Bracing S

PACIFIC EARTHQUAKE ENGINEERING
RESEARCH CENTER
PEER 2010/104
APRIL 2010
PACIFIC EARTHQUAKE ENGINEERING
Simulating the Inelastic Seismic Behavior of
Steel Braced Frames
Including the Effects of Low-Cycle Fatigue
Yuli Huang
and
Stephen A. Mahin
University of California, Berkeley
SimuIating the IneIastic Seismic Behavior
of SteeI Braced Frames
IncIuding the Effects of Low-CycIe Fatigue
YuIi Huang
Department of Civil and Environmental Engineering
Stephen A. Mahin
Department of Civil and Environmental Engineering
PEER Report 2010/104
Pacific Earthquake Engineering Research Center
College of Engineering
April 2010
ABSTRACT
This report describes simulations oI the inelastic seismic behavior oI steel braced Irames, including
the eIIects oI low-cycle Iatigue. For steel braced Irames under complex loading conditions, a wide
variety oI behavior mechanisms and Iailure modes may occur Ior each type oI member and con-
nection. Thus, numerical models that assess the initiation and propagation oI Iailure during cyclic
loading need to account Ior multi-axial states oI material nonlinearity, local and global buckling,
and the exhaustion oI the ability oI the material to deIorm inelastically caused by low-cycle Iatigue.
Following a review oI existing material models Ior simulating structural steel deterioration,
a series oI investigations are conducted using Iinite element modeling techniques. A new, numeri-
cally eIIicient continuum damage mechanics material model capable oI simulating inelastic behav-
ior and deterioration oI mechanical properties because oI low-cycle Iatigue has been devised and
implemented in a Iinite element soItware LS-DYNA. Computational results obtained with this new
material model correlate well with test results Ior several beam-to-column connections, individual
braces, and braced Irame subassemblies. Recommendations Ior characterizing material properties
Ior these types oI analyses are developed and presented. A series oI analyses are presented that
evaluate and reIine several requirements Ior detailing and analyzing special concentrically braced
steel Irame buildings.
iii
ACKNOWLEDGMENTS
The work perIormed in this investigation was Iunded in part by the Consortium oI Universities Ior
Research in Earthquake Engineering, as part oI the CUREE-Kajima Joint Research Program on
Earthquake Engineering. The authors grateIully acknowledge the Iunding Ior this project Irom the
Kajima Corporation, Japan, and the leadership oI the Joint Oversight Committee in helping deIine
the direction and scope oI the research eIIorts undertaken. Funding was also provided by the Na-
tional Science Foundation under Grants CMS-0600625 and CMS-0619161, and by the University
oI CaliIornia at Berkeley through the Byron and Elvira Nishkian Chair ProIessorship oI Structural
Engineering.
The authors wish to acknowledge the collaboration oI Dr. Yoshikazu Sawamoto oI Kajima
Corporation on this work through sharing oI test data, providing advice, and Ior thoroughly review-
ing the models and the results obtained. The assistance oI Dr. Patxi Uriz in providing data on braced
Irames and helping interpret prior test results is grateIully acknowledged. The authors appreciate
the assistance oI Dr. John Hallquist, Dr. Leonard Schwer, and Mr. Jim Day oI Livermore SoItware
Technology Corporation.
The Iindings, conclusions, and observations contained herein are those oI the authors. They
do not necessarily represent the policy, recommendations, or Iindings oI the Consortium oI Uni-
versities Ior Research in Earthquake Engineering, the Kajima Corporation, the National Science
Foundation, or the University oI CaliIornia at Berkeley.
iv
CONTENTS
ABSTRACT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
ACKNOWLEDGMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
TABLE OF CONTENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
LIST OF TABLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Objectives and Scope oI Research . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 REVIEW OF EXISTING EXPERIMENTAL INVESTIGATIONS . . . . . . . . . . . . . . . . . 9
2.1 Investigations oI Conventional Braces . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Investigations oI Gusset Plates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Investigations oI Braced Frame Subassemblies . . . . . . . . . . . . . . . . . . . 13
2.4 Investigations oI Beam-Column Connections . . . . . . . . . . . . . . . . . . . . 15
2.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3 STRUCTURAL STEEL DETERIORATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.1 Plastic Behavior and Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2 Damage and Fracture under Monotonic Loading . . . . . . . . . . . . . . . . . . . 29
3.2.1 Local Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2.2 Global Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2.3 Relation between Local and Global Approaches . . . . . . . . . . . . . . . 31
3.3 Damage and Fracture under Cyclic Loading . . . . . . . . . . . . . . . . . . . . . 32
3.3.1 Manson-CoIIin Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3.2 Continuous Damage Models . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3.3 Simple Comparison oI Continuous Damage Models . . . . . . . . . . . . . 35
4 CYCLIC DAMAGE PLASTICITY MODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
v
4.1 Theory oI Plasticity and Damage Evolution . . . . . . . . . . . . . . . . . . . . . 39
4.1.1 Strain Rate Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.1.2 Elastic Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.1.3 Plastic Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.1.4 Nonlinear Isotropic/Kinematic Hardening Model . . . . . . . . . . . . . . 41
4.1.5 Damage Evolution Model . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.2 Implementation oI the Plasticity Model . . . . . . . . . . . . . . . . . . . . . . . 42
4.2.1 Backward Euler DiIIerence Scheme . . . . . . . . . . . . . . . . . . . . . 42
4.2.2 Elastic-Plastic Operator Split . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2.3 Cutting-Plane Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2.4 Closest-Point Projection Algorithm . . . . . . . . . . . . . . . . . . . . . 45
4.3 Implementation oI the Damage Model . . . . . . . . . . . . . . . . . . . . . . . . 47
4.4 Time-Integration Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.4.1 Integration Procedure Ior Brick Elements . . . . . . . . . . . . . . . . . . 48
4.4.2 Integration Procedure Ior Shell Elements . . . . . . . . . . . . . . . . . . 50
5 CALIBRATION, VALIDATION, AND APPLICATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.1 Material Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.1.1 Calibration Using Monotonic Data . . . . . . . . . . . . . . . . . . . . . . 51
5.1.2 Calibration Using Cyclic Data . . . . . . . . . . . . . . . . . . . . . . . . 52
5.2 Validation and Application to Components . . . . . . . . . . . . . . . . . . . . . . 64
5.2.1 Single Brace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.2.2 Single Beam-Column Connection . . . . . . . . . . . . . . . . . . . . . . 74
5.3 Validation and Application to Subassemblies . . . . . . . . . . . . . . . . . . . . . 82
5.3.1 Subassembly Subjected to Quasi-Static Loading . . . . . . . . . . . . . . 82
5.3.2 Subassembly Subjected to Dynamic Loading . . . . . . . . . . . . . . . . 85
6 STUDIES OF STEEL BRACED FRAME BEHAVIOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.1 Brace Proportions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.2 Connection Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
vi
6.3 Lateral Bracing Ior Beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
6.4 Estimate oI Interstory DriIt Demands . . . . . . . . . . . . . . . . . . . . . . . . . 124
7 CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
REFERENCES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
APPENDIX A: APPLYING PROPORTIONAL LOADS USING CONSTRAINTS. . . . . A.1
A.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.1
A.2 Experiments with Proportional Loading . . . . . . . . . . . . . . . . . . . . . . . A.2
A.3 Numerical Proportional Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . A.4
A.4 Linear Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.6
A.5 Nonlinear SoItening Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.7
A.5.1 Nonlinear Springs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.8
A.5.2 Biaxial Tension oI a Shell Element with Nonlinear Material . . . . . . . . A.10
A.6 Practical Example Ior Pushover Analysis . . . . . . . . . . . . . . . . . . . . . . . A.11
A.7 A Special Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.14
A.8 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.15
APPENDIX B: DISCUSSION ON INSTABILITY OF NEWMARK INTEGRATOR. . . B.1
B.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.1
B.2 Linear Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.1
B.3 Demonstration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.3
B.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.4
B.4.1 Source oI Initial Numerical Error . . . . . . . . . . . . . . . . . . . . . . B.4
B.4.2 Other Values oI Parameter . . . . . . . . . . . . . . . . . . . . . . . . . B.5
B.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.5
vii
LIST OF FIGURES
1.1 Schematic steel building comprising braced and moment-resisting Irames . . . . . . . 5
2.1 Experimental setup Ior conventional brace test (Yang and Mahin, 2005) . . . . . . . . 17
2.2 Gusset plate damage (Roeder et al., 2006) . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3 Test setup Ior SCBF (Uriz, 2005) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Target rooI displacement history oI SCBF test (Uriz, 2005) . . . . . . . . . . . . . . . 19
2.5 Brace buckling (Uriz, 2005) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.6 Brace Iailure (Uriz, 2005) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.7 Beam-column connection Iailure (Uriz, 2005) . . . . . . . . . . . . . . . . . . . . . . 22
2.8 Base shear versus rooI lateral displacement (Uriz, 2005) . . . . . . . . . . . . . . . . 22
2.9 Test setup Ior BW10 beam-column connection (Tanaka et al., 2000) . . . . . . . . . . 23
2.10 Test setup Ior N4 beam-column connection (Suita et al., 2000) . . . . . . . . . . . . . 23
3.1 Void accumulation in a tensile specimen (Puttick, 1959) . . . . . . . . . . . . . . . . . 27
3.2 Void nucleation, growth, and coalescence in ductile metals . . . . . . . . . . . . . . . 27
3.3 Nonlinear Bauschinger eIIect and cyclic hardening . . . . . . . . . . . . . . . . . . . 28
3.4 Ratcheting under asymmetric stress history . . . . . . . . . . . . . . . . . . . . . . . 28
3.5 Equivalent plastic strain to Iailure versus stress triaxiality . . . . . . . . . . . . . . . . 31
3.6 Damage evolution oI CVGM model . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.7 Damage evolution oI CDM model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.8 Manson-CoIIin plot oI CVGM model . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.9 Manson-CoIIin plot oI CDM model . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.1 Illustration oI kinematic hardening parameters . . . . . . . . . . . . . . . . . . . . . . 52
5.2 Experimental versus numerical results (SM490 cyclic test 1) . . . . . . . . . . . . . . 55
ix
5.6 Experimental versus numerical results (SS440 cyclic test 1) . . . . . . . . . . . . . . . 59
5.10 Experimental versus numerical results (monotonic loading) . . . . . . . . . . . . . . . 63
5.11 Global buckling oI brace specimen 5 (aIter Yang and Mahin, 2005) . . . . . . . . . . . 67
5.12 Local buckling oI brace specimen 5 (aIter Yang and Mahin, 2005) . . . . . . . . . . . 67
5.13 Crack initiation oI brace specimen 5 (aIter Yang and Mahin, 2005) . . . . . . . . . . . 68
5.14 Crack propagation oI brace specimen 5 (aIter Yang and Mahin, 2005) . . . . . . . . . 68
5.15 Hysteresis loops oI brace specimen 5 (aIter Yang and Mahin, 2005) . . . . . . . . . . 69
5.18 Peak loads oI brace specimen 5 (aIter Yang and Mahin, 2005) . . . . . . . . . . . . . . 72
5.21 Evolution and mesh sensitivity oI plastic strain (shell thickness t) . . . . . . . . . . . . 73
5.22 Finite element modeling oI BW10 specimen . . . . . . . . . . . . . . . . . . . . . . . 75
5.23 Hysteretic loop oI BW10 specimen . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.24 BW10 behavior comparison 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.27 Finite element modeling oI N4 specimen . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.28 Hysteretic loop oI N4 specimen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.29 Predicted local buckling and damage . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.30 Predicted Iracture in top Ilange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.31 Damage and Iracture oI brace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.32 Damage and Iracture oI beam-column connection . . . . . . . . . . . . . . . . . . . . 83
5.33 Experimental result Ior braced Irame . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.34 Numerical result Ior braced Irame . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.35 Subassembly model and ground motion . . . . . . . . . . . . . . . . . . . . . . . . . 86
x
5.36 Failure mode at time t 13.8 sec, global view . . . . . . . . . . . . . . . . . . . . . . 86
5.37 Failure mode at time t 13.8 sec, local view 1 . . . . . . . . . . . . . . . . . . . . . . 87
5.38 Failure mode at time t 13.8 sec, local view 2 . . . . . . . . . . . . . . . . . . . . . . 87
5.39 Base shear versus Iirst-story driIt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5.40 First-story driIt time history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
6.1 DeIormation history Ior reIerence brace . . . . . . . . . . . . . . . . . . . . . . . . . 100
6.2 Interstory driIt ratios to Iailure versus width-thickness ratios . . . . . . . . . . . . . . 100
6.3 InIluence oI width-thickness ratios b/t (KL/r = 51) . . . . . . . . . . . . . . . . . . . 101
6.4 InIluence oI slenderness ratio KL/r (b/t = 14.2) . . . . . . . . . . . . . . . . . . . . 102
6.5 InIluence oI number oI cycles per deIormation level . . . . . . . . . . . . . . . . . . . 103
6.6 InIluence oI tension/compression asymmetry ratio . . . . . . . . . . . . . . . . . . . . 104
6.7 DeIormation histories (Tremblay and Bouatay, 2002) . . . . . . . . . . . . . . . . . . 105
6.8 Critical width-thickness ratios b/t Ior east crustal events at distance . . . . . . . . . . 106
6.9 Critical width-thickness ratios b/t Ior west crustal events at distance . . . . . . . . . . 107
6.10 Critical width-thickness ratios b/t Ior west near-Iield events . . . . . . . . . . . . . . 108
6.11 Critical width-thickness ratios b/t Ior west subduction events . . . . . . . . . . . . . . 109
6.12 Damage evolution Ior diIIerent strengthening methods . . . . . . . . . . . . . . . . . 112
6.13 Details oI column: reIerence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.14 Details oI column: reduced depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.15 Details oI column: thickened web . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6.16 Details oI column: thickened Ilange . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6.17 Details oI reinIorcement: reIerence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
6.18 Details oI reinIorcement: stiIIener . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
6.19 Details oI reinIorcement: reinIorcing plates . . . . . . . . . . . . . . . . . . . . . . . 116
6.20 Details oI reinIorcement: T-shear tab . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
6.21 Details oI reinIorcement: welded Ilanges . . . . . . . . . . . . . . . . . . . . . . . . . 117
6.22 Details oI reinIorcement: stiIIener and welded Ilanges . . . . . . . . . . . . . . . . . . 117
6.23 Schematics oI two lateral bracing conIigurations . . . . . . . . . . . . . . . . . . . . . 120
6.24 Comparison oI Iirst-story hysteretic loops . . . . . . . . . . . . . . . . . . . . . . . . 120
xi
6.25 Damage modes Ior driIt to the right . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.26 Damage modes Ior driIt to the leIt . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6.27 Detail used to induce in-plane buckling oI braces . . . . . . . . . . . . . . . . . . . . 123
6.28 Analysis result oI in-plane buckling oI braces . . . . . . . . . . . . . . . . . . . . . . 123
6.29 Model building Iloor plan and braced Irame elevations . . . . . . . . . . . . . . . . . 125
6.30 Finite element model and reIined regions oI the building . . . . . . . . . . . . . . . . 125
6.31 Conversion to nth-mode SDF system . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.32 Monotonic and cyclic pushover curves Ior Mode 1 . . . . . . . . . . . . . . . . . . . 129
6.33 Interstory driIts by elastic spectral and response history analysis . . . . . . . . . . . . 130
6.34 Interstory driIts by modal pushover and response history analysis . . . . . . . . . . . . 131
A.1 WhiIIle-Tree load system (Harris and Muskivitch, 1980) . . . . . . . . . . . . . . . . A.3
A.2 Articulated system oI the WhiIIle-Tree . . . . . . . . . . . . . . . . . . . . . . . . . . A.4
A.3 Interpolation relation Ior displacements . . . . . . . . . . . . . . . . . . . . . . . . . A.5
A.4 A linear example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.6
A.5 Nonlinear springs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.8
A.6 Load versus displacement Ior nonlinear springs . . . . . . . . . . . . . . . . . . . . . A.9
A.7 Displacement histories oI nonlinear springs . . . . . . . . . . . . . . . . . . . . . . . A.9
A.8 Load histories oI nonlinear springs . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.10
A.9 Biaxial tension oI a shell element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.11
A.10 Strain histories oI nonlinear material . . . . . . . . . . . . . . . . . . . . . . . . . . . A.12
A.11 Stress histories oI nonlinear material . . . . . . . . . . . . . . . . . . . . . . . . . . . A.12
A.12 Pushover analysis model oI braced Irame . . . . . . . . . . . . . . . . . . . . . . . . A.13
A.13 Story shears versus rooI displacement oI braced Irame . . . . . . . . . . . . . . . . . A.13
A.14 A two-spring system and a 'dummy spring . . . . . . . . . . . . . . . . . . . . . . . A.15
B.1 System Ior demonstration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.3
B.2 Accelerations at Node 2 (undamped, = 1/2, = 1/4) . . . . . . . . . . . . . . . . . B.6
B.3 Accelerations at Node 3 (undamped, = 1/2, = 1/4) . . . . . . . . . . . . . . . . . B.7
B.4 Accelerations at Node 2 (damped, = 1/2, = 1/4) . . . . . . . . . . . . . . . . . . B.8
B.5 Accelerations at Node 3 (damped, = 1/2, = 1/4) . . . . . . . . . . . . . . . . . . B.9
xii
B.6 Accelerations at Node 2 ( = 1/2, = 1/4 + 0.01) . . . . . . . . . . . . . . . . . . . B.10
B.7 Accelerations at Node 3 ( = 1/2, = 1/4 + 0.01) . . . . . . . . . . . . . . . . . . . B.11
xiii
LIST OF TABLES
3.1 Predictions oI Iailure Ior an ideal Manson-CoIIin experiment . . . . . . . . . . . . . . 36
6.1 Numerical experiment matrix Ior brace proportion . . . . . . . . . . . . . . . . . . . . 99
6.2 Numerical results oI 'pass or Iail type loading histories . . . . . . . . . . . . . . . . 99
6.3 Elastic modal properties oI the three-story SCBF subassembly . . . . . . . . . . . . . 128
xv
1 Introduction
1.1 BACKGROUND
Engineers Irequently use various types oI computer analysis to assess the saIety oI engineered
structures. SimpliIied analysis methods are oIten employed in the design oI standard structures.
Modeling guidelines and acceptance criteria used in conjunction with such simpliIied methods are
intended to provide acceptable conservatismrelative to the saIety and serviceability oI the structure.
In some cases, however, more reIined analysis methods are required. Such reIined analysis is used
in situations where:
1. Existing structures are to be evaluated Ior loading or environmental conditions more severe
than initially considered in design. SimpliIied analysis methods not directly related to the
expected behavior oI as-built structural elements may not be able to quantiIy adequately the
true capability oI a structure to withstand more severe loading conditions, raising questions
regarding the need to retroIit or upgrade a structure Ior the new criteria.
2. An existing structure is to be architecturally or Iunctionally remodeled, requiring removal
or shiIting oI structural members, changing the overall size oI the building, or altering the
type and conIiguration oI the structural system.
3. Newstructural systems employ unusual Ieatures, details, or conIigurations where simpliIied
analysis methods may not be adequate or where the incorporation oI these new structural
elements introduce signiIicant uncertainties regarding the adequacy oI the proposed design.
Modern architecture oIten does not conIorm to past practices regarding gravity and lateral
load resisting systems and tends to incorporate unusual details or member proportions.
4. A precise evaluation oI the saIety or perIormance oI a structure is required Ior hypothetical
man-made or natural hazards. Today, various stakeholders (owners, occupants, government
oIIicials, insurance companies, Iinancial institutions) are requesting assessment oI the saIety
and the probable losses that might occur as a result oI a variety oI natural and man-made
hazards.
In cases such as these, commonly used simpliIied analysis methods may not be able to
give a true picture oI actual damage or losses. Thus, a more complex analysis is needed that can
simulate the various behavior modes that can occur at the element and structural level and track the
evolution oI damage Irom onset to eventual member and structural Iailure. Because several modes
oI behavior can potentially occur in all members and connections, the models must simulate to a
Iair degree oI accuracy those modes, and correctly identiIy those modes that control behavior and
ultimate Iailure.
Although this work Iocuses on the behavior oI components, it is recognized that the local
Iailure oI a part oI one member may not lead to the loss oI structural integrity oI that member or
oI the structure as a whole. Where collapse does occur, it may involve one or several local regions
oI the structure. In some cases, the Iailure oI one element may be critical to the overall stability oI
the structure. The complete structure, or a substantial portion oI the structure above (and below)
the element or region that initially Iails, may collapse. Situations where a local Iailure leads to
disproportionate damage or collapse oI large portions oI a structure are known as 'progressive
collapse.
Predicting the onset oI damage within a large and complex structure, and the consequence
oI such local damage on other elements and the structure as a whole, requires analytical models that
simulate to a Iair degree oI accuracy the behavior oI structural elements and systems undergoing
large inelastic deIormations. As such, available analysis procedures and models will be reviewed
and evaluated to assess their ability to predict the behavior and Iailure oI elements and structural
systems.
For structures subjected to earthquakes, material models need to account Ior cyclic plas-
ticity, including deterioration and eventual Iailure because oI low-cycle Iatigue. The deterioration
oI a member may alter not only the distribution oI Iorces and damage within a structural system,
but also the response history. Because response history inIluences the degree oI deterioration, ac-
curate material and component models that predict the response oI structures subjected to severe
earthquake shaking or other transient dynamic excitations are critical.
2
In seismic-resistant design, equivalent static or dynamic analysis based on elastic repre-
sentations oI member properties are commonly employed, even though signiIicant inelastic action
is expected. Engineers are increasingly interested in predicting inelastic response using nonlinear
static or nonlinear dynamic analysis procedures. For nonlinear static methods, a realistic structural
model is subjected to a Iixed distribution oI lateral Iorces, which monotonically increase in intensity
with time. Such static inelastic 'pushover analyses, Iairly common in design practice, provide a
design engineer with a general indication oI the distribution oI inelastic deIormations (and internal
Iorces) within the structure, the global load capacity oI the structure, and the lateral displacement
at which certain key events occur (initial yielding, initiation oI member Iailure, loss oI global sta-
bility, etc.). Nonlinear dynamic time history analysis are computationally more demanding, but
provide a more realistic prediction oI the response quantities oI interest.
Nonlinear time history analysis oI structures during design usually assume ideal ductile
member behavior. The computer model estimates global response quantities (interstory driIt, story
shears, etc.) and locally required plastic deIormation demands (plastic hinge rotations, plastic
strains, etc.) based on ideal ductile behavior. Although this assumption is quite broad, it is generally
accepted because the elements and connections can subsequently be detailed to develop predicted
inelastic demands, or iI the resulting details are objectionable or the structure does not meet accep-
tance criteria Ior the overall system behavior, the conIiguration, the proportions, and the details oI
the structural system and elements can be changed.
Members and connections in new and especially existing structures have Iinite ductility
capacities, however, and because a certain degree oI deterioration oI properties might be expected
generally, inIinitely ductile models may not be adequate Ior seismic or other abnormal loading
conditions, as inelastic demands will likely lead to deterioration and possible Iailure oI members.
These weakened members will, in turn, inIluence dynamic response and overall system stability.
In order to assess behavior oI an element or structure as it approaches Iailure under earth-
quake or other excitations, several types oI nonlinear behavior need to be considered. These relate
to material inelasticity, Iracture, low-cycle Iatigue, and local and global geometric nonlinearities.
In this research, the eIIects oI suddenly started, quasi-brittle Iracture are not considered.
Where issues oI Iracture mechanics need to be taken into account, a large number oI small-sized
Iinite elements are required. Such analysis models and methods do not lend themselves to the anal-
3
ysis oI a complete structural systemunder dynamic loading. Thus, it is assumed in this research that
a separate Iracture-mechanics-based analysis should be carried out on individual Iracture critical
regions. Where behavior is Iound to be vulnerable to quasi-brittle or ductile Iracture, this report
assumes that (1) a simpliIied material or damage model is devised that would mimic the mode oI
Iracture detected or (2) problematic details would be changed to mitigate the vulnerability. Re-
gardless, the research presented herein includes scenarios where members rupture because oI the
materials reaching and exceeding their ability to develop Iurther inelastic deIormations, either under
monotonic or cyclic loading. Such ruptures may appear similar to and have similar consequences
as Iracture, but they are diIIerent physical phenomena and require diIIerent analysis approaches.
Steel braced Irames are the Iocus oI this investigation. For such structures (see the Irame
designated as a braced Irame in Fig. 1.1), lateral load resistance is, in large part, because oI the
braced Irame acting as a vertically oriented truss. The braces, columns, beams, and connections are
subjected to signiIicant axial loads. Braces are generally expected to respond in the inelastic range
during moderate and severe earthquakes. At displacement levels associated with brace yielding or
buckling, signiIicant bending and shear demands can also develop in all members and connections.
Under these complex-loading conditions, a wide variety oI behavior mechanisms and Iailure modes
must be taken into account Ior each type oI member and connection. The particular behavior that
occurs and the consequence oI its occurrence depend on material properties, member proportions,
local details, and the applied loading or deIormations (which will depend not only on the externally
applied loads, but also on the behavior oI adjacent members and connections). Thus, models that
assess the potential Ior Iailure to propagate throughout the structure need to account Ior multi-axial
states oI material nonlinearity, local and global buckling, and low-cycle Iatigue.
This investigation considers elements that make up typical concentrically braced steel Irames.
These include conventional bracing members that may buckle laterally and locally during loading
in compression. Because braced Irames are oIten used as part oI a dual system, some attention will
be given to the modeling oI conventional steel beam-to-column connections. Additional issues
related to modeling and analysis in moment-resisting Irame components and connections are the
Iocus oI a parallel study carried out at StanIord University (Lignos and Krawinkler, 2007).
Generally speaking, three basic approaches will be taken to modeling structural members
oI the type examined in this report. These are reIerred to as (1) phenomenological models, (2)
4
Moment Frame Braced Frame
Rigid Beam-Column
Connection
Beam
Column
Pinned connection
Brace
Gusset plate
connection
Figure 1.1 Schematic steel building comprising braced and moment-resisting frames
physical theory models, and (3) Iinite element models.
Phenomenological Models. Phenomenological models use a simpliIied physical repre-
sentation oI the member in question. For example, a phenomenological model Ior a brace
usually consists oI a simple uniaxial, pin-ended truss element with a hysteretic axial Iorce-
axial deIormation relation deIined by a set oI generic rules that simply mimic the hysteretic
response observed Ior a brace. The user is responsible Ior inputting data that enables the
generic brace model to represent the particular design being analyzed. In the case oI a Ilex-
urally dominated beam, a common phenomenological model represents the member as an
elastic beam connected at each end to the element`s nodes by means oI inelastic rotational
springs. The rotational springs are simpliIied in terms oI rules that deIine the hysteretic
nature oI the moment-rotation characteristics oI the plastic hinge region at the end oI the
member. To model the eIIects oI spalling oI concrete, and yielding, buckling, and Iracture
oI steel, more and more complex rules are employed. Generally, these rules do not depend
on or attempt to track the actual occurrence oI concrete spalling, or steel yielding, buckling,
5
or Iracture. These models depend on having an extensive database oI experimental results
with which the numerical modeling parameters can be identiIied. Where various behav-
ior modes can occur (bending plus axial buckling, Ilexure plus shear, or Ilexure plus axial
loading) the rules become even more complex or are uncoupled, leading to questionable
results.
Physical Theory Models. In many cases, member behavior can be well represented by
certain established but simpliIied rules oI mechanics. For example, in a beam with low
axial load and shear, plastic hinges occur at easily identiIied locations. Outside oI these
regions, this beam will behave in a nearly elastic manner. Within the plastic hinge region,
the material may yield and undergo various actions that will aIIect its properties. In such
cases, it may be acceptable to make certain assumptions regarding the distribution oI strains
across a section and the distribution oI curvatures along the member, and track the actual
stress-strain history in individual material Iibers located across the section. In this way, cer-
tain types oI physical behavior can be tracked automatically in the analysis, so that complex
empirically calibrated rules are not needed. As shown later, this can be applied to a variety
oI beams, columns, and braces; however, some oI the assumptions may not be valid, as
the typical Iiber-based model incorporates only unidirectional material properties. Conse-
quently, multi-axial stress states and three-dimensional phenomena |such as local buckling,
section warping, or distortion (because oI shear or torsion), bar pull out, etc.| cannot be eas-
ily represented.
Finite Element Models. Although the phenomenological and physical theory models de-
scribed above are special Iorms oI Iinite elements, the more general case, where a member
and structure is divided up into a variety oI small shell or solid elements, allows the re-
sponse oI the structure in various modes to be simulated. Such Iinite element Iormulations
can account Ior multi-axial stress states, nonlinear material properties, and changes in ge-
ometry that occur during loading. Because oI the large number oI elements and the absence
oI simpliIying assumptions, these models are oIten computationally expensive and time-
consuming to develop and execute; however, they usually produce results close to those
expected oI real structures. Nevertheless, these models cannot reproduce behavior that is
6
not incorporated in the model (e.g., buckling because oI large displacements, material Iail-
ure, etc.), and they put much more reliance on having well-deIined multi-directional mate-
rial property models. As such, careIul calibration oI these models to test results is needed,
and the degree oI improvement in response prediction compared to the increased level oI
eIIort to perIorm these models should be careIully considered.
This investigation studies the ability oI Iinite element Iormulations to simulate accurately
the behavior modes oI interest in braces, beams, beam-to-column connections, and braced-Irame
subassemblies. Compared to phenomenological models, these models require more computational
eIIort, but incorporate more realistic physical representations oI members and materials, including
the initiation and evolution oI damage up to and including complete Iailure.
1.2 OB1ECTIVES AND SCOPE OF RESEARCH
The objectives oI the research presented in this report are as Iollows: (1) to review available in-
Iormation on the behavior and analysis oI members and subassemblies Irom steel concentrically
braced Irames where elements are subjected to deterioration; (2) to develop improved numerical
models that simulate adequately the behavior oI a steel braced Irames up to and including Iail-
ure; (3) to use the improved models to evaluate and reIine several requirements Ior detailing and
analyzing special concentrically braced steel Irame building.
The remaining chapters oI this report derive Irom the objectives listed above. Chapter 2
reviews existing experimental investigation. Chapter 3 reviews and evaluates material models on
structural steel deterioration. Chapter 4 develops an improved cyclic damage plasticity model.
Chapter 5 discusses calibration, validation and application oI the newdamage material model using
test results Irom beam-to-column connections, individual braces, and braced-Irame subassemblies.
Chapter 6 provides studies oI steel braced Irame behavior, including brace proportions, connection
details, lateral bracing Ior beams, and estimate oI interstory driIt demands. Concluding remarks and
recommendations are discussed in Chapter 7. An appendix includes a constraint method developed
in this research Ior pushover analysis and a discussion on instability oI the Newmark integrator.
Part oI the research presented in this report was undertaken as a subproject oI the Phase VI
CUREE/Kajima Collaborative Research Project on Iactors leading to the progressive collapse oI
7
structures. In this research, the cyclic behavior and modeling oI steel structural elements and con-
nections, with emphasis on those subjected to high axial loads, are examined as they approach and
reach Iailure. This work complements parallel work at StanIord University under the supervision
oI ProIessor H. Krawinkler on steel and reinIorced concrete elements and connections where axial
loads do not dominate the expected mode oI behavior (Lignos and Krawinkler, 2007), and work
at the University oI BuIIalo under the supervision oI ProIessor A. Reinhorn on large displacement
analysis oI structural systems (Sivaselvan et al., 2007).
8
2 Reviewof Existing ExperimentaI Investigations
This research Iocuses on braces, beams, columns, and connections that are part oI braced steel
Irames, where members are subjected to high axial loads alone or in combination with bending and
shear. Some aspects oI conventional moment Irame construction will also be addressed, as these
elements may be used in combination with braced Irames as a gravity load resisting system or a
backup lateral load resisting system. Both static and dynamic loading situations will be considered,
although, the Iocus will not be on cases where strain rate eIIects or where local inertial Iorces along
the length oI a member need to be considered. The analysis models considered will include the
eIIects oI material inelasticity, local and global buckling, and low-cycle Iatigue on the behavior and
Iailure oI various elements and structural systems. Thus experimental results that investigate such
behavior are oI interest. This chapter reviews experimental data Irominvestigations oI conventional
braces, gusset plates, braced-Irame subassemblies, and beam-column connections.
2.1 INVESTIGATIONS OF CONVENTIONAL BRACES
Tremblay (2002) compiled and interpreted the results oI more than 100 tests oI steel braces available
at that time. These tests include braces Iabricated Irom square hollow structural section (HSS)
sections, pipes, wide Ilange sections, double channels, single and double angles, and structural 'T
sections. This compilation does not include recent tests by Tremblay et al. (2003), Elchalakani et
al. (2003), Goggins et al. (2005), Yang and Mahin (2005), Fell et al. (2006), Han et al. (2007),
Tremblay et al. (2008), and Lehman et al. (2008).
Tremblay (2002) used this database to assess several useIul engineering relationships used
in design. For example, the initial buckling load was compared Ior the braces in the database to
theoretical predictions and to equations included in United States and Canadian building codes,
demonstrating that code values are oIten conservative; they tend to underestimate the buckling
capacity oI braces, especially Ior slender braces.
Importantly, Tremblay also examined several other simpliIied relations that may be used to
develop and interpret phenomenological models Ior braces. For example, relations exist that predict
the out-oI-plane displacement oI buckling braces and the deterioration oI the compressive buck-
ling capacity oI braces. For this later case, a relation was developed between the compressive load
capacities oI braces that have been shortened by a displacement equal to Iive times the displace-
ment at the onset oI buckling. Similar relations were developed Ior lower and higher normalized
displacements.
Tremblay also provided inIormation on the cyclic ductility that a brace may develop as a
Iunction oI loading history and brace slenderness. This ductility Iactor shows considerable vari-
ability and sensitivity to loading history. This variability must be considered when evaluating the
ability oI reIined analysis models to predict member Iailure. The reIined analysis may suggest a
higher level oI conIidence in the predicted deIormation capacities than that can be supported by
test data.
Also oI signiIicance is that the ductility values presented by Tremblay are cyclic values,
deIined as the sum oI the peak elongation displacement and compressive displacement divided by
the displacement at Iirst buckling. Thus, Ior a loading history where the amplitudes oI displacement
are equal in both tension and compression, the ductility values should be divided by two to estimate
the maximum shortening oI the brace prior to Iailure. The cyclic ductilities range Irom about 6 to
more than 20, corresponding to unidirectional ductilities Ior symmetric cycles oI about 3 to more
than 10. Thus, the braces considered in the database are not too ductile, with the stockier braces
having less ductility capacity.
A similar empirical investigation oI a nearly identical database has examined the energy
dissipation capacity oI braces and the deterioration oI energy dissipation eIIiciency and compres-
sion capacity with cumulative plastic shortening as a Iunction oI member type and slenderness ratio
(Lee and Bruneau, 2005). This research points out many diIIiculties in deIining useIul parameters
Ior design or analysis purposes.
Earlier studies oI brace test data by Lee and Goel (1987) and Ikeda et al. (1984) have
produced speciIic recommendations Ior two diIIerent types oI phenomenological model used Ior
braces (Jain and Goel, 1978; Ikeda et al., 1984). With parameters estimated empirically, hysteretic
10
loops can be predicted using the analytical models. As noted previously, these loops do not directly
account Ior any phenomena associated with yielding or buckling oI the brace. They only Iollow the
general shape oI the hysteretic loops based on empirically derived parameters. Ikeda et al. (1984)
suggested that the parameters are not well conditioned, however, and are, in Iact, very sensitive to
materials, section shape, loading history, etc. Ikeda and Mahin (1984) subsequently developed a
physical theory model Ior brace buckling that was also calibrated to available test data. This model
utilized a generalized plastic hinge based on classic plasticity at the center oI the member and a
large displacement Iormulation Ior the internal lateral displacements within the brace. It proved
diIIicult to properly model the behavior oI braces in the re-straightening phase Irom a previously
buckled conIiguration, and a variety oI empirical adjustments were imposed to improve accuracy.
Test data are needed to evaluate existing brace models and calibrate new ones. Because oI
their availably at the beginning oI this investigation, the primary experiment source oI data used is
a series oI eight nearly identical Hollow Structural Section (HSS) braces tested at the University
oI CaliIornia at Berkeley (Yang and Mahin, 2004; Uriz, 2005). The braces (see Fig. 2.1) consisted
oI square 6 6 3/8 HSS sections Iabricated Irom ASTM A500 Grade B steel having speciIied
minimum strengths oI F
y
46 ksi and F
u
58 ksi. Mill certiIicates showed that the specimens
were considerably stronger, with F
y
60 ksi and F
u
65 ksi. The ends oI the HSS sections were
slotted and welded to gusset plates at each end. The Iraming members to which the gusset plates
were attached at both ends were Iixed against rotation, but axial elongation or shortening oI the
brace was permitted. Some specimens were reinIorced locally near their ends where there was a
reduction oI the net area oI the brace because oI the slots used to Iasten the braces to the gusset
plates. The KL/r ratio Ior the braces, considering K = 1 and the length L extending Irom the Iaces
oI the Iraming members to which the gusset plates were attached, was 51 Ior all specimens.
Three diIIerent types oI axial displacement histories were imposed along the longitudinal
axis oI the brace. One oI the loading protocols consisted oI a cyclic displacement history based on a
nonlinear analysis oI a Special Concentrically Braced Frame (SCBF) that was subsequently tested
(Uriz, 2005). This would Iacilitate comparison oI hysteretic loops Ior individual braces and braces
tested as part oI a structure. A second loading protocol consisted oI a near-Iault type history, where
there was a modest cycle with compressive loading Iollowed by a large excursion in tension, Iol-
lowed, in turn, by a series oI smaller cycles about an oIIset position. A similar test was done where
11
the signs oI the displacement excursions were reversed (i.e., the large displacement excursion was
in compression). Finally, a third type oI protocol was used where a series oI symmetric, constant
amplitude displacement cycles at several levels were imposed until the brace Iailed. These test
results were previously used to assess damage accumulation models and the sensitivity oI Iailure
to the loading history (Uriz, 2005).
2.2 INVESTIGATIONS OF GUSSET PLATES
A comprehensive three-year literature review on the behavior oI gusset plates was completed in
2005 (Chambers and Ernst, 2005). Including more than 200 papers and reports, this review evalu-
ates the results in terms oI the capability oI modern codes and analysis methods to predict behav-
ior. The conclusions state that the amount oI inIormation available on gusset plates Ior conditions
similar to those Iound in current design practice, and Ior loads representative oI seismic loading
conditions, are too Iew to make an adequate assessment.
To help address this issue, a series oI tests are currently under way at the University Wash-
ington, Seattle, under the supervision oI ProIessors C. Roeder and D. Lehman (Roeder et al., 2006;
Lehman et al., 2008). These tests place the braces along the diagonal oI a single story, single
bay Irame. The braces include buckling restrained braces as well as conventional braces that are
allowed to buckle.
A variety oI gusset plate details are used. Initially, gusset plate connections were designed
by considering the conventional Whitmore section (eIIective width, Whitmore, 1952) and the uni-
Iorm Iorce method (Thornton, 1991; AISC, 2005a), which are basically strength-based design
methods and do not consider the inelastic deIormation demands expected during seismic loading
Irombrace buckling or Irame rotations. Currently, the investigation has expanded to include a num-
ber oI conIigurations and alternative design methods Ior the gusset plates. Previous test results have
shown that current gusset plate details tend to Iail prematurely along the weld connecting the plate
to the supporting beam and column, or initiate Iracture into the beam or column. Representative
damage observed in their tests is shown in Figure 2.2.
Aseries oI simple gusset plate tests were carried out at the University oI CaliIornia at Berke-
ley by Markarian and Mahin (2004). In these tests, only Irame Ilexural deIormations were imposed,
12
with no axial tensile or compression loads. The beams-column connections were pinned, consist-
ing oI a simple shear tab connection Irom the beam to the column. The addition oI the gusset
plate resulted in considerable moment transIer and stress states that do not appear in typical design
or analysis practices. These tests demonstrated a tendency Ior the beams and columns to exhibit
complex states oI stress, and Ior the connection to Iail in a brittle Iashion.
More recent tests have been conducted on gusset plate connections in a buckling-restrained
braced Irame (BRBF) system by Kishiki et al. (2008). In these tests, interaction between Iraming
components and the gusset plate was considered, but the inIluence oI bracing Iorces was ignored.
Beam-column Irame subassemblies with the gusset plate were subjected to cyclic lateral loading.
It was Iound that the eIIective length oI the beam was shortened by the presence oI the gusset plate
connections, indicating that the critical section oI the beamwas moved to the toe oI the gusset plate.
On the other hand, the eIIective length oI the column was hardly aIIected by the gusset-plate when
a rectangular hollow section (RHS) was used Ior the column.
2.3 INVESTIGATIONS OF BRACED FRAME SUBASSEMBLIES
Most tests oI braced Irames have used small-scale models or have employed details that are not
currently in use. Valuable data were obtained Irom Japanese pseudo-dynamic tests oI a Iull-scale,
six-story special concentrically braced Irame and eccentrically braced Irame. These tests were
perIormed at the Building Research Institute, Tsukuba, Japan, in the mid-1980s (Midorikawa et
al., 1988; Foutch et al., 1986). One-third scale models oI the same structures were tested on the
University oI CaliIornia at Berkeley shaking table (Whittaker et al., 1989).
Relatively Iew recent tests have been conducted on multistory subassemblies oI concentri-
cally braced Irames. One recent example was carried out by Uriz (2005). The Irame is shown in
Figure 2.3, and the displacement history imposed at the rooI oI the structure is shown in Figure
2.4. The Irame suIIered extensive damage to the braces in the lower level, the columns at the base
oI the building, and in the beam-column connections at the Iirst-story level. A wide variety oI be-
havior was observed, Irom yielding, local buckling, local tearing, brace Iracture, and column local
buckling and connection Iracture (Uriz, 2005). Following brace buckling, the strength oI the Irame
degraded signiIicantly Irom cycle to cycle. This specimen provides a good test oI the ability oI
13
various methods to predict behavior associated with members undergoing bending and axial load,
lateral buckling, and local buckling.
Uriz (2005) reported that the distribution oI lateral driIt was nearly equal Ior the top and
bottom stories at the beginning oI test. Both levels developed slight amounts oI buckling, but Iull
lateral buckles Iormed in the lower level, with local buckling occurring almost immediately (Fig.
2.5). Because oI the reduced load capacity oI the buckled braces, the Iorces in the upper (and lower)
level decreased, and no Iurther tendencies were observed Ior the upper level to buckle. This re-
sulted in a weak-story response, with nearly all oI the inelastic behavior and damage concentrated
in the lower level. This led to the complete Iracture oI the braces during the Iirst design level dis-
placement excursion (Fig. 2.6), with Iailure oI the lower-level beam-column connections occurring
soon thereaIter (Fig. 2.7).
The hysteretic loop relating the rooI lateral Iorce versus base shear is shown in Figure 2.8.
The points marked A, B, C, and D represent the initial tearing and Iinal complete Iracture oI the
braces in the lower story. Points E and F represent the Iracture oI the lower beam to column con-
nection, Iirst on one side oI the structure and then on the other. The initial buckling oI the braces
is easily identiIied by the sudden nonlinearity oI the system, with slight negative post-buckling
tangent stiIIness. Substantial deterioration oI the specimen load capacity occurred upon cycling at
the level that induced Iirst buckling oI the lower level braces. This buckling occurred at relatively
small driIts compared to driIts considered Ior design purposes in the United States. The braces
completely Iractured during subsequent cycles at this driIt level or during the Iirst excursion to the
design level.
A series oI Iour Iull-scale, two-story, single-bay, braced-Irame tests were perIormed at the
National Center Ior Research on Earthquake Engineering (NCREE) in Taiwan (Powell et al., 2008).
The Iirst test was a braced Irame with HSS tube braces in a multi-story X-brace conIiguration. It
provided an experimental examination oI the multi-story X-brace systemand oI the mid-span gusset
plate connections. At the same time, it was a conIirmation study Ior past research on corner gusset
plate connections perIormed at the University oI Washington.
Several tests oI BRBFs have recently been carried out at University oI CaliIornia at Berkeley
(Uriz, 2005). The tested subassemblies consist oI a portion oI a seven-story BRBF. Two conIigura-
tions oI braces were considered. These included a chevron conIiguration with each unbonded brace
14
having a single Ilat plate Ior the yielding core; in one brace the core was oriented horizontally and
in the other it was oriented vertically. The second conIiguration utilized a single unbonded brace
oriented across the diagonal oI the Irame. Two brace sizes were considered: one where the core
was the same size as used Ior the chevron conIiguration and another where the area was doubled
(and the core was cruciIorm in shape).
The buckling restrained braces (BRBs) perIormed quite well in these tests compared to
conventional braces; however, considerable yielding was noted throughout the surrounding Irame.
This included shear yielding oI the column webs, yielding oI the columns and gusset plates, and
yielding oI the beam to column connection regions. In addition, gusset plate buckling was noted.
Interestingly, this buckling occurred when the attached brace was in tension. The buckling was a
result oI the kinematic distortion oI the gusset plate that occurred as a result oI Irame action (the
gusset plate is squeezed as the brace is loaded in tension). Eventually, Iracture was noted in the
beams adjacent to the gusset plate. This Iracture appeared to be induced as a consequence oI Irame
bending behavior.
A series oI tests on BRBFs have been carried out at the NCREE in Taiwan (Tsai et al.,
2006). The basic diIIerences between these tests and the single diagonal brace test conducted by
Uriz (2005) is as Iollows: (1) a diIIerent type oI BRB is used; (2) a concrete-over-metal-deck Iloor
system is used; and (3) the Irame is subjected to two horizontal components oI motion using the
pseudo-dynamic test method, leading to bidirectional-bending eIIects in the columns. The speci-
men was able to undergo a substantial number oI simulated earthquakes. During a moderate level
test, a Iracture in the welds oI the gusset plate to the column was noted and required repair beIore
conducting the subsequent tests. The NCREE has also tested BRBFs in a three-story conIiguration
where the columns were concrete Iilled steel tubes (Tsai et al., 2004).
2.4 INVESTIGATIONS OF BEAM-COLUMN CONNECTIONS
The steel moment connection data used herein are Irom two sources. One is a series oI welded
steel beam-column connection tests (Tanaka et al., 2000) conducted by Kajima Corporation, Japan.
These tests incorporated box columns and beams with horizontally tapered haunches so that the
beam Ilanges widened to conIorm to the width oI the column. The columns were made oI welded
15
square tube section oI 40040019 (mm) and steel grade oI SS400 (F
y
280 MPa, F
u
430 MPa).
Beams were made oI rolled wide-Ilange section oI 5002001016 (mm) and steel grade oI SS400
(F
y
270 MPa, F
u
420 MPa). The distance Iromthe column axis to the load point at beamis 2000
mm. Various Iorms oI ductile behavior were observed with the specimens eventually developing
Iractures. Specimen BW10, which developed signiIicant local buckling prior to terminating the
test, is studied herein. The test setup Ior specimen BW10 is shown in Figure 2.9.
The other set oI data used here is Irom Suita et al., (2000), where a variety oI Japanese
post-Kobe style and U.S. post-Northridge style welded beam-to-column connections were tested.
This experimental study compared beams having uniIorm sections and reduced beam sections in
the plastic hinge region. Duplicate specimens Iabricated by two manuIacturers were tested to iden-
tiIy variability in perIormance with diIIerent heats oI steel and manuIacture. The beam-column
connections with a uniIorm Ilange section and no weld access hole (NWAH) Iailed because oI low-
cycle Iatigue. They are used in current research. The columns were made oI cold-Iormed square
tube sections oI 350 350 12 (mm) and steel grade oI BCR295 (F
y
295 MPa, F
u
400 MPa).
The beams were made oI rolled wide-Ilange sections oI 500 200 10 16 (mm) and steel grade
oI SN400B (F
y
235 MPa, F
u
400 MPa). The distance Irom the column axis to the load point
at beam is 3000 mm. The test setup Ior specimen N4 studied herein is shown in Figure 2.10.
An extensive database oI results oI steel beam-to-column connection tests has recently been
compiled by Lignos and Krawinkler (2007).
2.5 CONCLUDING REMARKS
Considerable experimental data exists on components such as conventional braces, beams, columns,
and connections, and on subassemblies including these components. Some oI this well-documented
data will be used to calibrate or validate the numerical models developed in this research.
16
(a) Specimen
(b) End connection
Figure 2.1 Experimental setup for conventional brace test (Yang and Mahin, 2005)
17
(a) Crack initiation and propagation
(b) Complete rupture
Figure 2.2 Gusset plate damage (Roeder !" #$%& 2006)
18
Figure 2.3 Test setup for SCBF (Uriz, 2005)
Figure 2.4 Target roof displacement history of SCBF test (Uriz, 2005)
19
(a) Global buckling
(b) Local buckling
Figure 2.5 Brace buckling (Uriz, 2005)
20
(a) Crack propagation
(b) Complete rupture
Figure 2.6 Brace failure (Uriz, 2005)
21
Figure 2.7 Beam-column connection failure (Uriz, 2005)
-5 -4 -3 -2 -1 0 1 2 3 4 5
-600
-400
-200
0
200
400
600

