1 + 3s )(1 + 3s )(1 + s )(1 + r )(− 1 + 3r )(− 1 + r ) (4. 22 e)
256

−9
N6 = (1 + s )(1 + 3s )(− 1 + s )(1 + 3r )(− 1 + 3r )(1 + r ) (4. 22 f)
256

N7 =
9
(1 + 3s )(− 1 + 3s )(− 1 + s )(1 + r )(1 + 3r )(− 1 + r ) (4. 22 g)
256
−9
N8 = (1 + s )(− 1 + 3s )(− 1 + s )(1 + 3r )(− 1 + 3r )(− 1 + r ) (4. 22 h)
256

N9 =
9
(1 + s )(1 + 3s )(− 1 + s )(1 + 3r )(− 1 + 3r )(− 1 + r ) (4. 22 i)
256
−9
N 10 = (1 + s )(1 + 3s )(− 1 + 3s )(1 + r )(1 + 3r )(− 1 + r ) (4. 22 j)
256

N 11 =
9
(1 + s )(− 1 + 3s )(− 1 + s )(1 + 3r )(− 1 + 3r )(1 + r ) (4. 22 k)
256
−9
N 12 = (1 + 3s )(− 1 + 3s )(− 1 + s )(1 + r )(− 1 + 3r )(− 1 + r ) (4. 22 l)
256
− 81
N 13 = (1 + s )(1 + 3s )(− 1 + s )(1 + r )(− 1 + 3r )(− 1 + r ) (4. 22 m)
256

N 14 =
81
(1 + s )(1 + 3s )(− 1 + s )(1 + r )(1 + 3r )(− 1 + r ) (4. 22 n)
256

N 15 =
81
(1 + s )(− 1 + 3s )(− 1 + s )(1 + r )(− 1 + 3r )(− 1 + r ) (4. 22 o)
256

56
 − 81
N 16 = (1 + s )(− 1 + 3s )(− 1 + s )(1 + r )(1 + 3r )(− 1 + r ) (4. 22 p)
256

The coordinates at any point within the element at load increment q are mathematically
expressed as,
n
t n (q ) (q )
x ( q ) = ∑ N k xk( q ) + ∑ teq k N kVnx k (4. 23)
k =1 2 k =1
n
t n (q ) (q )
y ( q ) = ∑ N k yk( q ) + ∑ teq k N kVny k (4. 24)
k =1 2 k =1
n
t n (q ) (q )
z ( q ) = ∑ N k zk( q ) + ∑ teq k N kVnz k (4. 25)
k =1 2 k =1
(q )
where teq is the equivalent shell thickness, and xk(q), yk(q), zk(q) are the Cartesian
coordinates of node k. The parameters Nk (k=1,..,n) are the shape functions, defined in
(4.22), as functions of r and s. The parameter Vn is a unit vector normal to the shell
surface, and the x, y, and z subscripts represent the vector components. In the same
manner, the equations to find the displacements at any point within the shell element are
given by,
n
t n (q ) (q )
u ( q ) = ∑ N k xk( q ) + ∑ teq k N kVnx k (4. 26)
k =1 2 k =1
n
t n (q ) (q )
v ( q ) = ∑ N k yk( q ) + ∑ teq k N kVny k (4. 27)
k =1 2 k =1
n
t n (q ) (q )
w( q ) = ∑ N k zk( q ) + ∑ teq k N kVnz k (4. 28)
k =1 2 k =1

(q ) ( q −1)
where Vni k = Vni k − Vni k is a vector normal to the shell element surface that stores the
increments of the direction cosines in the i = x, y or z direction. The elements of Vni k are
expressed as a function of the rotations at node k. However, a convenient way to
(1) (1)
determine Vni is by defining two initial orthogonal unit vectors V1k and V2 k for the

initial load increment, as depicted in Figure 4.4, given by,

57
 (1)
(1) e2 × Vn k
V1k = (1)
(4. 29)
e2 × Vn k
2

where e2 is a unit vector in the direction of the local y axis of the shell element. The
(1)
corresponding vector V2 k is then,

(1) (1) (1)

V2 k = Vn k × V1k (4. 30)

y, v Vn
x, u
γ V2

k
φ V1
teq

Figure 4.4. Vectors at node k

(q )
As shown in Figure 4.4, let φk and γk be the rotations of the normal vector Vn k around
(q ) (q )
the vectors V1k and V2 k , respectively, measured in the configuration corresponding to
load increment q. Thus, for small rotations, Vn is given by,
(q ) (q )
Vn k = −V2 k φk + V1k γ k (4. 31)

Since the formulation presented in this section is intended for large deformations, Vn must
be evaluated as,
( q −1) ( q −1) ( q −1)
Vn k = Vn k + ∫ − V2 k dφk + V1k dγ k (4. 32)
φ ,γ k

where the node rotations φ and γ are obtained from the previous analysis results. For the
first iteration, note that the rotations are equal to zero. The integration in Eq. (4.32) can

58
be performed using multiple steps as for the Euler method (Bathe and Bolourchi, 1982).
Also the integration can be performed in one step by using an orthogonal transformation
matrix of finite rotations (Argyris, 1982) having the following form,
(q ) ( q −1)
Vn k = TϕVn k (4. 33)

where the orthogonal transformation matrix is,

sin (ϕ ) 1 ⎛ sin (ϕ / 2) ⎞ 2
2

Tϕ = I 3 + S + ⎜⎜ ⎟ S
2 ⎝ ϕ / 2 ⎟⎠
(4. 34)
ϕ

in which,
⎡ 0 0 γk ⎤
ϕ = φ +γ , 2
k
2
k S = ⎢⎢ 0 0 − φk ⎥⎥
⎣⎢− γ k φk 0 ⎦⎥

Having the equations to evaluate the displacement components at any point in the shell
element (i.e., (4.26) to (4.28)), the displacement derivatives, in terms of natural
coordinates, are expressed in Cartesian coordinates as,

⎧ u, x ⎫ ⎧ u,r ⎫
⎪u ⎪ ⎪u ⎪
⎪ , y ⎪ ⎡ J −1 0 0 ⎤⎪ ⎪
,s

⎪⎪ u, z ⎪⎪ ⎢ m ⎪
⎥ ⎪ ,t ⎪⎪
u
⎨ ⎬=⎢ 0 J m−1 0 ⎥⎨ ⎬ (4. 35)
⎪ v, x ⎪ ⎢ 0 0
v,r
J m−1 ⎥⎦ ⎪ ⎪
⎪ M ⎪ ⎣ ⎪ M ⎪
⎪ ⎪ ⎪ ⎪
⎪⎩w, x ⎪⎭ ⎪⎩w,r ⎪⎭

where the comma denotes derivation, and the Jacobian matrix Jm is defined as,

⎡ ∂x ∂y ∂z ⎤
⎢ ∂r ∂r ∂r ⎥
⎢ ∂x ∂y ∂z ⎥
Jm = ⎢ ⎥ (4. 36)
⎢ ∂s ∂s ∂s ⎥
⎢ ∂x ∂y ∂z ⎥
⎢⎣ ∂t ∂t ∂t ⎥⎦

Using the displacement derivatives of Eq. (4.35), the linear strain-displacement matrix is,

59
 (q ) (q ) (q ) (q )
⎡h(q ) 0 0 g11k G1k k
g 21 G1k ⎤
⎢ k ,1 (q ) k (q ) (q ) k (q )
⎥
⎢ 0 hk(q, 2) 0 g12k G2 k
g 22 G2 ⎥
⎢ (q ) (q ) k (q ) k (q ) ⎥
(q ) ⎢ 0 0 hk(q,3) g13k G3k g 23 G3 ⎥ (4. 37)
BL = (q )
⎢h hk(q,1) 0 (g
k (q )
G2k
(q )
+ g12k G1k
(q ) (q )
)( k (q ) k (q )
g 21 G2 + g 22 G1 )
k (q ) k (q ) ⎥
⎢ k(q, 2)
(g ) ( )⎥
11
(q ) k (q ) k (q ) k (q ) k (q ) k (q ) k (q ) k (q ) k (q )
⎢ hk ,3 0 hk ,1 G +g G g 21 G3 + g 23 G1 ⎥
(g )( )
11 3 13 1
⎢ (q ) (q ) k (q ) k (q ) k (q ) k (q ) k (q ) k (q ) k (q ) k (q ) ⎥
⎣ 0 hk ,3 hk , 2 12 G3 + g13 G2 g 22 G3 + g 23 G2 ⎦

where
(q) (q)
hk( q,1) = J m−1i1 hk ,r + J m−1i 2 hk ,s Gik
(q)
( (q) (q)
)
= t J m−1i1 hk ,r + J m−1i 2 hk ,s + J m−1i 3 hk
(q)

(q ) 1 (q ) (q ) 1 (q )
g1ki = − teqV2ki g 2ki = teqV1ik
2 2

The nonlinear strain-displacement matrix is,

(q ) (q ) k (q ) k (q ) ⎤
⎡h(q ) 0 0 g11k G1k g 21 G1
⎢ k ,1 (q ) (q ) k (q ) k (q )
⎥
⎢ 0 hk(q,1) 0 g12k G1k g 22 G1 ⎥
⎢ (q ) (q ) k (q ) k (q ) ⎥
⎢ 0 0 hk(q,1) g13k G1k g 23 G1 ⎥
(q ) (q ) ( ) (q )
⎢h (q ) 0 0 g11k G2k k
g 21
q
G2k ⎥
(q ) ⎢ ⎥
k ,2
(q ) (q ) k (q ) k (q )
BNL = ⎢ 0 hk(q, 2) 0 g12k G2k g 22 G2 ⎥ (4. 38)
⎢ (q ) k (q ) k (q ) k (q ) k (q ) ⎥
⎢ 0 0 hk , 2 g 13 G 2 g 23 G2 ⎥
⎢h(q ) k (q ) k (q ) k (q ) k (q ) ⎥
0 0 g11 G3 g 21 G3
⎢ k ,3 (q ) (q ) (q ) k (q ) k (q )
⎥
⎢ 0 hk ,3 0 g11k G3k g 22 G3 ⎥
⎢ (q ) (q ) k (q ) k (q ) ⎥
⎣ 0 0 hk(q,3) g11k G3k g 23 G3 ⎦

The Cauchy stress matrix is,

⎡τ 11(q ) I 3 Sym. ⎤
⎢ ⎥
τ (q ) = ⎢τ 12(q ) I 3 τ 22(q ) I 3 ⎥ (4. 39)
⎢τ 13(q ) I 3 τ 31(q ) I 3 τ 33(q ) I 3 ⎥
⎣ ⎦

where I3 is a 3x3 identity matrix.

60
The Cauchy stress vector is,

τT
(q )
[
= τ 11
(q )
τ 22 (q ) τ 33(q ) τ 12 (q ) τ 13(q ) τ 23(q ) ] (4. 40)

Lastly, the element linear and nonlinear stiffness matrices are expressed as

k L(q ) = ∫ BL(q ) C (q ) BL(q )dV (q )

T
(4. 41)
V

(q ) (q ) (q ) (q )
= ∫ BNL τ BNL dV (q )
T
k NL (4. 42)
V

where C is the constitutive matrix at load increment q, expressed as follows,

⎡ E xeq
(q)
ν xy E xeq
(q)
0 0 0 0 ⎤
⎢ ⎥
⎢ν yx E yeq
(q) (q)
E yeq 0 0 0 0 ⎥
⎢ 0 0 0 0 0 0 ⎥
(q )
⎢ (q ) ⎥
C =⎢ 0 0 0 Gxyeq 0 0 ⎥ (4. 43)
⎢ 5 (q ) ⎥
⎢ 0 0 0 0 Gxyeq 0 ⎥
⎢ 6 ⎥
⎢ 0 5 (q ) ⎥
0 0 0 0 Gxyeq
⎢⎣ 6 ⎥⎦

where Exeq, Eyeq, Gxyeq are computed from Eqs. (4.8), (4.14), and (4.18), respectively. The
modulus of elasticity in the z direction is not considered, because the stresses in this
direction are negligible for plates and shells. The shear modulus of elasticity in the planes
x-z and y-z are affected by a factor of 5/6 to account for the parabolic variation of the
transverse shear strain through the thickness of the shell element (Cook, Malkus and
Plesha, 1989).

Equations (4.41) and (4.42) are integrated over the volume of the shell element. The
integration can be carried out using Gauss-Legendre sampling points, where the
weighting values are known. The number of sampling points depends on the type of
element; for a sixteen-node element, a 4x4x2 integration is recommended in order to
achieve a good approximation (i.e., the numerical integration is performed by using 32
sampling points). A detailed description of the nonlinear finite element analysis
formulation is provided by Bathe and Bolourchi (1982), and Bathe (1996).

61
The structural elastic stiffness matrix of a building, at load increment q, is expressed as,

K L(q ) = ∑ k L(q ) (4. 44)

ne

where ne is the number of shell elements in the model of the building. The structural
geometric nonlinear stiffness matrix of a building, at load increment q, is expressed as,
(q )
K NL = ∑ k NL
(q )
(4. 45)
ne

Finally, the equilibrium conditions in terms of incremental displacements ∆u for the

structure are represented by,

(K ( ) + K ( ) )⋅ ∆U ( ) = R (
L
q q
NL
i q +1)
(
− F (q +1) )(i −1)
(4. 46)

where R(q+1) is the vector of externally incremental applied loads on the building, R(q+1) -
F(q+1) represents the unbalanced forces in the building. i is the equilibrium iteration index
associated with the solution using iterative methods, such as the Newton-Raphson
method (Bathe, 1996). Note that iterative methods perform equilibrium iterations to
reduce or eliminate the unbalance force within each load increment. Thus, the solution of
Eq. (4.46) is reached when the vector of internal forces F(q+1) is equal to the vector of
external forces R. The vector of internal forces is evaluated as,

(F ( ) )(
q +1 i −1)
= ∫(
V )
( q +1 ) ( i −1 )
(B )
(q +1)T
L
(i −1)
(τˆ( ) )( ) (dV ( ) )(
q +1 i −1 q +1 i −1)
(4. 47)

If a single-step incremental load method is used to solve the nonlinear equilibrium Eq.
(4.46), no equilibrium iterations are performed. Thus, the internal force vector F(q+1) is
removed from Eq. (4.46), and the equilibrium equation reduces to the following
expression,

(K ( ) + K ( ) )⋅ ∆U = R (
L
q q
NL
q +1)
(4. 48)

Once the incremental displacements ∆U are obtained from the solution of Eq. (4.46) or
Eq. (4.48), the accumulated nodal displacements of the building are expressed as,

62
 (U ( ) )( ) = (U ( ) )(
q i q −1 i −1)
+ ∆U (i −1) and U (q ) = U (q−1) + ∆U (4. 49 a,b)

for iterative and single-step methods, respectively.

Knowing the nodal displacements of the building, the internal forces in the members are
computed using the strain-displacement matrices for the shell elements. The equations,
for estimating the axial force and bending moment in the studs, are presented in the next
section.

4.4. Estimating the Internal Forces in the Studs from the Equivalent Shell
Transforming SWP into equivalent shell elements simplifies the analysis process for CFS
buildings, because the designer can model a complete SWP using one equivalent shell
element that includes the sheathing and studs. As a result, the studs are not explicitly
included in the model of the SWP as frame elements. However, the internal forces in the
studs still need to be identified in order to review the gravity design, when PBD is being
carried out, or the capacity of the studs is evaluated.

The internal stresses at any point of the SWP can be computed through the strain-
displacement matrices using the nodal displacements from the structural analysis. The
procedure to estimate the axial forces in the studs is described in the following. First, the
locations of the studs within the equivalent shell elements need to be identified. Then the
axial stresses in the direction of the studs, σmy, are computed in the mid-plane of the shell
element, at several points along each stud within the equivalent shell element. In this
study, the axial stresses are computed at three points along the studs; that is, in the
bottom, the middle, and the top section of the stud element. The stresses could be
computed at more points, but the accuracy would not be improved much, because
typically the maximum forces and moments are registered at these three locations. In
terms of the finite element analysis, identifying a specific point within a shell element
consists of determining the Cartesian coordinates of the selected point and then its natural
coordinates (i.e., r, and s). Finally, having the stresses at the selected points of each stud,

63
the largest stress is used to determine the maximum axial force in the stud. The axial
force in the stud is estimated with the expression

PF = teqσ my wSF (4. 50)

where teq is the thickness of the equivalent shell, and σmy is the maximum stress in the y
direction at the mid-plane of the shell element at the bottom, middle, or top of the stud.
The parameter wSF is equal to sF/2 for the end-studs and sF for intermediate studs, where
sF is the distance between studs on centre. The internal bending moments in each stud,
computed from the equivalent shell element, are given by the following equation,
teq

M x = wSF ∫ 2t i z⋅i σ y ⋅ dz (4. 51)

− eq

where σy is the stress through the thickness of the equivalent shell element in the y
direction, at distance z from the mid-plane. The internal moments in each stud are also
computed at the bottom, middle and top section, so that the maximum moment for the
stud can be identified.

Although the material and geometric properties of the shell elements are set equivalent to
the properties of the SWP, their structural forms are different. The SWP consist of studs
and sheathing, similar to a ribbed plate, whereas the shell elements are flat plates with a
uniform thickness. Therefore, a variation in the prediction of the forces in the studs is
expected.

The accuracy of the axial force evaluation, based on the shell elements, which may not be
as good as that for the lateral displacements and forces, is affected by the following
factors. First, the vertical stiffness of a SWP is not uniform along its cross section,
because the stiffness is greater in the locations where the studs are placed and smaller at
the locations without studs; for the equivalent shell elements the vertical stiffness is
evenly distributed along its cross section. Second, SWP are subjected to forces that are
transmitted to the studs from the joists, where the distance between studs is typically
small (e.g., 406 mm, 609 mm). For the equivalent shell elements, the loads are

64
concentrated forces on the top four nodes that are typically evenly spaced at one-third of
the length of the equivalent shell element. Thus, depending on the aspect ratio of the
equivalent shell element the distance between nodes may be larger than the studs spacing.
This discrepancy can adversely affect the accuracy in the prediction of the axial forces.
Third, in this study, the stress used for estimating the axial forces is computed at three
discrete points along the height of the equivalent shell elements, in the locations of the
studs. Consequently, if any of these selected points coincide with an area of stress
concentration, the axial force results may be inaccurate. The stresses at any location
within an equivalent shell element are computed using the shape functions, even though
the points may not coincide with the nodes of the equivalent shell element.

4.5. Comparison of Results for SWP with SFEA and CFEA.

To demonstrate the effectiveness of SFEA for SWP, a comparison of results for nine
SWP with different lengths is carried out using SFEA and CFEA. Although CFEA is a
numerical approach for which the solution may not be exact, CFEA can provide accurate
results for most engineering problems. The accuracy of CFEA depends primarily on the
type, size, and number of elements used to model a structure. Even though CFEA is an
approximate approach, its precision is appropriate for most civil engineering problems.
Therefore, in this study CFEA is used as a benchmark for comparing the results from
SFEA.

The SFEA is carried out using a computer program described in Section 5.6, whereas the
CFEA is carried out using SAP2000 (2006) which is a commercial software widely used
in structural engineering practice. Nine SWP are analyzed with different lengths: 609 mm
(2 ft), 1219 mm (4 ft), 2438 mm (8 ft), 3657 mm (12 ft), 4876 mm (16 ft), 6095 mm (20
ft), 7314 mm (24 ft), 8533 mm (28 ft), and 9752 mm (32 ft). The height of all the SWP is
2438 mm (8 ft) and, therefore, the corresponding aspect ratios (i.e., height-to-length ratio)
of the foregoing SWP are 0.25, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, and 4.0, respectively. All
of the SWP are built with the same details and material properties: oriented strand board
(OSB) sheathing on one side, C-section studs and tracks 152C51– 1.18 mm (600S200-43

65
mils) with no lip, spaced at 406 mm (16 in) on centre, and with one end-stud. The C-
section profiles used for the top tracks are the same as those of the studs.

For the SFEA, one shell element and three frame elements are used to model each SWP.
The top track is modeled with three frame elements, since the shell elements have four
nodes on the top, but the bottom tracks are not included in the model because the bottom
nodes are fixed (i.e., the SWP is a cantilever). For the CFEA, the studs are modeled with
frame elements 203 mm in length, and the sheathing is modeled using square four-node
shell elements with a 203 mm side length. Thus, as the length of the SWP is increased,
the number of shell elements in the CFEA model increases as well, while the number of
shell elements for the SFEA remains the same. Shown in Figure 4.5 are both the
simplified and conventional analysis models. Given in Table 4.1 is the number of shell
and frame elements employed in each model (note the substantially smaller number of
shell elements in the SFEA model). The nodal loads for both models are calculated, based
on a uniformly distributed load on the top track of 17.23 kN/m, and are applied at the top
four nodes in the x and y directions, as depicted in Figure 4.5. Also, a 1.00 kN/m2 load is
applied on the face of the SWP to simulate wind loads.

Listed in Table 4.2 are the properties of the SWP employed in the calculation of the
equivalent shell elements. Although the Poisson ratios for the OSB sheathing are actually
0.16 and 0.23 in the x and y directions, respectively (Thomas, 2004), the Poisson ratio
value of 0.23 is here used in both directions because SAP2000 allows only one Poisson
ratio value in each plane. The number of studs in each SWP, nF, is presented in Table 4.3,
together with the properties of the equivalent shell elements. The equivalent properties
should be the same for all nine SWP. However, since the ratio of length-to-number of
studs is not the same for all the SWP, the equivalent properties of the SWP are in fact
different. The number of studs in the SWP is computed as the length of the SWP divided
by the spacing of the studs, plus one to account for the end-stud.