L
o
a
d

(
k
i
p
s
)
Displacement (in.)
A
B
C
D
E
F
Figure 2.8 Base shear versus roof lateral displacement (Uriz, 2005)
22
Figure 2.9 Test setup for BW10 beam-column connection (Tanaka !" #$%& 2000)
3000
1050 1200 750
175
410
410
350
500
500
500
CJP groove weld
radius cut RBS
beam: WF-5002001016, SN400B
guide for lateral brace
column: RHS-35035012, BCR295
unit:mm
diaphragm
Figure 2.10 Test setup for N4 beam-column connection (Suita !" #$%& 2000)
23
3 StructuraI SteeI Deterioration
For structural steel subjected to a severe cyclic loading history, such as that experienced during
a strong earthquake, several stages oI behavior commonly exist during the course oI material de-
terioration. Initially, it is assumed there are no macroscopic cracks or deIects, thus no stress or
strain singularities are associated with the material. The material is then loaded non-proportionally
and cyclically into the inelastic regime under asymmetric stress and strain histories oI varying am-
plitude. Deterioration then develops because oI imposition oI large plastic deIormation, resulting
in substantial energy dissipation. This causes progressive Iailure oI the material volume through
ductile damage and Iracture associated with microvoid nucleation, growth, and coalescence. AIter
complete Iailure oI a local material volume, a macroscopic crack is initiated eventually.
Damage resulting Irom plastic deIormation in ductile metals is mainly because oI the Ior-
mation oI microvoids, which initiate (nucleate) either as a result oI Iracturing or debonding oI
inclusions such as carbides and sulIides Irom the ductile matrix. The growth and coalescence oI
microvoids under increasing plastic strain progressively reduces the material`s capability to carry
loads and can result in complete Iailure. Figure 3.1 shows the section oI a tensile specimen dur-
ing the necking process (deIined as a mode oI tensile deIormation where relatively large amounts
oI strain localize disproportionately in a small region oI the material, Bridgman, 1952), clearly
demonstrating void accumulation at the specimen center (Puttick, 1959). Figure 3.2 schematically
illustrates the nucleation, growth, and coalescence oI such microvoids (Anderson, 1995). II the
initial volume Iraction oI voids is low, each void can be assumed to grow independently; upon
Iurther growth, neighboring voids interact. Plastic strain eventually concentrates along a sheet oI
voids, and local necking instabilities develop as shown schematically in the Iigure.
Aproper modeling oI this microvoid nucleation and growth mechanismis needed to predict
ductile Iailure in steel members and structures. In the context oI continuum mechanics, plasticity
and damage models are essential Ior the simulation. Plastic, damage, and Iracture behavior oI
structural steel under monotonic and cyclic loading are discussed in the Iollowing sections.
3.1 PLASTIC BEHAVIOR AND MODELS
Isotropic hardening and/or kinematic hardening are commonly used to describe the plastic behav-
ior oI metal-like materials under complex loading conditions. Prager (1956) and Ziegler (1959)
initiated the Iundamental Iramework used Ior kinematic hardening rules. Armstrong and Frederick
(1966) developed a nonlinear kinematic hardening rule that generalized its linear predecessor. In
this model, the kinematic hardening component is deIined to be an additive combination oI a purely
kinematic term (i.e., the linear Prager/Ziegler hardening law) and a dynamic recovery term, which
introduces the nonlinearity. When combined together with isotropic hardening, such a model can
account Ior the Iollowing observed phenomena: (1) The nonlinear Bauschinger eIIect (Fig. 3.3);
(2) cyclic hardening (Fig. 3.3); and (3) ratcheting (Fig. 3.4).
The Armstrong and Frederick`s rule was Iurther extended by Chaboche (1986, 1989), where
an additive decomposition oI the back stress was postulated. The evolution equation oI each back
stress component is similar to the work done by Armstrong and Frederick (1966). The advantages
oI this superposition are that a larger strain range can be realistically modeled, and a more accurate
description oI ratcheting is provided. These Ieatures allowmodeling oI inelastic behaviors oI metals
that are subjected to cycles oI load, resulting in signiIicant inelastic deIormation and, possibly, low-
cycle Iatigue Iailure. Discussion oI these plasticity models can be Iound in Lemaitre and Chaboche
(1990).
26
Figure 3.1 Void accumulation in a tensile specimen (Puttick, 1959)
(a) Inclusions in a ductile matrix (b) Void nucleation
(c) Void growth (d) Strain localization between voids
(e) Necking between voids (f) Void coalescence and fracture
|aIter Anderson (1995)|
Figure 3.2 Void nucleation, growth, and coalescence in ductile metals
27

S
t
r
e
s
s
Strain
1
2
3
1
2
Figure 3.3 Nonlinear Bauschinger effect and cyclic hardening

S
t
r
e
s
s
Strain
1 2 3
1 2
ratchet strain
mean stress
Figure 3.4 Ratcheting under asymmetric stress history
28
3.2 DAMAGE AND FRACTURE UNDER MONOTONIC LOADING
Two alternative approaches are generally considered Ior modeling material Iailure: local approaches
and global approaches. Local approaches Iocus on issues related to micromechanics, whereas
global approaches address issues oI Iracture mechanics.
3.2.1 Local Approaches
The local approach to Iracture can be deIined very generally as the combination oI the Iollowing
two Iactors: (1) the computation oI local stress and deIormation values in the most loaded zones
oI a component or structure; and (2) the use oI predeIined empirical models corresponding to var-
ious Iracture mechanisms, such as cleavages, ductile Iracture, Iatigue, creep, stress-corrosion, etc.
(Rousselier, 1987). The parameters needed Ior these empirical models are obtained by calibration
to experimental data.
Many damage models have been developed since the initial micromechanics studies oI Mc-
Clintock (1968) and Rice and Tracey (1969). The models can be categorized into two classes: (1)
void volume Iraction models and (2) continuum damage mechanics models. In the Iirst group,
Iailure is predicted when void volume Iraction reaches a critical value. In the second group, the
material is considered Iractured when the reduction oI the eIIective area exceeds a critical value.
Both types oI models can be written in the Iorm oI stress-modiIied critical plastic strain:
Damage evolution D =
_
F () G(
p
) dt (3.1a)
Failure criteria D = D
c
(3.1b)
where is the stress tensor, F is the stress modiIication Iunction,
p
is the plastic strain rate tensor,
G is the plastic strain rate Iunction, D represents the damage oI the material, and D
c
is the value
oI the critical damage parameter at Iailure. For example, the Rice and Tracey (1969) model can be
written as
F () = exp (1.5
m
/
eq
) G(
p
) =

p
(3.2)
where
m
= : 1/3 is the mean stress,
eq
=
_
3/2 is the equivalent stress,
m
/
eq
is the
stress triaxiality, and

p
is the equivalent plastic strain rate, deIined as

p
=
_
2
3

p
:
p
(3.3)
29
For the porous metal plasticity model (GTN model) developed by Gurson (1977), and Tvergaard
and Needleman (1984), a sophisticated yield Iunction is developed and the void growth part is given
by
F () = 1 G(
p
) =
p
v
=
p
: 1 (3.4)
where
p
v
is the volumetric plastic strain rate, and 1 is the second-order identity tensor 1 =
i j
e
i
e
j
.
For the ductile damage model proposed by Lemaitre (Lemaitre, 1992; DuIailly and Lemaitre,
1995), which is based on continuum damage mechanisms (CDM) introduced by Kachanov (1958),
the damage evolution Iunction becomes
F () =
_
Y
S
_
t
G(
p
) =

p
(3.5)
where S is a material constant in energy density units, t is a dimensionless material constant, and
Y is the internal energy density release rate, calculated as
Y =
1
2
: [D
e
]
1
: (3.6)
where D
e
represents the Iourth-order elasticity tensor
D
e
= 1 1 + 2G(I
1
3
1 1) (3.7)
where and G are the Lame constants, and I is the Iourth-order symmetric identity tensor.
I =
1
2
[
ik
jl
+
il
jk
] e
i
e
j
e
k
e
l
(3.8)
The Rice and Tracey (1969) model and the Lemaitre (1992) model predict very similar trends in
the case oI proportional loading, as seen in Figure 3.5, where the equivalent plastic strain to Iailure

p
f
versus stress triaxiality
m
/
eq
is plotted Ior the two models. In the case oI nonproportional
loading, Marini et al. (1985) showed that the two models also give similar results. Recently,
Steglich et al. (2005) investigated the relationship between the CDM and the GTN models.
3.2.2 Global Approaches
Global approaches are based on asymptotic continuum mechanics analysis. Under some situations,
single- or dual-parameters can uniquely characterize the crack tip condition. Well-known single-
parameter variables are stress intensity K, J-integral, and CTOD (crack tip opening displacement).
30
0.0 0.3 0.6 0.9 1.2
0.0
0.3
0.6
0.9
1.2
1.5
1.8
Stress triaxiality
m
/
eq