Given in Table 4.4 are the x, y and z displacements of node 1 predicted by CFEA and
SFEA; these results are also plotted in Figure 4.6. The maximum axial forces and

66
bending moment in the SWP studs, listed in Table 4.5 and shown in Figures 4.7 and 4.8,
are determined from the most critical stud in each SWP, regardless the location of the
studs; specifically, the maximum axial forces for the nine SWP analyzed were found at
the bottom of the end-studs, while the maximum bending moments were found at the
bottom of the middle studs.

Figure 4.5. SWP model for (a) SFEA and (b) CFEA

Table 4.1 Number of shell and frame elements in the modeling of the SWP

SWP
CFEA SFEA
Aspect ratio
mm Shell Frame Total Shell Frame Total
0.25 18 20 38 1 3 4
0.50 54 42 96 1 3 4
1.00 108 75 183 1 3 4
1.50 162 108 270 1 3 4
2.00 216 141 357 1 3 4
2.50 270 174 444 1 3 4
3.00 324 207 531 1 3 4
3.50 378 240 618 1 3 4
4.00 432 273 705 1 3 4

67
Table 4.2 Properties of the sheathing and framing studs employed for determining the properties of
the equivalent shell elements

Sheathing Stud
Ma- Ex Ey G t ν E G t A I
terial MPa MPa MPa mm MPa MPa mm mm2 mm4
OSB 1983 9917 925 11.11 0.23 203000 77000 1.18 292.49 9.7×105

Table 4.3 Properties of the equivalent shell elements

SWP Exeq Eyeq Geq teq

nF
Aspect ratio MPa MPa MPa mm
0.25 2 315.97 1676.74 58.44 175.85
0.50 4 315.97 1676.74 58.44 175.85
1.00 7 282.65 1489.10 56.31 182.51
1.50 10 271.71 1428.03 55.57 184.92
2.00 13 266.28 1397.78 55.20 186.17
2.50 16 263.03 1379.73 54.97 186.94
3.00 19 260.87 1367.73 54.82 187.45
3.50 22 259.33 1359.18 54.72 187.82
4.00 25 258.18 1352.78 54.63 188.10

Table 4.4 Displacements (mm) of node 1 for the SWP, predicted with SFEA and CFEA

SWP
CFEA SFEA
Aspect ratio
mm x y z x y z
0.25 7.78 0.56 6.87 14.00 1.55 6.20
0.50 5.62 0.50 6.83 7.01 0.88 6.19
1.00 4.66 0.34 7.65 5.21 0.59 6.20
1.50 4.42 0.24 7.69 4.76 0.47 6.18
2.00 4.31 0.15 7.49 4.58 0.38 6.17
2.50 4.27 0.07 7.34 4.48 0.30 6.17
3.00 4.24 -0.02 7.29 4.43 0.24 6.18
3.50 4.24 -0.10 7.29 4.42 0.19 6.19
4.00 4.25 -0.18 7.30 4.34 0.15 6.08

68
 Table 4.5 Maximum axial forces in the studs of the SWP from CFEA and SFEA

SWP CFEA SFEA

Aspect ratio Axial Moment Axial Moment
mm kN kN kN-mm kN kN kN-mm
0.25 28.30 -37.18 903.62 32.51 -101.11 1716.81
0.50 27.57 -35.40 1001.38 34.52 -41.73 1151.63
1.00 22.56 -31.64 1092.84 23.39 -33.53 1160.44
1.50 20.02 -31.29 1151.67 21.38 -30.46 1157.63
2.00 17.95 -32.15 1197.59 18.09 -32.54 1155.16
2.50 16.01 -33.53 1221.19 15.40 -33.41 1152.76
3.00 14.15 -35.17 1230.24 14.59 -33.27 1151.05
3.50 12.36 -36.96 1229.16 13.48 -32.62 1149.51
4.00 10.64 -35.85 1224.37 17.64 -31.48 1255.51

X (CFEA) X (SFEA) Y (CFEA) Y (SFEA) Z (CFEA) Z (SFEA)

14

12

10

8
DISPLACEMENT (mm)

0
0 0.5 1 1.5 2 2.5 3 3.5 4

-2
SWP ASPECT RATIO (LENGTH / HEIGHT)

Figure 4.6. Displacement comparisons at node 1 with SFEA and CFEA

69
 Tension end-stud (CFEA) Tension end-stud (SFEA) Compression end-stud (CFEA) Compression end-stud (SFEA)

60

40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4
AXIAL FORCE (kN)

-20

-40

-60

-80

-100

-120
SWP ASPECT RATIO (LENGTH / HEIGHT)

Figure 4.7. Maximum axial forces in the studs of the SWP

STUD Max (CFEA) STUD Max (SFEA)

1800.00

1600.00

1400.00

1200.00
MOMENT (kN mm)

1000.00

800.00

600.00

400.00

200.00
0 0.5 1 1.5 2 2.5 3 3.5 4
SWP ASPECT RATIO (LENGTH / HEIGHT)

Figure 4.8. Maximum bending moment in studs of the SWP

70
Figure 4.6 indicates good agreement for the in-plane displacements of the SWP computed
using SFEA and CFEA. However, the predicted out-of-plane displacements are different
because, even though the sheathing is attached to the flange of the studs, SAP2000 (2006)
models the shell and frame elements in the same plane, so that the offset of the sheathing
from the centerline of the studs is not considered. Conversely, the equivalent shell
elements consider the offset of the sheathing. In this case, however, the differences are
not significant since the sheathing is relatively thin compared to the depth of the studs.

The predictions of the maximum axial force in the studs of the SWP are in good
agreement, as shown in Figure 4.7, except for the studs of the SWP with an aspect ratio
equal to 0.25. The predictions of the bending moments depicted in Figure 4.8 are
acceptable, except for the two SWP with aspect ratio of 0.25 and 4.0, respectively; the
error may be caused by the narrow shape of the shell elements in one direction, for which
the cubic shape functions are not suitable. Therefore, based on the comparisons of the
nine SWP, it is concluded that the SFEA model predicts reasonably accurate results when
the aspect ratio for the equivalent shell elements is in the range of 0.5 to 3. For panels
with aspect ratios outside of this range, it is recommended to use more than one
equivalent shell element to model each panel.

In the analysis of the nine SWP, the loads were applied on the same four top nodes in
both CFEA and SFEA models, leads to good agreement in the prediction of the axial
forces. Typically, the CFEA model has a node on the top of each stud, whereas the SFEA
model has only four nodes on the top of the SWP, regardless of the number of studs. To
test how the accuracy of the analysis results is affected when uniformly distributed loads
are applied on the top of the panel, the analysis of the nine SWP was carried out again.
This time, the loads in the x and y directions are uniformly distributed on the top of the
panel. Listed in Table 4.6 are the displacements and axial forces obtained from the
CFEA. The displacements in the z direction and the maximum bending moments are not
in Table 4.6 since they have the same magnitude as those presented in Table 4.4. The
reason is that in both cases, the loads acting in the z direction are not applied on the nodes
but on the face of each shell element.

71
 Table 4.6 Displacements of node 1 and axial forces of the SWP, predicted with CFEA

SWP
Aspect Displacements (mm) Axial force (kN)
ratio
mm x y Tens Com
0.25 7.78 0.59 28.28 -37.15
0.50 5.62 0.52 27.61 -35.32
1.00 4.66 0.41 23.17 -31.00
1.50 4.41 0.38 21.72 -29.55
2.00 4.30 0.37 21.08 -28.92
2.50 4.24 0.36 20.73 -28.58
3.00 4.20 0.36 20.52 -28.38
3.50 4.18 0.35 20.39 -28.25
4.00 4.16 0.35 20.29 -28.16

Tension end-stud (CFEA) Tension end-stud (SFEA) Compression end-stud (CFEA)

Compression end-stud (SFEA) Tension end-stud UDL (CFEA) Compression end-stud UDL (CFEA)

60

40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4
AXIAL FORCE (kN)

-20

-40

-60

-80

-100

-120
SWP ASPECT RATIO (LENGTH / HEIGHT)

Figure 4.9. Maximum axial forces in the studs of the SWP

72
The displacements in the x and y direction in Tables 4.6 and 4.4 show a small difference
that can be neglected. Shown in Figure 4.9 are the predicted axial forces in the studs
listed in Table 4.6 and also the curves in Figure 4.7 in order to appreciate the difference
in the predictions. Although the difference in the predicted axial forces for the SWP with
concentrated loads and uniformly distributed load is not significant, the tendency is
different. This means that in the modeling of a building, the aspect ratio of the shell
elements may affect the prediction of the axial forces in the studs due to the different
distribution of the loads in the equivalent shell elements.

4.6. Stiffness Degradation Model for SWP

Experimental investigations, such reported by COLA-UCI (2001), Fulop and Dubina
(2004), NAHBRC (1997), and Branston et al. (2004, 2006), have shown that the load-
displacement relationship for SWP is nonlinear. In addition, it is observed in the
aforementioned experimental investigations that the loss of stiffness and the failure of
SWP is primarily due to looseness of the sheathing-to-framing connections. Therefore,
the nonlinear relationship between load and displacement of SWP needs to be accounted
for in the finite element model of analysis.

Figure 4.10. Characterization of the SWP loss in strength

73
Shown in Figure 4.10 is the proposed nonlinear model to characterize the stiffness
degradation of SWP subjected to lateral forces: Ki is the initial stiffness of the SWP,
tangent to Pr = 0, and is computed with SFEA using the equations in Section 4.2; K a(q ) is
the secant stiffness corresponding to the load increment, q; PR is the lateral strength of the
SWP, computed using the method described in Chapter 3 (PR can also be obtained from
experimental investigations or design tables, such as AISI (2004)); Pa is the magnitude of
the total lateral force applied on the top of the panel at load increment q; λ(q) is the
stiffness degradation coefficient that characterizes the nonlinear behaviour of the SWP
under the applied lateral forces until failure, and is defined by,
β
(q ) ⎛ Pa(q ) ⎞
λ = 1 − ⎜⎜ ⎟⎟ (4. 52)
⎝ PR ⎠
where β=(1.5 in/sC) is a stiffness degradation nonlinearity exponent, calibrated in
accordance with experimental results from Branston et al. (2004) and COLA-UCI (2001).
Described in the Appendix C is the calibration of β. It is observed from the results of the
experimental investigations that the nonlinear load-displacement relationship for a SWP
is primarily influenced by the sheathing material and edge screw spacing; for reasons of
simplicity, in this study β was calibrated considering the screw spacing only.

When the stiffness degradation coefficient is computed, using Eq. (4.52), λ(q) is equal to
unity at the initial load increment, which indicates that the SWP lateral stiffness has not
yet been affected. When the applied load Pa is equal or greater than the maximum
strength PR, λ(q) becomes zero. This indicates that the SWP has completely lost its
stiffness to resist lateral deformation induced by the applied lateral loads. However, the
modulus of elasticity of the CFS studs remains unchanged in the y direction of the shell
element, which means the equivalent shell may still have sufficient stiffness contributed
by the studs to allow the SWP to continue to carry vertical loads.

Having the stiffness degradation coefficient λ(q) calculated at load increment q, the
degraded stiffness of the SWP for the next load increment of the analysis is calculated by
first multiplying the modulus of elasticity of the sheathing by λ(q). Then, the reduced

74
modulus of elasticity is used to calculate the modulus of elasticity of the equivalent shell
element in accordance with the method described in Section 4.2. Afterwards, the
constitutive matrix for the sixteen-node shell element is formed using Eq. (4.43).

4.7. Prediction of the Nonlinear Response of SWP Subjected to Lateral Loading.

Experimental results for single SWP, such as from Branston et al. (2004), NAHBRC
(1997) and COLA-UCI (2001), are used to validate the accuracy of the proposed method
for the analysis of SWP. The accuracy of the predicted response is generally correlated
with the material properties and the geometric dimensions of the components. As not all
properties are reported in the foregoing literature, the material properties adopted in the
evaluation may not match with those of the tested materials. In this study, the geometric
gross properties of the steel studs are computed for each case, and the material properties
of the sheathing and steel being used in the calculations are defined in each case.

Branston et al. (2004) conducted a series of experimental investigations of SWP with

different sheathing materials. Shown in Figures 4.11 to 4.16 are predicted and
experimental force-displacement curves for the SWP tested by Branston et al. (2004).
The predicted force-displacement curves were obtained using the SFEA and stiffness
degradation models proposed by this study in the foregoing. The sheathing materials
investigated were 11 mm (7/16 in) OSB, and 12.5 mm (1/2 in) DFP, fastened on one side
of the panel with No. 8 screws (diameter = 4.06 mm) spaced every 305 mm (12 in) in the
field. The screw spacing at the edge of the SWP is indicated at the foot of each figure.
The same 92S41-1.12 mm (362S162-44 mils) C-shape CFS studs were used for all the
SWP test specimens. The studs were spaced 610 mm (24 in) on centre, and double back-
to-back C-shape studs were placed at the ends of the panel. Three edge screw spacings
(sC) of 76 mm (3 in), 102 mm (4 in) and 152 mm (6 in) were investigated. The tested
SWP were 2438 mm (8 ft) in height and 1219 mm (4 ft) in length.

75
 Experimental Experimental Experimental SFEA

20

18

16

14
Lateral load (kN)

12

10

21B 21C
6
21A

0
0 10 20 30 40 50 60 70 80
Lateral displacement (mm)

Figure 4.11. Predicted vs. experimental (Branston et al., 2004) curves of OSB SWP (sC = 152mm)

Experimental Experimental Experimental SFEA

30

25

20
Lateral load (kN)

21B

15
21A
23C

10

0
0 10 20 30 40 50 60 70 80
Lateral displacement (mm)

Figure 4.12. Predicted vs. experimental (Branston et al., 2004) curves of OSB SWP (sC = 102mm)

76
 Experimental Experimental Experimental SFEA

35

30

25
Lateral load (kN)

20

25C

15
25B
25A

10

0
0 10 20 30 40 50 60 70 80
Lateral displacement (mm)

Figure 4.13. Predicted vs. experimental (Branston et al., 2004) curves of OSB SWP (sC = 76 mm)

Experimental Experimental Experimental SFEA

25

20
Lateral load (kN)

15

10

11B 11A II
5

0
0 10 20 30 40 50 60 70 80 90
Lateral displacement (mm)

Figure 4.14. Predicted vs. experimental (Branston et al., 2004) curves of DFP SWP (sC = 152mm)

77
 Experimental Experimental Experimental Experimental SFEA

35

30

25
5D
5A
Lateral load (kN)

20

15

10
5B
5C
5

0
0 10 20 30 40 50 60 70 80 90
Lateral displacement (mm)

Figure 4.15. Predicted vs. experimental (Branston et al., 2004) curves of DFP SWP (sC = 102mm)

Experimental Experimental Experimental SFEA

40

35

30

25
Lateral load (kN)

20
13B

15

13A
10
13C

0
0 10 20 30 40 50 60 70 80
Lateral displacement (mm)

Figure 4.16. Predicted vs. experimental (Branston et al., 2004) curves of DFP SWP (sC = 76 mm)

78
The National Association of Home Builders Research Center (NAHBRC, 1997)
published a report that presents experimental results for long shear wall panels that are
12190mm (40 ft) in length and 2438mm (8ft) in height. Figure 4.17 presents a
comparison of the experimental results with those predicted herein for the force-
displacement curve of a long SWP without openings. NAHBRC (1997) does not specify
the dimensions of the studs, but it states that “The wall framing is consistent with usual
construction practices.” Therefore, 92S41-1.12 mm (362S162-44mils) C-shape CFS studs
were assumed in this comparison. OSB sheathing of 11.11 mm (7/16 in) thickness was
fastened on one side of the SWP using No. 8 screws spaced at 305 mm (12in) in the field
and at 152mm (6in) at the edge of the panel. GWB of 12.7 mm (1/2 in) thickness was
fastened on the other side of the SWP using No. 6 screws spaced at 254 mm (10 in) in the
field and at 178 mm (7 in) at the edge.

SFEA Experimental

200

180

Wall
160

140
Lateral Load (kN)

120

100

80

60

40

20

0
0 10 20 30 40 50 60
Lateral displacement (mm)

Figure 4.17. Predicted vs. experimental (NAHBRC, 1997) curves of a 40 ft OSB SWP (sC = 152mm)

Plotted in Figures 4.18 to 4.20 are comparisons of experimental versus predicted force-
displacement curves for SWP with different characteristics. The experimental curves
were obtained from investigations carried out by COLA-UCI (2001). The three SWP that

79
are compared have the same dimensions: 2438 mm (8 ft) in length and 2438 mm (8 ft) in
height. The studs used in the test were 89S41-0.86 mm (350S162-33mils) spaced 610 mm
(24 in) on centre; double end-studs were used. The sheathing used in the comparisons,
presented in Figures 4.10 and 4.19, was 4-ply Structural I Rated (STR) of 11.91mm
(15/32 in) thickness. The sheathing was fastened to the framing on one side only with No.
8 screws spaced 305 mm (12 in) in the field and either 152 mm (6 in) or 51 mm (2 in) at
the edge. The SWP, corresponding to Figure 4.20, had the same characteristics as the
SWP corresponding to Figure 4.18, except that the sheathing material used for the test in
Figure 4.20 was OSB of 11.11 mm (7/16 in) thickness.

Predicted Experimental Experimental

35

30

Group 14A
25
Lateral load (kN)

20

15

10

0
0 10 20 30 40 50 60
Lateral displacement (mm)

Figure 4.18. Prediction vs. experimental (COLA-UCI, 2001) curves of STR SWP (sC = 152mm)

The following material properties were used for SFEA of the SWP. The shear modulus of
elasticity for STR was 8490 MPA (COMPLY, 1999), whereas for OSB and DFP it was
925 MPa and 825 MPa (Okasha, 2004), respectively. The modulus of elasticity associated
with STR, OSB, and DFP were 675 MPa (COMPLY, 1999), 9917MPa (OSB, 1995), and
10445 MPa (CANPLY, 2003), respectively. For GWB, the Young and shear modulus of

80
elasticity were 1290 MPa (GA-235, 2002) and 561 MPa, respectively. Furthermore, the
yielding strength of the studs for all the calculations was 230 MPa.
SFEA Experimental Experimental

80

70

60
Group 16A

50
Lateral load (kN)

40

30

20

10

0
0 10 20 30 40 50 60
Lateral displacement (mm)

Figure 4.19. Predictied vs. experimental (COLA-UCI, 2001) curves of STR SWP (sC = 51mm)

SFEA Experimental

30

25

20
Lateral load (kN)

15
Group 17A

10

0
0 10 20 30 40 50 60
Lateral displacement (mm)

Figure 4.20. Prediction vs experimental (COLA-UCI, 2001) curves of OSB SWP (sC = 152mm)

81
The comparisons in this section are for SWP having different characteristics and
properties, such as different type of sheathing, length, screw spacing and so on. The
force-displacement curves for the SWP were obtained experimentally from the different
researchers mentioned. The predicted force-displacement curve was obtained using SFEA
and the stiffness degradation model proposed in this chapter. In general, the predictions
are in good agreement with the experimental results for all the SWP. However, the
difference in the shapes of the predicted and experimental curves for DFP is evident. The
shape of the curve is significantly affected by the factor β (nonlinearity exponent),
included in the stiffness degradation model defined in Eq. (4.52). For simplicity, the
stiffness degradation model was calibrated without accounting for the type of sheathing
material, thus it may be more appropriate for some type of sheathing materials than for
others. Another factor that significantly affects the prediction of the force-displacement
curve is the lateral strength of the SWP that establishes the failure load, which is the
highest point in the force-displacement response curve. Therefore, the accuracy in the
lateral strength calculation also affects the prediction of the nonlinear behaviour of SWP.

82
 Chapter 5
Pushover Analysis for Performance-Based Design
Assessment of CFS Buildings

5.1. Introduction
The purpose of the Performance-Based Design (PBD) is to design buildings that satisfy
the performance objectives specified by the designer and the governing code(s) of
practice. The application of PBD assessment does not follow a fixed procedure, however,
nor is it the same procedure for all structural systems, since it depends on the building
characteristics and designer’s preferences. For instance, several methods for carrying out
the seismic assessment of buildings for PBD are available in the literature, such as those
discussed in Section 2.2. The complexity of the application of each method is different
and so are their capabilities; as a result, some of the methods might not be suitable for
analyzing certain types of structures. The linear methods, particularly, have limited
applications as they can produce incorrect results when used to analyze structures having
a nonlinear response. From the methods discussed in Section 2.2, the nonlinear static
procedure, better known as pushover analysis, is selected in this study for conducting the
PBD assessment of CFS buildings. Pushover analysis is simpler than dynamic analysis;
yet, accurate results can be obtained for structures with a predominant first mode of
vibration (Krawinkler and Seneviratna, 1998). For the medium height and relatively high
lateral stiffness of low and mid-rise CFS buildings, the framework structures are expected
to have a predominant first mode of vibration. Therefore, reasonably correct results can
be expected from pushover analysis of cold-formed steel (CFS) buildings.

A performance objective involves the combination of a seismic hazard and a performance

level. The seismic hazards are represented by site spectra, which are determined from a
geotechnical study of the site or from seismic maps, provided by FEMA 450 (2003) and
USGS (2002). Typically, geotechnical studies are costly, and so, for conventional

83
structures, seismic maps are preferred. A performance level is a measure of the maximum
damage that a building is allowed to undergo, where the level of damage is associated
with the cost of the building.

PBD assessment involves transforming each selected performance objective, to be

satisfied by the structure, into a tangible parameter that can be characterized in a seismic
analysis. This study adopts a spectrum-based pushover analysis approach, also known as
the force-based approach, in which the performance objectives are transformed into target
base shears. From the maximum target base shear, the overall lateral load that the
building is going to be subjected to can be determined. When the lateral loads applied on
the building are equal to the target base shear for any one of the governing performance
objectives, the shear wall panels (SWP) are checked to determine if they comply with the
acceptance criteria corresponding to that performance level. In this chapter, the procedure
for carrying out the PBD assessment of CFS buildings is described.

The equations to transform the performance objectives into spectral accelerations, and
then into target base shears are described in Sections 5.2 and 5.3, respectively. In Section
5.4, the limit lateral drifts defining the acceptance criteria for the SWP are identified, and
the limit drift ratios are obtained for each performance level. In Section 5.5 the procedure
for carrying out the PBD assessment of CFS buildings is presented, and in Section 5.6 a
computer program created for this study is described.

5.2. Spectrum-Based PBD Assessment of CFS Buildings

In this thesis, the spectrum-based approach is preferred over the displacement-based
approach for the PBD assessment of CFS buildings. Typically, such design is carried out
by using design spectra from seismic codes. The design spectra are used to compute the
target base shears employed to calculate the pattern of lateral loads applied on the
building. Conversely, if the displacement-based approach is chosen for carrying out the
PBD assessment of a building, lateral displacement is used as a target parameter. In this
approach, arbitrary lateral load increments or lateral displacements are applied on the

84
building, until the target displacement is reached or until the structure collapse occurs.
Then, with the total load applied on the building during the analysis, the acceleration is
computed and compared with the site design spectra to determine if the building satisfies
the site requirements. In this approach, the designer does not know whether the building
satisfies the site requirements until the analysis is complete. This can lead to unnecessary
load or displacement increments. On the other hand, the spectrum-based approach is
directly consistent with the seismic codes. A discussion on the advantages and
disadvantages of both approaches is available in the literature (Grierson et al., 2006).
Although Grierson et al. (2006) focus their discussion on steel frameworks, the principles
of PBD assessment also apply to CFS buildings.
(PO) (PO) (PO)
Sa = 0.6 (S DS /T0 ) T + 0.4 S DS
(PO)
Sa = SDS
Spectral response accleration, S a, (g)
(PO)

(PO) (PO)
Sa = S D1 / T

(PO) (PO)
Sa =( S D1
T L) / T
(PO)
SD1

(PO) (PO) (PO)

0.0 T0 TS 1.0 TL
Period, T (sec)

Figure 5.1. General response spectrum (FEMA 450, 2003)

For spectrum-based PBD assessment, each performance objective is associated with a

particular design earthquake. The corresponding base shear is adopted as the damage
target parameter, which represents the maximum base shear that a building can undergo
during that earthquake. The target base shear for a building is computed as a function of
the spectral acceleration (Sa), structural seismic weight (W), and gravity acceleration
constant (g) for each selected performance objective as follows: For each of multiple
performance objectives possible for a building (see Table 2.1), the base shear is computed
by the following expression:

W
VbPO = S aPO (5.1)
g

85
Figure 5.1 signifies a general response spectrum provided by FEMA 450 (2003), in which
the spectral acceleration is given as a function of the earthquake hazard associated with
the performance objective under consideration, and is defined by the following equations:
PO
S DS
S aPO = 0.60 PO
T + 0.4 S DS
PO
for T ≤ T0PO (5.2 a)
T0

S aPO = S DS
PO
for T0PO ≤ T ≤ TSPO (5. 2 b)

S DPO1
S aPO = for TSPO ≤ T ≤ TL (5. 2 c)
T

S DPO1 TL
S aPO = for T > TL (5. 2 d)
T2
where T is the fundamental period of the structure, which can be found by computing the
eigenvalues of the building or by using the simplified method described in Appendix A.
The parameter TL is the ground long-period transition period, provided in regional hazard
seismic maps by FEMA 450 (2003). Finally, the parameters TSPO and T0PO are the ground
characteristic period and a fraction of the ground characteristic period, respectively; both
are shown in Figure 5.1, and computed using the following expressions:

S DPO1
TSPO = PO
(5. 3)
S DS

T0PO = 0.2TSPO (5.4)

PO
where S DS and S DPO1 are the design acceleration parameters, expressed as,

2 PO
PO
S DS = S MS (5.5)
3

2 PO
S DPO1 = SM1 (5. 6)
3
PO
in which S MS and S MPO1 are the spectral response accelerations parameters for a short
period (0.2 second) and one second period, respectively, given by,
PO
S MS = FaPO S SPO (5. 7)

86
 S MPO1 = FvPO S1PO (5. 8)

where Fa and Fv are factors to adjust the spectral acceleration in the short-period range
and one- second period range, respectively, for the site class, and are found from Tables
5.2 and 5.3 (FEMA 450, 1997). The site class is defined as A, B, C, D, E, and F for hard
rock, rock, very dense soil and soft rock, stiff soil, soft soil, and soils requiring site-
specific evaluations, respectively. S SPO and S1PO are the site response acceleration
parameters corresponding to a short period and a one-second period, respectively, for a
structure with 5% damping. These parameters are obtained from regional seismic hazard
maps such as FEMA 450 (2003) and USGS (2002). However, these documents provide
response acceleration parameters for earthquakes having only 2% and 10% probability of
exceedance in a fifty-year period. The response acceleration parameters for earthquakes
with 20% and 50% probability of exceedance in a fifty-year period are computed by
using equations provided by FEMA 273 (1997). Table 5.1 provides sample values of S S

and S1 for two locations in California and Nevada, USA (Gong, 2003).

Once the target base shears for a building are computed, based on the selected
performance objectives, the distribution of the lateral loads applied on the building are
computed by the equations described in the next section.

Table 5.1. Performance objective site parameters, Gong (2003)

Site Performance Earthquake Ss S1

a
Location Level Level (g) (g)
Latitude OP 50% / 50 0.126 0.061
36.9°N, IO 20% / 50 0.209 0.100
Longitud LS 10% / 50 0.290 0.140
120°W CP 2% / 50 0.500 0.230
Latitude OP 50% / 50 0.109 0.035
41°N, IO 20% / 50 0.180 0.058
Longitud LS 10% / 50 0.250 0.080
115.2°W CP 2% / 50 1.100 0.410
a
Eccedance probability / years

87
 Table 5.2. Values of Fa as a function of the Site Class and Mapped Short-Period Maximum
considered earthquake spectral acceleration, FEMA 450 (2003)

Site Mapped Maximum Considered Earthquake Spectral Response

Class Acceleration at Short Periods
Ss ≤ 0.25 Ss = 0.50 Ss = 0.75 Ss = 1.00 Ss ≥ 1.25
A 0.8 0.8 0.8 0.8 0.8
B 1.0 1.0 1.0 1.0 1.0
C 1.2 1.2 1.1 1.0 1.0
D 1.6 1.4 1.2 1.1 1.0
E 2.5 1.7 1.2 0.9 0.9
F a a a a a
NOTE: Use linear interpolation for intermediate values of SS.
a
Site-specific geotechnical investigation and dynamic site response analyses shall be performed.

Table 5.3. Values of Fv as a Function of the Site Class and Mapped 1 Second Period Maximum
considered earthquake spectral acceleration, FEMA 450 (2003)

Site Mapped Maximum Considered Earthquake Spectral Response

Class Acceleration at Short Periods
Ss ≤ 0.10 Ss = 0.20 Ss = 0.30 Ss = 0.40 Ss ≥ 0.50
A 0.8 0.8 0.8 0.8 0.8
B 1.0 1.0 1.0 1.0 1.0
C 1.7 1.6 1.5 1.4 1.3
D 2.4 2.0 1.8 1.6 1.5
E 3.5 3.2 2.8 2.4 2.4
F a a a a A
NOTE: Use linear interpolation for the intermediate values of S1.
a
Site-specific geotechnical investigation and dynamic site response analyses shall be performed.

5.3. Spectrum-Based Pushover Analysis for CFS Buildings.

For the PBD assessment of CFS buildings, a pushover analysis is employed to carry out
the seismic analysis for reasons of practicality. In the spectrum-based pushover analysis,
adopted in this study, the building is subjected to incrementally applied lateral loads, until
the target base shear is reached or the structure collapses. Throughout this analysis

88
process, the gravity loads are maintained constant on the building. Upon applying a
lateral load increment on the structure, the structural analysis is carried out by using a
suitable method for the type of structure under consideration. In this study, the structural
analysis is performed by using the finite element method, presented in Chapter 4, with
account for both geometric and material nonlinearities of the elements. From the
structural analysis results, the incremental nodal displacements are obtained, and then the
accumulated total displacements up to the current stage of the loading process are found.

The maximum base shear computed for the selected performance objective by Eq. (5.1) is
expressed as,

[ ]
Vb max = max VbPO (5. 9)

The lateral load increments are applied on the building until Vbmax is reached or the
structure collapses. Each load increment is a small portion of Vbmax, and selecting the
appropriate size for the lateral load increment is important to obtain the correct analysis
results. If the size of the selected load increment is excessively small, the program used in
the calculations might not be able to handle the required precision, causing incorrect
results. On the contrary, if the size of the load increment is excessively large, the
response curve from the analysis can be in significant error relative to the correct curve.
In order to achieve correct results, the size of the load increment is varied depending on
the type of structure. For example, Grierson et al. (2005) have applied a 5% increment of
the total lateral loads for analysis of steel frameworks. To determine the appropriate size
of the load increment for CFS buildings, the OSB shear wall panel analyzed in Section
4.7, and presented in Figure 4.10, is analyzed by varying the size of the load increment to
be, 10%, 3.5%, 2%, 1% and 0.5% of the total seismic load. Figure 5.2 exhibits the
response curves for the different sizes of the load increments. It is noted that the
difference in the response for the load increments ranging from of 0.5% to 3.5% is not
significant. Moreover, the curves for the load increments of 1% and 0.5% are almost the
same. Therefore, in this thesis a 1% load increments are used for the pushover analysis.

89
 SFEA 1% SFEA 3.5% SFEA 2% SFEA 1% SFEA 0.5%

16

14

12

10
Lateral load (kN)

0
0 5 10 15 20 25 30 35 40 45 50
Lateral displacement (mm)

Figure 5.2. Sensitivity tests of the response curve for different sizes of lateral load increments.

The seismic loads are distributed over a building’s height with a pre-defined profile,
determined as,

Fx = CvxVb max (5. 10)

where Fx is the lateral seismic load to be applied on the building, at height x from the
building base. Cvx is the lateral load distribution coefficient which represents the lateral
load pattern (FEMA 273, 1997) such that,

wx hxκ
Cvx = nt
(5. 11)
∑w h
z =1
κ
z z

where wz and wx are portions of the total seismic weight of the structure, corresponding to
storey levels z and x, respectively; hz and hx are the height from the building base to
storey levels z and x, respectively; and nt is the number of storeys in the building. The
exponent κ is a function of the structure fundamental period T; for structures with T≤

90
0.5sec κ=1, and for structures with T≥ 2.5sec κ=2. For the intermediate values of T, a
linear interpolation is used to find κ. Note that the coefficients Cvx found from Eq. (5.11)
over all nt storeys satisfy the condition,
nt

∑C
z =1
vx , z =1 (5. 12)

In the PBD assessment of a CFS building, the acceptance criteria of the shear wall panels
(SWP) are checked when the base shear of the building is equal to the target base shear
for any selected performance objective, VbPO. Thus, if the structural elements satisfy the
deformation and strength requirements, the acceptance criteria for the specified
performance objective are satisfied.

5.4. Performance Acceptance Criteria for CFS Buildings

The acceptance criteria determine if a building is properly designed to satisfy the
demands imposed by the seismic hazards. The acceptance criteria limit the damage that
the elements in a building are allowed to undergo, and are expressed in terms of strength,
displacement, deformations, stresses, and so on. In PBD assessment, the acceptance
criteria are given as a function of the performance levels. If FEMA criteria (FEMA 273,
1997) are adopted, as herein, the damage in the building goes from low to high for the
OP, IO, LS and CP performance levels, respectively.

In CFS buildings, the SWP’s are the primary structural elements in resisting the lateral
loads. Consequently, acceptance criteria based on SWP lateral drift is established.
However, the strength of each SWP also needs to be checked to prevent premature failure
prior to reaching the limit drift. In some cases, the lateral drift of a SWP might be less
than the limit value but the applied lateral loads exceed the SWP strength, or vice versa.
Therefore, both the lateral drift and strength of the SWP need to be checked to determine
if the panel has been designed appropriately.

91
FEMA 273 (1997) provides acceptance criteria defined by “limit” drift ratios for different
types of lateral-load resisting structural systems such as steel frameworks, concrete walls,
masonry walls, and wooden stud walls. In Chapter 8 of FEMA 273 (1997), the
magnitudes of the limit drift ratios are associated with three performance levels for
several types of SWP. For instance, the limit drift ratios for SWP with plywood sheathing
and wood framing are 1.4%, 2.6%, and 3.0% for the IO, LS, and CP performance levels,
respectively.

Chapter 2 of FEMA 273 (1997) includes descriptions of the expected damages and
corresponding limit drift ratios for different structural systems corresponding to the OP,
IO, LS and CP performance levels. Typically, these limit drift ratios are adopted to
establish the target displacements, which are likely to be experienced by a structure in a
seismic event. These ratios have been used in different studies (Hassan et al., 2002; and
Filiatrault and Folz, 2002) to calculate the target displacements for displacement-based
pushover analysis. For SWP with wood framing, the limit drift ratios are 3%, 2% and 1%
for the CP, LS and IO performance levels, respectively (FEMA 273, 1997). However, no
such limit drift ratios, neither as acceptance criteria nor for estimating target
displacements, are provided for CFS SWP.

Although wood and CFS shear wall panels are built in a similar fashion, their force-
displacement curves are not similar, and so the limit drift ratios for wood-framed SWP
cannot be used for CFS-framed SWP. Although there is no evidence of studies in the
literature where the behaviour of SWP with wood and steel framing is compared. SEAOC
and COLA-UCI (2001) have carried out extensive testing of SWP with wood and steel
framing under the same testing conditions. Shown in Figure 5.3 is the superposition of
the cycling test curves of SWP with wood and steel framing with structural sheathing
(STR 1) and two different edge screw spacings. The curves show that the wood-framed
SWP has greater ductility and smaller strength, and therefore, that limit drift values for
wood SWP are not suitable for modelling CFS SWP. Instead, in this thesis, the limit drift
ratios for CFS shear walls are estimated from experimental data (Branston et al., 2006).

92
 Figure 5.3. Comparison of the behaviour of SWP with wood and CFS framing for different edge
screw spacing: (a) 152mm and (b) 51mm (SEAOC and COLA-UCI, 2001)

Figure 5.4. Normalized force versus deformation ratio for wood elements (Fig. 8-1, FEMA 273, 1997)

FEMA 273 (1997) establishes the relationship between performance levels using the
normalized bilinear force-deformation curve for SWP depicted in Figure 5.4. In addition,
FEMA 273 (1997) provides the equations for determining the normalized deformation
associated with the three performance levels CP, LS and IO. The equations are expressed
as a function of the ductility d of the SWP, which is the normalized lateral deformation of
the SWP at the CP performance level:

93
 ∆ CP ∆ y = d (5. 13)

∆ LS ∆ y = 1.0 + 0.8(d − 1) (5. 14)

∆ IO ∆ y = 1.0 + 0.2(d − 1) (5. 15)

where ∆PL is the deformation of the SWP at the selected performance level CP, LS or IO,
and ∆y is the deformation at yielding of the SWP. The equation for the normalized
deformation corresponding to the OP performance level is not provided by FEMA 273
(1997). For steel frameworks this deformation is generally computed at the onset of
yielding in the elements (Hasan et al., 2002). For SWP it has been reported that forty
percent of the yield displacement (0.40∆y) represents the service load level (Chen et al.,
2006). Accordingly, this study adopts forty percent of the yielding deformation as being
representative of the OP performance level for CFS SWP. The limit drift ratios for SWP
are given by the following equation:

(
LDR(%) = ∆ PL ∆ y ) ∆h
y
× 100 (5. 16)

where LDR is the limit drift ratio for the selected performance level, given as a
percentage; h is the height of the SWP; and (∆ PL ∆ y ) is the normalized deformation
evaluated using Eqs. (5.13) to (5.15). Table 5.4 presents the yielding deformation and
normalized deformation at the failure of the SWP, found from the experimental testing
conducted by Branston et al. (2006). With this information, the limit drift ratios are
evaluated for the four performance levels and presented in the last row of Table 5.4.

The limit drifts for the performance objectives need to be computed before the pushover
analysis of a CFS building is initiated. The limit drifts for a SWP are evaluated by
multiplying the SWP height by the corresponding performance level LDR (i.e., 2.5%,
2.1%, 1.0% and 0.20% for the CP, LS, IO and OP performance levels, respectively). If
the SWP lateral drifts computed by the pushover analysis are smaller than the
corresponding limit drifts, the performance criteria are satisfied.

94
 Table 5.4. Determination of the limit drift ratios for CFS shear wall panels

Branston et al (2006)
Disp. at Normalized displacement
d
yielding Drift ratio
Specimen
(∆y) (∆failure/∆y) (5.13) (5.14) (5.15)
mm CP LS IO CP LS IO OP
4 – A,B,C -16.2 3.8 3.83 3.26 1.57 2.54% 2.17% 1.04% 0.27%
6 – A,B,C -15.5 4.1 4.08 3.46 1.62 2.59% 2.20% 1.03% 0.25%
8 – A,B,C -12.3 4.7 4.72 3.98 1.74 2.38% 2.01% 0.88% 0.20%
10 – A,B,C -18.5 3.1 3.07 2.66 1.41 2.33% 2.02% 1.07% 0.30%
Average negative values

12 – A,B,C -11.2 5.1 5.13 4.30 1.83 2.36% 1.98% 0.84% 0.18%
14 – A,B,C,D -15.5 4.1 4.05 3.44 1.61 2.57% 2.19% 1.02% 0.25%
16 – A,B,C -15.8 4.4 4.44 3.75 1.69 2.88% 2.43% 1.09% 0.26%
18 – A,B,C -18.4 3.3 3.34 2.87 1.47 2.52% 2.17% 1.11% 0.30%
20 – A,B,C -12.8 6.3 6.33 5.26 2.07 3.32% 2.76% 1.08% 0.21%
22 – A,B,C -8.2 5.9 5.86 4.89 1.97 1.97% 1.64% 0.66% 0.13%
24 – A,B,C -8.8 4.6 4.62 3.90 1.72 1.67% 1.41% 0.62% 0.14%
26 – A,B,C -8.9 4.4 4.39 3.71 1.68 1.60% 1.36% 0.61% 0.15%
28 – A,B,C -15.9 4.7 4.68 3.94 1.74 3.05% 2.57% 1.13% 0.26%
30 – A,B,C -15.6 3.8 3.78 3.22 1.56 2.42% 2.06% 1.00% 0.26%
32 – A,B,C -16.9 4.0 4.00 3.40 1.60 2.77% 2.36% 1.11% 0.28%
34 – A,B,C,D -16.5 3.8 3.77 3.22 1.55 2.55% 2.18% 1.05% 0.27%
4 – A,B,C 14.9 4.9 4.85 4.08 1.77 2.96% 2.49% 1.08% 0.24%
6 – A,B,C 16.2 4.2 4.21 3.57 1.64 2.80% 2.37% 1.09% 0.27%
8 – A,B,C 10.1 6.4 6.41 5.33 2.08 2.66% 2.21% 0.86% 0.17%
10 – A,B,C 17.3 3.8 3.79 3.23 1.56 2.69% 2.29% 1.11% 0.28%
Average positive values

12 – A,B,C 13.1 5.0 5.02 4.22 1.80 2.70% 2.27% 0.97% 0.21%
14 – A,B,C,D 19.3 3.5 3.51 3.01 1.50 2.78% 2.38% 1.19% 0.32%
16 – A,B,C 22.6 2.8 2.79 2.43 1.36 2.59% 2.25% 1.26% 0.37%
18 – A,B,C 27.0 2.3 2.27 2.02 1.25 2.51% 2.23% 1.39% 0.44%
20 – A,B,C 16.1 4.5 4.54 3.83 1.71 3.00% 2.53% 1.13% 0.26%
22 – A,B,C 7.5 7.4 7.38 6.10 2.28 2.27% 1.88% 0.70% 0.12%
24 – A,B,C 8.1 5.4 5.38 4.50 1.88 1.79% 1.50% 0.62% 0.13%
26 – A,B,C 10.6 4.6 4.55 3.84 1.71 1.98% 1.67% 0.74% 0.17%
28 – A,B,C 15.7 3.9 3.90 3.32 1.58 2.51% 2.14% 1.02% 0.26%
30 – A,B,C 10.7 6.0 6.01 5.01 2.00 2.64% 2.20% 0.88% 0.18%
32 – A,B,C 15.7 4.3 4.34 3.67 1.67 2.79% 2.36% 1.07% 0.26%
34 – A,B,C,D 17.3 4.0 4.01 3.41 1.60 2.85% 2.42% 1.14% 0.28%
Average rounded to two decimals 2.53% 2.15% 0.99% 0.24%
Average rounded to one decimal 2.5% 2.1% 1.0% 0.2%
Standard deviation 0.004 0.003 0.002 0.001
Coefficient of variation 0.156 0.156 0.196 0.297

In addition to checking the limit drifts of the SWP, it is necessary to check their lateral
strength. To determine if the lateral strength of the SWP PR is adequate, the lateral
strength of the SWP is computed in accordance with Eq. (3.1) and compared to the lateral
forces Pa found from the analysis. Thus, Pa must be smaller than PR; otherwise, the
design is not adequate and the SWP needs to be redesigned to satisfy the strength

95
requirements. To design or redesign a SWP, the factors that most affect their lateral
strength must be considered so that the strength can be effectively increased. The
following are typical modifications that have a major impact on the lateral strength of
SWP: increasing the sheathing thickness, attaching sheathing on the two sides of the
SWP (if it is being used only on one side), reducing the spacing of the screws for the
sheathing-to-framing connections at the edge of the SWP, and increasing the diameter of
the screws. The modifications that are considered to have a minor impact on the lateral
strength of a SWP include: increasing the thickness and depth of the CFS studs, reducing
the spacing between studs, and reducing the spacing of the screws for the sheathing-to-
framing connections in the field of the SWP. The influence of these factors on the lateral
strength of a SWP have been determined from experimental tests (Branston et al., 2006;
Serrette, 2002), and confirmed by the method described in Chapter 3.