E
q
u
i
v
a
l
e
n
t

p
l
a
s
t
i
c

s
t
r
a
i
n

t
o

f
a
i
l
u
r
e

f
p
Rice and Tracey (1969)
Lemaitre (1992) = 0.3
Figure 3.5 Equivalent plastic strain to failure versus stress triaxiality
A well-known dual-parameter Iormulation is based on the addition oI a T-stress parameter that
characterizes the crack tip constraint.
All these parameters are deIined at the global level oI the crack medium, within the Irame-
work oI Iracture mechanics. They are applicable to a number oI situations where it is not necessary
to know the exact state oI stress or damage in the vicinity oI the crack tip (ranging Irom a two-
dimensional, almost elastic medium, with only a small plastic zone relative to the crack size; to
a three-dimensional medium subjected to proportional loading or cyclic loading in Iatigue). On
the other hand, this approach may prove deIicient in some cases, because oI the size oI the cracks,
pronounced overall plasticity during ductile Iracture, or loading history eIIects.
3.2.3 Relation between Local and Global Approaches
A systematic comparative and parametric study oI local and global models was reported by Xia
and Shih (1995) using the representative volume element (RVE) methodology. Here, the size oI
the elements in the Iracture process layer (the Iracture elements representing the crack) in the local
approach is the key parameter linking the local and global approaches, and an approximate equality
31
exists:
J
Ic
= d
0
(3.9)
where J
Ic
is the critical J-integral under the Mode I condition, d is the size oI the elements,
0
is the Ilow stress, and is a Iactor oI order unity, suggesting that the element size in local ap-
proaches should be on the order oI J
Ic
/
0
or CTOD to obtain consistent results. This may make
local approaches inIeasible Ior large structural systems or subassemblies.
3.3 DAMAGE AND FRACTURE UNDER CYCLIC LOADING
In this section, local approaches (as deIined in previous section) to modeling damage and Irac-
ture under cyclic loading are examined in more detail. This approach is compared to the popular
Manson-CoIIin rule Ior low-cycle Iatigue. The traditional Manson-CoIIin rule 'increments the
damage state only at the end oI each cycle. In contrast, the micromechanics and damage plasticity
approach 'ramps the damage continuously during cycling.
3.3.1 Manson-Coffin Rule
The Manson-CoIIin rule is a popular model Ior low-cycle Iatigue because oI its simplicity. Gener-
ally, it is written in the Iorm
p
2
=
f
(2N
f
)
c
(3.10)
where
p
/2 is the amplitude oI plastic strain, N
f
is the number oI cycles,
f
is the ductility
coeIIicient, and c is the ductility exponent. In 1953, Manson, recognizing the Iorm oI Equation
(3.10) relating Iatigue liIe and plastic strain range, suggested that the magnitude oI 1/c was 'in
the neighborhood oI three (Manson, 1953). CoIIin (1954) showed that Ior practical purposes the
Iatigue property c is approximately equal to 1/2, and that
f
is related to the monotonic Iracture
ductility
f
(Tvernelli and CoIIin, 1959). Actually, c commonly ranges Irom0.5 to 0.7 Ior most
metals, with 0.6 as a representative value.
Despite a large amount oI work to generalize this law to multi-axial states oI stress (e.g.,
Morrow, 1964) and to complex histories oI loading (Manson et al., 1971), it remains a model
generally only applied to uniaxial periodic loadings. Still, a wide variety oI tests oI structure,
component, and material specimens have demonstrated the general validity oI the Manson-CoIIin
32
relation, and the range Ior the coeIIicient c cited above.
Recently, Uriz (2005) utilized the Manson-CoIIin rule together with the Palmgren-Miner
linear hypothesis Ior damage accumulation (Palmgren, 1924; Miner, 1945) and the rainIlow cycle
counting method (Matsuishi and Endo, 1968) to predict the low-cycle Iatigue behavior oI uniax-
ially loaded struts. This approach mimics the rupture oI uniaxially loaded material Iibers used to
represent the cross section oI a member. The rainIlow cycle counting method was simpliIied so
that only several recent cycles were considered. This allows computation oI damage during the
response rather than just at the end. AIter parameter calibration it was shown that the prediction
matched the experiment quite well. Note that the continuous damage models discussed in this chap-
ter do not rely on cycle counting algorithms, but rather damage continuously accumulating based
on the damage mechanics models presented.
3.3.2 Continuous Damage Models
Not much attention has been given to the possibility oI incorporating damage into cyclic plasticity
by means oI micromechanics. Recent works on porous metal plasticity are those oI Leblond et al.
(1995), Besson and Guillemer-Neel (2003), and Cedergren et al. (2004). These models introduce
nonlinear kinematic hardening into the GTN model. As Iar as continuum damage mechanics is
concerned, Pirondi and Bonora (2003) introduced uniaxial conditions to model stiIIness recovery
during tension-compression cyclic loading. Kanvinde and Deierlein (2004) extended the Rice and
Tracey (1969) Iormulation to incorporate a cyclic void growth model (CVGM). They revised the
material characteristic D
c
in Equation (3.1).
D
c
= D
c0
exp(
p
) (3.11)
so that the critical void volume Iraction decreases as the equivalent plastic strain increases, and
then revising the stress modiIication Iunction as
F() =
|
m
|
m
exp(1.5
m
/
eq
) (3.12)
so that the void volume Iraction decreases (heals) in compression.
Lemaitre (1992) introduced a relative simple modiIication Ior damage evolution in cyclic
33
loading
D =
_
_
_
Y
S
_
t
1
> 0
0 otherwise
(3.13)
where
1
is the maximum principal stress. Here, damage does not accumulate when all principle
stresses are in compression. In the present investigation, this is Iurther revised Ior simpler imple-
mentation as:
D =
_
_
_
Y
S
_
t

p

m
eq
>
1
3
0 otherwise
(3.14)
This simpliIication has negligible eIIect Ior most states oI stress.
To schematically illustrate the diIIerence between the CVGM and CDM models, prelimi-
nary results oI an analysis oI an individual brace (presented later in more detail in Section 5.2.1) are
shown in Figures 3.6 and 3.7. From Figure 3.6 it can be seen that Ior the cyclic void growth model,
the damage D increases during some portions oI the deIormation cycle and then heals in other
portions, and the critical damage parameter D
c
decreases with pseudo-time |according to Equation
(3.11)|. In contrast, Ior the same example, the continuum damage model given by Equation (3.14)
only increases with increasing pseudo-time, and the critical damage parameter D
c
is a constant.
The numerical values oI the critical damage parameters, D
c0
Ior the CVGM and D
c
Ior the CDM
models will, in general, diIIer. For the sake oI this illustration, they have been loosely calibrated
to tail at about the same time.
It is also worth mentioning that although the critical equivalent-plastic-strain approach
(F () = 1, G(
p
) =

p
) can be used in proportional loading cases where the triaxiality is con-
stant and known, it is unsuitable Ior arbitrary cases oI cyclic loading. As illustrated in Table 3.1
(and approximately in Figs. 3.8 and 3.9), Iatigue liIe (number oI cycles to Iailure) will tend to be
underestimated iI large strain amplitude data are used to calibrate the parameters (by a Iactor oI
4 in the table); conversely, the Iatigue liIe will tend to be overestimated iI small strain amplitude
data are used to calibrate the parameters. This is an inherent and important drawback oI applying
the critical equivalent-plastic-strain criterion to cyclic loading. This drawback is one oI the rea-
sons why 'stress-modiIied critical plastic strain criteria, such as CVGM and CDM, are typically
used to simulate low-cycle Iatigue Iailure behavior under arbitrarily varying strain histories. The
other reason is the triaxiality-independence oI critical equivalent plastic strain criterion; the eIIect
34
oI triaxial constraint on the initiation oI rupture is a well-known phenomenon.
3.3.3 Simple Comparison of Continuous Damage Models with Manson-Coffin Rule
Although the basic trends predicted by the Manson-CoIIin rule have been veriIied by considerable
low-cycle Iatigue data, the rule increments the damage state only at the end oI each cycle. This rule
is not suitable Ior the case where the number oI cycles to Iailure is less than 10 because this will
not allow a point in a material to Iracture until the end oI a Iull cycle. Continuous damage models
resolve this diIIiculty. Thus, it is useIul to use the Manson-CoIIin rule as a reIerence and to compare
the results oI diIIerent continuous damage models to the results predicted by the Manson-CoIIin
rule.
For this comparison, the cyclic void growth model in Equation (3.11) and (3.12) and the
simpliIied continuum damage mechanics model in Equation (3.14) are evaluated Ior low-cycle Ia-
tigue and compared against the Manson-CoIIin rule. A uniaxial stress condition model is subjected
to a series oI constant-amplitude strain cycles considering several amplitudes oI maximum strain.
These cyclic deIormation histories are imposed until rupture oI the material occurs. In this way,
standard Manson-CoIIin type plots can be prepared Ior the analysis results, and these can be com-
pared directly with the most basic Manson-CoIIin criteria Ior an ideal experiment, as shown in
Figures 3.8 and 3.9. Both models predict results that agree with those computed with the Manson-
CoIIin rule, with the ductility exponent c ranging Irom 0.5 to 0.7; i.e., corresponding to typical
values Ior metals.
Note that the low-cycle Iatigue criterion based on critical equivalent plastic strain results
in a Iixed ductility exponent c equal to 1. As such, it cannot predict the correct trend oI low-
cycle Iatigue Ior metals. Thus, although the eIIective plastic strain criterion can be calibrated Ior
a particular material and specimen conIiguration subjected to a speciIic loading protocol, the same
Iailure criterion might be inappropriate Ior diIIerent loading histories at other locations within the
same structure.
Considering that the CDM result in Figure 3.9 is more accurate than the CVGM result in
Figure 3.8, and that the CDM is comparatively easy to implement and numerically more eIIicient,
the simpliIied CDM model in Equation (3.14) is chosen herein as the damage evolution model
Ior Iurther investigation. It is believed that some underlying relationship should be satisIied when
35
continuous damage models match the Manson-CoIIin relation, suggesting that deeper investigation
is needed.
This chapter systematically reviews material models Ior structural steel subject to plasticity, dam-
age, and Iailure, and subject to cyclic loading history. Among several models reviews, the so-
phisticated low-cycle Iatigue model based on damage mechanics is chosen because oI its ability to
reproduce the Manson-CoIIin relation with reasonable accuracy.
Table 3.1 Predictions of failure for an ideal Manson-Coffin experiment
Calibrated to large strain amplitude data
Strain amplitude Failure Cycle Test/EPS
Test 2
EPS
0.20
2
1.00
Test 32
EPS
0.05
8
4.00
Calibrated to small strain amplitude data
Strain amplitude Failure Cycle Test/EPS
Test 2
EPS
0.20
8
0.25
Test 32
EPS
0.05
32
1.00
EPS critical equivalent-plastic-strain criterion

36
-10 0 10 20 30 40 50 60 70 80
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0

D
a
m
a
g
e

D
Pseudo Time
Cyclic Void Growth Model
Damage Evolution D
Critical Damage D
c
Fatigue life exhausted
Figure 3.6 Damage evolution of CVGM model
0 10 20 30 40 50 60 70 80
0.0
0.2
0.4
0.6
0.8
1.0

D
a
m
a
g
e

D
Pseudo Time
Continuum Damage Mechanics Model
Damage Evolution D
Critical Damage D
c
= 0.5
Fatigue life exhausted
Figure 3.7 Damage evolution of CDM model
37
0.1 1 10 100
0.01
0.1
1
CVGM = 0.2
CVGM = 0.4
Equivalent Plastic Strain
Manson-Coffin c = -0.55

P
l
a
s
t
i
c

s
t
r
a
i
n

a
m
p
l
i
t
u
d
e

p
/

2
Number of cycles to failure N
Figure 3.8 Manson-Coffin plot of CVGM model
1 10 100 1000
0.01
0.1
1
CDM t = 1
CDM t = 2
CDM t = 3
Equivalent Plastic Strain

P
l
a
s
t
i
c

s
t
r
a
i
n

a
m
p
l
i
t
u
d
e

p
/

2
Number of cycles to failure N
Figure 3.9 Manson-Coffin plot of CDM model
38
4 CycIic Damage PIasticity ModeI
This chapter develops a cyclic damage plasticity model based on the review oI material models in
the previous chapter. First, the theory oI plasticity and damage evolution is presented. Next, imple-
mentation oI plasticity and damage models using the cutting-plane algorithm or the closest-point
algorithm is discussed. Finally, examples oI the time-integration procedure Ior brick and shell ele-
ments is shown. The cyclic damage plasticity model is used in the next two chapters Ior calibration,
validation, application, and studies oI steel braced Irame components and subassemblies.
4.1 THEORY OF PLASTICITY AND DAMAGE EVOLUTION
Components oI the damage plasticity model, including strain rate decomposition, yield criterion,
isotropic and kinematic hardening behavior, and damage evolution are summarized in this section.
4.1.1 Strain Rate Decomposition
The rate-oI-deIormation tensor is written in the additive Iorm oI the elastic and plastic strain rate
components as
=
e
+
p
(4.1)
4.1.2 Elastic Behavior
The elastic behavior is modeled as isotropic hypoelasticity, i.e., the corotational rate oI the Cauchy
stress tensor is calculated Irom the elastic strain rate tensor
e
as
= D
e
:
e
(4.2)
where D
e
represents the Iourth-order elasticity tensor
D
e
= 1 1 + 2G(I
1
3
1 1) (4.3)
where and G are the Lame constants, 1 is the second-order identity tensor, and I is the Iourth-order
symmetric identity tensor.
1 =
i j
e
i
e
j
(4.4a)
I =
1
2
[
ik
jl
+
il
jk
] e
i
e
j
e
k
e
l
(4.4b)
4.1.3 Plastic Behavior
The plastic behavior is modeled as pressure-independent plasticity. II the Huber-von Mises yield
condition is adopted, the yield surIace is deIined by the Iunction
F =
y
= 0 (4.5)
where
y
is uniaxial yield stress, and is the eIIective von Mises stress, with respect to the eIIective
deviatoric stress tensor
s
e
= dev[] = s (4.6)
where s is deviatoric stress tensor and is the back stress tensor, which is decomposed into multiple
tensor components
=
j
(4.7)
The eIIective von Mises stress is deIined as
(s
e
) =
_
3
2
s
e
: s
e
=
_
3
2
s
e
(4.8)
The model Iollows the associated Levy-Saint Venant plastic Ilow rule, i.e., the plastic strain
rate tensor is deIined as

p
=
F
=
3
2
s
e

=
3
2
n
(4.9)
where

is the plastic consistency parameter, and n is the Ilow direction
n =
s
e

(4.10)
Note that Ior the sake oI implementation simplicity, n is not a unit vector; its norm is
n =
_
2
3
(4.11)
40
Adoption oI the above Iorm oI von Mises yield criterion results in

=

p
, where

p
is the
equivalent plastic strain rate deIined as

p
=
_
2
3

p
:
p
(4.12)
4.1.4 Nonlinear Isotropic/Kinematic Hardening Model
The size oI the yield surIace
y
is a Iunction oI the equivalent plastic strain
p
Ior materials that
either cyclically harden or soIten.
y
=
y
(
p
) (4.13)
where the equivalent plastic strain is deIined as an accumulation

p
(t) =
_
t
0

p
() d (4.14)
The evolution oI the back stress tensor components are oI the Armstrong-Frederick type
(Armstrong and Frederick, 1966), deIined as

j
=
2
3
C
j

p

j

p
= [C
j
n
j
j
]

p
(4.15)
where C
i
and
i
are material parameters. The recall term
j

p
introduces the nonlinearity in the
evolution law. The law can be degenerated into the linear kinematic one by speciIying only one
component with = 0.
4.1.5 Damage Evolution Model
Based on the principle oI strain equivalence (Lemaitre, 1971): 'Any strain constitutive equation
Ior a damage material may be derived in the same way as Ior a virgin material except that the
usual stress is replaced by the eIIective stress, the stress tensor (in the damaged material)
D
is
calculated Irom the eIIective stress tensor as
D
= (1 D) (4.16)
where Dis the damage variable, whose rate is given by Lemaitre`s model (Lemaitre, 1992; DuIailly
and Lemaitre, 1995)
D =
_
_
_
Y
S
_
t

p

p
>
p
d
and

m
eq
>
1
3
0 otherwise
(4.17)
41
where
m
= : 1/3 is the mean stress,
eq
=
_
3/2 is the equivalent stress,
m
/
eq
is the
stress triaxiality,
p
d
is damage threshold, S is a material constant in energy density units, t is a
dimensionless material constant, and Y is the internal energy density release rate, calculated as
Y =
1
2
: [D
e
]
1
: (4.18)
4.2 IMPLEMENTATION OF THE PLASTICITY MODEL
In numerical analysis, the state variables oI a material model such as stress, and elastic and plastic
strain are updated Irom one discrete time to the next, known as 'time integration oI state variables.
The corresponding discretization and the update procedure oI these state variables, i.e., implemen-
tation oI the plasticity model is presented in this section.
4.2.1 Backward Euler Difference Scheme
First, the state variables are discretized using a backward Euler diIIerence scheme. Subscript i
represent the step number; ()
i
`s are variables at time t
i

p
i+1
=
p
i
+ (4.19a)
i+1
=
i
+ D
e
: (
p
) (4.19b)
p
=
3
2
n
i+1
(4.19c)
j
= [C
j
n
i+1

j
(
j
)
i+1
] (4.19d)
4.2.2 Elastic-Plastic Operator Split
Next, the state variables are updated using a two-step algorithm: (1) an elastic trial predictor Iol-
lowed by (2) a plastic corrector that perIorms projection oI the trial states onto the yield surIace.
The elastic predictor obtains the trial elastic states by Ireezing plastic Ilow during the time step as

p,trial
i+1
=
p
i
(4.20a)
trial
i+1
=
i
+ D
e
: (4.20b)
p,trial
= 0 (4.20c)
trial
j
= 0 (4.20d)
42
The plastic return-mapping corrector then updates the states as

p
i+1
=
p,trial
i
+ (4.21a)
i+1
=
trial
i+1
D
e
:
p
(4.21b)
p
=
3
2
n
i+1
(4.21c)
j
= [C
j
n
i+1

j
(
j
)
i+1
] (4.21d)
The plastic corrector can be implemented either using the cutting-plane algorithm or the
closest-point algorithm. For the theory oI these two algorithms, see Simo and Hughes (1998). The
development oI these two algorithms Ior the damage plasticity model is shown below.
4.2.3 Cutting-Plane Algorithm
For the cutting-plane algorithm, in the k-th iteration, the states are integrated Irom ()
(k)
to ()
(k+1)
,
with linearization about ()
(k)
:

p,(k+1)
i+1
=
p,(k)
i+1
+
2
(k)
(4.22a)
(k+1)
i+1
=
(k)
i+1
3Gn
(k)
i+1
(k)
(4.22b)
(
j
)
(k+1)
i+1
= (
j
)
(k)
i+1
+
_
C
j
n
(k)
i+1

j
(
j
)
(k)
i+1
_
(k)
(4.22c)
y
_

p,(k+1)
i+1
_

y
_

p,(k)
i+1
_
+
y

p
_

p,(k)
i+1
_
(k)
(4.22d)
Solution oI the yield Iunction
0 = F
(k+1)
i+1
F
(k)
i+1
+
F
(k)
(k)
i+1
+
F
(k)
(k)
i+1
+
F
y
(k)
(k)
y,i+1
=F
(k)
i+1
+
3
2
n
(k)
i+1
:
_
3Gn
(k)
i+1
(k)
_
+
y

p
_

p,(k)
i+1
_
(k)
3
2
n
(k)
i+1
:
_
C
j
n
(k)
i+1

j
(
j
)
(k)
i+1
_
(k)
(4.23)
obtains
(k)
=
F
(k)
i+1
3G +
y

p
_

p,(k)
i+1
_
+
j
C
j

3
2
n
(k)
i+1
:
_
j
(
j
)
(k)
i+1
_
_
(4.24)
Because linearization is used to update the state except Ior
y
_

p,(k+1)
i+1
_

y
_

p,(k)
i+1
_
+
y

p
_

p,(k)
i+1
_
(k)
(4.25)
43
the yield Iunction can be calculated that
F
(k+1)
i+1
=
y
_

p,(k)
i+1
_
+
y

p
_

p,(k)
i+1
_
(k)

y
_

p,(k+1)
i+1
_
(4.26)
II the isotropic hardening Iunction
y
(
p
) is deIined using a piecewise linear curve, and
(0)
is small enough, the linearization oI
y
is accurate. Then, the error oI F
(1)
i+1
above may be
ignored. This leads to a non-iterative algorithm:
=
2
(0)
=
F
trial
i+1
k
(4.27)
where
k = 3G +
y

p
_

p
i
_
+
j
C
j

3
2
n
trial
i+1
:
_
j
(
j
)
i
_
_
(4.28)
whereby the states are updated as

p
i+1
=
p
i
+ (4.29a)
i+1
=
trial
i+1
3Gn
trial
i+1
(4.29b)
(
j
)
i+1
= (
j
)
i
+
_
C
j
n
trial
i+1

j
(
j
)
i
_
(4.29c)
And because
i+1
=
1
k
_
_
3Gn
trial
i+1
+
3
2
n
trial
i+1
i+1
:
_
j
(
j
)
i
_
_
_
_
(4.30)
where
n
trial
i+1
i+1
=
n
trial
i+1
s
trial
e,i+1
:
s
trial
e,i+1
i+1
=
I
3
2
n
trial
i+1
n
trial
i+1

trial
i+1
:
_
2G
_
I
1
3
1 1
__
=
2G

trial
i+1
_
I
1
3
1 1
3
2
n
trial
i+1
n
trial
i+1
_
(4.31)
then Irom
i+1
=
trial
i+1
3Gn
trial
i+1
(4.32a)
d
i+1
= D
e
: d
i+1
3G[n
trial
i+1
d + dn
trial
i+1
] (4.32b)
=
_
_
D
e
3Gn
trial
i+1

i+1
3G
n
trial
i+1
i+1
_
_
: d
i+1
(4.32c)
the algorithm consistent tangent moduli can be obtained as
i+1
i+1
= D
i+1
= 1 1 + 2G
i+1
(I
1
3
1 1) 3G
i+1
n
trial
i+1
n
trial
i+1

D
i+1
(4.33)
44
where
i+1
= 1
3G

trial
i+1
(4.34a)
i+1
=
3G
k
(1
i+1
) (4.34b)
D
i+1
=
9G
2
k
trial
i+1
n
trial
i+1

_
_
_
I
3
2
n
trial
i+1
n
trial
i+1
_
:
_
j
(
j
)
i
_
_
_
_
(4.34c)
4.2.4 Closest-Point Projection Algorithm
For the closest-point projection algorithm, in iteration number k, the states are integrated Irom()
(0)
to ()
(k)
, with linearization about ()
(k)
:

p,(k)
i+1
=
p
i
+
(k)
(4.35a)
s
(k)
i+1
= s
trial
i+1
3Gn
(k)
i+1
(k)
(4.35b)
(
j
)
(k)
i+1
= (
j
)
i
+
_
C
j
n
(k)
i+1

j
(
j
)
(k)
i+1
_
(k)
(4.35c)
resulting in
s
(k)
e,i+1
= s
(k)
i+1

(k)
i+1
=
_
_
s
trial
i+1

j
(
j
)
i
1 +
j
(k)
_
_
3G +
j
C
j
1 +
j
(k)
_
(k)
n
(k)
i+1
=
(k)
i+1

_
_
3G +
j
C
j
1 +
j
(k)
_
(k)
n
(k)
i+1
(4.36)
where
(k)
i+1
= s
trial
i+1

j
(
j
)
i
1 +
j
(k)
(4.37)
From the yield Iunction
0 = F
(k)
i+1
=
(k)
i+1

y
_

p,(k)
i+1
_
(4.38)
obtain
s
(k)
e,i+1
=
(k)
i+1
n
(k)
i+1
=
y
_

p,(k)
i+1
_
n
(k)
i+1
(4.39)
Hence, the direction n
(k)
i+1
is determined exclusively in terms oI
(k)
i+1
n
(k)
i+1
=
(k)
i+1

(k)
i+1
where
(k)
i+1
=
_
3
2
_
_
_
_
(k)
i+1
_
_
_
_
(4.40)
45
ThereIore,
y
_

p,(k)
i+1
_
n
(k)
i+1
=
(k)
i+1
n
(k)
i+1

_
_
3G +
j
C
j
1 +
j
(k)
_
(k)
n
(k)
i+1
(4.41)
and the Iollowing nonlinear scalar equation is obtained
0 = g
_
(k)
_
=
(k)
i+1

_
_
3G +
j
C
j
1 +
j
(k)
_
(k)

y
_

p,(k)
i+1
_
(4.42)
and
g
(k)
_
=
3
2
n
(k)
i+1
:
_
j
(
j
)
i
(1 +
j
(k)
)
2
_
_
3G
j
C
j
(1 +
j
(k)
)
2

y

p
_

p,(k)
i+1
_
(4.43)
Advancing to the next iteration by Newton`s method
(k+1)
=
(k)
_
g
(k)
_
_
1
g
_
(k)
_
(4.44)
AIter convergence, the states are updated as

p
i+1
=
p
i
+
(k+1)
(4.45a)
s
i+1
= s
trial
i+1
3Gn
(k+1)
i+1

(k+1)
(4.45b)
(
j
)
i+1
=
(
j
)
i
+C
j
(k+1)
n
(k+1)
i+1
1 +
j
(k+1)
(4.45c)
Because
i+1
is Iixed in the plastic corrector
0 =
g
i+1
=
g
()

i+1
+
g
s
trial
i+1
:
s
trial
i+1
i+1
=
g
()

i+1
+
3
2
n
i+1
: 2G
_
I
1
3
1 1
_
=
g
()

i+1
+ 3Gn
i+1
(4.46)
So

i+1
=
3G
k
n
i+1
, where k =
g
().
46
Then Irom
n
i+1
i+1
=
n
i+1
i+1
:
i+1
i+1
=
I
3
2
n
i+1
n
i+1

i+1
:
_
_
2G
_
I
1
3
1 1
_
+
j
(
j
)
i
(1 +
j
)
2

i+1
_
_
=
I
3
2
n
i+1
n
i+1

i+1
:
_
_
2G
_
I
1
3
1 1
_
+
3G
k
j
(
j
)
i
n
i+1
(1 +
j
)
2
_
_
=
2G

i+1
_
I
1
3
1 1
3
2
n
i+1
n
i+1
_
+
3G
k
i+1
_
I
3
2
n
i+1
n
i+1
_
:
_
j
(
j
)
i
n
i+1
(1 +
j
)
2
_
_
(4.47)
and
i+1
=
trial
i+1
3Gn
i+1
(4.48a)
d
i+1
= D
el
: d
i+1
3G[n
i+1
d + dn
i+1
] (4.48b)
=
_
D
el
3Gn
i+1

i+1
3G
n
i+1
i+1
_
: d
i+1
(4.48c)
the algorithm consistent tangent moduli is obtained as
i+1
i+1
= D
i+1
= 1 1 + 2G
i+1
(I
1
3
1 1) 3G
i+1
n
i+1
n
i+1

D
i+1
(4.49)
where
i+1
= 1
3G

i+1
(4.50a)
i+1
=
3G
k
(1
i+1
) (4.50b)
D
i+1
=
9G
2
k
i+1
_
I
3
2
n
i+1
n
i+1
_
:
_
j
(
j
)
i
n
i+1
(1 +
j
)
2
_
_
(4.50c)
4.3 IMPLEMENTATION OF THE DAMAGE MODEL
The damage state is updated using a trapezoidal scheme
D
i
=
1
2
_
_
Y
i
S
_
t
+
_
Y
i+1
S
_
t
_

p
i
(4.51)
47
where
Y
i
=
1
2
i
: [D
e
]
1
:
i
(4.52)
The stress in the damaged material is then updated as
D,i+1
=
i+1
(1 D
i+1
) (4.53)
Because oI the major symmetry oI the elastic moduli, the algorithmconsistent tangent mod-
uli Ior the stress (in the damaged material) is derived as Iollows:
Y
i+1
i+1
=
1
2
i+1
: [D
e
]
1
:

i+1
i+1
+
1
2
i+1
i+1
: [D
e
]
1
:
i+1
=
i+1
: [D
e
]
1
:

i+1
i+1
(4.54)
then
D
i
i+1
=
1
2
_
_
Y
i
S
_
t
+
_
Y
i+1
S
_
t
_

i+1
+
t
2
_
Y
i+1
S
_
t1
1
S
Y
i+1
i+1
=
1
2
_
_
Y
i
S
_
t
+
_
Y
i+1
S
_
t
_

i+1
+
t
2S
_
Y
i+1
S
_
t1
i+1
: [D
e
]
1
:

i+1
i+1
(4.55)
and Iinally
D,i+1
i+1
=

i+1
i+1
(1 D
i+1
)
i+1

D
i
i+1
(4.56)
4.4 TIME-INTEGRATION EXAMPLES
The pseudo-code Ior the time-integration procedure using a non-iterative cutting-plane algorithm
Ior brick and shell elements is listed step-by-step below. For the sake oI simplicity and clarity, there
is only one back stress tensor component and isotropic hardening is not included. The complete
cyclic damage plasticity model with two back stress tensor components, piecewise linear isotropic
hardening, and damage evolution was implemented as LS-DYNAmat153 (matdamage3, LSTC,
2007). A simpliIied version with one back stress tensor component, linear isotropic hardening,
and critical equivalent plastic strain damage criterion was implemented as LS-DYNA mat165
(matplasticnonlinearkinematic, LSTC, 2007).
4.4.1 Integration Procedure for Brick Elements
1. Compute incremental average strain
avg
=

1
+
2
+
3
3
(4.57)
48
2. Compute incremental hydrostatic stress
p = 3K
avg
(4.58)
3. Compute trial stress
trial
trial
i
=
i
+p + 2G(
i

avg
) (i = 1, 2, 3) (4.59a)
trial
i
=
i
+p +G
i
(i = 4, 5, 6) (4.59b)
4. Compute trial hydrostatic stress
p
trial
=
trial
1
+
trial
2
+
trial
3
3
(4.60)
5. Compute trial deviatoric eIIective stress s
trial
e
s
trial
e,i
=
trial
i
p
trial

i
(i = 1, 2, 3)
s
trial
e,i
=
trial
i

i
(i = 4, 5, 6)
(4.61)
6. Compute second invariant oI trial eIIective deviatoric stress
J
trial
2
=
_
s
trial
e,1
_
2
+
_
s
trial
e,2
_
2
+
_
s
trial
e,3
_
2
2
+
_
s
trial
e,4
_
2
+
_
s
trial
e,5
_
2
+
_
s
trial
e,6
_
2
(4.62)
7. Compute trial Mises stress

trial
=
_
3J
trial
2
(4.63)
8. Compute Ilow direction n
n
i
=
s
trial
e,i

trial
(i = 1, 2, 3, 4, 5, 6)
(4.64)
9. Compute equivalent plastic strain increment

p
=

trial

y
3G +C
3
2
_
3
i=1
n
i
i
+
6
i=4
2n
i
i
_
_
(4.65)
10. Update total equivalent plastic strain

p
=
p
+
p
(4.66)
49
11. Update back stress
i
=
i
+ (Cn
i

i
)
p
(i = 1, 2, 3, 4, 5, 6)
(4.67)
12. Update stress
i
=
trial
i
3Gn
i

p
(i = 1, 2, 3, 4, 5, 6)
(4.68)
4.4.2 Integration Procedure for Shell Elements
1. For plane stress iteration number j = 1, compute Iully elastic normal strain
(1)
3
=

1
(
1
+
2
) (4.69)
2. Follow the procedure oI brick element and compute normal stress
(i)
3
;
3. For iteration number j = 2, compute Iully plastic normal strain
(2)
3
= (
1
+
2
) (4.70)
4. Follow the procedure oI brick element and compute normal stress
(i)
3
;
5. Compute estimated normal strain
( j+1)
3
=
( j1)
3

( j)
3

( j1)
3
( j)
3

( j1)
3
( j1)
3
(4.71)
6. Compare
( j)
3
and
( j+1)
3
, check convergence and check normal stress condition
( j)
3
= 0; iI
either check does not satisIy its tolerance, increase j, and go to step 4.
7. Update state.
The theory, implementation, and examples oI the cyclic damage plasticity models are discussed
in detail. The model was implemented by the authors in Iinite element soItware LS-DYNA as
Material 153 (matdamage3, LSTC, 2007). This newly developed material model is used in the
Iollowing chapters.
50
5 CaIibration, VaIidation, and AppIication
This chapter Iocuses on calibration, validation, and application oI the new damage material model
using experimental results Irom beam-to-column connections, individual braces, and braced-Irame
subassemblies. Material calibration procedures using monotonic and cyclic experimental data is
discussed, Iollowed by the application oI the calibrated material model to components and sub-
assemblies oI steel braced Irame.
5.1 MATERIAL CALIBRATION
The material model developed in Chapter 4 has several plasticity and damage properties. These
properties can be calibrated using monotonic and/or cyclic experimental data, discussed as Iollows.
5.1.1 Calibration Using Monotonic Data
For monotonic loading in uniaxial tension test, the stress-plastic strain curve represented by the
model is given by
=
y
+
C
1
1
[1 exp (
1
p
)] +
C
2
2
[1 exp (
2
p
)] + H
p
(5.1)
where is the uniaxial stress,
p
is the uniaxial plastic strain,
y
is the initial yield stress, C
1
,
1
,
C
2
, and
2
are kinematic hardening parameters, and H is linear isotropic hardening parameter. II
i
= 0, the nonlinear term
C
i
i
[1 exp (
i
p
)] is reduced to a linear termC
i
p
. Equation (5.1) may
be used to determine the plasticity parameters iI only monotonic tension test data is available.
The physical meaning oI kinematic hardening parameters C
1
and
1
can be illustrated by
speciIying C
2
= H = 0; then Equation (5.1) becomes
=
y
+
C
1
1
[1 exp (
1
p
)] (5.2)
making the initial slope (derivative

p
=0
) equal to C
1
, and the saturated yield stress equal to
y
+C
1
/
1
(Fig. 5.1).
y
1
1
C
1
C
1
Figure 5.1 Illustration of kinematic hardening parameters