The vertical strength of the SWP is evaluated and compared with the vertical forces that
result from the gravity and lateral loads applied on the structure. Although the sheathing
can contribute to the vertical strength of the SWP, it is assumed that only the studs resist
the vertical forces. This assumption is widely adopted in current practice and design
standards (AISI, 2004). The studs’ axial strength is computed according to the standard
for design of CFS structural members (S136-01, 2002). The effective length coefficients
described in Section 3.4 that consider the lateral support of the sheathing are considered.

5.5. Procedure of PBD Assessment for CFS buildings

In this section, a step-by-step description of the procedure for carrying out the PBD
assessment of CFS buildings is provided to summarize the proposed methodology. Also
presented in Figure 5.5 is a flowchart illustrating the described procedure for conducting
pushover analysis for the PBD assessment of CFS buildings. The step numbers annotated
in the flowchart match the step numbers of the PBD assessment procedure described in
the following.

96
1. Establish the CFS building information. Specifically, create the preliminary
design of the building, which may be obtained from a design derived for
gravitational loads only. Determine the cross-sectional properties of the steel studs
and material properties of the structural sheathing. Also, establish the following
seismic design information:
a. Select the performance objectives to be satisfied by the building from
Table 2.1. Custom performance objectives can be formed by combining
the expected performance level for a building with seismic hazards.
b. Establish the seismic parameters of the site, such as response acceleration
parameters Ss and S1, site coefficients Fa and Fv, and long-period transition
period TL. These values are available in Tables 5.1, 5.2 and 5.3 (FEMA
450, 2003; USGS, 2002).

2. Determine the applicable gravity loads, Wg, as the summed magnitude of the dead
and live loads: the dead loads are computed from the self-weight of the elements
and components in the building; the live loads are computed from the applicable
code, such as the National Building Code of Canada (NBCC, 2005), or the
ASCE/SEI 7 (2005).

3. Evaluate the lateral strength of all the SWP for the building, PR, in accordance
with the analytical method presented in Chapter 3, using the properties of the
SWP established in Step 1.

4. Compute the fundamental period of the structure, T, by employing the procedure

described in Appendix A. Alternatively, the period can be computed by applicable
eigenvalue analysis.

5. Compute the target base shear, VbPO , for each performance objective: based on the
seismic parameters in Step 1, first calculate the seismic spectral acceleration using
Eq. (5.2) for each performance objective; then compute the target base shear for
each performance objective using Eq. (5.1).

97
6. Evaluate the maximum target base shear, Vb max , using Eq. (5.7).

7. Determine the lateral-load distribution coefficients, C vx , using Eq. (5.9); then,

compute the lateral force, Fx , in each storey using Eq. (5.10).

8. Determine the limit drift of each SWP for each performance objective, DriftPO, by
multiplying the height of each SWP by the drift ratio of the performance level
under consideration (i.e., by 2.5%, 2.1%, 1.0% and 0.6% for the CP, LS, IO and
OP performance levels, respectively).

9. Initially set the lateral strength reduction coefficient λ = 1, and the load increment
index q = 1.

10. Determine the constitutive properties, Exeq, Eyeq, and Gxyeq, of the equivalent shell
elements in accordance with Eqs. (4.10), (4.14), and (4.18), respectively.

11. Form the linear and nonlinear stiffness matrices for all elements, by using Eqs.
(4.41) and (4.42), respectively. Then, form the linear and nonlinear structural
stiffness matrices by using Eqs. (4.44) and (4.45), respectively.

12. If q=1, apply the totality of the gravity loads on the building. Otherwise, go to
step 13

13. Apply a lateral load increment on each floor, ∆Fx i , which is taken by this study to
be 1% (i.e., ∆=0.01) of the lateral load found from Step 7 (the magnitude of the
load increment remains constant throughout the pushover analysis).

14. Determine the nodal displacements and reactions of the building by solving Eq.
(4.46) or Eq. (4.48).

98
15. Calculate the building base shear at load increment q, Va(q )
The base shear is computed by summing the lateral load increments applied on the
building.

16. If the accumulated base shear is equal to the target base shear of any performance
objective, or greater (i.e., Va( q ) − VbPO < ∆Fx i ), go to step 17. Otherwise, go to step
20.

17. Determine the SWP lateral-drift, using the nodal displacements from step 14.
Verify if the SWP satisfy the acceptance criteria, by comparing the lateral drift of
the SWP (i.e., inter-storey drift) to the limit drift for the performance objective
being verified.

18. Determine the SWP lateral force, using the nodal displacements and element
stiffness matrices. Verify if the SWP lateral strength satisfy the acceptance
criteria, check if the lateral forces in the SWP have not exceeded their lateral
strength.

19. If the building base shear is equal or greater than the maximum base shear, the
analysis is complete. Otherwise, go to step 20.

20. Update the stiffness degradation coefficient λ, using Eq. (4.52) to reflect the SWP
loss of stiffness due to the lateral forces on the CFS building. Update the
equivalent shell elements constitutive matrix C, using Eq. (4.43). Increase the
load increment index q. Proceed to Step 11.

99
 CFS Building information:
Preliminary Design, and building
location parameters, Performance
Objectives (PO),
1
Apply a lateral load increment:
13 ∆Fx i = 0.01Fx i ,
2 Determine the applicable
gravitational loads, Wg i = Story

3 Evaluate the lateral strength of 14 Solve the structural stiffness

the SWP, PR matrices, and determine the
nodal displacements

4 Determine the fundamental

period of the structure, T 15 Calculate the building base
shear at the load increment q:
Stories
Va( q ) = q ∑ ∆F xi
5 Calculate the seismic base i =1
W
shears: VbPO = S aPO
g
PO = OP, IO, LS, CP. 16
No
Va(q) ≥ VbPO
6 Vb max = max V [ ]b
PO

Yes
7 Determine the vertical
distribution load coefficents, 17
Cvx, and lateral forces: Determine DriftSWP,
Fx i = C vx iVb max , i = story and compare it to
DriftPO

8 Calculate the SWP limit drifts 18

ratios for each PO, DriftPO Determine Pa, and
compare it to PR

9 Initiate the lateral strength

reduction coefficient and load
increment λ =1.0 , and q = 1
20 19
Update λ, C, and q = Va( q ) ≥ Vb max
PO

10 Transform the SWP into q+1 No

Equivalent Shell elements

Yes
12
Form structure Yes End: Building
11 q =1 Apply the totality
stiffness matrix,
of the vertical PBD assessment
K = KL + KNL No load, Wg completed

Figure 5.5. Flowchart for the PBD assessment of CFS buildings

100
5.6. Computer Program for the PBD assessment of CFS Buildings
As part of this study, a computer program has been developed to conduct the PBD
assessment of three-dimensional CFS buildings. The program is developed on the basis
the Structural Optimization and Design Analysis software (SODA, 1999), originally
developed at the University of Waterloo. The software, SODA, has the capability to
conduct the structural analysis and design for three-dimensional structural steel
frameworks. In this study, new modules to carry out the PBD assessment of CFS
buildings have been implemented. Thus, the resulting program is capable of conducting
pushover analysis and the PBD assessment for CFS buildings that can be modeled using
shell and frame elements as that demonstrated in Chapter 6.

To conduct the PBD assessment of CFS buildings using the computer program, the input
file must contain the geometry information of the building such as nodal coordinates,
member connectivity, boundary conditions, and material and cross-sectional properties of
structural frame elements and panels. In addition, response acceleration parameters of the
site and applicable loads are required. Based on the foregoing input information, the
computer program calculates the lateral strength of the SWP, the fundamental period of
vibration of the building, limit lateral drifts for the SWP, and the target base shears for
the performance objectives. The program also determines the constitutive properties of
the equivalent shell elements used to model the SWP. The program conducts the
pushover analysis of the building. The SWP inter-storey drifts and lateral forces are
computed at each performance objective, and compared to the limit inter-storey drifts and
SWP strength, to determine if the SWP satisfy the acceptance criteria.

The output files generated by the computer program contain the following information:
• Weight and the fundamental period of the building
• Target base shear for each performance objective
• Frame-element end forces
• Nodal displacements
• Support reactions
• SWP lateral strength

101
• SWP lateral force for each performance objective
• Studs axial strength
• Studs axial force for each performance objective
• Limit inter-storey drift for each performance objective
• SWP inter-storey drift for each performance objective
• Stiffness degradation factor for each performance objective

102
 Chapter 6
Examples

6.1. Introduction
The principal objective of this chapter is to present examples for the PBD assessment of
CFS buildings using the spectrum-based pushover analysis procedure described in
Section 5.5. Two different building models are created, and linear and pushover analyses
are carried out for each model by using simplified finite element analysis (SFEA). The
results from the linear and the pushover analyses are compared to determine the
differences. Moreover, the results from the linear analyses are compared to the results
from conventional finite element analysis (CFEA). The CFEA of the CFS buildings are
conducted by using SAP2000 (2006). The purpose of the comparison between linear
SFEA and linear CFEA is to determine the extent of the simplifications in SFEA that
affect the accuracy of the results. The results from the pushover analyses are not
compared, because to the author’s knowledge there is no commercial software which is
capable of carrying out the PBD assessment of CFS buildings.

In addition to presenting the practical application of the methodology developed in this

study for the PBD assessment of CFS buildings, the examples serve to illustrate the
nonlinear response of CFS buildings (as predicted by SFEA). Consequently, problems
that may arise when designing a CFS building are identified, and recommendations can
be made to avoid or reduce such problems.

6.2. Comparison of Linear SFEA and CFEA

The computer program developed in this study for carrying out the SFEA of CFS
buildings employs the equivalent shell element formulation, described in Chapter 4. The
modeling of CFS buildings using SFEA involves transforming all the panels in the

103
building, including those for walls, floors and roof, into equivalent shell elements. Each
equivalent shell element is modeled with a sixteen-node shell element. Based on the
comparison of the results from SFEA and CFEA presented in Section 4.4, it is
recommended to use shell elements with a length-to-height ratio between 0.5 and 3.0.
Otherwise, more than one shell element should be used to model the panels.

To transform the panels into flat shell elements of equivalent properties, the sheathing
and studs are accounted for in the calculations, but not the top and bottom tracks. The
reason for this is that the tracks do not provide lateral strength to the SWP, their primary
function being to distribute the loads on the studs and maintain the integrity of the panels.
However, if the tracks need to be included in the building model, they should be added as
frame elements. As described in Section 5.6, the computer program created for this study
has the capability to analyze frame elements in combination with shell elements.
However, the material nonlinear behaviour of frame elements is not accounted for. For
gravity wall panels, only the properties of the studs are accounted for when determining
the equivalent properties of the shell elements. The floor and roof panels are transformed
into equivalent shell elements using the same equations as that for SWP, which are
presented in Section 4.2. If a floor or roof is built as a flat slab without joists, the
constitutive properties of the slab and equivalent shell element are equal.
For the linear CFEA of CFS buildings, the studs, tracks and joists are modeled with frame
elements, while the sheathing is modeled with meshes of shell elements.

6.2.1. Example 1: Analysis of an Isolated SWP Using SFEA and CFEA

The objective of analyzing the isolated SWP in Figure 6.1 is to determine how the two
proposed simplifications of the SFEA affect the accuracy of the results. The two
simplifications involve 1) the transformation of the SWP into an equivalent shell element,
and 2) the modeling of the panel using one sixteen-node shell element.

104
 Figure 6.1. SWP model for CFEA

The SWP is built with 152 mm (6 in) depth studs (i.e., 152S51-1.37mm), and 11.11 mm
(7/16in) OSB sheathing. The SWP, in Figure 6.1, has fixed support. CFEA is carried out
using both SAP2000 (2006) and the computer program mentioned in Section 5.6; the two
analyses are called CFEA and CFEA1, respectively. A total of 57 frame elements are
included in each model to represent the studs and top track. In the CFEA model a total of
108 four-node shell elements are employed to model the sheathing, whereas in the
CFEA1 model only 12 sixteen-node shell elements are employed to model the sheathing.
A larger number of four-node shell elements is employed in the model because, usually,
the sixteen-node shell elements offer more accuracy due their higher order shape
functions. In fact, both models have the same number of nodes so that similar results can
be achieved with both models.

The SFEA analyses are carried out in two different ways: for the SFEA1 analysis the
SWP is transformed into equivalent shell elements modeled with 12 sixteen-node shell
elements and the top track is modeled with 12 frame elements; for the SFEA2 analysis
the SWP is modeled with one sixteen-node element, and three frame elements. The
numbers of elements for each model, as well as their size, are presented in Table 6.1. For
all cases, a 1.0 kN/m2 out-of-plane load is applied on the surface of the panels. In-plane

105
nodal forces of magnitude of 1 kN and 0.1 kN in the x and y direction, respectively, are
applied on nodes 1 and 2. Although in reality it is very unlikely to have a situation in
which a structure is subjected to combined wind and earthquake loads, in this example
both lateral and out-plane loads are applied on the SWP to compare the results obtained
from different types of analysis.

Table 6.1. Quantity of elements

Quantity of elements Dimension (mm)

Analysis
Frame Shell Frame
Model Shell
Track Stud width height Track Stud
CFEA 108 12 45 203 271 203 271
CFEA1 12 12 45 609 813 203 271
SFEA1 12 12 0 609 813 203 0
SFEA2 1 3 0 2436 2439 609 0

Presented in Table 6.2 are the results obtained from the four SWP analyses for the
displacements at node 1, the end-stud axial forces and the maximum bending moment in
the centre stud. The CFEA results are taken as the benchmark to compare the other three
analyses results. Since CFEA1 is carried out in the same fashion as CFEA, the results are
similar. The displacements in the x and y directions from SFEA1 are larger than those
from CFEA, but the axial forces are similar. Also, the displacements and axial forces
from SFEA2 are larger than those from CFEA. The difference in the displacements from
SFEA1 and SFEA2 is small, but the difference for the stud axial forces is relatively large.

Table 6.2. Comparison of results for Example 1

Displacements
Analysis Forces in end-stud
(Node 1), mm
Model
x y z Comp, kN Tension, kN Moment, kN-mm
CFEA 2.24 0.21 8.90 13.50 12.51 1526.43
CFEA1 2.28 0.22 8.95 13.72 12.71 1530.01

106
 SFEA1 2.47 0.27 5.89 14.41 14.20 1859.28
SFEA2 2.58 0.39 6.12 18.81 17.59 1741.04
According to the results in Table 6.2, the use of equivalent shell element(s) decreases the
accuracy of the x and y displacements and the bending moments, and also the accuracy of
the axial forces in a minor way. On the contrary, the displacements are less sensitive to
the number of equivalent shell elements in the model, while the axial forces are more
sensitive to the number of elements. As mentioned in Section 4.6, the difference in the
displacements in the z direction is caused by the SAP2000 model, which models the
sheathing at the mid-plane of the studs.

6.2.2. Example 2: Three-storey Building

Shown in Figure 6.2 is a three-storey CFS building. The typical floor plan of the building
is shown in Figure 6.3. All the SWP in the building are built with cold-formed studs
152S51–1.73 mm (600S200-68 mils), sheathed on both sides with 12.5 mm (1/2 in) DFP.
The studs in SWP 2 are spaced 650 mm on centre, while the studs in the rest of the panels
are spaced 600 mm on centre. Although, the stud spacing selected for this example is not
common in practice, is selected for reasons of practicality. In all the SWP a single end-
stud is used. No. 8 screws are used to attach the sheathing to the framing at 102 mm and
305 mm spacing on the edge and in the field of the SWP, respectively. The size and
spacing of the screws are included to evaluate the lateral strength of the SWP in
accordance with the procedure described in Chapter 3. All the SWP on the first storey
have pin supports. The floor panels consist of a concrete slab of 127 mm thickness,
supported by 254 mm (10 in) deep CFS joist, designated 254S64-1.37 mm (1000S250-54
mils). The height for all the storeys is 2850 mm, and the pitched roof has 17% slope. For
this example, the structural details of the roof panels are identical to the details of the
floor panels. Listed in Table 6.3 are the properties of the sheathing, studs, and floor and
roof joists used in the building. In addition, the axial strength of the studs is presented in
the table, which is computed in accordance with S136-02 (2002).

107
 Figure 6.2. Three-storey CFS building

Figure 6.3. Typical plan and element identification of the three-storey CFS building

108
A CFEA of the CFS building is conducted using SAP2000 (2006), where the sheathing is
modeled using shell elements that are 600 or 650 mm in length by 950 mm in height, and
the studs are modeled with C-section frame elements that are 950 mm in length. Thus, a
total of 1232 four-node shell elements and 1647 frame elements are used to model the
three-storey CFS building. The sheathing and framing elements are joined at their mid-
plane so that the offset between the shell elements and studs’ flange is neglected. In the
SFEA model of the CFS building, each SWP is modeled by an equivalent shell element,
while each floor and roof panel is modeled with six equivalent shell elements. The floor
and roof transversal joists in Figure 6.2 are modeled with three frame elements each.
Therefore, a total of 51 sixteen-node shell elements and 72 frame elements are used to
model the building. Table 6.4 records the properties for the equivalent shell elements for
the floor and roof panels in the SFEA model, evaluated in accordance with Section 4.2.

Table 6.3. Properties of the sheathing and CFS framing material

Properties Sheathing DFP Properties Stud Joist

Ex 1167 E 203000
Ey (MPa) 10445 G (MPa) 77000
G 825 Fy 345
t (mm) 12.7 t (mm) 1.73 1.37
νx 0.23 A (mm ) 2
472 515
νy 0.23 I (mm4) 1.63x106 4.5x106
Screw (mm) 4.06 Pn 60.03 Not
(kN)
Tn 97.38 applicable

The CFS building is subjected to both gravitational and lateral loads. The magnitudes of
the factored gravity loads (Dead + Live load) are 3.24 kN/m2 and 1.32 kN/m2 for the
floors and roof, respectively. The factored gravity loads are uniformly applied on the
floor and roof elements, whereas the lateral seismic loads are applied on the top of the
SWP along the x direction, as shown in Figure 6.2. The lateral seismic loads are
computed by assuming that the building is located in California at latitude 41.0° N and
longitude 115.2°W; the detailed calculations of the seismic loads are shown in Example 4

109
(in Section 6.3.1). The live loads and the factored load combinations are taken from
ASCE/SEI 7 (2005) for a housing building.

Shown in Table 6.5 are the displacements obtained from SFEA and CFEA for linear
analysis. For comparison, only the displacements of four joints (N1 to N4) are presented.
Figure 6.3 illustrates the locations of the selected joints N1 to N4. The displacement of
each joint is given in the x, y and z directions. The lateral inter-storey drift computed at
joints N1 to N4 is also indicated in Table 6.5.

Table 6.4. Constitutive properties of the equivalent shell elements

SWP Floor
Exeq Eyeq Geq teq Exeq Eyeq Geq teq
Panel
MPa MPa MPa mm MPa MPa MPa mm
Storeys 1 and 2
1, 3 159.03 1482.85 52.69 195.72
2 153.85 1432.09 52.11 197.89
5054.47 5057.21 1986.36 652.66
7, 9 159.03 1482.85 52.69 195.72
5, 6 167.19 1563.08 53.58 192.48
Storey 3
1, 3 155.41 1447.33 52.29 197.22 5041.63 5044.27 1984.67 653.21
2, 4, 5, 6 Same as for storeys 1 and 2

The displacements in the x direction predicted from SFEA and CFEA are in good
agreement. The displacements along the y direction are small, since no loads are applied
in this direction. Conversely, the difference in the displacements of the floor panels in the
z direction at joints N2 and N4 is large. The reason is that in the modeling with CFEA,
the studs and joists are modeled at the mid-plane of the shell elements. This yields a
smaller value of the moment of inertia of the combined cross-section, and consequently,
the resulting deflections are larger. The SFEA accounts for the offset of the sheathing
from the centroid of the studs.

110
The lateral strengths PR of the SWP in Table 6.6 are evaluated with the procedure
described in Chapter 3. The table also lists the lateral force Pa in each SWP that is
obtained from SFEA and CFEA. In SFEA, the lateral forces in the SWP are predicted by
adding up the lateral forces at the nodes of the SWP, which are computed by multiplying
the stiffness matrices of the shell elements by the nodal displacements of the building. In
SAP2000, it is not possible to obtain the forces in the nodes; instead the lateral forces are
computed using the average shear force in the shell elements obtained from the CFEA.
The average shear force is multiplied by the length of the SWP to determine the lateral
force in the SWP.

Table 6.5. Displacement and lateral drift at selected joints, mm

SFEA CFEA
Joint X Y Z Drift X Y Z Drift
Storey 1
N1 5.94 -0.04 1.25 5.94 6.01 0.00 1.59 6.01
N2 5.96 -0.02 -0.45 5.96 5.99 0.00 -0.90 5.99
N3 6.01 0.00 -0.79 6.01 5.99 0.00 -1.33 5.99
N4 5.96 0.03 -0.69 5.96 6.00 0.00 -0.94 6.00
Storey 2
N1 11.78 -0.05 1.67 5.84 12.08 0.01 2.08 6.07
N2 11.80 -0.02 -0.63 5.84 12.04 0.00 -1.22 6.05
N3 11.87 0.00 -1.12 5.86 12.02 0.00 -1.85 6.03
N4 11.80 0.05 -1.07 5.84 12.03 0.00 -1.28 6.03
Storey 3
N1 16.60 -0.09 1.66 4.82 17.18 0.03 2.08 5.10
N2 17.13 -0.07 -0.59 5.33 17.88 0.00 -1.03 5.84
N3 17.34 -0.05 -1.15 5.47 18.09 0.00 -1.92 6.07
N4 17.24 0.01 -1.18 5.44 18.29 0.00 -1.15 6.26

The lateral forces listed on Table 6.6 are obtained from the SFEA and CFEA, and are
computed by multiplying the average lateral force on the SWP by its length. Although
this approach is not exact it provides acceptable results, and integration of forces over the
SWP area is not required. The results in Table 6.6 indicate that the force in each SWP is

111
less than the SWP strength; That is, the seismic design of the SWP is appropriate for
resisting the lateral loads since none of the SWP fail. The lateral forces in SWP 1 and 3
are larger on the first storey than on the top storey, because the forces are transmitted
from top to bottom. On the contrary, SWP 4 and 5 on the third storey have the largest
force while the SWP on the first storey have the smallest force. Although no lateral forces
are applied on SWP 4 and 5, lateral deformation is induced in the SWP by the flexural
deformation of roof and floor panels under the gravity loads, as shown in Figure 6.4. As a
result, small lateral forces are found in SWP 4 and 5. No shear deformation is produced
in SWP 2 since it is located in the center bay of the building. Therefore, the lateral force
in SWP 2 is zero. In general, all the predictions of the lateral forces in the SWP are in
good agreement with the results obtained from CFEA. The lateral strength for SWP 1, 3,
7 and 9 is larger for the SWP in the third storey than for the SWP in the first and second
storeys because the average height of the SWP in the third storey is larger.

Table 6.6. SWP lateral strength PR and lateral force Pa, kN

SWP Storey 1 Storey 2 Storey 3

PR Pa PR Pa PR Pa
SFEA CFEA SFEA CFEA SFEA CFEA
1, 3 212.80 177.44 160.80 212.80 146.59 122.40 226.75 88.79 84.12
7, 9 212.80 177.60 163.20 212.80 149.66 120.00 226.75 94.50 87.12
4,5 128.21 1.99 0.48 128.21 3.72 1.92 128.21 4.72 3.36

Presented in Table 6.7 is the maximum compression force in the most critical stud of
each SWP. For all the SWP the most critical stud is an end stud: for SWP 1, 3, 7, 9, 4 and
5 the end-stud located towards the centre of the building is critical; for SWP 2 both end-
studs have the same maximum compression force. The axial forces in the studs are
estimated using Eq. (4.50) after the analysis of the building is carried out.

The compression forces predicted from SFEA and CFEA are presented in Table 6.7, and
demonstrate a large difference for SWP 1 and 3 in the third storey for the example
building. The size of the shell elements also affect the accuracy, as described in Example
1. Since only one shell element is used to model each SWP, the SWP 1, 3 and 9 are each

112
4800 mm in length and 2850 mm in height; ( i.e., a length-to-height ratio = 1.68).
Moreover, the predicted compression forces in the end-studs for SWP 1, 3, 7 and 9 are
larger than the studs’ strength, so those studs need to be reinforced by using double end-
studs.

Figure 6.4 Building deformation due to gravity loads

Table 6.7. Compression force in the most critical stud of each SWP, kN

SFEA CFEA
SWP
Storey 1 Storey 2 Storey 3 Storey 1 Storey 2 Storey 3
1, 3 64.27 41.16 19.36 61.13 30.67 12.99
2 0.00 0.00 3.09 0.00 0.00 8.28
4, 5 3.65 2.71 1.55 5.33 3.62 1.50
7, 9 65.94 43.53 13.82 63.64 34.69 14.96
Load-Bearing Wall
8, 10 28.94 18.31 10.88 39.31 18.91 2.79
6, 11 27.31 17.40 13.82 31.01 16.09 4.33

Table 6.8 presents the maximum tension force in the most critical stud of each SWP. The
difference is small in the predicted forces in storeys 1 and 2 for SWP 1 and 3 from both

113
analysis approaches. However, the difference is large for the SWP in storey 3. The studs
in SWP 4, 5, 6, 8, 10 and 11 do not have a tension force.

Table 6.8. Tension force in the most critical stud of each SWP, kN

SFEA CFEA
SWP
Storey 1 Storey 2 Storey 3 Storey 1 Storey 2 Storey 3
1, 3 89.14 46.12 28.14 65.61 26.36 5.43
2 8.45 4.92 1.13 15.87 9.28 2.31
7, 9 72.96 41.37 12.89 37.61 13.79 1.87

When the SFEA and CFEA results are compared, the differences in the displacements in
the z direction and the axial forces in the studs are large. However, the purpose of the
SFEA is to facilitate the PBD assessment of CFS buildings. In this context, the SWP
lateral displacements and lateral forces are employed as acceptance criteria. Since these
quantities are in good agreement with both approaches, the SFEA can be employed with
confidence for carrying out the preliminary performance-based design of this CFS
building.

6.2.3. Example 3: Five-Storey Building

The building in Figure 6.5 is analyzed using SFEA and CFEA for the loading indicated in
the x direction, and the results from both methods are compared. The dimensions of the
building typical plan and the identification of the panels are illustrated in Figure 6.6. The
storey height is 2850 mm for each of the five storeys. The primary load-carrying elements
in the building are the load-bearing panels and SWP, built with studs 152S41–1.73 mm
(600S162-68 mils) spaced at 609 mm on centre. The SWP 1, 5, 13, and 17 are covered
with 11.11 mm OSB sheathing on the two sides, and SWP 3, 6, 12, 15, 19, and 20 are
covered with the same type of sheathing on one side. The sheathing is attached with No.
8 screws at 152 mm and 305 mm spacing on the edge and in the field of the SWP,
respectively. The rest of the panels are load-bearing panels, with no sheathing. All the
SWP on the first floor are pin-connected to the foundation. In addition, the panels that

114
intersect at a ninety-degree angle (e.g., panels 3 and 27 in Figure 6.6) are joined at the top
and bottom nodes, but not at intermediate nodes. Therefore, the vertical and lateral forces
that those panels receive come from the floor. The floor consists of a 127-mm thick
concrete slab, supported by the load-bearing panels and SWP. This example CFS
building has several differences compared to the building in Example 2: different
geometry and size, different material properties, SWP with sheathing on the two sides of
the framing, and gravity wall panels are used. Listed in Table 6.9 are the properties of the
sheathing, studs and joist for the analysis. The compression and tension strengths of the
studs, Pn and Tn, respectively, are computed based on S136-02 (2002), and presented in
the last two rows of the table.

Figure 6.5. Five-storey CFS building

115
 Figure 6.6. Typical plan and element identification of the five-storey CFS building

The CFEA of the CFS building is carried out using SAP2000 (2006); the sheathing is
modeled using 600 mm × 950 mm shell elements, and the studs are modeled with 640-
mm long C-section frame elements. A total of 3480 four-node shell elements and 2580
frame elements are employed to model the CFS building. For the SFEA of the building,
each one of the 34 wall panels and SWP are transformed into a single shell element, and
the floors and roof are divided into 18 shell elements: 12 of 3000 mm × 4000 mm side
length, and 6 of 3000 mm × 1500 mm side length. Thus, the CFS building is modeled
with a total of 260 equivalent shell elements. Listed in Table 6.10 are the constitutive
properties of the equivalent shell elements in the SFEA, evaluated from the formulation
in Section 4.2. The parameters Exeq and Gxyeq for the wall panels are equal to zero since
these panels have no sheathing.

Table 6.9. Sheathing and CFS framing material properties

Properties OSB Sheathing Properties Studs Joists

Ex (MPa) 1983 E (MPa) 203000
Ey 9917 G 77000

116
 G 925 Fy 345
t (mm) 11.11 t (mm) 1.73 1.37
νx 0.23 A (mm2) 425 479
νy 0.23 I (mm4) 1.39x106 3.91x106
Screw (∅) (mm) 4.06 Pn 63.77 Not
(kN)
Tn 87.73 applicable

Table 6.10 Constitutive properties of the equivalent shell elements

Exeq Eyeq Gxyeq teq

Panels
MPa MPa MPa mm
Shear wall
407.65 2095.43 99.92 205.70
1, 5, 13, 17
Shear wall
312.00 1623.21 57.58 178.49
3, 6, 12, 15, 19, 20
Load-Bearing
2, 4, 8, 14, 16 18, 21, 0.00 4421.06 0.00 41.13
22, 23
Load-Bearing
7, 9, 10, 11, 23, 24, 25, 0.00 4584.81 0.00 41.13
26
Floors and Roof 24500 24500 10500 127.00

The CFS building is subjected to both gravitational and lateral loads: the magnitudes of
the factored gravity loads (Dead + Live loads) are 4.00 kN/m2 and 1.60 kN/m2 for the
intermediate floors and the roof, respectively. The factored gravity loads are uniformly
applied on the floor and roof elements, whereas the lateral seismic loads are applied as
point loads on the top of the SWP, as depicted in Figure 6.5. The seismic load applied on
each frame is distributed among the SWP (e.g., the load of 104.58 kN applied on the fifth
storey is equally distributed to SWP 1 and 13). The lateral seismic loads are computed by
assuming that the building is located in California at latitude 36.9°N and longitude
120°W (the detailed calculations of the seismic loads are shown in Example 5 in Section
6.3.2). The live loads and load combinations are determined from ASCE/SEI 7 (2005) for
an office building.

117
Table 6.11 presents the displacements found from the SFEA and CFEA linear analyses of
the CFS building. For comparison, only the x, y and z displacements of the four joints N1
to N4 shown in Figure 6.6 are listed. Also given in Table 6.11 are the inter-storey drifts at
joints N1 to N4.

It is noted in Table 6.11 that the displacements in the x and z directions are slightly larger
by SFEA, and the displacements in the y direction tend to zero since no forces are applied
in that direction. The y displacements at joints N1 and N4 are positive because of the
deformation pattern of the building. Shown in (b)
Figure 6.7(a) is the building deformation in the fifth storey caused by the lateral loads, the
storeys below follow the same deformation pattern except the first storey which has no
lateral deformation. The difference in the deformation in the z direction is caused by the
small number of shell elements employed to model the floor.

Table 6.11. Displacement and lateral drift at selected joints (mm)

SFEA CFEA
Joint X Y Z Drift X Y Z Drift
Storey 1
N1 9.76 0.00 3.54 9.76 9.35 0.00 3.14 9.35
N2 9.76 0.00 -0.80 9.76 9.34 0.00 -0.79 9.34
N3 9.76 0.00 -1.05 9.76 9.34 0.00 -0.62 9.34
N4 9.77 0.00 -4.14 9.77 9.35 0.00 -3.44 9.35
Storey 2
N1 23.08 0.01 5.47 13.32 21.87 0.01 4.87 12.52
N2 23.05 0.00 -1.53 13.29 21.84 0.00 -1.37 12.50
N3 23.05 0.00 -1.83 13.29 21.84 0.00 -1.13 12.50
N4 23.08 0.01 -6.52 13.31 21.87 0.01 -5.44 12.52
Storey 3
N1 37.71 0.01 6.40 14.63 35.44 0.01 5.69 13.57
N2 37.69 0.00 -2.09 14.64 35.41 0.00 -1.95 13.57
N3 37.69 0.00 -2.34 14.64 35.41 0.00 -1.50 13.57
N4 37.72 0.01 -7.78 14.64 35.45 0.02 -6.47 13.58
Joint Storey 4

118
 N1 52.18 0.02 6.64 14.47 48.67 0.02 5.88 13.23
N2 52.13 0.00 -2.45 14.44 48.62 0.00 -2.28 13.21
N3 52.13 0.00 -2.62 14.44 48.62 0.00 -1.76 13.21
N4 52.18 0.02 -8.21 14.46 48.67 0.02 -6.79 13.22
Storey 5
N1 64.86 0.04 6.57 12.68 60.04 0.04 5.79 11.37
N2 64.78 0.00 -2.64 12.65 59.96 0.00 -2.44 11.34
N3 64.78 0.00 -2.67 12.65 59.95 0.00 -1.89 11.33
N4 64.86 0.04 -8.21 12.68 60.03 0.04 -6.74 11.36

(a)

(b)
Figure 6.7 a) Fifth storey displacement in the y direction, b) Building inter-storey drift

119
The lateral loads accumulate from top to bottom. Thus the lateral load in the first storey is
larger than in the storeys above, the lateral load in the second storey is larger than in the
storeys above, and so on. However, the inter-storey drifts listed in Table 6.11 do not
follow the same tendency, since the largest inter-storey drift is obtained for the third
storey. The reason is that SWP undergo shear and bending deformation, and for this
building the bending deformation is significant. Shown in Figure 6.1(b)
Figure 6.7 is the deformed shape of the building, and the inter-storey drift for SWP 1.

Table 6.12 presents the lateral strengths of the SWP, and their lateral forces predicted
from SFEA and CFEA. Upon comparing the lateral strengths and forces, it is observed
that the SWP in storeys 1 and 2 are not properly designed to resist the seismic forces
since the lateral forces in the SWP are larger than their strengths.

From Table 6.12 it is noted that SWP 1, 5, 13 and 17 have the largest lateral forces, with
the largest being on the SWP in the first storey. The small lateral forces in SWP 3, 6, 12,
15, 19, and 20 are caused by the deformation of the building, as illustrated in Figure 6.4
for Example 2. The lateral forces in the SWP predicted by SFEA and CFEA, and
described in Table 6.12, are in good agreement.

Table 6.12. SWP lateral strength PR and lateral force Pa on the, kN

Pa
SFEA CFEA
SWP PR
Storey Storey
1 2 3 4 5 1 2 3 4 5
1, 5 101.13 140.41 126.87 101.13 82.69 43.03 133.71 112.74 93.59 69.01 40.62
13, 17 101.13 140.54 127.74 108.65 85.31 46.00 136.38 115.70 99.08 71.73 44.34
3, 6 48.85 4.31 10.54 12.96 13.06 12.01 2.46 7.32 9.98 10.53 10.29
12, 15 48.85 6.39 16.60 22.45 25.00 25.42 4.74 14.55 21.15 24.72 26.64
19, 20 48.85 1.87 4.98 7.00 8.11 8.52 0.61 2.49 3.96 5.04 5.58

Table 6.13 exhibits a notable difference in the compression force in the most critical stud
of each panel predicted by SFEA and CFEA. According to the results of Example 1, and

120
as explained in Section 4.4, the number of equivalent shell elements affects the accuracy
in the prediction of the axial forces. Since SFEA provides an approximation of the
maximum axial forces in the studs, these values are not intended for design purposes. The
reason is that in some cases, the error in the prediction is large (e.g., twice as large).
Moreover, according with the results, the end studs in SWP 1, 5, 13 and 17 have failed as
a consequence of the lateral loading, so those studs need to be reinforced.

Table 6.13. Compression force in the most critical stud of each SWP, kN

SFEA CFEA
SWP Storey Storey
1 2 3 4 5 1 2 3 4 5
1, 5 137.89 106.95 77.31 50.85 25.32 92.10 63.03 42.98 26.31 12.25
13, 17 225.24 140.00 85.41 49.27 35.74 128.44 78.55 45.94 21.39 4.04
3, 6, 17.58 12.08 7.87 5.62 3.79 44.29 34.02 21.99 11.94 4.75
12, 15 5.15 0.00 0.00 7.38 12.85 11.44 8.46 5.55 3.14 1.25
19, 20 15.06 11.49 8.83 6.21 5.23 18.79 14.92 11.12 7.33 3.58
Load-Bearing Wall
2, 4 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
14, 16 26.60 14.10 8.62 4.33 0.00 61.41 44.88 28.29 14.54 5.09
23, 24 70.86 54.95 37.89 22.27 9.13 50.32 35.28 21.08 9.70 1.99
25, 26 7.25 4.07 1.58 0.00 14.94 7.32 6.14 4.49 2.82 1.19
7, 9 19.25 15.03 10.93 6.88 3.35 12.52 11.12 8.94 6.23 3.05
10, 11 16.95 16.95 12.49 8.14 3.65 10.84 9.25 7.14 4.70 2.07
8, 18 12.91 8.85 7.97 6.84 1.73 44.26 34.00 21.79 11.91 4.75
21, 22 10.13 5.25 4.28 3.49 0.00 21.83 17.73 13.18 8.64 4.18
27, 28 39.09 34.08 25.26 15.65 7.08 26.41 20.93 15.85 10.55 5.06
33, 34 40.98 28.71 18.98 11.35 5.45 16.53 11.34 7.80 4.82 1.95
29, 30 40.14 32.29 23.49 14.34 5.29 25.64 20.14 15.21 10.29 5.31
31, 32 36.22 28.81 20.37 11.90 3.99 22.84 17.11 12.37 8.03 3.88

Presented in Table 6.14 is the maximum tension force in the most critical stud of each
panel. The SWP that do not have tension force are not listed in the table. The prediction
of the tension forces from SFEA and CFEA differ.

121
 Table 6.14. Tension force in the most critical stud of each SWP, kN
SFEA CFEA
SWP Storey Storey
1 2 3 4 5 1 2 3 4 5
1, 5 198.66 120.55 71.94 35.88 28.75 119.06 69.32 38.31 15.89 1.13
13, 17 110.02 85.66 63.29 44.13 0.00 78.45 52.04 35.65 22.85 12.85
3, 6, N/A N/A N/A N/A 1.07 N/A N/A N/A N/A 3.24
12, 15 11.66 14.14 13.93 12.92 6.92 22.76 17.47 9.96 4.27 1.53
19, 20 N/A N/A N/A N/A 3.79 N/A N/A N/A N/A N/A
Load-Bearing Wall
2, 4 13.96 13.12 11.45 10.66 5.32 46.41 32.96 19.40 8.71 2.33
25, 26 47.22 35.55 26.34 18.22 7.08 34.06 21.62 10.59 4.38 1.60

Similar to Example 2, the analysis results from SFEA and CFEA display differences,
where the largest differences are found for the axial forces in the studs. Also, it is
determined that analysis by SAP2000 does contribute to the differences in the results,
since SAP2000 does not account for the offset of the sheathing and studs. Despite the
differences in the stud axial forces, the lateral displacements and lateral forces in the
SWP are consistently found by both approaches. Therefore, conducting the pushover
analysis by SFEA can provide appropriate results for carrying out the PBD assessment of
CFS buildings, i.e., as lateral displacements and forces are the two key parameters in that
regard.

6.3. Performance-Based Design Assessment of CFS Buildings

In this section, the PBD assessment of the CFS buildings analyzed in Examples 2 and 3 is
carried out in accordance with the methodology described in Chapter 5, where a single-
mode spectrum-based pushover analysis is employed. After the lateral loads in the
building corresponding to the specified performance objectives are computed, the
pushover analysis is conducted. For the pushover analysis, the totality of the gravity loads
are applied on the building, and a lateral load increment of 1% of the maximum target
base shear is also applied. Then, the structural analysis is carried out. Before applying the
next lateral load increment, the lateral stiffness of the SWP is updated by multiplying the

122
constitutive properties of the SWP by the lateral strength reduction factor, as explained in
Section 4.6. When the lateral load applied on the building is equal to the target base shear
corresponding to any of the specified performance objectives, the analysis is “paused” to
check the acceptance criteria of the SWP. If all the specified performance objectives are
satisfied, the PBD assessment of the building is completed.

The lateral stiffness of the sheathing on the SWP is updated as the lateral loads on the
building increase. When the lateral stiffness of a SWP is zero, it is considered that the
sheathing fails. In this case, the SWP is treated as a wall panel with no lateral stiffness,
and only the vertical stiffness of the studs is considered for subsequent load increments.
However, if one or several SWP fail, the performance objective in consideration is not
satisfied, and the lateral design of the building needs to be improved accordingly.

The results from the pushover analyses applied to assess the PBD of the example CFS
buildings are not validated, since no commercial software capable of performing this task
is available to allow for a comparison of results. However, the SWP inter-storey drifts,
lateral forces and stud forces found from both the linear analysis and the pushover
analysis, are compared to determine the difference in the results from both approaches.

6.3.1. Example 4: Three storey building

The PBD assessment of the building shown in Figure 6.2 is carried out. The material
properties and geometry of the building described in Example 2 are adopted as the
preliminary design for this example. According to the analysis results of Example 2, the
SWP lateral strength is greater than required and, so, the SWP can be redesigned to
reduce the lateral strength and gain a more economical design. In addition, a double end-
stud is required in several SWP. However, the design is not modified, so that the results
from the pushover analysis in this example can be compared later to the results from the
linear elastic analysis. In addition, the building is subjected to the same set of seismic
loads as shown in Figure 6.2, applied in load increments of 1%. The gravity loads are
maintained constant on the building throughout the analysis. Also, the stiffness

123
degradation model (Section 4.6) is used to reduce the lateral stiffness of the SWP after
each load increment.