5.1.2 Calibration Using Cyclic Data
For uniaxial stress cyclic loading, the evolution oI back stress is given by

i
= C
i

p

i
|
p
| = ( nC
i
/
i
) n
i

p
(5.3)
where
i
`s are the uniaxial back stress components,
p
is the uniaxial plastic strain rate, and n =
sign(
p
) is the plastic Ilow direction. Equation (5.3) is transIormed as

i
i
nC
i
/
i
= n
i

p
(5.4)
II the plastic strain increases or decreases monotonically, direction n is constant. Integration oI
Equation (5.4) results in
ln

i
nC
i
/
i
i0
nC
i
/
i
= n
i
_
p

p
0
_
(5.5)
where the subscripts 0 denote the initial values. Equation (5.5) can be written as
i
= nC
i
/
i
+ (
i0
nC
i
/
i
) exp
_
n
i
_
p

p
0
__
(5.6)
For a stabilized symmetric cycle, the amplitude and peaks oI back stress are
i
/2 and
i
/2, respectively. Similarly, the amplitude and peaks oI plastic strain are
p
/2 and
p
/2,
respectively. Substitute them into Equation (5.5) and obtain
i
/2 nC
i
/
i
i
/2 nC
i
/
i
=
1
exp [2 (
i
p
/2)]
(5.7)
52
which leads to the relation between back stress and plastic strain amplitudes
i
2
=
C
i
i
tanh
_
p
2
_
(5.8)
For the tension halI cycle, direction n = 1, then Equation (5.6) becomes
i
=
C
i
_
C
i
i
tanh
_
p
2
_
+
C
i
i
_
exp
_
i
_
p
+

p
2
__
(5.9)
For the tension halI cycle oI a stabilized symmetric cycle, when the cyclic hardening is saturated,
the stress-plastic strain curve is given by
=
y
+
C
1
_
C
1
1
tanh
_
p
2
_
+
C
1
1
_
exp
_
1
_
p
+

p
2
__
+
C
2
_
C
2
2
tanh
_
p
2
_
+
C
2
2
_
exp
_
2
_
p
+

p
2
__ (5.10)
Equation (5.10) may be used to determine the kinematic hardening parameters using cyclic test
data. Then, isotropic hardening parameters can be obtained by removing the kinematic hardening
part oI the stress-equivalent plastic strain curve. The material properties may be Iurther adjusted
by running numerical tests and using a least-square nonlinear Iitting process to improve Iidelity.
The above calibration process may result in widely diIIerent sets oI kinematic hardening
parameters (C
1
,
1
, C
2
, and
2
) depending upon the strain range oI interest. Interest in response
with a larger range oI strain tends to result in lower kinematic hardening modules. Using a series
oI material tests oI Japanese SM490 steel (Fujimoto et al., 1985) and SS440 steel (Nishimura et
al., 1994) as examples, calibration results in two sets oI parameters. Parameters in Set A are C
1
=
30000,
1
= 300, C
2
= 2000, and
2
= 15, and parameters in Set B are C
1
= 4000,
1
=
400, C
2
= 500, and
2
= 0. The results are shown in Figures 5.2 to 5.5 Ior SM490 steel and
Figures 5.6 to 5.9 Ior SS440 steel, both Ior cases oI cyclic loading. Compared to the experimental
results, the numerical results predicted by the parameters in Set A are more accurate than those in
Set B, in which the ranges oI strain are relative small. For larger ranges oI strain, however, the
numerical results presented with the parameters in Set A are not as good as those in Set B. This is
demonstrated in Figure 5.10, which compares results Ior monotonic tensile tests. Thus, there are
two options: a user may choose Set A to obtain round corners during smaller strain cycles, or Set
B Ior better agreement during larger strain cycles or monotonic straining. Other sets oI parameters
53
may be determined by 'trial and error to provide a balance oI characteristics more appropriate Ior
a particular application.
The best method oI calibrating damage parameters is unknown because oI current lack oI
detailed low-cycle Iatigue data. Additional research and material testing is needed.
II no speciIic input is provided by the user, 'deIault values are provided in the implemen-
tation oI the damage plasticity model to maintain backward compatibility with LS-DYNA Material
104 (LSTC, 2007). The 'deIault values Ior S and t are
y
/200 and 1, respectively, where
y
is
the initial yield stress. The eIIect oI diIIerent values oI parameter t on the rate oI deterioration can
be seen in the simulated Manson-CoIIin curves plotted in Figure 3.9.
54
-0.10 -0.05 0.00 0.05 0.10
-800
-400
0
400
800
S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical
(a) Parameters Set A
-0.10 -0.05 0.00 0.05 0.10
-800
-400
0
400
800
S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical
(b) Parameters Set B
|Experimental results by Fujimoto et al. (1985)|
Figure 5.2 Experimental versus numerical results (SM490 cyclic test 1)
55
-0.10 -0.05 0.00 0.05 0.10
-800
-400
0
400
800
S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical
-0.10 -0.05 0.00 0.05 0.10
-800
-400
0
400
800
S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical
56
-0.10 -0.05 0.00 0.05 0.10
-800
-400
0
400
800
S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical
-0.10 -0.05 0.00 0.05 0.10
-800
-400
0
400
800
S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical
57
-0.10 -0.05 0.00 0.05 0.10
-800
-400
0
400
800

S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical
-0.10 -0.05 0.00 0.05 0.10
-800
-400
0
400
800

S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical
58
-0.03 -0.02 -0.01 0.00 0.01 0.02 0.03
-600
-400
-200
0
200
400
600
S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical
-0.03 -0.02 -0.01 0.00 0.01 0.02 0.03
-600
-400
-200
0
200
400
600
S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical
|Experimental results by Nishimura et al. (1994)|
Figure 5.6 Experimental versus numerical results (SS440 cyclic test 1)
59
-0.03 -0.02 -0.01 0.00 0.01 0.02 0.03
-600
-400
-200
0
200
400
600
S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical
-0.03 -0.02 -0.01 0.00 0.01 0.02 0.03
-600
-400
-200
0
200
400
600
S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical
60
-0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06
-600
-400
-200
0
200
400
600
S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical
-0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06
-600
-400
-200
0
200
400
600
S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical
61
-0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06
-600
-400
-200
0
200
400
600
S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical
-0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06
-600
-400
-200
0
200
400
600
S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical
62
0.0 0.1 0.2 0.3 0.4 0.5
0
200
400
600
800
1000
1200

S
t
r
e
s
s

(
M
P
a
)
Strain
Experimental
Numerical Set A
Numerical Set B
(a) SM490 steel
0.0 0.1 0.2 0.3
0
100
200
300
400
500
600
700
800

S
t
r
e
s
s

(
M
P
a
)
Plastic strain
Experimental
Numerical Set A
Numerical Set B
(b) SS440 steel
Figure 5.10 Experimental versus numerical results (monotonic loading)
63
5.2 VALIDATION AND APPLICATION TO COMPONENTS
Finite element analyses were done to assess the ability oI the newly developed damage plasticity
model to simulate the hysteretic behavior oI braces, braced-Irame subassemblies, and beam-to-
column connections. The Iocus oI this investigation was to evaluate the prediction oI local buckling
and the evolution oI damage because oI low-cycle Iatigue. Individual braces and beam-column con-
nection subassemblies are described in this section and braced-Irame subassemblies are described
in Section 5.3.
All Iinite element models were built using Iully integrated shell elements (Engelmann et
al., 1989; Simo and Hughes, 1986; Pian and Sumihara, 1985) and implicit time integration. This
type oI shell element is based on a combined co-rotational and velocity-strain Iormulation. An
embedded element coordinate system that deIorms with the element is deIined in terms oI the
nodal coordinates (LSTC, 2007). The Mindlin theory oI plates and shells (Mindlin, 1951) is used
to determine the velocity oI any point in the shell.
Shell elements instead oI solid elements were selected Ior these studies because the optimum
element type determined here Ior single braces is adopted Ior subsequent analyses oI braced-Irame
subassemblies and beam-column connections. Because the computational eIIort Ior a complete
braced Irame is substantially greater than Ior an individual brace, eIIicient Iinite element models
are preIerred. A shell element is more time-consuming than a solid element in that shell elements
require zero-normal stress iterations. But many more solid elements are required through the wall
thickness to capture the combined membrane and plate actions. Because shell and solid elements
have the same number oI degrees-oI-Ireedom per element, more total CPU time is expected Ior
models developed with solid elements than those with shell elements.
The choice oI shell elements instead oI beam elements is based on the consideration that
beam elements assume that plane sections remain plane and the sectional coordinates oI integration
points on plane sections remain constant during the course oI analysis. This assumption makes it
impossible to model local buckling oI a tube-section brace using beamelements. For shell elements,
a tube section is built with several elements. ThereIore, there are suIIicient degrees-oI-Ireedom
related to local buckling deIormation.
Crack initiation and propagation is modeled by element erosion (element removal). Once
64
the material point damage parameter exceeds a critical damage state corresponding to Iracture, it
is considered Iailed and is removed Irom a model. Although the shell element size is much larger
than the material characteristic length oI deterioration ( 0.01 in., Kanvinde and Deierlein, 2004),
the gradients oI equivalent plastic strain and damage variable contours are small with respect to the
characteristic length beIore crack initiation. ThereIore, it is acceptable to have shell element sizes
larger than the characteristic length, but a mesh convergence test is required. Mesh convergence
Ior a single brace is examined in Section 5.2.1 using progressively reIined Iinite element meshes.
5.2.1 Single Brace
AIinite element model was built Ior the single-brace test specimens discussed previously in Section
2.1 (Yang and Mahin, 2005). Because the yield stress and ultimate stress were the only material
properties known Ior the braces, the damage evolution constants Ior the steel material needed to be
calibrated using the overall results Ior individual braces. The applied displacement history Ior each
experiment was extracted and prescribed as a time-varying displacement boundary condition at one
end oI the brace and the other end was Iixed. No initial imperIection was introduced in the Iinite
element model. The buckling was expected to be trigged by perturbations oI numerical truncation
or round-oII error.
The analysis results are shown in Figures 5.11 to 5.14. Global buckling (Fig. 5.11), local
buckling and damage localization (Fig. 5.12), crack initiation (Fig. 5.13), and crack propagation
(Fig. 5.14) were successIully simulated, demonstrating that the cyclic damage plastic model is
suitable even though only basic plasticity and damage models were used. Comparisons between
experimental and numerical results Ior the axial Iorce axial displacement hysteretic curves are
shown in Figures 5.15 to 5.17, and the comparisons Ior peak loads in each halI cycle are shown in
Figures 5.18 to 5.20. By incorporating the damage model, the numerical analyses predict with a
Iair degree oI accuracy the strength deterioration and the resulting Iorce-displacement curve over
the whole loading history. This makes it reasonable to extend the modeling oI an individual brace
to a larger model oI a complete braced-Irame subassembly.
The mesh sensitivity oI plastic strain was validated using models with successively smaller
elements. Plastic strain was chosen as the index oI quality because it is the major source oI dam-
age. Four diIIerent element sizes in the reIined mesh region at brace mid-span (Fig. 5.12) were
65
evaluated. Figure 5.21 shows the time history oI plastic strain Ior these Iour element sizes. Note
that both the equivalent plastic strain and the damage state are seen to converge reliably when the
element size is on the order oI shell thickness. Obviously aIter crack initiation, the gradients oI
both the equivalent plastic strain and the damage state are much higher. Mesh sizes larger than the
material characteristic length combined with element erosion Ior cracking will result in a larger en-
ergy release rate at the crack tip (Xia and Shih, 1995) and blunt the crack Iront to an unrealistic size,
resulting in an exaggerated ductile behavior oI the crack tip. But this is a local behavior oI the crack
tip, and beIore it behaves inelastically, the strength and stiIIness oI a typical brace member have sig-
niIicantly deteriorated because oI lateral and local buckling; thereIore, the exaggerated crack Iront
blunting when using large-size shell elements should have less inIluence on the Iorce-deIormation
perIormance. Note that mesh sizes at the same scale oI the material characteristic length may be
required Ior analyses oI local crack tip behavior. It is believed that the choice oI shell element at
sizes around the shell thickness achieves an overall model that is as simple as reasonably possible
with reasonable accuracy Ior analyses oI braces and braced-Irame subassemblies. This choice oI
shell element size is applied to all analyses in this research.
66
(dashed line Ior undeIormed shape)
Figure 5.11 Global buckling of brace specimen 5 (after Yang and Mahin, 2005)
Figure 5.12 Local buckling of brace specimen 5 (after Yang and Mahin, 2005)
67
Figure 5.13 Crack initiation of brace specimen 5 (after Yang and Mahin, 2005)
Figure 5.14 Crack propagation of brace specimen 5 (after Yang and Mahin, 2005)
68
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5
-600
-400
-200
0
200
400
600

A
x
i
a
l

l
o
a
d

(
k
i
p
s
)
Axial displacement (in.)
(a) Experimental result
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5
-600
-400
-200
0
200
400
600

A
x
i
a
l

l
o
a
d

(
k
i
p
s
)
(b) Numerical result
Figure 5.15 Hysteresis loops of brace specimen 5 (after Yang and Mahin, 2005)
69
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
-600
-400
-200
0
200
400
600

A
x
i
a
l

l
o
a
d

(
k
i
p
s
)
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
-600
-400
-200
0
200
400
600

A
x
i
a
l

l
o
a
d

(
k
i
p
s
)
70
-1.0 -0.5 0.0 0.5 1.0 1.5
-600
-400
-200
0
200
400
600

A
x
i
a
l

l
o
a
d

(
k
i
p
s
)
-1.0 -0.5 0.0 0.5 1.0 1.5
-600
-400
-200
0
200
400
600

A
x
i
a
l

l
o
a
d

(
k
i
p
s
)
71
0 1 2 3 4 5 6 7 8 9 10
-600
-400
-200
0
200
400
600

P
e
a
k

l
o
a
d
s

(
k
i
p
s
)
Cycle number
Experimental
Numerical
Figure 5.18 Peak loads of brace specimen 5 (after Yang and Mahin, 2005)
0 2 4 6 8 10 12 14 16
-600
-400
-200
0
200
400
600

P
e
a
k

l
o
a
d
s

(
k
i
p
s
)
Cycle number
Experimental
Numerical
72
0 2 4 6 8 10 12 14
-600
-400
-200
0
200
400
600

P
e
a
k

l
o
a
d
s

(
k
i
p
s
)
Cycle number
Experimental
Numerical
0 10 20 30 40 50 60 70
1
2
3
4
5
6
7
8

E
q
u
i
v
a
l
e
n
t

p
l
a
s
t
i
c

s
t
r
a
i
n
Pseudo Time
Element size
2.7 t
1.3 t
1.0 t
0.5 t
Figure 5.21 Evolution and mesh sensitivity of plastic strain (shell thickness t)
73
5.2.2 Single Beam-Column Connection
A Iinite element model was built Ior the welded steel beam-column connection subassembly tested
by Tanaka et al. (2000) and discussed previously in Section 2.4. HalI oI the specimen is modeled
and symmetry boundary conditions were applied. The Iinite element model is shown in Figure
5.22. The newly developed cyclic damage plasticity model was used. The parameters oI com-
bined isotropic and nonlinear kinematic hardening were calibrated against experimental data Irom
Nishimura et al. (1994).
Figure 5.23 shows the numerical result oI the load-displacement curve compared with the
experiment. Most oI the hysteretic Ieatures are well captured by the simulation. Some comparisons
oI the simulations and the experiments Ior the buckling shape are shown in Figures 5.24 to 5.26.
The simulation results match the experimental ones very well.
The last halI cycle oI the experiment displacement history starts Irom 60 mm to 80mm.
In the numerical analysis, the displacement history was extended with 10 cycles oI displacement
amplitude at 80 mm. Further local buckling was observed and the hysteretic loop stabilized, but
no low-cycle Iatigue-induced Iracture was observed. In order to apply and validate the material
model, the beam-column connection test by Suita et al. (2000) was also simulated and is discussed
below.
74
Figure 5.22 Finite element modeling of BW10 specimen
-100 -80 -60 -40 -20 0 20 40 60 80 100
-600
-400
-200
0
200
400
600
Experiment
Numerical

L
o
a
d

(
k
N
)
Displacement (mm)
Figure 5.23 Hysteretic loop of BW10 specimen
75
Figure 5.24 BW10 behavior comparison 1
76
77
78
The beam-column connections experiment by Suita et al. (2000) discussed in Section 2.4
was also analyzed. The specimens ruptured because oI low-cycle Iatigue in a region oI the beam
Ilanges where large plastic deIormation and local buckling occurred. Specimen N4 was numerically
simulated and the results are plotted in Figures 5.27 to 5.30.
The Iinite element model used is shown in Figure 5.27. Because there was no weld access
hole, the structural mesh was easily established. To reduce computational costs, only the region
near the Iace oI the beam-column connection is reIined. Because a single beam Iramed into the
column, and assumptions oI symmetry could not be used as in the case oI the column tested by
Tanaka et al. (2000), the whole column (instead oI halI) was modeled.
The newly developed cyclic damage plasticity model was used Ior the simulations. The ini-
tial estimate used Ior the material modeling parameters was almost identical to that used in BW10,
except that the initial yield stress was changed to match those obtained Irom coupon tests Ior Spec-
imen N4. Damage was calculated, and the elements eroded during the analysis as damage accu-
mulated.
Because no detailed cyclic material stress-strain data was available Ior this specimen to
calibrate the model, the deIault damage parameter set Ior the material was used as an initial estimate.
The 'deIault damage parameter S is around Fy/200 to F
u
/200, where F
y
and F
u
are the yield stress
and the ultimate stress oI the steel, respectively. For this test specimen (N4, Set B), the reported F
y
Ior the Ilange was 295 MPa and that Ior the web was 325 MPa.
Figure 5.28 presents and compares the numerical load-displacement curves with the exper-
imental data. Both buckling and damage Ieatures are well captured by the analysis. The buckled
shape is shown in Figure 5.29, and the Iinal Iracture pattern in the Ilange is shown in Figure 5.30.
These results match the Iinal damage description in Suita et al. (2000) very well.
79
Figure 5.27 Finite element modeling of N4 specimen
-0.08 -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

m
o
m
e
n
t

M
m

/

M
p
Rotation (rad)
Experimental
Numerical
Figure 5.28 Hysteretic loop of N4 specimen
80
Figure 5.29 Predicted local buckling and damage
Figure 5.30 Predicted fracture in top flange
81
5.3 VALIDATION AND APPLICATION TO SUBASSEMBLIES
5.3.1 Subassembly Subjected to Quasi-Static Loading
A Iinite element model was developed Ior the braced-Irame subassembly discussed previously in
Section 2.3, as an extension oI the approach used to model Ior the individual brace discussed in
Section 5.2.1. Details including the gusset plate and the shear tab in beam-column connection are
modeled. For the sake oI simplicity, no out-oI-plane oIIset oI the shear tab is modeled. The rooI
level displacement was prescribed. The base was Iixed and out-oI-plane constraints were imposed
on the column ends and beam mid-spans as they were in the experiment.
The analysis results are shown in Figures 5.31 and 5.32. The distribution oI damage within
the subassembly is accurately simulated (Fig. 5.31). The sequence and nature oI global and local
buckling and Iracture oI braces are very similar to those oI test results. In additional, the damage and
Iracture at the beam-column connection, right near the corner oI the shear tab, were predicted as well
(Fig. 5.32). Considering that the material parameters used were approximate, these results show
that the cyclic damage plasticity model predicts overall response to a reasonable degree and should
be useIul Ior damage evaluation oI the steel structures, especially iI the material properties are well
deIined. Figures 5.33 and 5.34 show the base shear-rooI displacement hysteresis curves Ior the
experimental and numerical analysis, respectively. Note that strength, stiIIness, and deterioration
in the overall behavior oI the braced Irame are well simulated.
82
Figure 5.31 Damage and fracture of brace
Figure 5.32 Damage and fracture of beam-column connection
83
-5 -4 -3 -2 -1 0 1 2 3 4 5
-600
-400
-200
0
200
400
600

L
o
a
d

(
k
i
p
s
)
Roof displacement (in.)
Figure 5.33 Experimental result for braced frame
-5 -4 -3 -2 -1 0 1 2 3 4 5
-600
-400
-200
0
200
400
600

L
o
a
d

(
k
i
p
s
)
Figure 5.34 Numerical result for braced frame
84
5.3.2 Subassembly Subjected to Dynamic Loading
The previous subsection simulated the subassembly subjected to quasi-static loading. To predict re-
sponse under dynamic loading, masses were assigned to the Iloor levels, and the LA23 ground mo-
tion was imposed. LA23 is one oI the 20 ground motions assembled Ior the SACproject (Somerville
et al., 1997) representing an exceedance probability oI 2 in 50 years, derived Irom the Iree-Iield
motion recorded at a distance oI 3.5 km during the 1989 Loma Prieta earthquake. The design base
shear capacity oI the structure, based on the methods used in ASCE 7-05, was used to determine
the weight oI a structure that the code would assume could be supported by the braced Irame. The
total weight was equally distributed between both Iloors. The model and ground motion are shown
in Figure 5.35.
The Iailure mode at time t 13.8 second is shown in Figures 5.36 to 5.38 Irom diIIerent
points oI view. The braces in the Iirst story buckled and restraightened repeatedly, causing the
bottom story to become a soIt story.
From the hysteretic loops shown in Figure 5.39 and Iirst-story driIt time history shown in
Figure 5.40, one can see that extremely large interstory driIts occur Ior the selected earthquake
motion. Comparison oI models with and without damage incorporated are also shown in Figures
5.39 and 5.40. For the model without damage, the structure does not collapse; Ior the one with
damage, substantial deterioration oI strength and stiIIness occurs and the structure collapses. The
signiIicant diIIerence suggests the importance oI considering damage in the simulation.
Computational results obtained with the new cyclic damage plasticity model correlate well with the
test results Ior individual braces, beam-to-column connections, and braced-Irame subassemblies.
These applications oI the model illustrate the importance oI considering material nonlinearity, local
and global buckling, and the exhaustion oI the ability oI the material to deIorm inelastically caused
by low-cycle Iatigue.
85
Figure 5.35 Subassembly model and ground motion
Figure 5.36 Failure mode at time t 13.8 sec, global view
86
Figure 5.37 Failure mode at time t 13.8 sec, local view 1
Figure 5.38 Failure mode at time t 13.8 sec, local view 2
87
-4 -2 0 2 4 6 8 10
-800
-600
-400
-200
0
200
400
600
800

B
a
s
e

s
h
e
a
r

(
k
i
p
s
)
First story drift (%)
LA23
With damage
Without damage
Figure 5.39 Base shear versus first-story drift
0 4 8 12 16 20
-5
0
5
10
15
20

F
i
r
s
t

s
t
o
r
y

d
r
i
f
t

(
%
)
Time (s)
LA23
With damage
Without damage
Figure 5.40 First-story drift time history
88
6 Studies of SteeI Braced Frame Behavior
For Special Concentrically Braced Frame (SCBF) systems, energy dissipation is achieved by ten-
sion yielding and compression buckling oI the braces, and all other components are intended to
remain essentially elastic. To achieve desirable ductile behavior in the braces, the AISC Seismic
Provisions Ior Structural Steel Buildings (AISC, 2005b) have several requirements Ior proportions
and details oI SCBF systems. Requirements that are closely related to studies oI steel braced Irame
behavior in this chapter include:
Slenderness of brace. Bracing member slenderness is quantiIied as the slenderness ra-
tio KL/r, where K is the eIIective length Iactor, L is the laterally unbraced length oI the
member, and r is the governing radius oI gyration. Brace slenderness is limited to ensure
adequate compressive strength and resistance to the cyclic degradation oI the brace.
Compactness of section. Sections are classiIied as compact, noncompact, or slender-
element sections. Compact sections are those capable oI developing a Iully plastic stress
distribution and oI possessing a rotation capacity oI approximately three beIore the onset
oI local buckling (AISC, 2005b). Compactness oI section is quantiIied in terms oI width-
thickness ratios = b/t. For Ilanges oI rectangular hollow structural sections, width b is
the clear distance between webs less the inside corner radius on each side, and the thick-
ness t is the design wall thickness. The post-buckling perIormance oI the braces is also
dependent on the compactness oI the members. Higher width-to-thickness ratios may lead
to earlier and more severe local buckling, which in turn results in premature rupture oI the
brace and lower energy dissipation capacity.
Connections. Connections are responsible to combine and transmit Iorces between the
bracing element and Iraming beams and columns. Requirements Ior these connection de-
tails help realize ductility and energy dissipation oI the system by preventing premature
Iailure modes.
Beams. For V-type and inverted V-type Irames, beams must be continuous between columns
and must be designed Ior vertical unbalanced loads associated with yielding oI the tensile
brace and buckling oI the compressive brace. Beam top and bottom Ilanges must be braced
at the intersection oI the braces and beam.
This chapter Iirst looks at the eIIects oI brace proportions (slenderness ratio KL/r and width-
thickness ratios b/t) on the Iatigue liIe capacity oI bracing members subjected to cyclic deIormation
histories. Here, Iatigue liIe demand is how much the brace needs to deIorm inelastically to achieve
expected system perIormance. In contrast, Iatigue liIe capacity is how much the brace is able to
deIorm inelastically beIore Iailure occurs Irom low-cycle Iatigue. Fatigue liIe demand can Ior ex-
ample be speciIied as the demanded deIormation history oI a brace corresponding to two interstory
driIt cycles oI 2, Iollowed by three interstory driIt cycles oI 1.
Aparametric study oI Iatigue liIe demand Ior SCBF systems with diIIerent structural period,
brace slenderness and compactness, ground motion characteristics, and so on, is not considered in
this research. In the current research, only Iatigue liIe capacity is investigated. Factors aIIecting
Iatigue liIe capacity are discussed in Section 6.1. AIter Iatigue liIe demand and capacity are both
well understood, one can compare themIor a given structure and to avoid Iracture, adjust the design
parameters oI the structure such that the Iatigue liIe demand is satisIied by the Iatigue liIe capacity.
Code requirement oI braces proportions (slenderness oI member and compactness oI section) may
be similarly determined by statistical comparison oI Iatigue liIe demand and capacity oI diIIerent
structures.
Next, the eIIect oI connection details on damage accumulation in the column is studied.
Although the tensile, Ilexural, and compressive strength requirement Ior connection in the AISC
Seismic Provisions helps Iorces transIer and ensure system integrity, high local stress and large
plastic deIormation in the column leading to Iracture may aIIect the perIormance. Some connection
details may result in crack initiation and propagation in the column, as observed in Uriz (2005) (Fig.
2.7), and simulated in Section 5.3 (Fig. 5.32). Because damage control oI columns is essential to
maintain integrity oI the gravity load resisting system, it is necessary to evaluate and compare
damage evolution in columns Ior diIIerent connection details. To reduce the damage accumulation
90
in the column, improved connection details are recommended in Section 6.2.
The position oI lateral bracing Ior beams in V-type and inverted V-type braced Irames is
discussed next. The most common position Ior lateral bracing is at the intersection oI braces and
beam, i.e., both top and bottom Ilanges are laterally braced at the section passing through the work-
ing point oI braces at beams. As shown in Section 6.3, however, lateral bracing at this position
may result in local buckling oI gusset plates. More appropriate positions and methods to compen-
sate Ior problems detected Ior currently recommended lateral bracing positions are suggested and
evaluated.
Finally, Chen et al. (2008) Iound that Ior low-rise braced Irames with short Ioundational
periods, estimates oI interstory driIt based on elastic spectra diverge Irom nonlinear response his-
tory analysis results Ior intense ground motions. An alternative interstory driIt estimation method
Ior low-rise braced Irames that respond inelastically during strong earthquake ground shaking is
suggested based on the modal pushover analysis procedure. This method is discussed in Section
6.4.
The Iollowing questions are answered Ior all the Iour studies oI steel braced Irame behavior
in this chapter.
1. Is the Iatigue liIe capacity oI braces a Iunction oI the slenderness ratios KL/r or the width-
thickness ratios b/t, or a Iunction oI both?
2. Which connection details mitigate damage accumulation in the columns?
3. Is there a more appropriate position oI lateral bracing Ior beams than the intersection oI
braces and beam?
4. It is possible to better estimate the interstory driIt Ior low-rise braced Irames?
6.1 BRACE PROPORTIONS
The proportions oI the braces in a SCBF system, including their slenderness and compactness, are
critical when considering compressive buckling strength and, more importantly, the postbucking
hysteresis perIormance oI brace members. It is well established that overall instability oI braces,
including overall compressive buckling strength and cyclic strength degradation, are mostly de-
91
pendent on slenderness ratios KL/r (Bruneau et al., 1998; Galambos, 1998). In addition, the
width-thickness ratio b/t has been shown to be a signiIicant Iactor Ior Iracture because oI local
instability, including local buckling and low-cycle Iatigue (Goel, 1992; Hassan and Goel, 1991;
Tang and Goel, 1989). Despite these advances, it is still unclear whether and how the slenderness
ratios KL/r aIIect the Iatigue liIe capacity oI braces, which is closely related to local instability.
Tremblay (2000) demonstrated that Irames with very slender braces designed Ior compres-
sion strength behavior behave well because oI the system overstrength inherent with their tension
yielding, which results in a reduction in the Iatigue liIe demand. Higher overstrength Iactor reduc-
ing the demand oI Iatigue liIe can be explained intuitively by Newmark and Hall`s energy preserva-
tion Ior short-period systems (Newmark and Hall, 1982): iI the strength increases, the deIormation
and ductility demand decreases. Besides the overstrength Iactor, it is believed that Iatigue liIe de-
mand also depends on the period and the hysteresis oI the structure, the characteristics oI ground
motion, and the contribution oI Iraming actions, etc. A separate systematic research on Iatigue liIe
demand is expected to be carried out and not considered in this research.
Regarding the Iatigue liIe capacity, dependence on diIIerent Iactors has been proposed in
existing literature: dependence on slenderness ratios KL/r only, dependence on width-thickness
ratios b/t only, and dependence on both. Goel and his colleagues showed that the Iatigue liIe
capacity oI a brace member depends on both slenderness ratios KL/r and width-thickness ratios
(b
2t)/t (equivalent to b/t in AISCSpeciIications). The Iatigue liIe capacity is generally constant
when slenderness ratios KL/r are smaller than 60 but increases with an increase in slenderness
ratios KL/r when it is larger than 60 (Tang and Goel, 1987, 1989). They give:
N
f
=
_
_
262
(b
/d
) (60)
[(b
2t)/t]
2
(KL/r 60)
262
(b
/d
) (KL/r)
[(b
2t)/t]
2
(KL/r > 60)
(6.1)
where N
f
is the accumulated tensile deIormation normalized by the tensile yield deIormation, the
width b
is the clear distance between webs, and the depth d
is the clear distance between Ilanges.