The three-storey building is considered to be ordinary according to the categories

established by SEAOC (1995), so that the building must satisfy the four performance
objectives described in Table 2.1. The building is assumed to be located in California at
latitude 41.0° N and longitude 115.2°W on a site class B. The corresponding seismic
parameters needed to calculate the spectral acceleration for the four performance
objectives are obtained from Table 5.1. The four performance objectives are the
following: OP for a 50%/50 year earthquake, IO for a 25%/50 year earthquake, LS for a
10%/50 year earthquake, and CP for a 2%/50 year earthquake.

The gravity dead loads are computed from the building’s self-weight, and the gravity live
loads are obtained from ASCE/SEI 7 (2005) for residential buildings: Dead load (D) =
1.1 kN/m2 ; Live load floor (L) = 1.92 kN/m2 ; Live load roof (Lr) = 0.96 kN/m2

The ASCE/SEI 7 (2005) describes two load combinations that include earthquake loads
(E):

1 .2 ( D ) + 1 .0 ( E ) + L + 0 .2 S (6. 1)

0.9(D ) + 1.0(E ) + 1.6 H (6. 2)

where H is the load due to earth pressure, ground pressure, or bulk material pressure, and
S is the load due to snow. Bulk material pressure and snow load are not applicable for this
example because the building is not going to be built underground or in a city with risk of
snow. The most critical load combination is chosen to determine the factored loads to be
applied on the building As such, the factored gravity load applied on the floors of the
building is computed using Eq. (6.1), such that,

1.2(1.1) + 1.92 = 3.24kN / m 2

Also from Eq. (6.1), the factored gravity load applied on the roof of the building is
computed by,

124
 1.2(1.1) = 1.32kN / m 2

The seismic weight of the building is W=405 kN. The natural period of the building is
T=0.324 second in the x direction, which is computed according to the procedure
described in Appendix A.

Table 6.15 gives the target base shears corresponding to the specified four performance
objectives, which are computed as functions of the zone seismic parameters and natural
period of the building. The variables from the second to the ninth column are defined in
Section 5.2. After the base shears are computed, the lateral load for the pushover analysis
is defined by the maximum target base shear Vb max=730.70 kN. The lateral loads are
applied on the building in 1% increments of 7.307 kN, until the maximum target base
shear is reached or the structure collapse occurs. The factors to adjust the spectral
acceleration in the short period and one second period range, Fa and Fv, are provided by
Tables 5.2 and 5.3, respectively. Listed in Table 6.16 are the lateral loads for each storey,
Fa, computed by multiplying the maximum target base shear by the lateral load
distribution coefficient Cvx for each storey, which is computed using Eq. (5.63), for κ =
1.0 since the building’s period is smaller than 0.5 second. The variables h, w, and Cvx in
Table 6.16 are the inter-storey height, storey weight, and lateral load distribution
coefficient, respectively.

Table 6.15. Target base shears, kN

SS S1 SMS SM1 SDS SD1 To Ts Sa Vb

PO
(g) (sec) (g) (kN)
OP 0.109 0.035 0.109 0.035 0.073 0.023 0.064 0.321 0.073 71.75
IO 0.180 0.058 0.180 0.058 0.120 0.039 0.064 0.322 0.120 118.91
LS 0.250 0.080 0.250 0.080 0.167 0.053 0.064 0.320 0.167 164.01
CP 1.100 0.410 1.100 0.410 0.733 0.273 0.075 0.373 0.733 730.70

125
 Table 6.16. Lateral loads in the building

Storey, x h (m) w (kN) Cvx Fx (kN)

1 2.85 309.23 0.16 113.29
2 5.70 312.91 0.31 229.27
3 8.55 309.69 0.53 388.14
Σ 931.83 1.00 730.70

Table 6.17 presents the inter-storey drifts and limit drifts for the SWP related to each
performance objective, which are computed by the limit drift ratios developed in Section
5.4 (i.e., 0.2%, 1.0%, 2.1% and 2.5% for the OP, IO, LS and CP performance levels). The
limit inter-storey drifts are obtained by multiplying the lateral drift ratios by the height of
the SWP. SWP 1 and 3 exhibit the largest lateral drifts, because the seismic loads are
applied on these panels. Yet, the limit drifts are not exceeded, so that the SWP are well
designed to resist the seismic loads.

Table 6.17. SWP inter-storey drift, mm

Performance objective

OP (limit=4.8) IO (limit=24.0) LS (limit=50.4) CP (limit=60.0)

SWP
Storey Storey Storey Storey
1 2 3 1 2 3 1 2 3 1 2 3
1 0.34 0.39 0.43 0.76 0.77 0.76 1.22 1.19 1.10 21.40 14.84 9.44
3 0.36 0.40 0.43 0.78 0.78 0.76 1.25 1.20 1.11 21.50 14.89 9.46
6 0.34 0.39 0.43 0.75 0.76 0.76 1.22 1.18 1.11 21.38 14.81 9.48
7 0.35 0.39 0.43 0.76 0.77 0.75 1.23 1.19 1.10 21.43 14.85 9.41
9 0.35 0.39 0.43 0.77 0.77 0.75 1.24 1.19 1.10 21.47 14.87 9.41

The results from the pushover analysis at the CP performance level are compared to the
results from the linear analysis to determine the influence of each type of analysis. At the
CP performance level, the building is subjected to the same loading condition as that for
the linear analysis. The difference in the displacements predicted from linear and
pushover analysis is significantly large. The inter-storey drift for the pushover analysis is
more than triple that for the linear analysis. The reason is that the pushover analysis

126
accounts for the geometric nonlinearities and the stiffness degradation of the SWP. Figure
6.8 is a comparison of the inter-storey drifts and lateral displacements for SWP 1,
obtained from linear and pushover analysis.

Table 6.18 indicates the lateral strengths PR and lateral forces Pa for the SWP. The lateral
strengths are computed in accordance with Section 3.3 of Chapter 3, and the lateral forces
are obtained from the SFEA (in Section 4.3). Since the lateral forces in the SWP are
smaller than their strengths, the performance objectives are deemed to be satisfied. The
largest forces are obtained for the CP performance objective, since it is associated with
the maximum target base shear. The forces from linear analysis are slightly larger than
the forces from pushover analysis. Yet, the difference is small, indicating that the
prediction of the lateral forces acting on the SWP is not affected by the method of
analysis (i.e., linear or pushover analysis).

Figure 6.8. Inter-storey drifts and displacements of the SWP 1: Linear and pushover analysis

In Table 6.18, the summation of lateral forces in the SWP is different to the lateral loads
applied on the building, i.e., for the CP performance objective the summation of forces is
677 kN while the lateral force applied on the building is 730 kN. The reason for the
difference is that the panels in the y direction (i.e., perpendicular to the SWP 1, 3, 7 and

127
9) are taking a proportion of the lateral loads: i.e., SWP 2 is taking 23.79 kN, SWP 4 is
taking 14.85 kN, and SWP 5 is taking 14.77 kN.

Table 6.18. SWP lateral strength PR and lateral force Pa, kN

Lateral force, Pa

OP IO LS CP
SWP PR
Storey Storey Storey Storey
1 2 3 1 2 3 1 2 3 1 2 3
1 9.14 7.66 4.31 17.89 14.77 8.60 26.66 21.91 12.89 171.46 140.23 85.71
212.80
3 9.29 7.77 4.18 18.13 14.92 8.47 26.97 22.08 12.76 171.71 140.41 85.66
212.80
7 8.14 6.62 4.47 16.93 14.02 9.10 25.64 21.34 13.67 166.81 138.51 87.89
212.80
9 8.43 6.75 4.43 17.23 14.16 9.06 25.95 21.49 13.64 167.02 138.67 87.86
212.80

Listed in Table 6.19 is the stiffness degradation coefficient, represented by λ, for each
SWP, which is computed using Eq. (4.52). When λ is equal to zero the SWP has failed;
for this example none of the SWP failed.

Table 6.19. SWP stiffness degradation coefficient λ

OP IO LS CP

SWP Storey Storey Storey Storey

1 2 3 1 2 3 1 2 3 1 2 3
1 0.69 0.71 0.77 0.60 0.63 0.71 0.54 0.57 0.66 0.08 0.14 0.30
2 0.93 0.95 0.93 0.92 0.95 0.91 0.91 0.94 0.90 0.85 0.90 0.80
3 0.69 0.71 0.78 0.60 0.63 0.71 0.54 0.57 0.66 0.08 0.14 0.30
4 0.82 0.76 0.75 0.82 0.76 0.75 0.82 0.76 0.75 0.79 0.74 0.75
5 0.84 0.77 0.75 0.84 0.77 0.74 0.84 0.77 0.74 0.85 0.75 0.71
7 0.70 0.73 0.77 0.61 0.64 0.70 0.55 0.58 0.65 0.09 0.15 0.30
9 0.70 0.72 0.77 0.61 0.64 0.70 0.54 0.58 0.65 0.09 0.15 0.30

The maximum compression force in the most critical stud of each SWP, resulting from
the gravity and seismic loads applied on the building, are provided in Table 6.20. The
studs in the building have not exceeded their axial compression strength for the LS and

128
CP performance objectives, which is 60.03 kN as shown in Table 6.3. The compression
forces predicted by pushover analysis are smaller than the forces predicted by linear
analysis, especially for SWP 2, 4, 5 and 6.

Table 6.20. Compression force in most critical stud of each SWP, kN

Performance objective

SWP OP IO LS CP
Storey Storey Storey Storey
1 2 3 1 2 3 1 2 3 1 2 3
1 12.96 8.08 7.21 14.71 9.12 6.42 26.81 16.84 11.90 40.62 28.65 10.44
2
1.23 0.91 0.53 0.96 0.91 0.67 1.90 1.82 1.32 0.00 0.00 4.24
3
12.73 8.00 7.07 14.57 9.06 6.26 26.69 16.75 11.64 40.40 28.56 10.08
4
3.16 1.91 0.81 3.21 1.98 0.87 6.41 3.95 1.74 6.58 4.80 2.79
5
3.17 1.91 0.81 3.22 1.97 0.87 6.44 3.93 1.74 6.59 4.69 2.79
6
8.32 6.07 4.49 8.91 6.40 5.18 17.88 12.81 10.37 53.73 31.73 23.77
7
22.93 12.46 4.73 24.14 13.29 4.70 44.16 24.56 8.74 48.88 30.89 16.08
8
13.48 7.85 2.95 14.15 8.02 3.31 28.32 16.06 6.64 59.53 32.74 18.93
9
23.12 12.51 4.67 24.25 13.31 4.63 44.26 24.59 8.62 48.49 30.82 15.32
10
13.49 7.82 2.96 14.16 8.01 3.30 28.33 16.03 6.60 59.60 32.74 18.83
11
8.24 6.07 4.48 8.93 6.40 5.18 17.91 12.81 10.37 53.58 31.65 23.83

Table 6.21 indicates the maximum tension force in the most critical stud of each SWP, as
a result of the gravity and seismic loads applied on the building. SWP 1, 3, 7 and 9
exhibit one end-stud in compression and the other in tension. Also, SWP 2 has studs in
tension for the CP performance objective. None of the studs in the SWP exceeds the axial
tension strength, which is 97.38 kN. The stud forces predicted by linear analysis are
slightly larger than the forces predicted by pushover analysis.

129
 Table 6.21. Tension force in most critical stud of each SWP, kN

Performance objective

SWP OP IO LS CP
Storey Storey Storey Storey
1 2 3 1 2 3 1 2 3 1 2 3
1 2.36 0.00 0.00 6.70 3.59 0.95 12.79 6.91 1.86 70.05 38.80 19.42
2 2.45 0.00 0.00 6.85 3.61 0.91 13.08 6.95 1.79 70.81 38.88 19.43
3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 49.26 29.44 22.47
7 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 49.29 28.93 22.43
9 2.36 0.00 0.00 6.70 3.59 0.95 12.79 6.91 1.86 70.05 38.80 19.42

From the results of the spectrum-based pushover analysis, it is observed that the building
satisfies all seismic requirements. The four performance objectives are deemed to be
satisfied since the inter-storey drifts and lateral strengths for the SWP are smaller than the
limit values. Therefore, the PBD assessment for seismic forces is completed, and the
lateral design of the building does not require modification.

According to the pushover and linear analyses results, the prediction of the lateral forces
in the SWP, and forces in the studs are not significantly affected by the type of analysis
used. Conversely, the difference in the prediction of the lateral displacements by both
approaches is significantly large. Therefore, CFS buildings can be designed for the
strength limit state, according to seismic codes, with sufficient accuracy by linear
analysis. However, the PBD assessment of a CFS building must not be carried out by
linear analysis since the acceptance criteria are established as functions of the
displacements. In summary, then, if linear analysis is employed to assess the seismic
behaviour of a CFS building, the results can lead to inappropriate seismic designs.

6.3.2. Example 5. Five-Storey Building

The PBD assessment of the building in Example 3 is conducted with the seismic loads
applied in the x and y directions (see Figure 6.5). Since the building is for offices, the
performance objectives for ordinary structures described in Table 2.1 can be used, but a

130
higher level of performance is desirable. Consequently, it is designed to satisfy the
performance objectives for hazardous facilities (SEAOC, 1995): OP for a 10%/50 year
earthquake and IO for a 2%/50 year earthquake. These performance objectives provide
the CFS building with higher performance since the considered earthquakes are the
largest possible, according to code requirements. Thus, by satisfying the OP and IO
performance levels it is “assured” that the building is not going to collapse, since these
levels result in less damage than that of the CP performance level. The building will be
built in California on a site class B, and the site coordinates are latitude 36.9°N and
longitude 120°W. The seismic weight (W) of the building is 1238.24 kN, and its natural
period (T) is 2.05 seconds, as computed by the procedure described in Appendix A.

The maximum target base shear (Vb max) is listed in Table 6.22 as 251.78 kN, representing
the maximum lateral loads applied on the building for the pushover analysis. In Table
6.23, the seismic loads for each storey are defined by the lateral load distribution
coefficients Cvx (Eq. 5.11), for exponent κ =1.75 since the building’s natural period is
larger than 0.5 seconds and smaller than 2.5 seconds.

Table 6.22. Target base shears in the x direction

SS S1 SMS SM1 SDS SD1 To Ts Sa Vb

PO
(g) (sec) (g) (kN)
OP 0.290 0.140 0.290 0.140 0.193 0.093 0.097 0.483 0.102 339.19
IO 0.500 0.230 0.500 0.230 0.333 0.153 0.092 0.460 0.179 592.07

Table 6.23. Lateral loads in the building in the x direction

Storey, x h (m) w (kN) Cvx Fx (kN)

1 2.85 642.03 0.054 31.80
2 5.70 642.03 0.122 72.00
3 8.55 642.03 0.196 116.12
4 11.40 642.03 0.275 163.00
5 14.25 633.23 0.353 209.15
Σ 2090.06 1.0 592.07

131
The dead gravity loads are calculated from the self-weight of the building, and the live
gravity loads are specified by ASCE/SEI 7 (2005) for office buildings: Dead load (D) =
1.3 kN/m2 ; Live load (L) = 2.40 kN/m2 ; Live load roof (Lr) = 0.96 kN/m2 .

The factored gravity loads applied on the first two storeys of the building are computed
by Eq. (6.1) as 1.2(1.1) + 2.4 = 3.72kN / m 2 . Also from Eq. (6.1), the factored gravity load

applied on the roof of the building is computed as 1.2(1.1) = 1.32kN / m 2 .

Table 6.24 presents the inter-storey drifts of the SWP, which are computed in the local x
direction of the SWP (see Figure 4.2). In the first row of the table are the limit drifts
associated with the OP and IO performance levels, computed by multiplying the height of
the panels by the limit drift ratios in Section 6.4. According to Table 6.24, both
performance objectives are not satisfied, since the SWP inter-storey drifts are much
larger than the limit drifts. Comparing the lateral drifts predicted by the pushover analysis
and those predicted by the linear analysis, it can be noted that the lateral drifts obtained
with pushover analysis are much larger. The inter-storey drifts and lateral displacements
for SWP 1 predicted by both approaches are depicted in Figure 6.9.

Table 6.24. SWP inter-storey drift in the x direction, mm

OP (Limit: 5.70 mm) IO (Limit: 28.5 mm)

SWP Storey Storey
1 2 3 4 5 1 2 3 4 5
1 11.25 11.97 11.80 10.29 8.11 41.11 53.60 38.29 25.71 17.27
5 11.25 11.97 11.79 10.29 8.11 41.07 53.55 38.26 25.71 17.28
13 11.25 11.96 11.78 10.29 8.09 41.10 53.56 38.27 25.71 17.24
17 11.25 11.95 11.79 10.28 8.09 41.06 53.51 38.25 25.70 17.24

132
 Figure 6.9. Inter-storey drifts and displacements of the SWP 1: Linear and pushover analysis

Listed in Table 6.25 is the stiffness degradation coefficient λ for each SWP, which is
computed using Eq. (4.52). For this example, the SWP in the first and second storeys
have failed for the IO performance objective since their λ value is equal to zero, and the
SWP on the third storey are at a point of imminent failure.

Table 6.25. SWP stiffness degradation coefficient λ

OP IO
SWP Storey Storey
1 2 3 4 5 1 2 3 4 5
1 0.13 0.17 0.22 0.30 0.45 0.00 0.00 0.02 0.10 0.28
5 0.13 0.17 0.22 0.30 0.45 0.00 0.00 0.02 0.10 0.28
13 0.13 0.17 0.21 0.29 0.43 0.00 0.00 0.02 0.10 0.28
17 0.13 0.17 0.21 0.29 0.43 0.00 0.00 0.02 0.10 0.28

The lateral strengths and lateral forces for the SWP are listed in Table 6.26. Comparing
the SWP strengths and lateral forces, it is observed that lateral force acting on the SWP in
the first story is larger than the SWP strength. Therefore, those SWP have failed; it is for
this reason that the lateral displacements are so large. Also, it can be noted in Table 6.26
that the lateral forces predicted by the linear analysis are slightly larger than the forces
predicted by the pushover analysis.

133
 Table 6.26. SWP lateral strength PR and lateral force Pa, kN

Lateral force, Pa
PR OP IO
SWP
Storey Storey
1 2 3 4 5 1 2 3 4 5
1 101.13 69.63 60.97 52.57 39.31 20.18 132.34 101.89 96.69 76.59 41.38
5 101.13 69.60 60.95 52.55 39.30 20.17 132.15 101.90 96.67 76.54 41.30
13 101.13 69.31 60.98 53.16 40.60 22.27 130.61 101.42 96.12 76.41 41.85
17 101.13 69.29 60.97 53.15 40.59 22.27 130.48 101.42 96.10 76.39 41.86

Table 6.27 indicates the maximum compression force in the most critical studs of each
panel, caused by the gravity and seismic loads. In several cases, the force in the studs is
greater that their strength. For all the SWP, as was found in Example 3, the largest
compression force is registered at the end studs. The compression forces in the studs from
the pushover analysis are larger than the compression forces from linear analysis;
especially for the end-studs in the SWP that have failed, SWP 1, 5, 13 and 17, because
these studs are subjected to larger forces due to the over-turning of the building.

Table 6.27. Compression force in the most critical stud of each SWP, kN

OP IO
Panel Storey Storey
1 2 3 4 5 1 2 3 4 5
1 65.53 52.61 39.46 26.12 13.51 200.47 47.74 46.59 38.86 16.93
2 0.21 0.10 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00
3 23.00 16.65 13.03 8.16 2.64 24.23 17.14 9.44 5.95 3.52
4 0.22 0.10 0.08 0.06 0.00 0.12 0.00 0.00 0.00 0.00
5 65.54 52.63 39.48 26.14 13.51 201.00 47.94 46.72 38.92 16.90
6 23.04 16.68 13.00 8.15 2.64 24.05 17.22 9.53 5.82 3.53
7 36.79 32.04 24.11 14.94 5.90 38.53 32.25 25.92 16.75 7.29
8 24.42 15.25 13.27 11.58 2.76 23.65 13.88 13.20 12.01 0.00
9 36.80 31.99 24.06 14.89 5.87 38.62 32.21 25.81 16.62 7.20
10 40.94 34.10 24.80 14.85 5.29 47.79 38.52 26.53 15.33 6.24
11 40.90 34.05 24.75 14.80 5.32 47.68 38.15 26.26 15.20 6.29
12 15.15 11.27 9.71 9.31 8.80 8.19 6.66 4.02 2.86 10.86

134
 13 119.52 87.24 53.25 33.56 22.71 397.18 90.17 64.08 41.66 35.89
14 38.05 18.57 12.11 7.33 0.00 49.43 25.98 15.44 7.02 0.00
15 15.12 11.24 9.67 9.25 8.75 8.02 6.54 3.91 2.72 10.67
16 38.05 18.57 12.11 7.33 0.00 49.43 25.98 15.44 7.02 0.00
17 119.50 87.20 53.21 33.56 22.71 397.20 90.26 64.08 41.66 35.93
18 24.42 15.25 13.27 11.58 2.76 23.65 13.88 13.20 12.01 0.00
19 25.08 17.85 13.29 9.63 7.81 21.68 15.13 10.82 9.20 0.37
20 25.09 17.86 13.31 9.65 7.83 21.66 15.12 10.83 9.25 0.37
21 2.20 1.34 1.15 0.93 0.00 0.19 0.08 0.08 0.07 0.00
22 2.20 1.34 1.15 0.93 0.00 0.19 0.08 0.08 0.07 0.00
23 67.97 50.86 35.24 20.34 8.08 116.86 85.17 60.01 35.58 15.80
24 67.95 50.92 35.32 20.42 8.03 116.91 85.60 60.37 35.86 15.64
25 32.27 22.73 15.47 14.89 11.61 34.52 22.41 15.67 1.59 21.69
26 32.26 22.71 15.45 14.89 11.61 34.23 22.23 15.63 1.58 21.72
27 75.71 57.30 39.78 24.56 11.64 83.62 66.81 46.26 27.10 12.48
28 75.71 57.30 39.79 24.57 11.65 83.64 66.76 46.25 27.12 12.49
29 78.95 62.01 44.47 26.58 9.48 87.32 67.47 47.20 28.59 10.80
30 78.96 62.06 44.53 26.65 9.52 87.37 67.74 47.40 28.74 10.93
31 79.91 62.85 44.33 25.99 8.65 84.83 66.41 45.23 25.74 8.45
32 79.95 62.93 44.40 26.05 8.68 84.89 66.71 45.46 25.89 8.52
33 98.44 74.80 50.82 28.22 11.49 117.77 88.64 58.67 32.15 11.65
34 98.45 74.81 50.83 28.22 11.48 117.72 88.82 58.75 32.15 11.60

Table 6.28 provides the maximum tension forces in the most critical stud of each panel.
According to the results obtained, the tension forces in several studs are much larger than
their axial strength, especially for the studs in the SWP that have failed.

Table 6.28 Tension force in the most critical stud of each SWP, kN

OP IO
Panel Storey Storey
1 2 3 4 5 1 2 3 4 5
1 81.99 49.62 29.61 3.03 N/A 335.93 82.51 60.85 32.38 26.00
2 11.40 14.12 13.34 N/A N/A 28.76 30.66 21.43 17.98 13.23
3 N/A N/A N/A N/A 4.26 N/A N/A N/A N/A 3.01

135
 4 11.40 14.12 13.34 N/A N/A 28.76 30.66 21.43 17.98 13.23
5 81.90 49.61 29.60 3.00 N/A 335.10 82.08 60.67 32.33 26.05
6 N/A N/A N/A N/A 4.26 N/A N/A N/A N/A 3.09
12 5.12 7.83 N/A N/A 5.85 12.68 15.13 13.28 11.41 7.19
13 34.07 27.08 N/A N/A N/A 149.35 42.63 39.71 33.44 13.21
15 5.17 7.89 N/A N/A 5.89 12.93 15.37 13.56 11.66 7.35
17 34.05 27.03 N/A N/A N/A 148.70 42.42 39.68 33.44 13.23
25 26.66 19.49 3.26 N/A 4.90 67.50 48.53 38.66 27.00 11.66
26 26.63 19.47 3.27 N/A 4.86 67.70 48.37 38.59 26.98 11.61

The pushover analysis in the y direction is conducted in the same fashion as the analysis
in the x direction. However, since the period of the building in the y direction differs from
the period in the x direction, the y-direction target base shears and lateral loads are
different. The fundamental period of the CFS building in the y direction is 1.11 seconds.
Tabulated in Table 6.29 are the target base shears for the specified two performance
levels. Table 6.30 tabulates the lateral loads applied on each storey of the building. The
pushover analysis is carried out by the same procedure as that for the previous examples.

Table 6.29. Target base shears for analysis in the y direction

SS S1 SMS SM1 SDS SD1 To Ts Sa Vb

PL
(g) (sec) (g) (kN)
OP 0.290 0.140 0.290 0.140 0.193 0.093 0.097 0.483 0.102 339.19
IO 0.500 0.230 0.500 0.230 0.333 0.153 0.092 0.460 0.179 592.07

Table 6.30. Lateral loads in the building for analysis in the y direction

Storey, x h (m) w (kN) Cvx Fx (kN)

1 2.85 642.03 0.054 31.80
2 5.70 642.03 0.122 72.00
3 8.55 642.03 0.196 116.12
4 11.40 642.03 0.275 163.00
5 14.25 633.23 0.353 209.15
Σ 2090.06 1.0 592.07

136
Table 6.31 lists the lateral drifts for the SWP for the OP performance level. The lateral
drifts for the IO performance level are not presented because the building has collapsed
before reaching the associated lateral load level. Since the building inter-storey drifts are
larger than the limit drift for the OP performance level and the building failed before
reaching the IO lateral load level, the OP and IO performance objectives have not been
satisfied. Therefore, the design of the building must be improved to resist the seismic
forces in the y direction (i.e., additional reinforcement of the SWP is required).

Table 6.31. SWP inter-storey drift, mm

OP (Limit: 5.7 mm)

SWP Storey
1 2 3 4 5
3 18.36 13.19 11.10 7.94 4.43
6 18.36 13.19 11.10 7.94 4.43
12 18.36 13.19 11.10 7.94 4.43
15 18.36 13.19 11.10 7.94 4.43
19 18.35 13.20 11.10 7.94 4.42
20 18.35 13.20 11.10 7.94 4.43

Table 6.32 gives the SWP lateral strengths PR, lateral forces Pa, and the stiffness
degradation coefficient λ found from the pushover analysis for the OP performance level.
It can be observed in the table that λ is close to zero for the SWP in the first and second
storeys.

Table 6.32. SWP lateral strength PR, lateral force Pa, kN, and stiffness degradation coefficient, λ

Lateral force, Pa λ
OP OP
SWP PR
Storey Storey
1 2 3 4 5 1 2 3 4 5
3 48.84 40.61 34.75 31.81 26.59 18.97 0.05 0.08 0.10 0.14 0.21
6 48.84 40.27 33.40 29.32 22.22 11.09 0.05 0.09 0.12 0.18 0.31
12 48.84 40.26 33.40 29.32 22.23 11.10 0.05 0.09 0.12 0.18 0.31

137
 15 48.84 40.60 34.74 31.81 26.59 18.97 0.05 0.08 0.10 0.14 0.21
19 48.84 40.52 34.88 32.27 27.43 20.22 0.05 0.08 0.10 0.13 0.20
20 48.84 40.05 32.77 28.16 20.27 9.21 0.05 0.10 0.13 0.20 0.34

Listed in Table 6.33 are the maximum compression and tension forces in the most critical
stud of each panel for the OP performance objective. The studs in several SWP have
failed.

Table 6.33. Compression force in most critical stud of each SWP, kN

OP, Compression OP, Tension

Panel Storey Storey
1 2 3 4 5 1 2 3 4 5
1 77.12 54.28 34.62 20.57 8.05 N/A N/A N/A N/A N/A
2 4.16 3.23 2.79 2.07 1.16 N/A N/A N/A N/A 0.15
3 13.75 13.10 12.42 11.00 8.25 N/A N/A N/A N/A N/A
4 5.43 3.07 2.26 1.38 0.00 N/A N/A N/A N/A 0.16
5 28.50 23.64 20.13 16.50 9.26 N/A N/A N/A N/A N/A
6 20.19 16.82 15.44 12.85 8.89 N/A N/A N/A N/A 8.94
7 53.86 42.18 28.90 16.10 5.56 N/A N/A N/A N/A N/A
8 3.52 2.98 2.68 2.17 0.60 N/A N/A N/A N/A 0.14
9 32.41 24.43 16.71 11.71 5.10 N/A N/A N/A N/A N/A
10 53.89 42.24 28.93 16.11 5.55 N/A N/A N/A N/A N/A
11 32.43 24.37 16.65 11.66 5.02 N/A N/A N/A N/A N/A
12 20.19 16.82 15.44 12.85 8.89 N/A N/A N/A N/A 8.94
13 77.20 54.22 34.56 20.53 8.03 N/A N/A N/A N/A N/A
14 4.16 3.24 2.79 2.08 1.16 N/A N/A N/A N/A 0.15
15 13.75 13.10 12.42 11.00 8.25 N/A N/A N/A N/A N/A
16 5.42 3.08 2.27 1.39 0.00 N/A N/A N/A N/A 0.16
17 29.16 22.87 19.76 16.14 8.99 N/A N/A N/A N/A N/A
18 3.57 2.88 2.46 1.84 1.00 N/A N/A N/A N/A 0.13
19 18.61 16.81 15.78 13.77 10.25 N/A N/A N/A N/A 6.16
20 21.32 17.32 15.05 11.65 3.63 N/A N/A N/A N/A 7.73
21 3.52 2.98 2.68 2.17 0.60 N/A N/A N/A N/A 0.14
22 3.57 2.88 2.46 1.84 1.00 N/A N/A N/A N/A 0.13

138
 23 34.90 25.74 17.90 10.63 3.17 N/A N/A N/A N/A N/A
24 28.26 20.96 14.29 7.95 3.36 N/A N/A N/A N/A N/A
25 34.86 25.84 18.07 10.79 3.30 N/A N/A N/A N/A N/A
26 28.21 21.00 14.30 7.96 3.42 N/A N/A N/A N/A N/A
27 73.62 59.16 43.59 26.51 12.89 N/A N/A N/A N/A 0.18
28 104.08 75.47 48.45 24.97 10.50 N/A N/A N/A N/A 2.43
29 74.41 60.78 44.37 27.03 9.67 N/A N/A N/A N/A N/A
30 87.67 67.17 46.40 26.69 9.47 N/A N/A N/A N/A N/A
31 74.40 60.80 44.38 27.03 9.68 N/A N/A N/A N/A N/A
32 87.66 67.15 46.38 26.67 9.44 N/A N/A N/A N/A N/A
33 73.62 59.17 43.58 26.50 12.88 N/A N/A N/A N/A 0.18
34 104.14 75.51 48.42 24.95 10.49 N/A N/A N/A N/A 2.44

According to the PBD assessment of the building, conducted in the x and y directions, the
building has not satisfied the OP and IO performance objectives. Therefore, the SWP in
the building must be reinforced to increase both the building lateral and vertical strength
capacities. To increase the building lateral strength, the screw spacing can be reduced, but
if this is not sufficient then the number of SWP must be increased. To increase the
vertical strength of the SWP, thicker studs can be used.

Although CFS buildings may be lighter than buildings built with other materials, such as
steel, concrete and masonry, the live loads that are imposed on them are the same. As
such, the lower weight of CFS buildings results in smaller seismic loads. However, an
adequate number of SWP must be used in the building to resist the lateral loads.

Since CFS members are thin-walled, their axial strength is very limited. This may cause
problems during the PBD of CFS buildings. In particular, it is important to pay strict
attention to the design of studs for mid-rise CFS buildings.

139
 Chapter 7
Conclusions and Recommendations

7.1. Conclusions
The principal contribution of this study is the development of a methodology for carrying
out the Performance-Based Design (PBD) assessment of low and mid-rise buildings built
with Cold-Formed Steel (CFS) shear wall systems. The proposed methodology includes
the establishment of the acceptance criteria for Shear Wall Panels (SWP), a simplified
nonlinear finite element technique for spectrum-based pushover analysis, and a stiffness
degradation model for the SWP.

A primary difficulty to conduct the PBD assessment of a structural system is establishing

the acceptance criteria associated with the performance levels of the structural elements.
Since SWP are the primary lateral load-resistant elements in CFS buildings, the
acceptance criteria are established according to the lateral drift and lateral strength of the
SWP. For CFS SWP, the lateral drift ratios are 2.5%, 2.1%, 1.0% and 0.2% for the CP,
LS, IO and OP performance levels, respectively. The limit drift ratios for CFS SWP are
determined from experimental data, obtained by Branston et al. (2006).

A method for computing the lateral strength of the SWP is devised. The proposed method
is used to evaluate the lateral strength of SWP with different sheathing and framing
materials, panel dimensions, and construction details such as fastener spacing. The
comparisons of the results from the proposed method and extensive experimental tests
carried out by different researchers, are in good agreement. The method includes the
formulation for estimating the lateral drift of SWP at the ultimate load level. According to
comparisons made with experimental results, the proposed method yields better accuracy
for lateral strength than for lateral displacements. Of the iterative and simplified
procedures proposed for evaluating the lateral strength of SWP, the simplified procedure
requires significantly less computational effort than the iterative one. As such, the

140
simplified procedure for evaluating the ultimate strength coefficient is an excellent
alternative to design tables, such as AISI (2004). The comparison of experimental tests
and predicted lateral strengths and displacements were in good agreement; the average of
the ratios of experimental-to-predicted lateral strengths is 1.05 with a standard deviation
of 0.11, while for lateral displacements the ratio average is 1.13 and the standard
deviation is 0.19.

The predicted linear analysis results from Simplified Finite Element Analysis (SFEA) and
Conventional Finite Element Analysis (CFEA) for isolated SWP of various lengths are in
good agreement, except that the out-of-plane displacements are different because the
CFEA models the sheathing in a way that does not fully represent the real conditions of
the SWP. Whereas, the shell element properties employed for SFEA are equivalent to the
SWP properties. It was found that the ratio length-to-height of the (sixteen-node) shell
elements for modeling SWP should be in the range of 0.5 and 3.0. Comparisons of SWP
force-displacement curves from experimental investigations and from SFEA using the
proposed stiffness degradation model are in good agreement, except some difference
exists in the shape of the response curve for SWP with Douglas Fir Plywood (DFP)
sheathing material. This difference is caused primarily by the nonlinearity exponent in the
stiffness degradation model, calibrated using different types of wood sheathing materials.

It is determined that the prediction of the displacements is affected more by the

equivalent shell concept than by the number of elements in the model. However, the
opposite exists for the prediction of the studs’ axial forces. The comparison of the results
of linear SFEA and CFEA for example three-storey buildings are in good agreement for
both the lateral displacements and lateral forces in the SWP. However, some differences
exist for the axial forces in the studs and displacements in the z direction: for SWP in the
direction of analysis, the difference between the predicted stud axial forces is large, while
the difference for the displacements in the z direction is large. It is not recommended to
use the predicted stud forces for final design. Nevertheless, these differences do not
otherwise affect the PBD assessment, since the SWP acceptance criteria are evaluated
solely as functions of inter-storey drifts and lateral forces.

141
For linear analysis of an example five-storey building, the results from SFEA and CFEA
show similar tendencies in the predicted displacements and SWP lateral forces. However,
there are large differences in the predicted forces in the studs due to the limitations of the
model proposed in Section 4.4. Another reason of the difference in the predictions is that
since the models created in SAP2000 cannot match the conditions as accurately as the
SFEA models, they are not reliable to validate the results of SFEA. Therefore, specialized
software for finite element analysis should be employed to validate the results of SFEA in
future studies.

For the example three-storey building considered, the lateral and axial strengths of SWP
are sufficient to resist the applied loads. For the five-storey building, the SWP and studs
in the first and second storey failed since those elements did not have the sufficient
strength to resist the loads applied on the building. Although only two different building
models were analyzed by this study, it is quite likely that similar behaviour may occur for
other CFS buildings. An appropriate design should include a large enough number of
SWP and load-bearing wall panels that result in a suitable distribution of the gravity loads
in the building, and sufficient lateral strength.

Linear analysis can be consistently employed for carrying out the lateral seismic strength
design of SWP since there is no significant difference in their lateral forces predicted by
linear and pushover analysis. This suggests that linear analysis may be effectively used
for preliminary sizing of CFS buildings. However, the results of this study suggest that
linear analysis greatly underestimates the lateral drifts and the stud axial forces for SWP
and, therefore, that pushover analysis is required to carry out the PBD assessment of CFS
buildings.

The main feature of SFEA is the limited number of elements it requires to model a CFS
building (e.g., for the example five-storey building, CFEA required 3480 four-node shell
elements and 2580 frame elements, whereas SFEA required only 260 sixteen-node shell
elements). The reduction in pushover analysis computing time using SFEA is very

142
significant. It can be concluded that the proposed methodology provides an efficient and
practical technique for the PBD assessment of CFS buildings.

7.2. Recommendation for Future Work

The use of pushover analysis to assess the seismic behaviour and performance-based
design of CFS buildings has notable advantages over prescriptive design based on
seismic codes. The proposed methodology for carrying out the PBD assessment of CFS
buildings can be extended to include account for other types of materials and elements,
such steel columns and girders. Combining different types of structural elements may
provide other advantages, such as the capability for designing high-rise composite hot-
rolled and cold-formed steel buildings. In addition, though not explicitly dealt with
herein, the PBD process itself can be optimized to find more economical and safer
designs. Mentioned in the following are aspects of this work that extend or increase the
robustness of the proposed methodology developed by this study.

• Develop optimization criteria to implement the performance-based design process

for CFS buildings.
An optimization algorithm can improve the methodology for the PBD of CFS
buildings, producing more efficient designs. Different approaches are found in the
literature for carrying out the PBD optimization of structural systems, such as
those published by Gong (2003) and Grierson et al. (2006) for hot-rolled steel
buildings, which may form a basis for PBD optimization of CFS buildings.

• Investigate the interaction of cold-formed steel shear wall panels with other
structural elements, such as steel columns and beams.
In construction practice, it is common to combine CFS SWP with structural steel
elements. The interaction of these two material types in an integrated structural
design context presents an interesting challenge. For example, the aforementioned
method for carrying out the PBD of hot-rolled steel buildings can be combined

143
 with the CFS methodology presented in this study to provide for the PBD of
composite hot-rolled and cold-formed steel high-rise buildings.

• Extend the single-mode pushover analysis employed by this study to multi-mode

or modal pushover analysis.
The single-mode pushover analysis is appropriate for buildings with a
predominant fundamental period of vibration, such that the building response is
not affected by the participation of higher-frequency vibration modes. If higher
modes do affect the response of a building, however, then modal pushover
analysis should be used to assess seismic behaviour. FEMA 450 (2003) has
established the basis for modal pushover analysis, which has been successfully
implemented in other studies for a variety of structural systems, such as by Gong
(2003) for multi-storey steel frames.

• Account for new generation PBD principles to establish performance objectives

based on risk due to earthquake-induced losses.
New generation PBD (ATC, 2005) provides for the establishment of performance
objectives in terms of losses, which are directly associated with cost.

• Account for the shear buckling mode of failure in cold-formed steel shear wall
panels.
Since some construction materials used as sheathing for CFS SWP are prone to
shear buckling, this mode of failure should be accounted for when determining the
lateral strength of CFS SWP.

144
 References
AISI. (2004). Standard for Cold-Formed Steel Framing-Lateral Design. American Iron

and Steel Institute, Washington, DC.

AISI. (1998). Shear Wall Design Guide. American Iron and Steel Institute, Technical

Data RG-9804, Washington, DC.

Argyris, J. H. (1982). An Excursion Into Large Rotations. Computer Methods In Applied

Mechanics and Engineering, Vol. 32(85-155).

ATC. (2005). Guidelines for Seismic Performance Assessment of Buildings, 25%

Complete Draft. Applied Technology Council; Redwood City, California.

ATC 58-2. (2003). Preliminary Evaluation of Methods for Defining Performance.

Bathe, K. J. (1996). Finite Element Procedures. Prentice Hall.

Bathe, K. J., and Bolourchi, S. (1982). A Geometric and Material Nonlinear Plate and

Shell Analysis. Computers and Structures; Vol. 11, pp. 23-48.

Bertero, R. D., and Bertero, V. V. (2002). Performance-based seismic engineering: the

need for a reliable conceptual comprehensive approach. Earthquake

Engineering and Structural Dynamics; Vol. 31, pp. 627–652.

Brandt, G. D. (1982). Rapid Determination of Ultimate Strength of Eccentrically Loaded

Bolt Groups. Engineering Journal, AISC.

Branston, A. E., Boudreault, F. A., Chen, C. Y. and Rogers, C. A. (2004). Light Gauge

Steel Frame / Wood Panel Shear Wall Test Data: Summer 2003. Research

145
 Report, Department of Civil & Applied Mechanics, McGill University,

Montreal, Canada.

Branston, A.E., Chen, C.Y., Boudreault, F.A., and Rogers, C.A., (2006). Testing of

Light-Gauge Steel Frame - Wood Structural Panel Shear Walls. Canadian

Journal of Civil Engineering, Vol. 33 No. 5, 573-587.

CANPLY. (2003). Plywood Design Fundamentals. Canadian Plywood Association;

Vancouver, BC.

Chopra, A. K., and Goel, R. K. (2002). A modal pushover analysis procedure for

estimating seismic demands for buildings. Earthquake Engineering and

Structural Dynamics; Vol. 31, pp. 561–582.

COLA-UCI. (2001). Report of a testing program of wall-framed walls with wood-

sheathed shear panels. Structural Engineers Association of Southern California,

COLA-UCI Light Frame Committee.

COMPLAY. (1999). Design Capacities of APA Performance Rated Structural – Use

Panels. The Engineered Wood Association; Tacoma, WA.

Cook, R. D, Malkus, D. S., and Plesha, M. E. (1989). Concepts and Applications of Finite

Element Analysis. John Wiley and Sons Ltd.

CSA S16-01.(2003). Handbook of Steel Construction. Canadian Institute of Steel

Construction. Toronto, Ontario.

ER-6151. (2003). Sure-board 200 series structural panels. ICC Evaluation Service, Inc.

Legacy report. Retrieved July 15, 2006, from http://www.icc-es.org/.

FEMA 273. (1997). NEHRP Guidelines for the Seismic Rehabilitation of Buildings.

Federal Emergency Management Agency.