Based on the test results oI Lee and Goel (1987) and Liu and Goel (1987), Tang and Goel (1987,
1989) observed that a larger slenderness ratio KL/r resulted in less severe local buckling and pro-
posed that overall buckling also aIIects the Iracture liIe capacity oI the bracing members. Based on
study oI the hysteretic behavior oI bracing members in Jain and Goel (1978), Tang and Goel (1987,
92
1989) also proposed that the eIIect oI the slenderness ratio KL/r is not considered in the Iatigue
liIe capacity when it is less than 60.
Lee and Goel (1987) proposed a reIined Iatigue liIe capacity Iormula, suggesting that the
ductility depends on only compactness:
f
= 1335
(46/F
y
)
1.2
[(b
2t)/t]
1.6
_
4b
/d
+ 1
5
_
(6.2)
where
f
is the accumulated eIIective tension excursion oI braces, deIined as
f
=
(0.1
1
+
2
) (6.3)
In a tension halI cycle oI the load-deIormation loops oI the braces,
1
is the tension deIormation
Irom the load reversal point to one third oI the tension yield strength oI the brace P
y
/3, and
2
is
the tension deIormation Irom one third oI the tension yield strength oI the brace P
y
/3 to the end
point oI the tension halI cycle.
Tremblay (2002) analyzed existing experimental data Ior braces with various slenderness
and compactness ratios. He proposed that the normalized range in a cycle deIormation history is a
linear Iunction oI the slenderness ratio KL/r.
f
= 2.4 + 8.3
kL
r
_
F
y
2
E
(6.4)
where
f
is the peak-to-peak deIormation range normalized by tensile yield deIormation. Tremblay
et al. (2003) later proposed that
f
= 1 +
2L
D
y
_
_
0.0455
_
b
t
d
t
_
0.1 _
KL
r
_
0.3
_
_
(6.5)
where D
y
is the tensile yield deIormation oI the braces. These two equations suggest either no or
little dependence on width-thickness ratios b/t.
Recently, Fell (2008) analyzed almost the same set oI experimental data Ior braces, but
proposed that the maximum range in a cyclic deIormation history is a Iunction oI width-thickness
ratios b/t:
f
=
L
2
_
_
b
/2 + md
_ _
b
t
_
_
2
(6.6)
where
f
is the peak-to-peak deIormation range oI braces, the width b
is the clear distance between

webs, and the depth d
is the clear distance between Ilanges. Based on experimental data, Fell

93
(2008) suggested the plastic hinge length is given by L
h
= b
/2 + md
, m = 0.5, = 0.9, and

= 0.53.
These conIlicting opinions regarding Iactors controlling Iatigue liIe capacity arise Iromsev-
eral reasons, the primary one being the characteristics oI the experimental data. A comparison oI
the statistical data in Tremblay (2002) and Fell (2008) demonstrates that the specimens with higher
ductility have larger slenderness ratios and smaller width-thickness ratios, and the specimens with
lower ductility have smaller slenderness ratios and larger width-thickness ratios.
It is diIIicult to distinguish the contributions oI these two parameters: does the ductility
comes Irom larger slenderness ratios or smaller width-thickness ratios? The dispersion oI ductility
may also be attributed to variation oI the ductility measure, variation oI the deIormation histories,
and probable variations oI the steel Iatigue properties. Clearly, more research is needed on this
issue.
In summary, in order to get a better understanding oI how proportions and deIormation
histories eIIect ductility, experiments with one parameter altered at each step in the study would
be necessary. This type oI experiment is conducted numerically and presented in this section. The
numerical parameter matrix is shown is Table 6.1. The Iirst two parameters are width-thickness
ratios b/t and slenderness ratios KL/r. For the hollow square section, the radius oI gyration is
proportional to the width b. Thus by changing thickness t and Iixing width b, it is possible to vary
b/t while keeping KL/r constant, and vice versa. The third parameter is the number oI cycles per
deIormation level. The Iourth parameter is the tension/compression asymmetry ratio. For a T/C
asymmetry ratio oI 3:1 and deIormation range oI 1.0, the peak tensile deIormation is 0.75 and peak
compressive deIormation is 0.25. The deIormation history imposed has an increasing amplitude
level series oI 0.5, 0.75, 1.0, 1.5, 2.0, 3.0, 4.5, 6.0, 7.5, 9.0, 10.5, and 12.0 times tensile yield
displacement (range series oI 1.0, 1.5, 3.0, 4.0, 6.0, 9.0, 12.0, 15.0, 18.0, and 21.0 times tensile
yield displacement). ThereIore, the Iatigue liIe can be uniquely identiIied Ior a given deIormation
level to Iailure, where 'Iailure is deIined here as cracking oI three sides oI a brace section and
signiIicant loss oI strength (Fig. 5.14).
Brace 1 in Table 6.1 used as the reIerence in these analyses is the same size as the sin-
gle brace described in the previous chapter (Section 5.2.1), with a b/t oI 14.2 and KL/r oI 50.
The number oI cycles per deIormation level Ior the reIerence case is two, the tension/compression
94
asymmetry ratio is 1:1. The corresponding deIormation history is shown in Figure 6.1, and the
deIormation level to Iailure Ior the reIerence brace is
f
/2 = 7.5 (Figs. 6.3b, 6.4b, 6.5b, 6.6b, and
Table 6.1). Note that
f
is the deIormation range to Iailure normalized by tensile yield deIorma-
tion, and
f
/2 is deIormation level (deIormation amplitude) to Iailure normalized by tensile yield
deIormation.
The numerical results are grouped and displayed in Figures 6.3 to 6.6, and the deIormation
levels to Iailure are summarized in Table 6.1. The inIluence oI width-thickness ratios b/t is shown
in Figure 6.3 demonstrating clearly that the width-thickness ratios b/t have a pronounced eIIect
on ductility capacity: the smaller the width-thickness ratios, the higher the ductility. As shown in
Figure 6.4, however, even though it is clear that specimens with smaller KL/r have higher com-
pressive buckling strength and tend to dissipate more energy within each cycle, it is more noticeable
that they have an almost identical deIormation level at Iailure. In other words, slenderness ratios
have little eIIect on ductility capacity.
The number oI cycles per deIormation level also plays an important role (Fig. 6.5). Ob-
viously, increasing the number oI cycles at a deIormation level causes more cumulative plastic
deIormation (and thus more damage). This suggests that Ior a given ground motion, structures
with shorter periods may experience more cycles, and thus more damage Ior the same maximum
driIt is expected. Finally, Ior the asymmetry deIormation history, bias in terms oI compression
leads to less ductility, as shown in Figure 6.6, which may be because compressive deIormation is
concentrated at the mid-span oI the brace where local buckling occurs, but tensile deIormation is
distributed along the length oI brace.
Based on numerical analysis, it is evident that Iatigue liIe capacity is heavily dependent
on width-thickness ratios b/t and deIormation histories. Slenderness ratios KL/r appear to have
negligible, iI any, eIIect on Iatigue liIe capacity. It is worth noting that slenderness ratios KL/r
may reduce Iatigue liIe demand (Tremblay 2000; Uriz 2005). CareIul experimental validation oI
these observations Irom numerical simulations is advised.
Based on the above discussion and evaluation on Iatigue liIe capacity, Ior its counterpart,
Iatigue liIe demand, the Iollowing important parameters are suggested. Fatigue liIe demand may
be speciIied as the demanded deIormation history oI a brace, e.g., corresponding to two interstory
driIt cycles oI 2 Iollowed by three interstory driIt cycles oI 1. In this sense, the Iatigue liIe
95
demand is, essentially, a loading protocol Ior braces. The Iour most important basic parameters Ior
realistic loading protocols are numbers, amplitudes, sequence, and asymmetry oI cycles.
These Iour parameters depend on many Iactors:
First, they depend on the intensity |e.g., peak ground acceleration (PGA) and duration| and
the Irequency content oI ground motion, which in turn depend on the magnitude oI the
earthquake, the distance Irom the epicenter, the soil type at the site, etc.
Second, they depend on the conIiguration, strength, stiIIness, and modal properties (peri-
ods and participation Iactors) oI the structure, as well as the contribution oI Iraming beam
and column members, and the eIIects oI response modiIication coeIIicients (R Iactors) and
system overstrength Iactors (
0
Iactors) (ASCE 7-05).
Third, they depend on the deterioration characteristics oI the structure. For braced Irames,
compressive strength, overstrength Iactor, and resistance to the cyclic degradation oI braces
may depend on the slenderness ratios KL/r as well as steel yield strength F
y
. All these
Iactors are equally important and interact with each other. For example, the period oI a
structure has a signiIicant eIIect on the response to near-Iault ground motions; compressive
strength, overstrength Iactor, and slenderness ratios KL/r may be not Iully independent oI
each other and they may be reduced to a smaller number oI independent Iactors.
There are many decisions and judgments to be made in developing a loading history that
is statistically representative oI the Iull range oI ground motions and structural characteristics. A
detailed discussion oI the development oI loading histories Ior steel moment Irames and wood-
Irame structures that are good guidelines Ior Iuture development oI Iatigue liIe demand Ior braced
Irame systems can be Iound in Krawinkler et al. (2000a, b). For braced Irame systems, it is rec-
ommended that several variations be considered, including consideration oI both 'ordinary and
near-Iault ground motions, the duration oI earthquake versus the period oI a structure, the slen-
derness ratios KL/r and resulting overstrength Iactor, and steel yield strength F
y
and hardening
properties. Other variations oI parameters should also be considered, with the cumulative damage
concepts kept in mind. Considering both a series oI response history analyses oI single-story braced
Irame systems and a statistical analysis oI the response history oI brace deIormation, a representa-
96
tive loading history can be developed with proper numbers, amplitudes, sequence, and asymmetry
oI cycles.
Once the deIormation history or the Iatigue liIe demand is developed, code requirements
Ior brace proportions can be determined. For example, iI the Iollowing assumptions are met:
1. The reIerence history in this parametric study (Fig. 6.1) is a typical deIormation history;
2. The plasticity and damage properties calibrated against Yang and Mahin (2005) in this re-
search are typical material properties Ior ASTM A500 Grade B steel;
3. The braces oI the braced Irames are oriented at 45 degrees, and the nominal yield strain oI
braces is
y
= 0.16 approximately;
then Irom a normalized deIormation level to Iailure Ior diIIerent width-thickness ratios b/t sum-
marized in Table 6.1, a relationship can be developed between interstory driIt ratios to Iailure and
width-thickness ratios b/t. For example, Ior b/t = 14.2, the deIormation level to Iailure is 7.5
times the yield deIormation. By assuming 45-degree orientation oI braces, the interstory driIt to
Iailure is 2
y
7.5 = 2.4. This relationship is shown in Figure 6.2. Based on this relationship,
in order to achieve interstory driIt capacity without Iailure at 2.5, the limiting width-thickness
ratio b/t is 13.1.
Alternatively, test protocols oI the 'pass or Iail type proposed by Tremblay and Bouatay
(2002) may be used to determine the limiting width-thickness ratios b/t. In this type oI test, a brace
system is said to be adequate iI it survives the applied displacement history. The largest width-
thickness ratio b/t oI survivals is the critical Iactor. Tremblay and Bouatay`s deIormation histories
do not include all the variations discussed herein. In particular, they are independent oI structural
properties (such as the slenderness ratios KL/r), which may be important based on observations by
Tremblay (2000); the histories were proposed with careIul consideration oI the individual eIIects
oI Iour ground motion types.
The Iour proposed deIormation histories adopted in current research represent (1) crustal
(intra-plate) earthquakes that occur in Eastern North America (rich in high Irequency), (2) crustal
earthquakes occurring along the PaciIic coast, (3) west coast near-Iield ground motions that are
characterized by large acceleration pulses leading to high ground velocity and displacement, and
97
(4) earthquakes occurring along the Cascadia subduction zone. These Iour deIormation histories are
shown in Figure 6.7. In order to show clearly the Iailure oI unsatisIying braces, a tensile excursion
is appended at the end oI the original histories. 'Failure is deIined here as cracking oI three
sides oI a brace section and a signiIicant drop oI strength, and the appended tensile excursion and
corresponding resistance response are purely Ior visual demonstration oI the Iailure ('strength lost
or 'strength recovered annotated in Figs. 6.8 to 6.11).
When a braced Irame is subjected to an interstory driIt history, one brace in the pair is in
tensile deIormation while the other is in compressive deIormation and vice versa iI deIor-
mation oI beam is ignored. Based on the previous parametric analysis in this section, compressive
asymmetry causes more damage. ThereIore, the deIormation histories applied to braces are the
ones biased in terms oI compression, i.e., only the braces with greater compressive deIormation are
analyzed, as they are related to the limiting width-thickness ratios.
Braces with diIIerent width-thickness ratios b/t are analyzed using these Iour 'pass or Iail
type loading protocols. The results are shown in Figures 6.8 to 6.11 Ior Iour loading histories,
respectively. For both west crustal events at distance and west near-Iield events, the critical width-
thickness ratios b/t are 13.2. Braces with width-thickness ratios less than 13.2 will survive the
applied displacement history. For east crustal events at distance and west subduction events, the
critical width-thickness ratios b/t are 16.4 and 15.9, respectively. The critical width-thickness ratios
Ior Iour loading histories are summarized in Table 6.2.
98
Table 6.1 Numerical experiment matrix for brace proportion
b/t KL/r No. oI cycle T/C
f
/2
1 14.2 51 2 1/1 7.5
2 8.1 51 2 1/1 10.5
3 28.2 51 2 1/1 6.0
4 14.2 32 2 1/1 7.5
5 14.2 70 2 1/1 7.5
6 14.2 100 2 1/1 7.5
7 14.2 51 1 1/1 9.0
8 14.2 51 3 1/1 6.0
9 14.2 51 2 3/1 7.5
10 14 51 2 1/3 6.0
Table 6.2 Numerical results of ~pass or fail type loading histories
Event b/t Pass/Fail Peak DeI. Peak IDR
15.3 Pass
East crustal events at distance 16.4 Pass 5D
y
1.6
17.6 Fail
12.5 Pass
West crustal events at distance 13.2 Pass 7D
y
2.2
14.2 Fail
12.5 Pass
West near-Iield events 13.2 Pass 12D
y
3.8
14.2 Fail
14.9 Pass
West subduction events 15.9 Pass 5D
y
1.6
17.0 Fail
Peak Interstory DriIt Ratios are calculated by assuming that the braces are 45-degree
oriented and the nominal yield strain oI braces is
y
= 0.16
99
-15
-10
-5
0
5
10
15
-4.5
-3.0
-1.5
0.0
1.5
3.0
4.5
A
p
p
r
o
x
i
m
a
t
e

i
n
t
e
r
s
t
o
r
y

d
r
i
f
t

(
%
)

N
o
r
m
a
l
i
z
e
d

d
i
s
p
l
a
c
e
m
e
n
t

(

D
/
D
y

)
Time
0.5
0.75
1.0
1.5
2.0
3.0
4.5
6.0
7.5
9.0
10.5
12.0
Figure 6.1 Deformation history for reference brace
5 10 15 20 25 30
0
1
2
3
4
13.1

I
n
t
e
r
s
t
o
r
y

d
r
i
f
t

r
a
t
i
o

a
t

f
r
a
c
t
u
r
e

o
f

b
r
a
c
e
s

(
%
)
width-thickness ratios b/t
2.5
Figure 6.2 Interstory drift ratios to failure versus width-thickness ratios
100
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
Normalized deformation (D/Dy)
(a) b/t = 8.1
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(b) b/t = 14.2
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(c) b/t = 28.2
Figure 6.3 Influence of width-thickness ratios b/t (KL/r = 51)
101
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(a) KL/r = 32
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(b) KL/r = 51
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(c) KL/r = 70
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(d) KL/r = 100
Figure 6.4 Influence of slenderness ratio KL/r (b/t = 14.2)
102
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(a) 1 cycle per deIormation level
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(b) 2 cycles per deIormation level
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(c) 3 cycles per deIormation level
(b/t = 14.2, KL/r = 51)
Figure 6.5 Influence of number of cycles per deformation level
103
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(a) tension:compression1:3
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(b) tension:compression1:1
-15 -10 -5 0 5 10 15
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
(c) tension:compression3:1
(b/t = 14.2, KL/r = 51)
Figure 6.6 Influence of tension/compression asymmetry ratio
104
-6
-4
-2
0
2
4

N
o
r
m
a
l
i
z
e
d

d
e
f
.

(

D
/
D
y

)
Time
(a) east crustal events at distance
-8
-6
-4
-2
0
2
4
6

N
o
r
m
a
l
i
z
e
d

d
e
f
.

(

D
/
D
y

)
Time
(b) west crustal events at distance
-15
-12
-9
-6
-3
0
3
6

N
o
r
m
a
l
i
z
e
d

d
e
f
.

(

D
/
D
y

)
Time
(c) west near-Iield events
-6
-4
-2
0
2
4
6

N
o
r
m
a
l
i
z
e
d

d
e
f
.

(

D
/
D
y

)
Time
(d) west subduction events
Figure 6.7 Deformation histories (Tremblay and Bouatay, 2002)
105
-6 -4 -2 0 2 4 6
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength recovered
(a) b/t = 15.3, pass
-6 -4 -2 0 2 4 6
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength recovered
(b) b/t = 16.4, pass
-6 -4 -2 0 2 4 6
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength lost
(c) b/t = 17.6, Iail
Figure 6.8 Critical width-thickness ratios b/t for east crustal events at distance
106
-8 -6 -4 -2 0 2 4 6 8
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength recovered
(a) b/t = 12.5, pass
-8 -6 -4 -2 0 2 4 6 8
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength recovered
(b) b/t = 13.2, pass
-8 -6 -4 -2 0 2 4 6 8
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength lost
(c) b/t = 14.2, Iail
Figure 6.9 Critical width-thickness ratios b/t for west crustal events at distance
107
-15 -12 -9 -6 -3 0 3 6 9
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength recovered
(a) b/t = 12.5, pass
-15 -12 -9 -6 -3 0 3 6 9
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength recovered
(b) b/t = 13.2, pass
-15 -12 -9 -6 -3 0 3 6 9
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength lost
(c) b/t = 14.2, Iail
Figure 6.10 Critical width-thickness ratios b/t for west near-field events
108
-8 -6 -4 -2 0 2 4 6 8
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength recovered
(a) b/t = 14.9, pass
-8 -6 -4 -2 0 2 4 6 8
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength recovered
(b) b/t = 15.9, pass
-8 -6 -4 -2 0 2 4 6 8
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

N
o
r
m
a
l
i
z
e
d

l
o
a
d

(
P
/
P
y
)
strength lost
(c) b/t = 17.0, Iail
Figure 6.11 Critical width-thickness ratios b/t for west subduction events
109
6.2 CONNECTION DETAILS
The AISC Provisions place a great deal oI emphasis on brace-to-gusset connection details. Al-
though the connection between the brace and the Irame is strengthen by the gusset, the 'rigid zone
created by the beam-column connection and the gusset plate generate more deIormation and dam-
age to a member immediately surrounding the zone: the column and the beam. This may result in
cracks in column Ilanges and webs, as seen in the inverted-V type SCBF test discussed in previous
chapter, or cracks in beam Ilanges and webs, as seen in the BRBF test in Uriz (2005).
Several methods Ior reducing damage accumulation in the column are analyzed in this sec-
tion, including:
1. Reduced depth (Fig. 6.14). Reducing the depth oI column has two eIIects. First, the stiII-
ness and strength oI the column is reduced; and second, Ior a given curvature, the strain at
the outer Iiber will decrease with decreasing section depth. These two eIIects counteract
each other and together may help alleviate damage in the column.
2. Thickened web (Fig. 6.15). Thickening the web increases the stiIIness and strength oI the
column. StiIIness and a stronger column will have less deIormation and damage.
3. Thickened Ilanges (Fig. 6.16). Thickening the Ilanges has a similar eIIect as the thickened
web. Because the Ilange has more outer Iibers, thickening the Ilanges are expected to be
more eIIective.
4. StiIIener (Fig. 6.18). There is a tensile component oI interIace Iorce between the beam and
the column. A stiIIener may distribute the tensile Iorce and reduce damage at the column.
5. ReinIorcing plates at the Ilanges (Fig. 6.19). The reinIorcing plates locally strengthen the
damage zone oI the column.
6. T-shear tab (Fig. 6.20). A T-section shear tab is a variation oI the reinIorcing plate, but
easier to manuIacture and install.
7. Welded Ilanges (Fig. 6.21). Welding the Ilanges oI a beam and a column provides a more
uniIorm interIace Iorce transIer than does a shear tab.
110
8. StiIIener and welded Ilanges (Fig. 6.22). Combine the stiIIener and welded Ilanges and
create a standard moment connection.
The two-story SCBF used in this numerical analysis is identical to the one discussed in the
previous chapter, which was subjected to a displacement history at the top oI the Irame. The critical
damage state oI steel material was 0.5 (typical value, Lemaitre, 1992), beyond which is considered
Iailure. The history oI maximum damage at the column was recorded and compared in Figure 6.12.
ModiIication methods listed in the legend are sorted by damage state, with the largest damage state
history at the top.
The numerical analysis shows that strengthening the column uniIormly (reduced depth,
thickened web, and thickened Ilanges) is, in general, less eIIective than strengthening the column
locally at the damage zone. ReinIorcing plates are the most eIIective method. UnIortunately, this
is the least Ieasible method because it is diIIicult to access the weld near the beam web. The use
oI a stiIIener and a combination oI stiIIener and welded Ilanges are good candidates Ior eIIective
damage reduction methods.
The combination oI stiIIener and welded Ilanges results in more damage than using the
stiIIener by itselI because the connection becomes more rigid. As the displacement history at the top
oI the Irame is prescribed, rigid connections lead to more deIormation at the column. For structures
subjected to earthquake ground motion, more rigid connections lead to higher stiIIness, which
generally results in less overall deIormation. ThereIore the diIIerence between using a stiIIener and
a combination oI a stiIIener and welded Ilanges are expected to be less. Finally, the combination
oI a stiIIener and welded Ilanges is a standard moment connection, which can be easily designed
by structural engineers. Based on these considerations, the combination oI a stiIIener and welded
Ilanges are recommended.
111
0 20 40 60 80
-6
-4
-2
0
2
4
6

R
o
o
f

d
i
s
p
l
a
c
e
m
e
n
t

(
i
n
.
)
Pseudo-time (s)
(a) Displacement history
0 10 20 30 40 50 60 70 80
0.0
0.1
0.2
0.3
0.4
0.5

D
a
m
a
g
e

a
t

c
o
l
u
m
n
Pseudo-time (s)
Reference
Reduced depth (85%)
Welded flange
Thicken web (300%)
Thicken flange (200%)
T shear tab
Stiffener + welded flange
Stiffener
Reinforceing plate
(b) Damage history
Figure 6.12 Damage evolution for different strengthening methods
112
Figure 6.13 Details of column: reference
Figure 6.14 Details of column: reduced depth
113
Figure 6.15 Details of column: thickened web
Figure 6.16 Details of column: thickened flange
114
Figure 6.17 Details of reinforcement: reference
Figure 6.18 Details of reinforcement: stiffener
115
Figure 6.19 Details of reinforcement: reinforcing plates
Figure 6.20 Details of reinforcement: T-shear tab
116
Figure 6.21 Details of reinforcement: welded flanges
Figure 6.22 Details of reinforcement: stiffener and welded flanges
117
6.3 LATERAL BRACING FOR BEAMS
For V-type and inverted V-type braced Irames, beam top and bottom Ilanges must be laterally
braced at the intersection oI the braces and beam. Bracing helps the beams resist rotation about
the longitudinal axis (lateral-torsional buckling), in accordance with the AISC Seismic Provisions.
Although AISC speciIies the required brace strength and stiIIness, the term 'intersection oI braces
and beam is not precisely deIined. In an idealistic model, braces are pin-connected to the beam
at the mid-span point. But in reality, the brace-gusset-beam connection is not a point but a region.
Thus, there are at least two bracing conIigurations: one where the lateral bracing is at the middle
oI the beam, and another where two lateral bracing members are provided, one at each end oI the
gusset plate (Fig. 6.23). Note that the lateral bracing members at the top Ilanges and continuous
bracing by the slab are not shown in the Iigures. Although the Iirst conIiguration is most common in
structural engineering practice, the second one might be more appropriate. For example, consider:
1. For special moment Irames, the Seismic Provisions (Section 9.8) require that lateral bracing
members Ior beams shall be placed at locations where changes in the cross section occur
and where analysis indicates Iormation oI a plastic hinge. For SCBFs, at the edges oI the
gusset plate, the cross section changes because the gusset plates enhance beam bending
stiIIness. Moreover, because oI the vertical unbalanced load, it is possible that a plastic
hinge in the beam will Iorm at the ends oI the gusset plate during inelastic deIormation oI
the braced Irame. To this end, placing lateral bracing members at the gusset plate edges
would be consistent with practices Ior special moment Irames.
2. Gussets are typically detailed in accordance with Seismic Provisions Commentary Section
C13.1. This results in a Iixed-Iixed boundary with a sidesway buckling conIiguration Ior
the gussets, with coeIIicient K = 1.2 (AISC 2006). As the brace-gusset-beam connection
is not a rigid region out-oI-plane, the conIiguration with lateral bracing members at the
edge oI the gusset (Fig. 6.23b) is more likely to satisIy the Iixed-end condition, typically
assumed in design practice.
The analysis results oI braced Irames with the two types oI bracing conIigurations subjected
to earthquake ground motion are compared. Structural parameters oI the braced Irame are described
118
in Section 6.4. The SAC LA32 ground motion that was randomly chosen and is large enough to
buckle the braces, was applied to the structure. The comparison oI Iirst-story hysteretic loops is
shown in Figure 6.24, and a comparison oI behaviors is shown in Figures 6.25 and 6.26. As can
be seen, the gusset with a lateral bracing member at the middle buckled out oI plane, which then
led to lateral torsional buckling oI the beam, as well as a plastic hinge at the beam right next to
the connection (Figs. 6.25b, 6.26b). The conIiguration with two lateral bracing members at the
edges oI the gusset provided suIIicient stiIIness against torsional deIormation oI the beam and the
beam-column connection; thereIore, energy dissipation was localized at the braces, which is ideal
Ior SCBFs.
Abraced Irame with only one lateral bracing member at the middle, but with a thicker gusset
design with a Iixed-pin boundary with a sidesway buckling conIiguration was also analyzed. The
result is similar to the two-lateral-bracing-member conIiguration, with buckling localized in braces.
Alternatively, the in-plane buckling connection might be used to reduce twisting Iorces on
the beam. An example oI this kind oI connection (RutherIord and Chekene, 2008) is shown in
Figure 6.27. In order to eliminate the out-oI-plane buckling mode oI the gusset-brace-gusset com-
ponent, the out-oI-plane bending stiIIness at the ends oI the component has to be increased. Here,
stiIIeners are used to strengthen the gusset and prevent it Irom compression buckling. Thicken-
ing oI the gusset and incorporating a gusset stiIIener eIIectively increase the out-oI-plane radius oI
gyration, thereby increasing its compression buckling strength. Thus, the in-plane bucking connec-
tion is also helpIul when only one lateral bracing member at the middle is Ieasible. The analysis
results oI the same braced Irame but with the in-plane buckling connection included are shown is
Figure 6.28. The energy dissipation is again localized in the braces, and buckling oI the gusset is
prevented.
119
(a) At the middle oI the beam (b) At each end oI the gusset plate
Figure 6.23 Schematics of two lateral bracing configurations
-20 -15 -10 -5 0 5 10
-2000
-1000
0
1000
2000

F
i
r
s
t

s
t
o
r
y

s
h
e
a
r

(
k
i
p
s
)
First story drift (in.)
One bracing member
Two bracing members
Figure 6.24 Comparison of first-story hysteretic loops
120
(a) One lateral bracing member at the middle
(b) Two lateral bracing members at the edge oI gusset plate
Figure 6.25 Damage modes for drift to the right
121
(a) One lateral bracing member at the middle
(b) Two lateral bracing members at the edge oI gusset plate
Figure 6.26 Damage modes for drift to the left
122
Figure 6.27 Detail used to induce in-plane buckling of braces
Figure 6.28 Analysis result of in-plane buckling of braces
123
6.4 ESTIMATE OF INTERSTORY DRIFT DEMANDS
For moment Irames, and tall braced Irames, which have relatively long Iundamental periods Ialling
into the Newmark and Hall displacement-preserved range (Newmark and Hall, 1982), the elastic
spectral analysis may give a reasonably conservative estimate oI average driIt demands Ior inelastic
systems. In other words, Ior short-period braced Irames, the estimation based on elastic spectra is
relatively poor. It is not clear whether equivalent nonlinear modal analysis, namely modal pushover
analysis (MPA), can be used to realistically estimate interstory driIt Ior low-rise braced Irames. In
this section, a three-story SCBF structure is evaluated using nonlinear response history analysis
(RHA) and MPA, and the results compared. A new method Ior pushover analysis using a multiple
point constraint is derived based on the principle oI virtual work and is presented in Appendix A.
The nonlinear static procedure or pushover analysis (FEMA 273, FEMA 356) has been
adopted by structural engineers as a building evaluation tool Ior seismic demand estimations. Based
on structural dynamics theory, while retaining the simplicity oI existing pushover analysis, the
MPA procedure was developed to include modal contributions that may be signiIicant to seismic
demands (Chopra and Goel, 2002). The procedure has been applied to SAC buildings (Goel and
Chopra, 2004), height-wise regular generic Irames (Chintanapakdee and Chopra, 2003) and irreg-
ular generic Irames (Chintanapakdee and Chopra, 2004). The goal oI this section is to bridge the
gap between computationally intensive RHA and eIIiciency-demanding design practice based on
a comparison oI results Ior low-rise braced Irames with short Iundamental periods. The results in
terms oI maximum interstory driIt ratios (IDRs) are compared, and based on these analyses, the
method to estimate interstory driIt demands is suggested.
The model building (DASSE, 2007) developed to investigate the SCBF systems is a rectan-
gular structure with Iour perimeter braced-Irame subassemblies. Braces are arranged as a double-
story X-type. The building was designed in conIormance with the provisions oI 2006 IBC, ASCE
7-05, and AISC341-05. Atypical building Iloor plan and Irame elevation are shown in Figure 6.29.
A Iinite element model was developed Ior the braced-Irame subassembly, using shell elements and
the cyclic damage plasticity model. The Iinite element model is shown in Figure 6.30.
First, the model building was subjected to the sixty ground motions that were assembled Ior
the SAC project (Somerville et al., 1997), representing seismic events ranging Irom the Irequently
124
Figure 6.29 Model building floor plan and braced frame elevations
occurring to those that are very rare, and was evaluated using RHA procedure. In all 60 analyses,
the maximum interstory driIt occurs in the Iirst story, thus, hereinaIter maximum interstory driIt
ratios reIer to those ratios in the Iirst story. In the RHA, a standard Newmark integrator is adopted.
Discussion on the instability oI the Newmark integrator can be Iound in Appendix B.
Figure 6.30 Finite element model and refined regions of the building
125
V
bny
V
bn
1
k
n
u
rny
u
rn
1 k
n
Idealized
Actual
(a) An nth-mode pushover curve and its idealization
V
bny
/M
n
*
F
sn
/L
n
1
(
n
)
2
D
ny
=u
rny
/(
n
rn
)
D
n
1
(
n
)
2
(b) Force-deIormation curve Ior nth-mode SDF system
Figure 6.31 Conversion to nth-mode SDF system
126
Then the model building was evaluated using MPA:
1. The natural Irequencies
i
, and modes
i
, were computed with modal analysis.
2. For the nth mode, the base-shear-rooI displacement, V
bn
-u
rn
, a pushover curve by nonlinear
static analysis oI the building using lateral Iorce distributions s = m
n
was developed.
The structural deIormation history is saved. A new pushover method using constraints
(Appendix A) is adopted in this research.
3. The V
bn
-u
rn
pushover curve was converted to the Iorce-deIormation, F
sn
/L
n
-D
n
, relation Ior
the nth-mode equivalent inelastic single-degree-oI-Ireedom system with unit mass. This
step is illustrated in Figure 6.31, where M
n
is the nth-mode eIIective modal mass.
4. The Iorce-deIormation hysteresis relation Ior the nth-mode single-degree-oI-Ireedom sys-
tem was idealized. In this analysis, a peak-oriented stiIIness deteriorating hysteresis type
was chosen based on a comparison oI the monotonic and cyclic pushover curves Ior Mode
1, shown in Figure 6.32.
5. The peak deIormation D
n
oI the nth-model inelastic single-degree-oI-Ireedom system de-
Iined by the hysteretic Iorce-deIormation relation developed in Step 4 with response history
analysis was computed.
6. The peak rooI displacement u
rn
associated with the nth-mode inelastic system Irom u
rn
=
rn
D
n
was computed.
7. From the structural deIormation history obtained in Step 2, the story displacements vector
at rooI displacement equal to u
rn
was extracted.
8. Steps 3 to 7 were repeated Ior all modes to be included.
9. The total responses by combining the peak modal responses using an appropriate model
combination rule was determined. The SRSS rule was used in this paper. Some modal
properties are listed in Table 6.3.
127
Table 6.3 Elastic modal properties of the three-story SCBF subassembly
n T
n
(s)
n

n
V
bn
/M
n
(in/sec
2
)
1 0.521 [0.32, 0.71, 1.00]
T
1.26 200
2 0.204 [1.00, 0.95, 1.00]
T
0.33 1300
3 0.146 [1.00, 0.80, 0.25]
T
0.26 4000
In Step 5, the behavior oI the second and third modal inelastic single-degree-oI-Ireedom
systems were essentially elastic because oI their relative high yield Iorces compared to the Iirst
mode (Table 6.3).
Maximuminterstory driIts (in Story 1) Iromelastic spectral analysis and RHAare compared
in Figure 6.33. Estimates oI interstory driIt based on elastic spectra substantially diverge Irom
nonlinear RHA results Ior intense ground motions. This is similar to the observations reported by
Chen et al. (2008). Maximum interstory driIts (in Story 1) Irom MPA and RHA are then compared
in Figure 6.34. In contrast to estimates using elastic spectra, good agreement is observed between
predictions using MPA as proposed above (with inelastic spectra) and simulations using RHA,
Ior MPA oI both the Iirst mode and the Iirst three modes. The results oI MPA Ior the Iirst three
modes are slightly better than those Ior the Iirst mode in terms oI the coeIIicient oI determination
R
2
(0.9568 versus 0.9651). The reason why the improvement is small is again because oI the elastic
behavior oI the second and third modal single-degree-oI-Ireedomsystems, as well as smaller modal
contribution Iactors Ior higher modes. Thus, the dominant interstory driIt response is that Ior the
Iirst mode.
As demonstrated above, Ior low-rise braced Irames with short Iundamental periods that do
not Iall into the Newmark and Hall displacement-preserved range, the MPA procedure estimates
interstory driIt with reasonable accuracy.
A series oI analyses are presented that evaluate and reIine several requirements Ior detailing and
analyzing special concentrically braced steel Irame buildings. Numerical simulations demonstrate
that the Iatigue liIe capacity oI braces is heavily dependent on width-thickness ratios and deIor-
128
mation histories. Recommendations are presented Ior developing Iatigue liIe demand or loading
protocols Ior use in numerical and experimental investigations. Improved beam-column connec-
tion details are recommended to reduce the damage accumulation. More appropriate lateral bracing
positions Ior beams in V-type and inverted V-type braced Irames are suggested and evaluated. A
method to estimate interstory driIt demands Ior low-rise braced Irames is suggested based on the
modal pushover analysis procedure.
-10 -8 -6 -4 -2 0 2 4 6 8 10
-1500
-1000
-500
0
500
1000
1500

B
a
s
e

s
h
e
a
r

(
k
i
p
s
)
Monotonic
Cyclic
Figure 6.32 Monotonic and cyclic pushover curves for Mode 1
129
0.1 1 10 100
0.1
1
10
100

I
n
t
e
r
s
t
o
r
y

d
r
i
f
t

f
r
o
m

R
H
A

(
i
n
.
)
Interstory drift from elastic spectrum (1 mode) (in.)
R
2
= 0.4538
(a) Elastic spectrum oI 1 mode
0.1 1 10 100
0.1
1
10
100

I
n
t
e
r
s
t
o
r
y

d
r
i
f
t

f
r
o
m

R
H
A

(
i
n
.
)
Interstory drift from elastic spectra (3 modes) (in.)
R
2
= 0.4642
(b) Elastic spectra oI 3 modes
Figure 6.33 Interstory drifts by elastic spectral and response history analysis
130
0.1 1 10 100
0.1
1
10
100

I
n
t
e
r
s
t
o
r
y

d
r
i
f
t

f
r
o
m

R
H
A

(
i
n
.
)
Interstory drift from MPA (1 mode) (in.)
R
2
= 0.9651
(a) MPA oI 1 mode
0.1 1 10 100
0.1
1
10
100

I
n
t
e
r
s
t
o
r
y

d
r
i
f
t

f
r
o
m

R
H
A

(
i
n
.
)
Interstory drift from MPA (3 modes) (in.)
R
2
= 0.9658
(b) MPA oI 3 modes
Figure 6.34 Interstory drifts by modal pushover and response history analysis
131
7 ConcIusions
This research Iocuses on simulations oI the inelastic seismic behavior oI steel braced Irames includ-
ing the eIIects oI low-cycle Iatigue. Numerical models that assess the initiation and propagation oI
Iailure during cyclic loading need to account Ior multi-axial states oI material nonlinearity, local
and global buckling, and the inability oI the material to deIorm inelastically because oI low-cycle
Iatigue. This study simulates Iailures caused by low-cycle Iatigue using Iinite element models.
Compared to phenomenological models, Iinite element models require more computational eIIort
but incorporate more realistic physical representations oI members and materials, including the ini-
tiation and evolution oI damage through complete Iailure. In addition to examining the inelastic
behavior and Iailure oI traditional steel beam-to-column connections, members and connections
in concentrically braced Irames were investigated to better understand the inIluence oI global and
local buckling on member deterioration and Iailure.
Following a review oI existing material models Ior simulating structural steel deterioration,
a series oI investigations are conducted using Iinite element modeling techniques. Finite element
methods can directly account Ior complex states oI stress and changes in deIormed shape. Material
models are critical Ior constitutive behavior at integration points oI the Iinite element models; how-
ever, available material models tend to emphasize behavior associated with ideal ductile response
or with Iailure occurring under monotonic loading conditions (e.g., during metal-Iorming processes
or vehicle collision). These models are not suitable Ior progressive collapse analysis under cyclic
loading where the consequence oI this adverse behavior on the subsequent response or integrity oI
the structure is oI interest.
ThereIore, a new, numerically eIIicient continuumdamage mechanics material model capa-
ble oI simulating inelastic behavior and deterioration oI mechanical properties because oI low-cycle
Iatigue has been devised and implemented in a Iinite element soItware LS-DYNA (LSTC 2007).
Computational results obtained with this new material model correlate well with test results Ior
several beam-to-column connections, individual braces, and braced-Irame subassemblies. These
applications oI the Iinite element model to realistic cases involving progressive collapse illustrate
the importance oI material deterioration and rupture. UnIortunately, the ability oI the material
model to predict ultimate behavior depends heavily on the material modeling properties speciIied.
Recommendations Ior characterizing material properties Ior these types oI analysis are developed
and presented.
Aseries oI analyses are presented that evaluate and reIine several requirements Ior detailing
and analyzing special concentrically braced steel Irame buildings, demonstrating that the Iatigue
liIe capacity oI braces is heavily dependent on width-thickness ratios and deIormation histories.
Member slenderness ratios are shown to have negligible eIIect on Iatigue liIe capacity. ThereIore,
recommendations are presented Ior developing Iatigue liIe demand or loading protocols Ior use in
numerical and experimental investigations. Next, damage evolution in gusseted beam-to-column
connections is evaluated and compared Ior diIIerent connection details, and improved connection
details are recommended to reduce the damage accumulation. The position oI lateral bracing mem-
bers Ior beams in V-type and inverted V-type braced Irames are also examined. More appropriate
positions and methods to compensate Ior problems detected Ior currently recommended lateral
bracing member positions are suggested and evaluated. Finally, Ior low-rise braced Irames that
respond inelastically during strong earthquake ground shaking, an alternative method to estimate
interstory driIt demands is suggested based on the modal pushover analysis procedure.
Additional research work is needed to develop improved guidelines Ior modeling. These
depend on having high-quality data on the cyclic hysteretic properties oI the materials being an-
alyzed and on their low-cycle Iatigue characteristics. In addition, more experience with the new
continuum damage mechanics model is required, including analysis oI systems with high axial
loads in members, considering a wider range oI behavior. For such elements to become more use-
Iul and reliable, it is essential to obtain additional high-quality data Ior evaluation and calibration.
When such models can be implemented with conIidence, they can be used to improve structural
details, assess the behavior oI structures subjected to unusual loading, and provide a mechanism Ior
calibrating simpler numerical (phenomenological or physical theory) models that can be applied to
large structures where computational costs prohibit the use oI Iinite element models.
134
REFERENCES
American Institute oI Steel Construction, Inc. (AISC) (2005a). Steel Construction Manual, AISC,
Chicago, IL.
American Institute oI Steel Construction, Inc. (AISC) (2005b). Seismic Provisions for Structural Steel
Buildings, ANSI/AISC 341-05, AISC, Chicago, IL.
American Institute oI Steel Construction, Inc. (AISC) (2006). Seismic Design Manual, American
Institute oI Steel Construction, Inc. and Structural Steel Educational Council, Chicago, IL.
Anderson, T. L. (1995). Fracture Mechanics, Fundamentals and Applications, Second Edition, CRC
Press.
Armstrong, P. J., and Frederick, C. O. (1966). 'A Mathematical Representation oI the Multiaxial
Bauschinger EIIect. CEGBReport RD/B/NN/731, Berkeley Laboratories, R&DDepartment, CA.
Besson, J., and Guillemer-Neel, C. (2003). 'An Extension oI the Green and Gurson Models to Kine-
matic Hardening. Mechanics of Materials, 35, 118.
Bridgman, P. W. (1952). Large Plastic Flow and Fracture, McGraw-Hill.
Bruneau, M., Uang, C. M., and Whittaker, A. (1998). Ductile Design of Steel Structures, McGraw
Hill.
Cedergren, J., Melin, S., and Lidstrom, P. (2004). 'Numerical Modelling oI P/M Steel Bars Sub-
jected to Fatigue Loading using an Extended Gurson Model. European Journal of Mechanics
A/Solids, 23, 899908.
Chaboche, J.-L. (1986). 'Time-Independent Constitutive Theories Ior Cyclic Plasticity. International
Journal of Plasticitv, 2(2), 149188.
Chaboche, J.-L. (1989). 'Constitutive Equations Ior Cyclic Plasticity and Cyclic Visco-Plasticity.
International Journal of Plasticitv, 5(3), 247302.
Chambers, J. J., and Ernst, C. J. (2005). 'Brace Frame Gusset Plate Research: Phase I Literature
Review. American Institute oI Steel Construction.
Chen, C.-H., Lai, J.-W., and Mahin, S. A. (2008). 'Numerical Modeling and PerIormance Assess-
ment oI Concentrically Braced Steel Frames. Proceedings of the 2008 Structures Congress, April
2426, 2008, Vancouver, BC, Canada.
Chintanapakdee, C., and Chopra, A. K. (2003). 'Evaluation oI Modal Pushover Analysis using
Generic Frames. Earthquake Engineering and Structural Dvnamics, 32, 417442.
Chintanapakdee, C., and Chopra, A. K. (2004). 'Seismic Response oI Vertically Irregular Frames:
Response History and Modal Pushover Analyses. Journal of Structural Engineering, ASCE,
130(8), 11771185.
Chopra, A. K., and Goel, R. K. (2002). 'AModal Pushover Analysis Procedure Ior Estimating Seismic
Demands Ior Buildings. Earthquake Engineering and Structural Dvnamics, 31, 561582.
Clarke, M. J., and Hancock, G. J. (1990). 'A Study oI Incremental-Iterative Strategies Ior Non-linear
Analyses. International Journal for Numerical Methods in Engineering, 29, 13651391.
CoIIin, L. F. (1954). 'A Study oI the EIIects oI Cyclic Thermal Stresses on a Ductile Metal. Tran-
sations, ASME, 76, 931950.
CrisIield, M. A. (1981). 'A Fast Incremental/Iteration Solution Procedure that Handles Snap-
Through`. Computers and Structures, 13, 5562.
DASSE (2007). 'Cost Advantages oI Buckling Restrained Braced Frame Buildings. DASSE Design
Inc, private communication.
DuIailly, J., and Lemaitre, J. (1995). 'Modeling Very Low Cycle Fatigue. International Journal of
Damage Mechanics, 4, 153170.
Elchalakani, M., Zhao, X.-L., and Grzebieta, R. (2003). 'Tests on Cold-Formed Circular Tubular
Braces under Cyclic Axial Loading. Journal of Structural Engineering, ASCE, 129(4), 507514.
Englemann, B. E., Whirley, R. G., and Goudreau, G. L. (1989). 'ASimple Shell Element Formulation
Ior Large-Scale Elastoplastic Analysis. Analvtical and Computational Models of Shells, ASME,
New York.
Fell, B. V., Kanvinde, A. M., Deierlein, G. G., Myers, A. M., and Fu, X. (2006). 'Buckling and
Iracture oI concentric braces under inelastic cyclic loading. SteelTIPS, Technical InIormation
and Product Service, Structural Steel Educational Council, Moraga, CA.
Fell, B. V. (2008). 'Large-Scale Testing and Simulation oI Earthquake-Induced Ultra Low Cycle
Fatigue in Bracing Members Subjected to Cyclic Inelastic Buckling. Doctoral Dissertation, Uni-
versity oI CaliIornia at Davis.
Foutch, D. A., Goel, S. C., and Roeder, C. W. (1986). 'Seismic Testing oI Full-Scale Steel Building
Part I. Journal of the Structural Division, ASCE, 113, 21112129.
Fujimoto, M., Hashimoto, A., Nakagomi, T., and Yamada, T. (1985). 'Study on Fracture oI Welded
Connections in Steel Structures under Cyclic Loads Based on Nonlinear Fracture Mechanics.
Journal of Structural and Construction Engineering, Transactions oI AIJ, 356, 93102.
Galambos, T. V. (1998). Guide to Stabilitv Design Criteria for Metal Structures, FiIth Edition, John
Wiley and Sons, Inc.
Goggins, J. M., Broderick, B. M., Elghazouli, A. Y., and Lucas, A. S. (2005). 'Experimental Cyclic
Response oI Cold-Formed Hollow Steel Bracing Members. Engineering Structures, 27(7),
977989.
Gurson, A. L. (1977). 'Continuum Theory oI Ductile Rupture by Void Nucleation and Growth: Part
I Yield Criteria and Flow Rules Ior Porous Ductile Media. Journal of Engineering Materials
and Technologv, ASME, 99, 215.
Goel, S. C. (1992). 'Cyclic Post Buckling Behavior oI Steel Bracing Members. Stabilitv and Ductilitv
of Steel Structures under Cvclic Loading, 75104, CRC Press, Boca Raton, FL.
Goel, R. K., and Chopra, A. K. (2004). 'Evaluation oI Modal and FEMA Pushover Analyses: SAC
Buildings. Earthquake Spectra, 20, 225254.
Han, S.-W., Kim, W. T., and Foutch, D. A. (2007). 'Seismic Behavior oI HSS Bracing Members
According to Width-Thickness Ratio under Symmetric Cyclic Loading. Journal of Structural
Engineering, ASCE, 133 (2), 264273.
136
Harris, H. G., and Muskivitch, J. C. (1980). 'Models oI Precast Concrete Large Panel Buildings.
Journal of the Structural Division, ASCE, 106, 545565.
Harris, H. G., and Sabnis, G. M. (1990). Structural Modeling and Experimental Techniques, CRC
Press LLC.
Hassan, O., and Goel, S. C. (1991). 'Seismic Behavior and Design oI Concentrically Braced Steel
Structures. Report UMCE 91-1, University oI Michigan, Department oI Civil and Environmental
Engineering, Ann Arbor, MI.
Hughes, T. J. R. (1987). Finite Element Method, Linear Static and Dvnamic Finite Element Analvsis,
Prentice-Hall, Englewood CliIIs.
Ikeda, K., Mahin, S. A., and Dermitzakis, S. N. (1984). 'Phenomenological Modeling oI Steel Braces
under Cyclic Loading. UCB/EERC-84/09, Earthquake Engineering Research Center, Berkeley,
CA.
Ikeda, K., and Mahin S. A. (1984). 'A ReIined Physical Theory Model Ior Predicting the Seismic
Behavior oI Braced Steel Frames. Report UCB/EERC-84/12, Earthquake Engineering Research
Center, Berkeley, CA.
Iserles, A. (1996). A First Course in the Numerical Analvsis of Differential Equations, Cambridge
University Press, New York.
Jain, A. K., and Goel, S. (1978). 'Hysteresis Models Ior Steel Members Subjected to Cyclic Buckling
or Cyclic End Moments and Buckling Users Guide Ior DRAIN-2D: EL9 AND EL10. Report
UMEE 78-6, University oI Michigan, College oI Engineering, Ann Arbor, MI.
Kachanov, L. M. (1958). 'Time oI Rupture Process under Creep Conditions. Ievestiva Akademii
Navk SSSR Odtelemie Tekhniheskikh Nauk, 8, 2631.
Kanvinde, A., and Deierlein, G. G. (2004). 'Micromechanical Simulation oI Earthquake Induced
Fractures in Steel Structures. Blume Earthquake Engineering Center Report No. 145, StanIord
University, StanIord, CA.
Kishiki, S., Yamada, S., and Wada, A. (2008). 'Experimental Evaluation oI Structural Behavior oI
Gusset Plate Connection in BRBFrame System. Proceedings of 14th World Conference on Earth-
quake Engineering, October 1217, 2008, Beijing, China.
Krawinkler, H., Gupta, A., Medina, R., and Luco, N. (2000a). 'Development oI Loading Histories
Ior Testing oI Steel Beam-to-Column Assemblies. SAC Background Report SAC/BD-00/10.
Krawinkler, H., Parisi, F., Ibarra, L., Ayoub, A., and Medina, R. (2000b). 'Development oI a Testing
Protocol Ior WoodIrame Structures. CUREE Publication No. W-02.
Leblond, J.-B., Perrin, G., and Devaux, J. (1995). 'An Improved Gurson-type Model Ior Hardenable
Ductile Metals. European Journal of Mechanics A/Solids, 14, 499527.
Lee, K., and Bruneau, M. (2005). 'Energy Dissipation oI Compression Members in Concentri-
cally Braced Frames: Review oI Experimental Data. Journal of Structural Engineering, ASCE,
131(4), 552559.
Lee, S., and Goel, S. C. (1987). 'Investigation oI Concrete-Filled Steel Tubular Cyclic Bending Buck-
ling Members. Report UMCE 87-3, University oI Michigan, Department oI Civil and Environ-
mental Engineering, Ann Arbor, MI.
137
Lehman, D. E., Roeder, C. W., Herman, D., Johnson, S., and Kotulka, B. (2008). 'Improved Seismic
PerIormance oI Gusset Plate Connections. Journal of Structural Engineering, ASCE, 134(6),
890901.
Lemaitre, J. (1971). 'Evaluation oI Dissipation and Damage in Metals Submitted to Dynamic Load-
ing. Mechanical Behavior of Materials. Proceedings of 1st International Conference, August
520, 1971, Kyoto, Japan, 540549.
Lemaitre, J. (1992). A Course on Damage Mechanics, Springer-Verlag.
Lemaitre, J., and Chaboche, J.-L. (1990). Mechanics of Solid Materials, Cambridge University Press.
Lignos, D., and Krawinkler, H. (2007). 'Contributions to Collapse Prediction oI Frame Structures.
Kajima-CUREE Joint Research Program, Phase VI: Investigation oI Factors Leading to Progres-
sive Collapse oI Structures, Category II: Analysis oI Structural Component Failure.
Liu, Z., and Goel, S. C. (1987). 'Seismic Behavior oI Hollow and Concrete-Filled Tabular Bracing
Members. Report UMCE 87-11, University oI Michigan, Department oI Civil and Environmental
Engineering, Ann Arbor, MI.
LSTC (2007). LS-DYNA Kevword Users Manual, Livermore SoItware Technology Corporation, CA,
US.
Manson, S. S. (1953). 'Behavior oI Materials Under conditions oI Thermal Stress. Heat Transfer
Svmposium, University oI Michigan Engineering Research Institute, 975.
Manson, S. S., HalIord, G. R., and Hirschberg, M. H. (1971). 'Creep Fatigue Analysis by Strain Range
Partitioning. Svmposium on Design for Elevated Temperature Environment, ASME, NASA,
TMX.
Marini, B., Mudry, F., and Pineau, A. (1985). 'Ductile Rupture oI A508 Steel under Nonradial Load-
ing. Engineering Fracture Mechanics, 22(3), 375386.
Markarian, A., and Mahin, S. (2004). 'Tests oI a Gusset Plate Connection Ior Concentric Braced
Frames. EERI Student Paper.
Matsuishi M, and Endo T. (1968). 'Fatigue oI Metals Subjected to Varying Stresses. Paper Presented
to Japan Society oI Mechanical Engineers.
McClintock, F. A. (1968). 'A Criterion Ior Ductile Fracture by the Growth oI Holes. Journal of
Applied Mechanics, 35, 363371.
Midorikawa, M., Nishiyama, I., and Yamanouchi, H. (1988). 'Static and Dynamic Analysis oI Full-
Scale Concentrically K-Braced Steel Building. Proceedings of 9th World Conference on Earth-
quake Engineering, Tokyo, Japan, 255260.
Mindlin, R. D. (1951). 'InIluence oI Rotary Inertia and Shear on Flexural Motions oI Isotropic Elastic
Plates. Journal of Applied Mechanics, 18, 3138.
Miner, M. A. (1945). 'Cumulative Damage in Fatigue. Journal of Applied Mechanics, ASME, 67,
A159A164.
Morrow, J. D. (1964). 'Cyclic Plastic Strain Energy and Fatigue oI Metals. Svmposium, ASTM,
Chicago.
138
Newmark, N. M. (1959). 'A Method oI Computation Ior Structural Dynamics. Journal of the Engi-
neering Mechanics Division, ASCE, 85, 6794.
Newmark, N. M., and Hall, W. J. (1982). Earthquake Spectra and Design, EERI Monograph Series,
Earthquake Engineering Research Institute, Oakland, CA.
Nishimura, N., Ono, K., Ikeuchi, T., and Shinke, T. (1994). 'Experimental Investigation on Hysteretic
Behavior oI Structural Steels in Plastic Range. Steel Construction Engineering, 1, 173182.
Okamoto, S., Nakata, S., Kitagawa, Y., Yoshimura M., and Kaminosono T. (1982). 'AProgress Report
on the Full-scale Seismic Experiment oI a Seven-story ReinIorced Concrete Building Part oI
the U.S.-Japan Cooperative Program. BRI Research Paper No. 94, Building Research Institute,
Ministry oI Construction, Japan.
Palmgren, A. (1924). 'Die Lebensdauer von Kugellagern. Zeitschrift des Jereines Deutscher Inge-
nieure, 68(14), 339341.
Pian, T. H. H., and Sumihara, K. (1985). 'Rational Approach Ior Assumed Stress Elements. Inter-
national Journal for Numerical Methods in Engineering, 20, 16851695.
Pirondi, A., and Bonora, N. (2003). 'Modeling Ductile Damage under Fully Reversed Cycling. Com-
putational Materials Science, 26, 129141.
Powell, J., Clark, K., Lehman, D. E., and Roeder, C. W. (2008). 'Test oI a Full Scale Concentrically
Braced Frame with Multi-Story X-Bracing. Proceedings of the 2008 Structures Congress, April
2426, 2008, Vancouver, BC, Canada.
Prager, W. (1956). 'A New Method oI Analyzing Stress and Strains in Work-Hardening Plastic
Solids. Journal of Applied Mechanics, 23, 493496.
Puttick, K. E. (1959). 'Ductile Fracture in Metals. Philosophical Maga:ine, 4, 964969.
Ramm, E. (1981). 'Strategies Ior Tracing the Nonlinear Response near Limit Points. Nonlinear Finite
Element Analvsis in Structural Mechanics, Edited by E. Wunderlich, E. Stein, and K. J. Bathe,
Springer-Verlag, Berlin.
Rice, J. R., and Tracey, D. M. (1969). 'On the Ductile Enlargement oI Voids in Triaxial Stress Fields.
Journal of the Mechanics and Phvsics of Solids, 17, 201217.
Roeder, C. W., Lehman, D. E., Johnson, S., Herman D., and Yoo J. H. (2006). 'Seismic PerIormance
oI SCBF Braced Frame Gusset Plate Connections. Proceedings of 4th International Conference
on Earthquake Engineering, October 1213, 2006, Taipei, Taiwan.
Rousselier, G. (1987). 'Ductile Fracture Models and Their Potential in Local Approach oI Fracture.
Nuclear Engineering and Design, 105, 97111.
Somerville, P. G., Smith, N. F., Punyamurthula, S., and Sun, J. (1997). 'Development oI Ground
Motion Time Histories Ior Phase 2 oI the FEMA/SAC Steel Project. SAC Background Report
SAC BD/97-04, SAC Steel Joint Venture, Sacramento, CaliIornia.
Simo, J. C., and Hughes, T. J. R. (1986). 'On the Variational Foundations oI Assumed Strain methods.
Journal of Applied Mechanics, 53, 16851695.
Simo, J. C., and Hughes, T. J. R. (1998). Computational Inelasticitv, Springer-Verlag.
139
Sivaselvan, M., Lavan, O., Dargush, G., Oren, L., and Reinhorn, A. (2007). 'Toward Analysis oI Full
System Failure oI Structures. Kajima-CUREE Joint Research Program, Phase VI: Investigation
oI Factors Leading to Progressive Collapse oI Structures, Category I: Evaluation and Development
oI Analytical Tools.
Steglich, D., Pirondi, A., Bonora, N., and Brocks, W. (2005). 'Micromechanical Modelling oI Cyclic
Plasticity Incorporating Damage. International Journal of Solids and Structures, 42, 337351.
Suita, K., Nakashima M., and Engelhardt M. D. (2000). 'Comparison oI Seismic Capacity between
Post-Northridge and Post-Kobe Beam-to-Column Connections. Behaviour of Steel Structures
in Seismic Areas. Proceedings of 3rd International Conference, August 2124, 2000, Montreal,
Canada, 527534.
Tanaka, N., Sawamoto, Y., and Saeki, T. (2000). 'Elasto-Plastic Behavior oI Beam-Column Connec-
tions. Proceedings of 12th World Conference on Earthquake Engineering, New Zealand.
Tang, X., and Goel, S. C. (1987). 'Seismic Analysis and Design Considerations oI Braced Steel Struc-
tures. Report UMCE 87-4, University oI Michigan, Department oI Civil and Environmental En-
gineering, Ann Arbor, MI.
Tang, X., and Goel, S. C. (1989). 'Brace Fractures and Analysis oI Phase I Structure. Journal of
Structural Engineering, ASCE, 115(8), 19601976.
Thomason, P. F. (1990). Ductile Fracture of Metals, Pergamon Press.
Thornton, W. A. (1991). 'On the Analysis and Design oI Bracing Connections. Proceedings oI Na-
tional Steel Contruction ConIerence, AISC, Chicago, IL, 26.126.33.
Tremblay, R. (2000). 'InIluence oI Brace Slenderness on the Seismic Response oI Concentrically
Braced Steel Frames. Behaviour of Steel Structures in Seismic Areas. Proceedings of 3rd Inter-
national Conference, August 2124, 2000, Montreal, Canada, 527534.
Tremblay, R. (2002). 'Inelastic Seismic Response oI Steel Bracing Members. Journal of Construc-
tional Steel Research, 58, 665701.
Tremblay, R., and Bouatay, N. (2002). 'Loading Protocols Ior the Seismic Testing oI Ductile Bracing
Members in Concentrically Braced Steel Frames. Proceeding of 12th European Conference on
Earthquake Engineering, London, UK, Paper No. 480.
Tremblay, R., Archambault, M. H., and Filiatrault, A. (2003). 'Seismic PerIormance oI Concentrically
Braced Steel Frames made with Rectangular Hollow Bracing Members. Journal of Structural
Engineering, ASCE, 129(12), 16261636.
Tremblay, R., Haddad, M., Martinez, G., Richard, J., and MoIIatt, K. (2008). 'Inelastic Cyclic Testing
oI Large Size Steel Bracing Members. Proceedings of 14th World Conference on Earthquake
Engineering, October 1217, 2008, Beijing, China.
Tsai, K. C., Hsiao, P. C., Chen, C. H., Lai, J. W., Lin, M. L., and Weng, Y. T. (2004). 'Pseudo Dynamic
Tests oI a Full Scale CFT/BRBComposite Frame., Proceedings of the 2004 Structures Congress,
May 2226, 2004, Nashville, TN.
Tsai, K. C., Weng, Y. T., Wang, K. J., Tsai, C. Y., and Lai, J. W. (2006). 'Bi-Directional Sub-Structural
Pseudo-Dynamic Tests oI a Full-Scale 2-Story BRBF, Part I: Seismic Design, Analytical and Ex-
perimental PerIormance Assessments. Proceedings of 8th U.S. National Conference on Earth-
quake Engineering, EERI, April 1822, 2006, San Francisco, CA.
140
Tvergaard, V., and Needleman, A. (1984). 'Analysis oI the Cup-Cone Fracture in a Round Tensile
Bar. Acta Metallurgica, 32(1), 157169.
Tvernelli, J. F., and CoIIin, L. F. (1959). 'A Compilation and Interpretation oI Cyclic Strain Fatigue
Tests on Metals. Transations, ASM, 51, 438453.
Uriz, P. (2005). 'Towards Earthquake Resistant Design oI Concentrically Braced Steel Structures.
Doctoral Dissertation, University oI CaliIornia at Berkeley.
Whitmore, R. E. (1952). 'Experimental Investigation oI Stresses in Gusset Plates. Bulletin No. 16,
Civil Engineering, The University oI Tennessee Engineering Experiment Station, Knoxville, TN.
Whittaker, A. S., Uang, C. M., and Bertero, V. V. (1989). 'Experimental Behavior oI a Dual Steel
System. Journal of the Structural Division, ASCE, 115.
Xia, L., and Shih, C. F. (1995). 'Ductile Crack Growth I. A Numerical Study Using Computa-
tional Cells with Microstructurally-Based Length Scales. Journal of the Mechanics and Phvsics
of Solids, 43(2), 233259.
Yang, F., and Mahin, S. (2005). 'Limiting Net Section Fracture in Slotted Tube Braces. Steel Tips,
Structural Steel Education Council, Orinda, CA.
Ziegler, H. (1959). 'A ModiIication oI Prager`s Hardening rule. Quarterlv of Applied Mechanics,
17, 5565.
141
Appendix A: AppIying ProportionaI Loads
Using Constraints
A.1 INTRODUCTION
Several structural analyses require a Iixed load pattern, with all load magnitudes varying with
a single scalar parameter. This is known as 'proportional loading. One example oI an analysis
that requires proportional loading is a pushover analysis in which loads are applied at some or
all stories. Another example is multiaxial compression or tension oI concrete specimens in
which two or three surIace loads are applied.
One source oI diIIiculty in perIorming such analysis is that part oI structure, or the whole
structure, may undergo soItening because oI mechanical deterioration and/or large deIormations.
For such conditions, the load Iactor must decrease to maintain equilibriumas the structure soItens.
To conduct these types oI analyses, a sophisticated load-adjusting scheme is necessary with an
accompanying solution procedure.
Several such schemes have been proposed, including the well known arc-length method
(CrisIield, 1981; Ramm, 1981). A comprehensive discussion oI load-adjusting schemes can be
Iound in Clarke and Hancock (1990).
Load-adjusting schemes are useIul Ior solving an equilibrium state with global instability,
which maniIests itselI in a global load-displacement response that has a negative tangent stiIIness
and a smooth equilibriumpath in load-displacement space. However, typical global load-adjusting
schemes may not work iI the instability is localized because oI a severe local nonlinearity and the
equilibrium path is no longer smooth. An example oI an analysis that may address the limitation
oI such a load-adjusting scheme would be an analysis oI a reinIorced concrete structure, where
tensile cracks oI the concrete result in randomly distributed local instabilities.
When an external load is applied at only one degree-oI-Ireedom oI a structure, it is con-
venient to prescribe incremental displacement at this degree-oI-Ireedom making the structural
response stable thereby making it easier to evaluate the load-displacement behavior oI the
structure. However, iI loads are to be applied at multiple degrees-oI-Ireedom, prescribing dis-
placements at all these degrees-oI-Ireedom is not applicable Ior proportional loads; a priori dis-
placements needed to maintain the proportional load pattern are unknown Ior a nonlinear structure.
For laboratory experiments, several well-developed methods Ior applying proportional
loads exist. Inspired by these laboratory methods, a new numerical method is developed Ior ap-
plication oI proportional loads. The degrees-oI-Ireedom corresponding to the prescribed propor-
tional loads are called 'target degrees-oI-Ireedom. In the proposed method, an additional control
degree-oI-Ireedom is introduced, and a multiple point constraint (MPC) involving the control
degree-oI-Ireedom and the target degrees-oI-Ireedom is Iormulated. By prescribing the displace-
ment oI the control degree-oI-Ireedom, proportional loads can be applied at target degrees-oI-
Ireedom without applying a load-adjusting scheme.
A.2 EXPERIMENTS WITH PROPORTIONAL LOADING
One method used by laboratories to apply proportional loads is through multiple actuators. Dis-
placement oI one actuator is prescribed and the resistance oI this actuator is measured. The loads
to be applied at other degrees-oI-Ireedom are calculated using predeIined load proportionality,
and then these Iorces are applied by other actuators to maintain the prescribed load distribution.
This method oI laboratory testing is described by Okamoto et al. (1982). However, this method is
not suitable Ior numerical analysis because it needs Ieedback control, which does not Iit into the
standard numerical analysis procedure.
Another method is the WhiIIle-Tree load system (Harris and Muskivitch, 1980; Harris and
Sabnis, 1999). This mechanical systeminvolves an articulated set oI load distribution bars and one
actuator to apply a single concentrated load (Fig. A.1).
Note that the articulated system oI the WhiIIle-Tree is statically determinate. To apply
proportional loading with proportions (p
1
, p
2
, p
3
, , p
n
), where n is the number oI the loads, a
WhiIIle-Tree load system may be designed as Iollows: Iirst, Loads 1 and 2 are combined with a
loading beam at one level below the test structure. At the level below this, another loading beam
A.2
Figure A.1 Whiffle-Tree load system (Harris and Muskivitch, 1980)
carries the previously combined Loads oI 1 and 2 on one end, and Load 3 on the other end with a
resultant oI the sumoI the Iirst three loads, as shown in Figure A.2. This procedure can be repeated
to include all loads.
Given that all loading beams are pin-connected, the proportionality oI length can be ob-
tained Irom moment equilibrium:
l
1
l
2
=
p
2
p
1
,
l
12
l
3
=
p
3
p
1
+ p
2
,
l
123
l
4
=
p
4
p
1
+ p
2
+ p
3
(A.1)
The loading beams can be Ilexible, and as long as the length proportionalities are maintained, the
load distribution will be maintained even iI the loading beams deIorm, translate, or rotate. II the
loading beams are stiII enough as to be considered rigid, the relations between displacements can
be derived.
From Figure A.3 we write
d
12
=
l
2
d
1
+ l
1
d
2
l
1
+ l
2
=
p
1
d
1
+ p
2
d
2
p
1
+ p
2
d
123
=
l
3
d
12
+ l
12
d
3
l
12
+ l
3
=
(p
1
+ p
2
) d
12
+ p
3
d
3
p
1
+ p
2
+ p
3
=
p
1
d
1
+ p
2
d
2
+ p
3
d
3
p
1
+ p
2
+ p
3
(A.2)
Continuing this derivation reveals that an interesting relation oI displacements implied by the
WhiIIle-Tree System:
d
1234
=
p
1
d
1
+ p
2
d
2
+ p
3
d
3
+ p
4
d
4
p
1
+ p
2
+ p
3
+ p
4
(A.3)
or
p
1
d
1
+ p
2
d
2
+ p
3
d
3
+ p
4
d
4
(p
1
+ p
2
+ p
3
+ p
4
) d
1234
= 0 (A.4)
A.3
d
1
p
1
p
2
p
3
p
4
l
1
l
2 d
2
l
12
l
3
d
12
d
3
l
123
l
4
d
123
d
4
d
1234
p
0
Figure A.2 Articulated system of the Whiffle-Tree
A.3 NUMERICAL PROPORTIONAL LOADING
Suppose a linear multiple point constraint is applied to the original structural system where the
proportional loads are applied:
(p
i
d
i
)
_
p
i
_
d
0
= 0 (A.5)
where p
i
is proportion oI load, d
i
is displacement, at the i-th degree-oI-Ireedom, and d
0
is the
displacement at the new degree-oI-Ireedom introduced by the multiple point constraint. Displace-
ment d
0
can be considered as a weighted average displacement:
d
0
=
(p
i
d
i
)
p
i
(A.6)
where the weight Ior d
i
is the corresponding load proportion p
i
. II this kinetic constraint is satisIied,
the application oI the principle-oI-virtual-work to the rigid constraint will lead to
P
0
d
0
+
(P
i
d
i
) = 0 (A.7)
A.4
l
1
l
2
d
1
d
2
d
12
Figure A.3 Interpolation relation for displacements
where P
i
`s are the proportional loads applied on the original structure by the constraint, and P
0
and (P
i
)`s are the loads on the constraints corresponding to d
0
and d
i
`s. Because the constraint
is rigid, the right-hand side oI Equation (A.7) the internal virtual work is zero. Because d
0
and d
i
`s are the corresponding virtual displacements, they are also required to satisIy the kinetics
constraint in Equation (A.5), i.e.,
(p
i
d
i
)
_
p
i
_
d
0
= 0 (A.8)
Combining Equations (A.7) and (A.8) and eliminating d
0
leads to
__
p
i
P
0