146
FEMA 274. (1997). NEHRP Commentary on the Guidelines for the Seismic

Rehabilitation of Buildings. Federal Emergency Management Agency.

FEMA 450. (2003). NEHRP Recommended provisions for Seismic Regulations for new

Buildings and other Structures. Federal Emergency Management Agency.

Filiatrault, A., and Folz, B. (2002). Performance-Based Seismic Design of Wood Framed

Buildings. Journal of Structural Engineering, ASCE; Vol. 128 No. 1, pp. 39-47.

Fulop, L., and Dubina, D. (2004). Performance of wall-stud cold-formed shear panels

under monotic and cyclic loading Part I: Experimental research. Thin-Walled

Structures; Vol. 42, pp. 321-338.

Fulop, L., and Dubina, D. (2004b). Performance of wall-stud cold-formed shear panels

under monotonic and cyclic loading Part II: Numerical modelling and

performance analysis. Thin-Walled Structures; Vol. 42, pp. 339-349.

GA-235. (2002). Gypsum Board Typical Mechanical and Physical Properties. Gypsum

Association; Washington, DC.

Gad, E.F., Chandler, A. M., Duffield, and C.F., Stark, G. (1999). Lateral Behaviour of

Plasterboard-Clad residential Steel Frames. Journal of Structural Engineering,

ASCE; pp. 32-39.

Ghobarah, A. (2001). Performance-based deign in earthquake engineering: State of

development. Engineering Structures; Vol. 23, pp. 878-884.

Gong, Y. (2003). Perfomance-Based Design of Steel Building Frameworks Under

Seismic Loading. University of Waterloo, PhD Thesis.

147
Grierson, D.E., Xu, L., and Liu, Y. (2005). Progressive-Failure Analysis of Buildings

Subjected to Abnormal Loading. Computer-Aided Civil and Infrastructure

Engineering; Vol. 20, pp. 155-171.

Grierson, D.E., Gong, Y., and Xu, L. (2006). Optimal performance-based seismic design

using modal pushover analysis. Journal of Earthquake Engineering; Vol. 10,

pp. 73-96.

Hasan, R., Xu, L., and Grierson, D. E. (2002). Push-over analysis for performance-based

seismic design. Computers and Structures; Vol. 80, pp. 483-2493.

Krawinkler, H., and Senervitna, G. D. P. K. (1998). Pros and cons of a pushover analysis

of seismic performance evaluation. Engineering Structures; Vol. 20, pp. 42-464.

Lee K., and Foutch D. A. (2002). Performance evaluation of new steel frame buildings

for seismic loads. Earthquake Engineering and Structural Dynamics; Vol. 31,

pp. 653-670.

Lee Y.K. (1999). Behavior of Gypsum-Sheathed Cold-Formed Steel Wall Stud Panels.

Oregon State University, PhD Thesis.

Liew, Y. R., and Chen, H. (2004). Direct analysis for performance-based design of steel

and composite structures. Prog. Struct. Engng Mater; Vol. 6, pp. 213-228.

NAHBRC. (1997). Monotonic tests of cold-formed steel shear walls with openings.

Report prepared by NAHB Research Center, Inc. AISI; Washington, DC.

NBCC, (2005). National Building Code of Canada. Canadian Commission on Building

and fire Codes. National Research Council; Ottawa, Ontario.

Okasha, A. F. (2004). Performance of steel frame / wood sheathing screw connections

subjected to monotonic and cyclic loading. University of McGill, MSc Thesis.

148
OSB. (1995). OSB Design Manual, Design Rated Oriented Strand Board. Structural

Board Association, Willowdale, ON.

S136-01. (2002). North American Specification for the Design of Cold-formed Steel

Structural Members. Canadian Standards Association.

SAP2000. (2006). Integrated software for structural analysis and design. Computers and

Structures Inc; Berkeley, California.

SEAOC. (1995). VISION 2000 – A Framework for Performance Based Earthquake

Engineering. Structural Engineers Association of California; Vol. 1.

SEAOC. (1997). Standard Method of Cyclic (reversed) Load Test for Shear Resistance of

Framed Walls for Buildings. Structural Engineers Association of California;

Whittier, USA.

SEAOC and COLA-UCI. (2001). Report of a Testing Program of Light-Framed Walls

with Wood-Sheathed Shear Panels. Structural Engineers Association of

California, COLA-UCI Light Frame Test Committee, and Department of Civil

Eng. University of California, Irvine. Research award: FEMA-DR-1008-8011.

SODA (1999). Structural Optimization Design Analysis. Acronym Software, Inc.

Waterloo, On., Canada.

Serrette, R. L., Morgan, K. A., and Sorhouet, M. A. (2002). Performance of Cold-Formed

Steel-Framed Shear Walls: Alternative Configurations. Santa Clara University;

Final Report LGSRG-06-02.

Serrette, R. L, and Ogunfunmi, K. (1996). Shear Resistance of Gypsum-Sheathed Light-

Gauge Steel Stud Walls. Journal of Structural Engineering, ASCE; pp. 383-

389.

149
SSMA. (2001). Product technical information. Steel Stud Manufacturers Association,

Report ICBO ER-4943P. Retrieved January 10, 2004, from

http://www.ssma.com.

Sureboard (2004). Product technical information. Sure-Board, Intermat, Inc.

http://www.sureboard.com

Telue, Y.K., and Mahendran, M. (2001). Behaviour of cold-formed steel wall frames

lined with plasterboard. Journal of Constructional Steel Research; Vol. 57, pp.

435-452.

Thomas, W.H. (2004). Poisson’s ratios of an oriented strand board. Wood Science and

Technology, Springer Berlin / Heidelberg; Vol. 37, pp. 259-268.

Troitsky, M., S. (1976). Stiffened Plates, Bending, Stability and Vibrations. Elsevier

Scientific Publishing Company.

USGS. (2002). Interpolated Probabilistic Ground Motion for the Conterminous 48 States,

2002 Data. U.S. Geological Survey Earthquake Hazards program. Retrived

August 2, 2006, from http://earthquake.usgs.gov/research/hazmaps/.

Zalka, K. A. (2001). A simplified method for calculation of the natural frequencies of

wall-frame buildings. Engineering Structures; Vol. 23, pp. 1544-1555.

Zou, X. K., and Chan, C. M. (2005). Optimal Seismic Performance-Based Design of

Reinforced Building using Nonlinear Pushover Analysis. Engineering

Structures; Vol. 27, pp. 1289-1302.

150
 Appendix A
Simplified method of calculating the structure period of
vibration

Zalka (2001) developed a simplified method for calculating the natural period of
vibration of structures built with frameworks and shear walls. It is assumed that the
buildings have a uniform layout in all the storeys, and uniform geometry along their
height. The method transforms each frame, including shear walls in the direction under
consideration, into a column with equivalent mass and stiffness. Then, the natural period
is computed for the equivalent column. Three types of building deformation are
considered in the analysis: shear, full-height bending, and full-height bending of
individual columns.

Several modifications are needed to use Zalka’s method for CFS buildings. Instead of
computing the frequency for each plane frame, the frequency is computed for the entire
three-dimensional building simultaneously. This is accomplished by taking into account
all the elements in the building, instead of only a plane frame. The stiffness equations are
modified to include the stiffness of the SWP made of CFS. The method still requires that
the building plan distribution is the same for all the storeys. The procedure to determine
the period of vibration of CFS buildings with frame elements and SWP is described in the
following.

1. The global shear stiffness associated with the building beams and SWP is,

nco
Eb I b ,i n p Gs ,i AS
K b = ∑12 + ∑12 (A. 1)
i li h i 1 .2 h

151
where: i is the number of the beams; the parameters nco and np are respectively the
numbers of columns and SWP in the building; the parameters Eb and Ib are respectively
the modulus of elasticity and moment of inertia for the beams; l is the SWP length; h is
the storey height; Gs is the sheathing shear modulus of elasticity; and AS is the sheathing
cross-sectional area.

2. The columns’ shear stiffness is given by,

nco
Ec I c ,i
K c = ∑12 (A. 2)
i h2

where Ic is the column i moment of inertia, and Ec is the column modulus of

elasticity.

3. The stiffness combination factor is calculated as,

Kc
rc = (A. 3)
Kc + Kb

4. The building’s shear stiffness is,

K = rc K b (A. 4)

5. The building’s shear frequency is given by,

1 rf2 K
f s2' = (A. 5)
(4 H )2 m

where: H is the building height; m is the mass density per unit of the building height;
and rf is a reduction factor whose value depends on the number of storeys, (see Table
A.1).

6. The bending frequency associated with the building full height is defined as,

0.313rf2 Ec I g
f g2 = (A. 6)
H 4m

where

152
 nco
I g = ∑ Ac ,i d 2f ,i (A. 7)
i

in which Ac is the column cross-sectional area, and df is the distance from the column
centroid to the floor plan centroid.

7. The shear stiffness effectiveness factor is,

f g2
s =
2
(A. 8)
f s2' + f g2

8. The building effective shear stiffness is,

Ke = s2 K j (A. 9)

9. The building lateral frequency associated with shear deformation is,

2
1 rf K e
f =
2
(A. 10)
s
(4 H )2 m
10. The building equivalent bending stiffness, taking into account the columns and SWP,
is given by,

⎛ ⎞
EI = EI Columns + EI SWP = Ec I c r + ∑ ⎜ ESy I S + ∑ E F I F ⎟ (A. 11)
np ⎝ Studs ⎠

where ESy is the sheathing modulus of elasticity in the y direction such that

t pl 3
IS = (A. 12)
12

11. The frequency associated with the bending stiffness of the building is defined as,

0.313r f2 EI
f =
b
2
(A. 13)
H 4m

12. The non-dimensional parameter is,

Ke
kp = H (A. 14)
EI

153
13. The building natural frequency is calculated as,

⎛ η s2 k p2 1⎞
f = s⋅ f ⎜ 2
− − 1 + ⎟ + f s2 (A. 15)
⎜ 0.313 5
b
s ⎟⎠
⎝

where η is obtained from Table A.2.

14. The building lateral period of vibration is,

1
T = (A. 16)
f

If the CFS building being analyzed is built with SWP only, without frame elements, the
procedure of analysis is slightly different. In fact, the stiffness equations are the only
difference with respect to the procedure described in the foregoing for buildings that also
include frame elements. To determine the natural frequency of vibration of a building
having only SWP, calculate the following quantities.

1. The global shear stiffness associated with the building’s SWP is,
np
Gs ,i As
K b = ∑12 (A. 17)
i 1.2h

2. The building equivalent bending stiffness including columns and SWP is,

⎛ ⎞
EI = EI SWP = ∑ ⎜ ESy I S + ∑ EF I F ⎟ (A. 18)
np ⎝ Studs ⎠

3. The frequency associated with the bending stiffness of the building is given by,

0.313r f2 EI
f =
b
2
(A. 19)
H 4m

4. The building lateral frequency associated with shear deformation is computed as,
2
1 rf K b
f =
2
(A. 20)
s
(4 H )2 m

154
5. The non-dimensional parameter is,

Kb
kp = H (A. 21)
EI

6. The building lateral frequency is defined as,

⎛ η s2 k p2 1⎞
f = s ⋅ f ⎜⎜ b
2
− − 1 + ⎟⎟ + f s2 (A. 22)
⎝ 0.313 5 s⎠

7. The building lateral period of vibration is,

1
T = (A. 23)
f

Table A. 1 Reduction factor rf, Zalka (2001).

n1 1 2 3 4 5 6 7
rf 0.493 0.653 0.770 0.812 0.842 0.863 0.879
n1 8 9 10 11 12 13 14
rf 0.892 0.902 0.911 0.918 0.924 0.929 0.934
n1 15 16 18 20 19 20 21
rf 0.938 0.941 0.947 0.952 0.961 0.967 0.98

n > 50 n /(n + 2.06)

1
Building number of storeys

Table A. 2. Frequency parameter, Zalka (2001)

kp ηs kp ηs kp ηs kp ηs kp ηs
0.0 0.5596 4.5 1.465 9.5 2.680 14.5 3.913 20 5.278
0.1 0.5606 5.0 1.586 10.0 2.803 15.0 4.036 30 7.769
0.5 0.5851 5.5 1.706 10.5 2.926 15.5 4.160 40 10.26
1.0 0.6542 6.0 1.827 11.0 3.049 16.0 4.284 50 12.76
1.5 0.7511 6.5 1.949 11.5 3.172 16.5 4.408 60 15.26
2.0 0.8628 7.0 2.070 12.0 2.295 17.0 4.532 70 17.26

155
2.5 0.9809 7.5 2.192 12.5 3.418 17.5 4.656 80 20.26
3.0 1.1014 8.0 2.313 13.0 3.542 18.0 4.781 90 22.76
3.5 1.2226 8.5 2.435 13.5 3.665 18.5 4.905 100 25.26
4.0 1.3437 9.0 2.558 14.0 3.789 19.0 5.029 >100 kp/4

156
 Appendix B
Calibration of the αb and αv coefficients
The purpose of the coefficients αb and αv, introduced in Section 3.3, is to transform the
elastic stiffness of the sheathing to the secant stiffness at the ultimate load and
displacement level. Thus, the framing contribution to the SWP lateral strength is
accounted for at the ultimate limit state. The coefficients αb and αv have been calibrated
using experimental data obtained by Branston et al. (2006) from monotonic and cyclic
testing of SWP. The results reported by Branston for a SWP length of 1219mm (4 ft)
have been used in the calibration. The results for other SWP lengths have not been used
because the variation in length of the SWP is accounted by the parameter η, which is less
than unity for SWP lengths smaller than 1219 mm (4ft), and larger than unity otherwise.

Listed in Table B.1 are the SWP stiffnesses at the ultimate load level, K, derived from
experimental data (Branston et al. 2006). Also, presented in the table are the elastic shear
and bending stiffnesses of the SWP, Kv and Kb, respectively. The calibration of the
coefficients αb and αv has been carried out “manually” since there are three variables that
affect the stiffness reduction of the SWP, and the author did not have the appropriate
software to make the calibration on a computer. These variables are screw spacing sC,
number of screws nC, and ultimate strength factor Cu.

Since the experimental stiffness of the SWP at the ultimate load level is known, as is the
elastic stiffness of the SWP, several equations which account for the variables that affect
the stiffness of the SWP were proposed. Then, it was found that the equations that fit best
the experimental data are Eqs. (3.10) and (3.11) repeated below,
1.8
⎛ Cu ⎞ ⎛ 6in ⎞
αV = ⎜⎜ ⎟⎟ ⋅ ⎜⎜ ⎟⎟
⎝ 3.3 ⋅ nC ⎠ ⎝ C ⎠s

157
 1.3⋅nC
2
⎛ 6 ⎞ ⎛ 6in ⎞ Cu
α B = ⎜⎜ ⎟⎟ ⋅ ⎜⎜ ⎟⎟
⎝ Cu ⎠ ⎝ s C ⎠

Presented in the fist column of Table B.1 is a description of the specimens tested by
Branston (2004), in the second to the fourth columns are the SWP experimental lateral
resistance, maximum displacement and ultimate lateral stiffness. The Kv and Kb values in
the table are the elastic shear and bending stiffnesses of the SWP, computed from
principles of mechanics. The ultimate lateral stiffness of the SWP, KS , is computed by
multiplying the elastic stiffnesses Kv and Kb by the coefficients αb and αv, respectively, as
described for Eq. (3.8). Listed in the last column is the ratio of the experimental and
predicted ultimate lateral stiffness. Also given are statistical results obtained, such as
average of the predicted-to-experimental ratio, the standard deviation and the correlation.

Table B.1 Calibration of coefficients αb and αv using experimental data (Branston, 2004)
Assembly maximum:
Resistance Disp. K Cu nC sC αv αb Kv Kb KS KS/K
Specimen kN/m mm kN/mm kN/mm
1 – A,B,C CSP 16.6 60.6 0.27 39.04 74 102 0.06 0.06 2.59 2.88 0.33 1.20
5 – A,B,C,D DFP 23.8 60.6 0.39 39.04 74 102 0.06 0.06 4.30 4.08 0.50 1.27
7 – A,B,C CSP 12.7 50.7 0.25 26.19 50 152 0.04 0.05 2.59 2.88 0.25 0.98
9 – A,B,C CSP 25.1 61 0.41 51.89 98 76 0.07 0.07 2.59 2.88 0.40 0.98
11 – A,B,C DFP 16 54.8 0.29 26.19 50 152 0.04 0.05 4.30 4.08 0.37 1.27
13 – A,B,C2 DFP 29.7 58.2 0.51 51.89 98 76 0.07 0.07 4.30 4.08 0.62 1.21
21 – A,B,C OSB 13.2 41.1 0.32 26.19 50 152 0.04 0.05 4.24 3.41 0.33 1.04
23 – A,B,C OSB 19.3 39.5 0.49 39.04 74 102 0.06 0.06 4.24 3.41 0.45 0.93
25 – A,B,C OSB 23.5 40.7 0.58 51.89 98 76 0.07 0.07 4.24 3.41 0.56 0.98
4 – A,B,C CSP 17.5 -15.3 0.33 39.04 74 102 0.06 0.06 2.59 2.88 0.33 1.00
6 – A,B,C DFP 22.6 -19.6 0.41 39.04 74 102 0.06 0.06 4.30 4.08 0.50 1.21
8 – A,B,C CSP 11.9 -10.6 0.26 26.19 50 152 0.04 0.05 2.59 2.88 0.25 0.96
10 – A,B,C CSP 26.2 -23.1 0.51 51.89 98 76 0.07 0.07 2.59 2.88 0.40 0.78
12 – A,B,C DFP 14.6 -13.4 0.31 26.19 50 152 0.04 0.05 4.30 4.08 0.37 1.20
14 – A,B,C,D DFP 29.7 -26.2 0.53 51.89 98 76 0.07 0.07 4.30 4.08 0.62 1.17
22 – A,B,C OSB 11.7 -10.5 0.31 26.19 50 152 0.04 0.05 4.24 3.41 0.33 1.06
24 – A,B,C OSB 17.2 -15.7 0.51 39.04 74 102 0.06 0.06 4.24 3.41 0.45 0.89
26 – A,B,C OSB 23.5 -22.4 0.67 51.89 98 76 0.07 0.07 4.24 3.41 0.56 0.85
Average 1.05
Standard Deviation 0.15
Correlation 0.85

158
 Appendix C
Calibration of the β exponent

In the stiffness degradation model presented in Section 4.6, the loss of lateral stiffness
of the SWP is determined as the ratio of the lateral loads applied on the SWP over their
lateral strength. Since the loss of stiffness of the SWP is not linear, the exponent β has
been introduced in Eq. 4.52 to characterize the nonlinear loss of SWP stiffness. It has
been observed in experimental tests (Rogers, 2004) that the nonlinear stiffness
degradation of SWP depends on several factors, the most important being the sheathing
material and the screw spacing on the edge of the panel. However, the parameter β is
calibrated taking into account the screw spacing only. The stiffness degradation model
is intended to be used for any type of wood sheathing, so the exponent β is calibrated
using the lowest force-displacement values obtained from the experimental testing for
different types of wood sheathing materials.

Since the ratio of lateral force over lateral strength (Pa/PR) in Eq. 4.52 is linear, the first
condition that β must satisfy to characterize the nonlinear behaviour similar to that
obtained from the experimental results is that it must be less than unity (otherwise the
shape of the curve would be convex). Another condition is that β must decrease as the
screw spacing increases because, as observed in the experimental data, the “radius” of
the force-displacement curve becomes smaller as the screw spacing increases. With
these two conditions in mind, it is determined that the screw spacing must be a
reciprocal parameter in the equation for β. Thus β can be expressed as,
β = x / sC (C. 1)

where sC is the screw spacing on the edge of the panel, and x is a constant to be found
from the experimental results. Depicted in Figures C.1, C.2, and C.3 are the
experimental plots for the normalized force-displacement curves for screw spacings of
76 mm (3 in), 101 mm (4 in), and 152 mm (6 in), respectively. Then, different values
for the constant x were tested, and the value that fits best with the experimental results

159
was found to be x = 38.1 mm (1.5 in). Figures C.1, C.2 and C.3 also show the
normalized force-displacement curve predicted for β.

The experimental data used to calibrate β accounted for SWP lengths of 609 mm (2 ft),
1219 mm (4 ft), and 2438 mm (8 ft), and height of 2438 mm (8ft). Also, the three types
of sheathing OSB, CSP and DFP were accounted for. For all SWP, the sheathing was
attached to the framing with No. 8 screws.

9A 9B 9C 13A 13B 13C 25A

25B 25C 33A 33B 33C Beta sc=3in

1.200

1.000

0.800
Normalized Lateral Force

0.600

0.400

0.200

0.000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Displacement

Figure C.1 Experimental (Branston et al., 2004) vs. predicted SWP response for screw
spacing sC=76 mm (3 in)

160
 1A 1B 1C 5A 5C 5D 17A
17B 17C 23A 23B 23C 27A 27B
27C 31A 31B 31C Beta sc=4 in

1.2

0.8
Normalized Lateral Force

0.6

0.4

0.2

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Displacement

Figure C.2 Experimental (Branston et al., 2004) vs. predicted SWP response for screw
spacing sC=101 mm (4 in)
7A 7B 7C 11A 11B 11C 15C 19A 19B
19C 21A 21B 21C 29A 29B 29C LAMB

1.2

0.8
Normalized Lateral Force

0.6

0.4

0.2

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Displacement

Figure C.3 Experimental (Branston et al., 2004) vs. predicted SWP response for screw
spacing sC=152 mm (6 in)

161