_
p
i
_
P
i
_
d
i
_
= 0 (A.9)
The above derivation is true even iI the additional degree-oI-Ireedom d
0
is Iurther constrained
such that d
0
= 0 (e.g., d
0
is under displacement control as a prescribed time history). As long as
all d
i
`s are at free degrees-oI-Ireedom, their variations d
i
`s are arbitrary. Thus their coeIIicients
must vanish:
p
i
P
0

_
p
i
_
P
i
= 0 i (A.10)
or more useIully
(P
1
: P
2
: : P
n
) = (p
1
: p
2
: : p
n
) iI

p
i
0 (A.11)
In other words, the constraint
_
(p
i
d
i
) (
_
p
i
) d
0
= 0 enIorces load proportionality
(P
1
: P
2
: : P
n
) = (p
1
: p
2
: : p
n
). The application oI the virtual work principle provides an
alternative support Ior the relation Iound Irom displacement interpolation in the previous section,
i.e., Equation (A.4) Ior the case oI Iour target degrees-oI-Ireedom. The above derivation proves
A.5
the validity oI the proposed numerical method oI applying proportional loads using multiple point
constraints. Because the prooI relies only on the linearity and rigidity oI the constraint and prin-
ciple oI virtual work but does not depend on the properties oI the original system it can be
applied to all linear and nonlinear structures.
A.4 LINEAR EXAMPLE
Consider a structure consisting oI two independent springs with stiIIness 1 and 2, respectively. A
load pattern oI proportionality 2:1 is applied to the springs. The goal is to evaluate the behavior
oI the structure. For this simple example, the loading can be directly applied because oI linearity,
e.g., 4 and 2 (Ior proportionality 2 : 1). When the displacements oI 4 and 1 are obtained, the
behavior can be evaluated (Fig. A.4):
P
1
= 1d
1
and P
2
= 2d
2
(A.12)
d
1
= 4
d
2
= 1
P
1
= 4
P
2
= 2
k
1
= 1
k
2
= 2
Figure A.4 A linear example
Next, the constraint method developed above is applied. In order to maintain the loading
proportionality oI 2 : 1, or p
1
= 2 and p
2
= 1, the linear constraint would be:
2d
1
+ 1d
2
3d
0
= 0 (A.13)
where again d
0
is displacement at the additional degree-oI-Ireedom introduced by this constraint.
As shown in the previous section, the application oI the principle oI virtual work to the rigid
constraint will lead to
P
0
d
0
P
1
d
1
P
2
d
2
= 0 (A.14)
A.6
Virtual displacements d
0
, d
1
, and d
2
are also required to satisIy the kinetics constraint in Equa-
tion (A.13), i.e.,
2d
1
+ 1d
2
3d
0
= 0 (A.15)
Combining Equations (A.14) and (A.15) and eliminating d
0
leads to
(2P
0
3P
1
) d
1
+ (P
0
3P
2
) d
2
= 0 (A.16)
where d
1
and d
2
are Iree degrees-oI-Ireedom, so their variation d
1
and d
2
are arbitrary. Thus
their coeIIicients must vanish:
2P
0
3P
1
= P
0
3P
2
= 0 or P
1
: P
2
= 2 : 1 (A.17)
The constraint 2d
1
+ 1d
2
3d
0
= 0 is rewritten as
g(d) = Qd

d
g
= 0 (A.18)
where Q = {2 1}, d = {d
1
d
2
}
T
, and

d
g
= 3d
0
. II the Lagrange multiplier method is used, then
the general equation to solve the constrained problem is
_
_
K Q
T
Q 0
_
_
_
_
d
_
=
_
_
P
d
g
_
_
(A.19)
For the above linear example, this equation becomes
_
_
1 0 2
0 2 1
2 1 0
_
_
_
_
d
1
d
2
_
=
_
_
0
0
3d
0
_
_
(A.20)
Solving this system oI equations with weighted average displacement d
0
= 3 |d
0
is introduced in
Equation (A.13)| will result in displacements d
1
= 4, d
2
= 1 and loads P
1
= 4, P
2
= 2. The loads
satisIy proportionality oI 2 : 1.
A.5 NONLINEAR SOFTENING EXAMPLES
In this section the multiple point constraint method is demonstrated Ior two inelastic systems: one
consists oI nonlinear springs and the other is the biaxial tension oI a shell element with nonlinear
material response.
A.7
A.5.1 Nonlinear Springs
Consider a system consisting oI three springs, all with a stiIIness oI 1000 (Fig. A.5). Spring 1 and
Spring 2 are linear. Spring 3 yields at a strength oI 50 and has a post-yield stiIIness oI 200.
P
1
P
2
P
3
d
1
d
2
d
3
Figure A.5 Nonlinear springs
Load proportionality at Springs 1, 2 and 3 are 1 : 2 : 3. ThereIore, the constraint is
d
1
+ 2d
2
+ 3d
3
6d
0
= 0 (A.21)
where d
0
is time-varying weighted average displacement prescribed as
d
0
= t/100 (A.22)
The system is solved by using LS-DYNA (LSTC, 2007), the constraint is applied using keyword
constrainedlinearglobal. The results are shown in Figures A.6 to A.8. Figure A.6 shows that
Spring 3 yields and soItens at about Iour second, but the other two springs unload elastically.
Figure A.7 shows consistent results: d
3
increases and d
1
and d
2
decrease around the same time.
The load proportionality 1 : 2 : 3 is kept constant and shown in Figure A.8.
A.8
0.00 0.05 0.10 0.15 0.20
0
10
20
30
40
50
60

L
o
a
d
Displacement
Spring 1
Spring 2
Spring 3
Figure A.6 Load versus displacement for nonlinear springs
0 2 4 6 8 10
0.00
0.04
0.08
0.12
0.16
0.20

D
i
s
p
l
a
c
e
m
e
n
t
Time (s)
d
d
1
d
2
d
3
Figure A.7 Displacement histories of nonlinear springs
A.9
0 2 4 6 8 10
0
10
20
30
40
50
60

L
o
a
d
Time (s)
P
1
P
2
P
3
Figure A.8 Load histories of nonlinear springs
A.5.2 Biaxial Tension of a Shell Element with Nonlinear Material
A single shell element system is subjected to biaxial tension (Fig. A.9). The size oI the shell
element is 1 1, and the thickness is 1. The Young`s modulus oI the material is 10, 000, and the
yield stress and post-yield modulus in uniaxial tension test is 50 and 2, 000, respectively.
The load proportionality speciIied in the two directions is 2 : 1. When the displacements
are small, the stress proportionality is approximately 2 : 1. The constraints are
2d
1
+ d
2
3d
0
= 0 and d
1
d
1
= d
2
d
2
= 0 (A.23)
or
2d
1
+ 2d
1
+ d
2
+ d
2
6d
0
= 0 (A.24)
where d
0
is the time-varying weighted average displacement prescribed as
d
0
= t/10000 (A.25)
A.10
d
1
P
1
d
2
P
2
d'
1
d'
2
Figure A.9 Biaxial tension of a shell element
The system is again solved using LS-DYNA. The results are shown in Figures A.10 and A.11.
Figure A.10 shows that
11
increases and
22
decreases aIter the material yields and soItens at
around 40 seconds. The approximate stress proportionality 2 : 1 is kept constant, as shown in
Figure A.11.
A.6 PRACTICAL EXAMPLE FOR PUSHOVER ANALYSIS
The proposed method applying proportional loads using multiple point constraints is Iurther il-
lustrated with a pushover analysis oI a three-story steel braced Irame. The prescribed Iirst-mode
load distribution is 0.32 : 0.71 : 1.00 (Table 6.3, Fig. A.12). The story shear-rooI displacement
response Ior each story solved in LS-DYNA is shown in Figure A.13. The story shear proportion-
ality, 2.03 : 1.71 : 1.00, is constant and consistent with the applied load distribution beIore and
aIter the highly nonlinear responses, such as buckling and Iracture oI the brace, and Iracture oI
the column.
A.11
0 20 40 60 80 100
0.000
0.003
0.006
0.009
0.012
0.015

S
t
r
a
i
n
Time (s)
d
0

11

22
Figure A.10 Strain histories of nonlinear material
0 20 40 60 80 100
0
15
30
45
60

S
t
r
e
s
s
Time (s)

11

22
Figure A.11 Stress histories of nonlinear material
A.12
Figure A.12 Pushover analysis model of braced frame
0 10 20 30 40 50 60
0
300
600
900
1200
1500

S
t
o
r
y

s
h
e
a
r
s

(
k
i
p
s
)
Shear in Story 1
Shear in Story 2
Shear in Story 3
Figure A.13 Story shears versus roof displacement of braced frame
A.13
A.7 A SPECIAL CASE
Consider the Iollowing constraint:
(p
i
d
i
) cd
0
= 0 (A.26)
where c is a non-zero constant. II
_
(p
i
d
i
) = 0, speciIically iI
_
p
i
= 0 and the d
i
`s are iden-
tical, then d
0
must be zero. Following the same procedure in Section A.3, it is still possible to
prove that the speciIied load proportions are satisIied. It is valid to apply a load at the degree-
oI-Ireedom corresponding to d
0
and maintain load proportionality. However, it is not possible to
prescribe a time-varying non-zero displacement at this degree-oI-Ireedom, as it will conIlict with
the constraint. Fortunately, the structural system can be extended by adding another 'dummy
degree-oI-Ireedom at which a non-zero stiIIness pivot is provided, e.g., a 'dummy linear spring
connecting this degree-oI-Ireedom and a Iixed degree-oI-Ireedom. The multiple point constraint
is then extended to the additional degree-oI-Ireedom such that
_
(p
i
d
i
) 0.
For example, iI loads are applied to the two identical springs (Fig. A.14) with proportion-
ality 1 : 1, the linear constraint would be
d
1
d
2
cd
0
= 0 (A.27)
To avoid the singular stiIIness when prescribing the displacement at d
0
, a 'dummy spring and
a 'dummy degree-oI-Ireedom are added at d
3
, shown in Figure A.14. For the new system, the
revised constraint is
d
1
d
2
+ d
3
d
0
= 0 (A.28)
Then the constraint conIlict can be solved.
A.14
P
1
P
2
P
3
d
1
d
2
d
3
Dummy spring
Figure A.14 A two-spring system and a ~dummy spring
A.8 CONCLUDING REMARKS
A numerical method Ior applying proportional loading using multiple point constraints is devel-
oped based on the principle oI virtual work. It is applicable Ior an arbitrary number oI loads, in
arbitrary directions, and Ior an arbitrary type oI structure. Some examples demonstrate the appli-
cation and the accuracy oI the multiple point constraint method. This method provides a stable
and eIIicient way to evaluate structures that may undergo soItening.
A.15
Appendix B: Discussion on InstabiIity of
Newmark Integrator
B.1 INTRODUCTION
The Newmark integrator (Newmark, 1959) is one oI the most popular classes oI time integrators.
It has been considered unconditionally stable iI parameters 1/2 and /2 are used (Hughes,
1987); = 1/4 and = 1/2 are the most widely used combination oI parameters. However, there
is one situation that has been generally ignored: iI the mass is zero, and parameters = 1/4 and
= 1/2 are used, the integrator becomes unstable.
Zero modal mass is not rare in structural dynamic analyses. It is common to assign zero
(distributed) mass density to Iinite elements but a lumped mass at the nodes. In this case, there
is no rotary inertia at the nodes, and modal analysis shows a mode with zero mass. II the whole
systemis solved without static condensation, the zero-mass mode is excited when Iorce is applied,
and instability will occur.
In this chapter, the linear stability oI the Newmark integrator Ior zero mass is analyzed,
and the instability is demonstrated with a numerical example.
B.2 LINEAR STABILITY ANALYSIS
Consider a linear single-degree-oI-Ireedom mass-damper-spring system, which can be described
by equation
ma
i+1
+ cv
i+1
+ ku
i+1
= 0 (B.1)
at time t
i+1
, where m is mass; c is the damping coeIIicient; k is stiIIness; and a, v, and u are
acceleration, velocity and displacement, respectively. II a Newmark integrator with parameters
= 1/4 and = 1/2 is applied, then
_
_
u
i+1
= u
i
+ v
i
h +
h
2
4
(a
i
+ a
i+1
)
v
i+1
= v
i
+
h
2
(a
i
+ a
i+1
)
(B.2)
where h is the integration time step. II a
i+1
Irom Equation (B.1) is substituted into (B.2), then
_
_
mu
i+1
= mu
i
+ mv
i
h
h
2
4
(cv
i
+ ku
i
+ cv
i+1
+ ku
i+1
)
mv
i+1
= mv
i

h
2
(cv
i
+ ku
i
+ cv
i+1
+ ku
i+1
)
(B.3)
This system oI equations can be rearranged into the recursion Iorm
y
i+1
= Ay
i
(B.4)
where vector y = [u, v]
T
and A is the ampliIication matrix
A =
1
4m + 2hc + h
2
k
_
_
4m + 2hc h
2
k 4hm
4hk 4m 2hc h
2
k
_
_
(B.5)
The eigenvalues oI the ampliIication matrix are
r =
[4m h
2
k] 2h
c
2
4mk
4m + 2hc + h
2
k
(B.6)
Linear stability oI recursion requires that these eigenvalues lie within the unit circle in the complex
plane, and those on the circle have multiplicities oI 1 (Iserles, 1996). In other words, iI r 1
and those roots with r = 1 are not repeated, then the integrator is stable. II mass is larger than
zero m > 0 and parameters 1/2 and /2 are used, the integrator is unconditionally stable
(Hughes, 1987).
But iI mass is zero (m = 0), the eigenvalues become
r
1,2
=
hk 2c
2c + hk
r
1
= 1 and r
2
= 1
2hk
2c + hk
(B.7)
II c > 0 and h > 0, then r
2
(1, 1), and there is only one eigenvalue r
1
= 1 lying on the
unit circle. Numerical error will propagate Irom one step to the next constantly. II the damping
coeIIicient is also zero, c = 0, then both eigenvalues become 1. They both lie on the unit circle
and the multiplicity is 2. For this case, the Newmark`s method becomes unstable. The numerical
error will grow linearly (instead oI exponentially).
B.2
B.3 DEMONSTRATION
Figure B.1 shows a system with three nodes connected by two truss elements. Node 1 is Iixed.
The mass at the Node 3 is 0.001. The stiIIness oI truss elements is 1. Loads are applied at Node 2
and 3, and P
2
= P
3
= 0.0016t, where t is time. The system stiIIness and mass matrixes are
K =
_
_
1.6 0.8
0.8 0.8
_
_
, M =
_
_
0 0
0 0.001
_
_
(B.8)
k
1
= 0.8 k
2
= 0.8
m
3
= 0.001
P
2
P
3
1 2 3
m
2
= 0
Figure B.1 System for demonstration
For the damped system, stiIIness proportional damped is applied using Rayleigh damping
c = aM + bK, with a = 0 and b = 0.005. This results in a damping ratio
1
= 5 Ior the
Iirst mode (
1
= 20). Modes oI the system are
1
= [0.5, 1]
T
, and
2
= [1, 0]
T
. Mode 1 has
non-zero modal mass (M
1
= 0.001), whereas Mode 2 has zero modal mass (M
2
= 0). Based on
the linear stability analysis in previous section, iI the Newmark integrator is used to solve this
dynamics system, Mode 1 is stable and Mode 2 is unstable. Because Mode 2 contributes only to
the displacement oI Node 2 (
22
= 0), Node 3 will be stable and Node 2 will be unstable.
The analytical solution oI acceleration Ior system m q + c q + kq = t is
q =
_
k
exp
n
t

n
_
1
2
sin
d
t (m 0)
0 (m = 0)
(B.9)
where
n
=
_
k/m, = c/(2m
n
) < 1 and
d
=
n
_
1
2
. By superposition, the accelerations
Ior undamped system in Figure B.1 are
a
2
= 0.06 sin 20t
a
3
= 0.12 sin 20t
(B.10)
B.3
and analytical solutions oI accelerations Ior = 5 under-damped system are
a
2
=
0.06
0.9975
exp(t) sin(20
0.9975t)
a
3
=
0.12
0.9975
exp(t) sin(20
0.9975t)
(B.11)
Figures B.2 and B.3 compare the results Ior the undamped system. Figures B.2a and B.3a
are the analytical solutions; accelerations at Node 2 and 3 are both sinusoidal curve. II zero mass
at Node 2 is speciIied, the error Ior Mode 2 will grow linearly, shown in Figure B.2b Ior a time
step oI 0.005. Figures B.2b and B.3b also show that the instability is at Node 2 only. This is
because Mode 2 has no component at Node 3. II a small mass is assigned at Node 2, then as
shown in Figure B.2c, the instability is diminished. The oscillation at Node 2 is expected because
high-Irequency vibration oI Mode 2 is not damped out. Figures B.4 and B.5 compare the results
Ior the damped system. Figures B.4a and B.5a show the analytical solution oI ever-decreasing
oscillation oI acceleration at Node 2 and 3. II the mass at Node 2 is zero, the absolute error oI Mode
2 will propagate constantly and the numerical solution oscillates about the analytical solution.
The numerical result Ior this case is shown in Figure B.4b. The high-Irequency part (Mode 2) oI
acceleration at Node 2 never decreases, even iI its low-Irequency part (Mode 1) at Node 2 does
because oI damping. II a small mass is assigned at Node 2, then the numerical result will quickly
converge to the analytical one, as observed in Figure B.4c.
B.4 DISCUSSION
B.4.1 Source of Initial Numerical Error
The truncation error is small Ior a second order integrator and it decreases with decreasing step
sizes. The round-oII error is negligible Ior double precision evaluation. ThereIore, they are not
the major source oI the initial numerical error that linearly grows, as observed in Figure B.2b,
or propagates constantly, as observed in Figure B.4b. One possible source oI the major initial
numerical error comes Irom the discrepancy oI zero initial velocity in the numerical solution and
non-zero initial velocity in analytical solution. For a system with non-singular mass, zero initial
velocity is commonly speciIied as the initial condition. This is compatible with the analytical
result. For a mode with zero mass, however, the analytical solution oI velocity may be non-zero. II
B.4
zero initial velocities in numerical solution are still adopted, then there can be an initial numerical
error. The error in velocity is ampliIied when diIIerentiated into acceleration.
B.4.2 Other Values of Parameter
II parameters = 1/2, > /2 are used, the eigenvalues oI the ampliIication matrix are
r =
2m + (2 1)h
2
k
_
h
_
c
2
4mk (4 1)h
2
k
2
_
2m + hc + 2h
2
k
(B.12)
II mass is zero, m = 0, the eigenvalues become
r
1,2
=
(2 1)hk
_
c
2
(4 1)h
2
k
2
2 + 2hk
(B.13)
II c > 0 and h > 0, then the integrator is unconditionally stable, even when m = 0. But iI c = 0,
then
r
1,2
=
(2 1) i
4 1
2
(B.14)
where i is the imaginary unit. It can be calculated that r = 1, thereIore a numerical error will
propagate Irom one step to the next constantly. It is interesting to Iind that iI = 1/2, > 1/4, the
Newmark integrator is stable even iI the mass is zero. The most commonly used parameter set
= 1/2, = 1/4 is, in Iact, the worse case iI mass is zero. Numerical results Ior parameters
= 1/2, = 1/4+0.01 are shown in Figures B.6 and B.7. One can observe Irom the Iigures that
even iI the mass is zero at Node 2, the results are stable provided is strictly larger than /2.
B.5 CONCLUDING REMARKS
II the mass is zero, the second order equation is inIinitely stiII (in Iact, it is a Iirst order equation, but
is solved as a second order one). Previously it was commonly believed that an implicit numerical
integrator was adequate to address this situation and that Newmark integrator is unconditionally
stable iI parameters 1/2 and /2 are used. This chapter analyzed the linear numerical
stability oI Newmark integrator, and demonstrated that even iI Newmark integrator is implicit, it
can be unstable Ior the case oI zero mass. It is better to apply a small amount oI mass and/or a
small amount oI damping to all degrees-oI-Ireedom in a structural system to avoid this instability.
B.5
0.0 0.2 0.4 0.6 0.8 1.0
-0.10
-0.05
0.00
0.05
0.10

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

2
Time
(a) Analytical result
0.0 0.2 0.4 0.6 0.8 1.0
-200
-150
-100
-50
0
50
100
150
200

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

2
Time
(b) Numerical result, zero mass m
2
= 0
0.0 0.2 0.4 0.6 0.8 1.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

2
Time
(c) Numerical result, small mass m
2
= 1 10
4
Figure B.2 Accelerations at Node 2 (undamped, = 1/2, = 1/4)
B.6
0.0 0.2 0.4 0.6 0.8 1.0
-0.2
-0.1
0.0
0.1
0.2

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

3
Time
0.0 0.2 0.4 0.6 0.8 1.0
-0.2
-0.1
0.0
0.1
0.2

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

3
Time
2
= 0
0.0 0.2 0.4 0.6 0.8 1.0
-0.2
-0.1
0.0
0.1
0.2

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

3
Time
2
= 1 10
4
Figure B.3 Accelerations at Node 3 (undamped, = 1/2, = 1/4)
B.7
0.0 0.2 0.4 0.6 0.8 1.0
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

2
Time
0.0 0.2 0.4 0.6 0.8 1.0
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

2
Time
2
= 0
0.0 0.2 0.4 0.6 0.8 1.0
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

2
Time
2
= 1 10
4
Figure B.4 Accelerations at Node 2 (damped, = 1/2, = 1/4)
B.8
0.0 0.2 0.4 0.6 0.8 1.0
-0.2
-0.1
0.0
0.1
0.2

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

3
Time
0.0 0.2 0.4 0.6 0.8 1.0
-0.2
-0.1
0.0
0.1
0.2

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

3
Time
2
= 0
0.0 0.2 0.4 0.6 0.8 1.0
-0.2
-0.1
0.0
0.1
0.2

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

3
Time
2
= 1 10
4
Figure B.5 Accelerations at Node 3 (damped, = 1/2, = 1/4)
B.9
0.0 0.2 0.4 0.6 0.8 1.0
-3
-2
-1
0
1
2
3

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

2
Time
(a) Numerical result, undamped, zero mass m
2
= 0
0.0 0.2 0.4 0.6 0.8 1.0
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

2
Time
(b) Numerical result, damped, zero mass m
2
= 0
Figure B.6 Accelerations at Node 2 ( = 1/2, = 1/4 + 0.01)
B.10
0.0 0.2 0.4 0.6 0.8 1.0
-0.2
-0.1
0.0
0.1
0.2

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

3
Time
(a) Numerical result, undamped, zero mass m
2
= 0
0.0 0.2 0.4 0.6 0.8 1.0
-0.2
-0.1
0.0
0.1
0.2

A
c
c
e
l
e
r
a
t
i
o
n

a
t

N
o
d
e

3
Time
(b) Numerical result, damped, zero mass m
2
= 0
Figure B.7 Accelerations at Node 3 ( = 1/2, = 1/4 + 0.01)
B.11

PEER REPORTS
PEER reports are available individually or by yearly subscription. PEER reports can be ordered at
http://peer.berkeley.edu/publications/peer_reports.html or by contacting the Pacific Earthquake Engineering Research Center, 1301
South 46
th
Street, Richmond, CA 94804-4698. Tel.: (510) 665-3448; Fax: (510) 665-3456; Email: peer_editor@berkeley.edu
PEER 2010/01 Structural Response and Cost Characterization of Bridge Construction Using Seismic Performance Enhancement
Strategies. Ady Aviram, Boidar Stojadinovi!, Gustavo J. Parra-Montesinos, and Kevin R. Mackie. March 2010.
PEER 2009/03 The Integration of Experimental and Simulation Data in the Study of Reinforced Concrete Bridge Systems
Including Soil-Foundation-Structure Interaction. Matthew Dryden and Gregory L. Fenves. November 2009.
PEER 2009/02 Improving Earthquake Mitigation through Innovations and Applications in Seismic Science, Engineering,
Communication, and Response. Proceedings of a U.S.-Iran Seismic Workshop. October 2009.
PEER 2009/01 Evaluation of Ground Motion Selection and Modification Methods: Predicting Median Interstory Drift Response of
Buildings. Curt B. Haselton, Ed. June 2009.
PEER 2008/10 Technical Manual for Strata. Albert R. Kottke and Ellen M. Rathje. February 2009.
PEER 2008/09 NGA Model for Average Horizontal Component of Peak Ground Motion and Response Spectra. Brian S.-J. Chiou
and Robert R. Youngs. November 2008.
PEER 2008/08 Toward Earthquake-Resistant Design of Concentrically Braced Steel Structures. Patxi Uriz and Stephen A.
Mahin. November 2008.
PEER 2008/07 Using OpenSees for Performance-Based Evaluation of Bridges on Liquefiable Soils. Stephen L. Kramer, Pedro
Arduino, and HyungSuk Shin. November 2008.
PEER 2008/06 Shaking Table Tests and Numerical Investigation of Self-Centering Reinforced Concrete Bridge Columns. Hyung
IL Jeong, Junichi Sakai, and Stephen A. Mahin. September 2008.
PEER 2008/05 Performance-Based Earthquake Engineering Design Evaluation Procedure for Bridge Foundations Undergoing
Liquefaction-Induced Lateral Ground Displacement. Christian A. Ledezma and Jonathan D. Bray. August 2008.
PEER 2008/04 Benchmarking of Nonlinear Geotechnical Ground Response Analysis Procedures. Jonathan P. Stewart, Annie
On-Lei Kwok, Yousseff M. A. Hashash, Neven Matasovic, Robert Pyke, Zhiliang Wang, and Zhaohui Yang.
August 2008.
PEER 2008/03 Guidelines for Nonlinear Analysis of Bridge Structures in California. Ady Aviram, Kevin R. Mackie, and Boidar
Stojadinovi!. August 2008.
PEER 2008/02 Treatment of Uncertainties in Seismic-Risk Analysis of Transportation Systems. Evangelos Stergiou and Anne S.
Kiremidjian. July 2008.
PEER 2008/01 Seismic Performance Objectives for Tall Buildings. William T. Holmes, Charles Kircher, William Petak, and Nabih
Youssef. August 2008.
PEER 2007/12 An Assessment to Benchmark the Seismic Performance of a Code-Conforming Reinforced Concrete Moment-
Frame Building. Curt Haselton, Christine A. Goulet, Judith Mitrani-Reiser, James L. Beck, Gregory G. Deierlein,
Keith A. Porter, Jonathan P. Stewart, and Ertugrul Taciroglu. August 2008.
PEER 2007/11 Bar Buckling in Reinforced Concrete Bridge Columns. Wayne A. Brown, Dawn E. Lehman, and John F. Stanton.
February 2008.
PEER 2007/10 Computational Modeling of Progressive Collapse in Reinforced Concrete Frame Structures. Mohamed M. Talaat
and Khalid M. Mosalam. May 2008.
PEER 2007/09 Integrated Probabilistic Performance-Based Evaluation of Benchmark Reinforced Concrete Bridges. Kevin R.
Mackie, John-Michael Wong, and Boidar Stojadinovi!. January 2008.
PEER 2007/08 Assessing Seismic Collapse Safety of Modern Reinforced Concrete Moment-Frame Buildings. Curt B. Haselton
and Gregory G. Deierlein. February 2008.
PEER 2007/07 Performance Modeling Strategies for Modern Reinforced Concrete Bridge Columns. Michael P. Berry and Marc
O. Eberhard. April 2008.
PEER 2007/06 Development of Improved Procedures for Seismic Design of Buried and Partially Buried Structures. Linda Al Atik
and Nicholas Sitar. June 2007.
PEER 2007/05 Uncertainty and Correlation in Seismic Risk Assessment of Transportation Systems. Renee G. Lee and Anne S.
Kiremidjian. July 2007.

PEER 2007/04 Numerical Models for Analysis and Performance-Based Design of Shallow Foundations Subjected to Seismic
Loading. Sivapalan Gajan, Tara C. Hutchinson, Bruce L. Kutter, Prishati Raychowdhury, Jos A. Ugalde, and
Jonathan P. Stewart. May 2008.
PEER 2007/03 Beam-Column Element Model Calibrated for Predicting Flexural Response Leading to Global Collapse of RC
Frame Buildings. Curt B. Haselton, Abbie B. Liel, Sarah Taylor Lange, and Gregory G. Deierlein. May 2008.
PEER 2007/02 Campbell-Bozorgnia NGA Ground Motion Relations for the Geometric Mean Horizontal Component of Peak and
Spectral Ground Motion Parameters. Kenneth W. Campbell and Yousef Bozorgnia. May 2007.
PEER 2007/01 Boore-Atkinson NGA Ground Motion Relations for the Geometric Mean Horizontal Component of Peak and
Spectral Ground Motion Parameters. David M. Boore and Gail M. Atkinson. May. May 2007.
PEER 2006/12 Societal Implications of Performance-Based Earthquake Engineering. Peter J. May. May 2007.
PEER 2006/11 Probabilistic Seismic Demand Analysis Using Advanced Ground Motion Intensity Measures, Attenuation
Relationships, and Near-Fault Effects. Polsak Tothong and C. Allin Cornell. March 2007.
PEER 2006/10 Application of the PEER PBEE Methodology to the I-880 Viaduct. Sashi Kunnath. February 2007.
PEER 2006/09 Quantifying Economic Losses from Travel Forgone Following a Large Metropolitan Earthquake. James Moore,
Sungbin Cho, Yue Yue Fan, and Stuart Werner. November 2006.
PEER 2006/08 Vector-Valued Ground Motion Intensity Measures for Probabilistic Seismic Demand Analysis. Jack W. Baker and
C. Allin Cornell. October 2006.
PEER 2006/07 Analytical Modeling of Reinforced Concrete Walls for Predicting Flexural and CoupledShear-
Flexural Responses. Kutay Orakcal, Leonardo M. Massone, and John W. Wallace. October 2006.
PEER 2006/06 Nonlinear Analysis of a Soil-Drilled Pier System under Static and Dynamic Axial Loading. Gang Wang and
Nicholas Sitar. November 2006.
PEER 2006/05 Advanced Seismic Assessment Guidelines. Paolo Bazzurro, C. Allin Cornell, Charles Menun, Maziar Motahari,
and Nicolas Luco. September 2006.
PEER 2006/04 Probabilistic Seismic Evaluation of Reinforced Concrete Structural Components and Systems. Tae Hyung Lee
and Khalid M. Mosalam. August 2006.
PEER 2006/03 Performance of Lifelines Subjected to Lateral Spreading. Scott A. Ashford and Teerawut Juirnarongrit. July 2006.
PEER 2006/02 Pacific Earthquake Engineering Research Center Highway Demonstration Project. Anne Kiremidjian, James
Moore, Yue Yue Fan, Nesrin Basoz, Ozgur Yazali, and Meredith Williams. April 2006.
PEER 2006/01 Bracing Berkeley. A Guide to Seismic Safety on the UC Berkeley Campus. Mary C. Comerio, Stephen Tobriner,
and Ariane Fehrenkamp. January 2006.
PEER 2005/16 Seismic Response and Reliability of Electrical Substation Equipment and Systems. Junho Song, Armen Der
Kiureghian, and Jerome L. Sackman. April 2006.
PEER 2005/15 CPT-Based Probabilistic Assessment of Seismic Soil Liquefaction Initiation. R. E. S. Moss, R. B. Seed, R. E.
Kayen, J. P. Stewart, and A. Der Kiureghian. April 2006.
PEER 2005/14 Workshop on Modeling of Nonlinear Cyclic Load-Deformation Behavior of Shallow Foundations. Bruce L. Kutter,
Geoffrey Martin, Tara Hutchinson, Chad Harden, Sivapalan Gajan, and Justin Phalen. March 2006.
PEER 2005/13 Stochastic Characterization and Decision Bases under Time-Dependent Aftershock Risk in Performance-Based
Earthquake Engineering. Gee Liek Yeo and C. Allin Cornell. July 2005.
PEER 2005/12 PEER Testbed Study on a Laboratory Building: Exercising Seismic Performance Assessment. Mary C. Comerio,
editor. November 2005.
PEER 2005/11 Van Nuys Hotel Building Testbed Report: Exercising Seismic Performance Assessment. Helmut Krawinkler,
editor. October 2005.
PEER 2005/10 First NEES/E-Defense Workshop on Collapse Simulation of Reinforced Concrete Building Structures. September
2005.
PEER 2005/09 Test Applications of Advanced Seismic Assessment Guidelines. Joe Maffei, Karl Telleen, Danya Mohr, William
Holmes, and Yuki Nakayama. August 2006.
PEER 2005/08 Damage Accumulation in Lightly Confined Reinforced Concrete Bridge Columns. R. Tyler Ranf, Jared M. Nelson,
Zach Price, Marc O. Eberhard, and John F. Stanton. April 2006.
PEER 2005/07 Experimental and Analytical Studies on the Seismic Response of Freestanding and Anchored Laboratory
Equipment. Dimitrios Konstantinidis and Nicos Makris. January 2005.

PEER 2005/06 Global Collapse of Frame Structures under Seismic Excitations. Luis F. Ibarra and Helmut Krawinkler.
September 2005.
PEER 2005//05 Performance Characterization of Bench- and Shelf-Mounted Equipment. Samit Ray Chaudhuri and Tara C.
Hutchinson. May 2006.
PEER 2005/04 Numerical Modeling of the Nonlinear Cyclic Response of Shallow Foundations. Chad Harden, Tara Hutchinson,
Geoffrey R. Martin, and Bruce L. Kutter. August 2005.
PEER 2005/03 A Taxonomy of Building Components for Performance-Based Earthquake Engineering. Keith A. Porter.
September 2005.
PEER 2005/02 Fragility Basis for California Highway Overpass Bridge Seismic Decision Making. Kevin R. Mackie and Boidar
Stojadinovi!. June 2005.
PEER 2005/01 Empirical Characterization of Site Conditions on Strong Ground Motion. Jonathan P. Stewart, Yoojoong Choi,
and Robert W. Graves. June 2005.
PEER 2004/09 Electrical Substation Equipment Interaction: Experimental Rigid Conductor Studies. Christopher Stearns and
Andr Filiatrault. February 2005.
PEER 2004/08 Seismic Qualification and Fragility Testing of Line Break 550-kV Disconnect Switches. Shakhzod M. Takhirov,
Gregory L. Fenves, and Eric Fujisaki. January 2005.
PEER 2004/07 Ground Motions for Earthquake Simulator Qualification of Electrical Substation Equipment. Shakhzod M.
Takhirov, Gregory L. Fenves, Eric Fujisaki, and Don Clyde. January 2005.
PEER 2004/06 Performance-Based Regulation and Regulatory Regimes. Peter J. May and Chris Koski. September 2004.
PEER 2004/05 Performance-Based Seismic Design Concepts and Implementation: Proceedings of an International Workshop.
Peter Fajfar and Helmut Krawinkler, editors. September 2004.
PEER 2004/04 Seismic Performance of an Instrumented Tilt-up Wall Building. James C. Anderson and Vitelmo V. Bertero. July
2004.
PEER 2004/03 Evaluation and Application of Concrete Tilt-up Assessment Methodologies. Timothy Graf and James O. Malley.
October 2004.
PEER 2004/02 Analytical Investigations of New Methods for Reducing Residual Displacements of Reinforced Concrete Bridge
Columns. Junichi Sakai and Stephen A. Mahin. August 2004.
PEER 2004/01 Seismic Performance of Masonry Buildings and Design Implications. Kerri Anne Taeko Tokoro, James C.
Anderson, and Vitelmo V. Bertero. February 2004.
PEER 2003/18 Performance Models for Flexural Damage in Reinforced Concrete Columns. Michael Berry and Marc Eberhard.
August 2003.
PEER 2003/17 Predicting Earthquake Damage in Older Reinforced Concrete Beam-Column Joints. Catherine Pagni and Laura
Lowes. October 2004.
PEER 2003/16 Seismic Demands for Performance-Based Design of Bridges. Kevin Mackie and Boidar Stojadinovi!. August
2003.
PEER 2003/15 Seismic Demands for Nondeteriorating Frame Structures and Their Dependence on Ground Motions. Ricardo
Antonio Medina and Helmut Krawinkler. May 2004.
PEER 2003/14 Finite Element Reliability and Sensitivity Methods for Performance-Based Earthquake Engineering. Terje
Haukaas and Armen Der Kiureghian. April 2004.
PEER 2003/13 Effects of Connection Hysteretic Degradation on the Seismic Behavior of Steel Moment-Resisting Frames. Janise
E. Rodgers and Stephen A. Mahin. March 2004.
PEER 2003/12 Implementation Manual for the Seismic Protection of Laboratory Contents: Format and Case Studies. William T.
Holmes and Mary C. Comerio. October 2003.
PEER 2003/11 Fifth U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced
Concrete Building Structures. February 2004.
PEER 2003/10 A Beam-Column Joint Model for Simulating the Earthquake Response of Reinforced Concrete Frames. Laura N.
Lowes, Nilanjan Mitra, and Arash Altoontash. February 2004.
PEER 2003/09 Sequencing Repairs after an Earthquake: An Economic Approach. Marco Casari and Simon J. Wilkie. April 2004.
PEER 2003/08 A Technical Framework for Probability-Based Demand and Capacity Factor Design (DCFD) Seismic Formats.
Fatemeh Jalayer and C. Allin Cornell. November 2003.

PEER 2003/07 Uncertainty Specification and Propagation for Loss Estimation Using FOSM Methods. Jack W. Baker and C. Allin
Cornell. September 2003.
PEER 2003/06 Performance of Circular Reinforced Concrete Bridge Columns under Bidirectional Earthquake Loading. Mahmoud
M. Hachem, Stephen A. Mahin, and Jack P. Moehle. February 2003.
PEER 2003/05 Response Assessment for Building-Specific Loss Estimation. Eduardo Miranda and Shahram Taghavi.
September 2003.
PEER 2003/04 Experimental Assessment of Columns with Short Lap Splices Subjected to Cyclic Loads. Murat Melek, John W.
Wallace, and Joel Conte. April 2003.
PEER 2003/03 Probabilistic Response Assessment for Building-Specific Loss Estimation. Eduardo Miranda and Hesameddin
Aslani. September 2003.
PEER 2003/02 Software Framework for Collaborative Development of Nonlinear Dynamic Analysis Program. Jun Peng and
Kincho H. Law. September 2003.
PEER 2003/01 Shake Table Tests and Analytical Studies on the Gravity Load Collapse of Reinforced Concrete Frames. Kenneth
John Elwood and Jack P. Moehle. November 2003.
PEER 2002/24 Performance of Beam to Column Bridge Joints Subjected to a Large Velocity Pulse. Natalie Gibson, Andr
Filiatrault, and Scott A. Ashford. April 2002.
PEER 2002/23 Effects of Large Velocity Pulses on Reinforced Concrete Bridge Columns. Greg L. Orozco and Scott A. Ashford.
April 2002.
PEER 2002/22 Characterization of Large Velocity Pulses for Laboratory Testing. Kenneth E. Cox and Scott A. Ashford. April
2002.
PEER 2002/21 Fourth U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced
Concrete Building Structures. December 2002.
PEER 2002/20 Barriers to Adoption and Implementation of PBEE Innovations. Peter J. May. August 2002.
PEER 2002/19 Economic-Engineered Integrated Models for Earthquakes: Socioeconomic Impacts. Peter Gordon, James E.
Moore II, and Harry W. Richardson. July 2002.
PEER 2002/18 Assessment of Reinforced Concrete Building Exterior Joints with Substandard Details. Chris P. Pantelides, Jon
Hansen, Justin Nadauld, and Lawrence D. Reaveley. May 2002.
PEER 2002/17 Structural Characterization and Seismic Response Analysis of a Highway Overcrossing Equipped with
Elastomeric Bearings and Fluid Dampers: A Case Study. Nicos Makris and Jian Zhang. November 2002.
PEER 2002/16 Estimation of Uncertainty in Geotechnical Properties for Performance-Based Earthquake Engineering. Allen L.
Jones, Steven L. Kramer, and Pedro Arduino. December 2002.
PEER 2002/15 Seismic Behavior of Bridge Columns Subjected to Various Loading Patterns. Asadollah Esmaeily-Gh. and Yan
Xiao. December 2002.
PEER 2002/14 Inelastic Seismic Response of Extended Pile Shaft Supported Bridge Structures. T.C. Hutchinson, R.W.
Boulanger, Y.H. Chai, and I.M. Idriss. December 2002.
PEER 2002/13 Probabilistic Models and Fragility Estimates for Bridge Components and Systems. Paolo Gardoni, Armen Der
Kiureghian, and Khalid M. Mosalam. June 2002.
PEER 2002/12 Effects of Fault Dip and Slip Rake on Near-Source Ground Motions: Why Chi-Chi Was a Relatively Mild M7.6
Earthquake. Brad T. Aagaard, John F. Hall, and Thomas H. Heaton. December 2002.
PEER 2002/11 Analytical and Experimental Study of Fiber-Reinforced Strip Isolators. James M. Kelly and Shakhzod M. Takhirov.
September 2002.
PEER 2002/10 Centrifuge Modeling of Settlement and Lateral Spreading with Comparisons to Numerical Analyses. Sivapalan
Gajan and Bruce L. Kutter. January 2003.
PEER 2002/09 Documentation and Analysis of Field Case Histories of Seismic Compression during the 1994 Northridge,
California, Earthquake. Jonathan P. Stewart, Patrick M. Smith, Daniel H. Whang, and Jonathan D. Bray. October
2002.
PEER 2002/08 Component Testing, Stability Analysis and Characterization of Buckling-Restrained Unbonded Braces
TM
. Cameron
Black, Nicos Makris, and Ian Aiken. September 2002.
PEER 2002/07 Seismic Performance of Pile-Wharf Connections. Charles W. Roeder, Robert Graff, Jennifer Soderstrom, and Jun
Han Yoo. December 2001.

PEER 2002/06 The Use of Benefit-Cost Analysis for Evaluation of Performance-Based Earthquake Engineering Decisions.
Richard O. Zerbe and Anthony Falit-Baiamonte. September 2001.
PEER 2002/05 Guidelines, Specifications, and Seismic Performance Characterization of Nonstructural Building Components and
Equipment. Andr Filiatrault, Constantin Christopoulos, and Christopher Stearns. September 2001.
PEER 2002/04 Consortium of Organizations for Strong-Motion Observation Systems and the Pacific Earthquake Engineering
Research Center Lifelines Program: Invited Workshop on Archiving and Web Dissemination of Geotechnical
Data, 45 October 2001. September 2002.
PEER 2002/03 Investigation of Sensitivity of Building Loss Estimates to Major Uncertain Variables for the Van Nuys Testbed.
Keith A. Porter, James L. Beck, and Rustem V. Shaikhutdinov. August 2002.
PEER 2002/02 The Third U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced
Concrete Building Structures. July 2002.
PEER 2002/01 Nonstructural Loss Estimation: The UC Berkeley Case Study. Mary C. Comerio and John C. Stallmeyer.
December 2001.
PEER 2001/16 Statistics of SDF-System Estimate of Roof Displacement for Pushover Analysis of Buildings. Anil K. Chopra,
Rakesh K. Goel, and Chatpan Chintanapakdee. December 2001.
PEER 2001/15 Damage to Bridges during the 2001 Nisqually Earthquake. R. Tyler Ranf, Marc O. Eberhard, and Michael P.
Berry. November 2001.
PEER 2001/14 Rocking Response of Equipment Anchored to a Base Foundation. Nicos Makris and Cameron J. Black.
September 2001.
PEER 2001/13 Modeling Soil Liquefaction Hazards for Performance-Based Earthquake Engineering. Steven L. Kramer and
Ahmed-W. Elgamal. February 2001.
PEER 2001/12 Development of Geotechnical Capabilities in OpenSees. Boris Jeremi . September 2001.
PEER 2001/11 Analytical and Experimental Study of Fiber-Reinforced Elastomeric Isolators. James M. Kelly and Shakhzod M.
Takhirov. September 2001.
PEER 2001/10 Amplification Factors for Spectral Acceleration in Active Regions. Jonathan P. Stewart, Andrew H. Liu, Yoojoong
Choi, and Mehmet B. Baturay. December 2001.
PEER 2001/09 Ground Motion Evaluation Procedures for Performance-Based Design. Jonathan P. Stewart, Shyh-Jeng Chiou,
Jonathan D. Bray, Robert W. Graves, Paul G. Somerville, and Norman A. Abrahamson. September 2001.
PEER 2001/08 Experimental and Computational Evaluation of Reinforced Concrete Bridge Beam-Column Connections for
Seismic Performance. Clay J. Naito, Jack P. Moehle, and Khalid M. Mosalam. November 2001.
PEER 2001/07 The Rocking Spectrum and the Shortcomings of Design Guidelines. Nicos Makris and Dimitrios Konstantinidis.
August 2001.
PEER 2001/06 Development of an Electrical Substation Equipment Performance Database for Evaluation of Equipment
Fragilities. Thalia Agnanos. April 1999.
PEER 2001/05 Stiffness Analysis of Fiber-Reinforced Elastomeric Isolators. Hsiang-Chuan Tsai and James M. Kelly. May 2001.
PEER 2001/04 Organizational and Societal Considerations for Performance-Based Earthquake Engineering. Peter J. May. April
2001.
PEER 2001/03 A Modal Pushover Analysis Procedure to Estimate Seismic Demands for Buildings: Theory and Preliminary
Evaluation. Anil K. Chopra and Rakesh K. Goel. January 2001.
PEER 2001/02 Seismic Response Analysis of Highway Overcrossings Including Soil-Structure Interaction. Jian Zhang and Nicos
Makris. March 2001.
PEER 2001/01 Experimental Study of Large Seismic Steel Beam-to-Column Connections. Egor P. Popov and Shakhzod M.
Takhirov. November 2000.
PEER 2000/10 The Second U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced
Concrete Building Structures. March 2000.
PEER 2000/09 Structural Engineering Reconnaissance of the August 17, 1999 Earthquake: Kocaeli (Izmit), Turkey. Halil Sezen,
Kenneth J. Elwood, Andrew S. Whittaker, Khalid Mosalam, John J. Wallace, and John F. Stanton. December
2000.
PEER 2000/08 Behavior of Reinforced Concrete Bridge Columns Having Varying Aspect Ratios and Varying Lengths of
Confinement. Anthony J. Calderone, Dawn E. Lehman, and Jack P. Moehle. January 2001.

PEER 2000/07 Cover-Plate and Flange-Plate Reinforced Steel Moment-Resisting Connections. Taejin Kim, Andrew S. Whittaker,
Amir S. Gilani, Vitelmo V. Bertero, and Shakhzod M. Takhirov. September 2000.
PEER 2000/06 Seismic Evaluation and Analysis of 230-kV Disconnect Switches. Amir S. J. Gilani, Andrew S. Whittaker, Gregory
L. Fenves, Chun-Hao Chen, Henry Ho, and Eric Fujisaki. July 2000.
PEER 2000/05 Performance-Based Evaluation of Exterior Reinforced Concrete Building Joints for Seismic Excitation. Chandra
Clyde, Chris P. Pantelides, and Lawrence D. Reaveley. July 2000.
PEER 2000/04 An Evaluation of Seismic Energy Demand: An Attenuation Approach. Chung-Che Chou and Chia-Ming Uang. July
1999.
PEER 2000/03 Framing Earthquake Retrofitting Decisions: The Case of Hillside Homes in Los Angeles. Detlof von Winterfeldt,
Nels Roselund, and Alicia Kitsuse. March 2000.
PEER 2000/02 U.S.-Japan Workshop on the Effects of Near-Field Earthquake Shaking. Andrew Whittaker, ed. July 2000.
PEER 2000/01 Further Studies on Seismic Interaction in Interconnected Electrical Substation Equipment. Armen Der Kiureghian,
Kee-Jeung Hong, and Jerome L. Sackman. November 1999.
PEER 1999/14 Seismic Evaluation and Retrofit of 230-kV Porcelain Transformer Bushings. Amir S. Gilani, Andrew S. Whittaker,
Gregory L. Fenves, and Eric Fujisaki. December 1999.
PEER 1999/13 Building Vulnerability Studies: Modeling and Evaluation of Tilt-up and Steel Reinforced Concrete Buildings. John
W. Wallace, Jonathan P. Stewart, and Andrew S. Whittaker, editors. December 1999.
PEER 1999/12 Rehabilitation of Nonductile RC Frame Building Using Encasement Plates and Energy-Dissipating Devices.
Mehrdad Sasani, Vitelmo V. Bertero, James C. Anderson. December 1999.
PEER 1999/11 Performance Evaluation Database for Concrete Bridge Components and Systems under Simulated Seismic
Loads. Yael D. Hose and Frieder Seible. November 1999.
PEER 1999/10 U.S.-Japan Workshop on Performance-Based Earthquake Engineering Methodology for Reinforced Concrete
Building Structures. December 1999.
PEER 1999/09 Performance Improvement of Long Period Building Structures Subjected to Severe Pulse-Type Ground Motions.
James C. Anderson, Vitelmo V. Bertero, and Raul Bertero. October 1999.
PEER 1999/08 Envelopes for Seismic Response Vectors. Charles Menun and Armen Der Kiureghian. July 1999.
PEER 1999/07 Documentation of Strengths and Weaknesses of Current Computer Analysis Methods for Seismic Performance of
Reinforced Concrete Members. William F. Cofer. November 1999.
PEER 1999/06 Rocking Response and Overturning of Anchored Equipment under Seismic Excitations. Nicos Makris and Jian
Zhang. November 1999.
PEER 1999/05 Seismic Evaluation of 550 kV Porcelain Transformer Bushings. Amir S. Gilani, Andrew S. Whittaker, Gregory L.
Fenves, and Eric Fujisaki. October 1999.
PEER 1999/04 Adoption and Enforcement of Earthquake Risk-Reduction Measures. Peter J. May, Raymond J. Burby, T. Jens
Feeley, and Robert Wood.
PEER 1999/03 Task 3 Characterization of Site Response General Site Categories. Adrian Rodriguez-Marek, Jonathan D. Bray,
and Norman Abrahamson. February 1999.
PEER 1999/02 Capacity-Demand-Diagram Methods for Estimating Seismic Deformation of Inelastic Structures: SDF Systems.
Anil K. Chopra and Rakesh Goel. April 1999.
PEER 1999/01 Interaction in Interconnected Electrical Substation Equipment Subjected to Earthquake Ground Motions. Armen
Der Kiureghian, Jerome L. Sackman, and Kee-Jeung Hong. February 1999.
PEER 1998/08 Behavior and Failure Analysis of a Multiple-Frame Highway Bridge in the 1994 Northridge Earthquake. Gregory L.
Fenves and Michael Ellery. December 1998.
PEER 1998/07 Empirical Evaluation of Inertial Soil-Structure Interaction Effects. Jonathan P. Stewart, Raymond B. Seed, and
Gregory L. Fenves. November 1998.
PEER 1998/06 Effect of Damping Mechanisms on the Response of Seismic Isolated Structures. Nicos Makris and Shih-Po
Chang. November 1998.
PEER 1998/05 Rocking Response and Overturning of Equipment under Horizontal Pulse-Type Motions. Nicos Makris and
Yiannis Roussos. October 1998.
PEER 1998/04 Pacific Earthquake Engineering Research Invitational Workshop Proceedings, May 1415, 1998: Defining the
Links between Planning, Policy Analysis, Economics and Earthquake Engineering. Mary Comerio and Peter
Gordon. September 1998.

PEER 1998/03 Repair/Upgrade Procedures for Welded Beam to Column Connections. James C. Anderson and Xiaojing Duan.
May 1998.
PEER 1998/02 Seismic Evaluation of 196 kV Porcelain Transformer Bushings. Amir S. Gilani, Juan W. Chavez, Gregory L.
Fenves, and Andrew S. Whittaker. May 1998.
PEER 1998/01 Seismic Performance of Well-Confined Concrete Bridge Columns. Dawn E. Lehman and Jack P. Moehle.
December 2000.

ONLINE REPORTS
The following PEER reports are available by Internet only at http://peer.berkeley.edu/publications/peer_reports.html
PEER 2010/106 Verification of Probabilistic Seismic Hazard Analysis Computer Programs. Patricia Thomas, Ivan Wong, and
Norman Abrahamson. May 2010.
PEER 2010/105 Structural Engineering Reconnaissance of the April 6, 2009, Abruzzo, Italy, Earthquake, and Lessons Learned. M.
Selim Gnay and Khalid M. Mosalam. April 2010.
PEER 2010/104 Simulating the Inelastic Seismic Behavior of Steel Braced Frames Including Low-Cycle Fatigue. Yuli Huang and
Stephen A. Mahin. April 2010.
PEER 2010/103 Post-Earthquake Traffic Capacity of Modern Bridges in California. Vesna Terzic and Boidar Stojadinovi!. March
2010.
PEER 2010/102 Analysis of Cumulative Absolute Velocity (CAV) and JMA Instrumental Seismic Intensity (I
JMA
) Using the PEER
NGA Strong Motion Database. Kenneth W. Campbell and Yousef Bozorgnia. February 2010.
PEER 2009/109 Simulation and Performance-Based Earthquake Engineering Assessment of Self-Centering Post-Tensioned
Concrete Bridge Systems. Won K. Lee and Sarah L. Billington. December 2009.
PEER 2009/108 PEER Lifelines Geotechnical Virtual Data Center. J. Carl Stepp, Daniel J. Ponti, Loren L. Turner, Jennifer N.
Swift, Sean Devlin, Yang Zhu, Jean Benoit, and John Bobbitt. September 2009.
PEER 2009/107 Experimental and Computational Evaluation of Current and Innovative In-Span Hinge Details in Reinforced
Concrete Box-Girder Bridges: Part 2: Post-Test Analysis and Design Recommendations. Matias A. Hube and
Khalid M. Mosalam. December 2009.
PEER 2009/106 Shear Strength Models of Exterior Beam-Column Joints without Transverse Reinforcement. Sangjoon Park and
Khalid M. Mosalam. November 2009.
PEER 2009/105 Reduced Uncertainty of Ground Motion Prediction Equations through Bayesian Variance Analysis. Robb Eric S.
Moss. November 2009.
PEER 2009/104 Advanced Implementation of Hybrid Simulation. Andreas H. Schellenberg, Stephen A. Mahin, Gregory L. Fenves.
November 2009.
PEER 2009/103 Performance Evaluation of Innovative Steel Braced Frames. T. Y. Yang, Jack P. Moehle, and Boidar
Stojadinovic. August 2009.
PEER 2009/102 Reinvestigation of Liquefaction and Nonliquefaction Case Histories from the 1976 Tangshan Earthquake. Robb
Eric Moss, Robert E. Kayen, Liyuan Tong, Songyu Liu, Guojun Cai, and Jiaer Wu. August 2009.
PEER 2009/101 Report of the First Joint Planning Meeting for the Second Phase of NEES/E-Defense Collaborative Research on
Earthquake Engineering. Stephen A. Mahin et al. July 2009.
PEER 2008/104 Experimental and Analytical Study of the Seismic Performance of Retaining Structures. Linda Al Atik and Nicholas
Sitar. January 2009.
PEER 2008/103 Experimental and Computational Evaluation of Current and Innovative In-Span Hinge Details in Reinforced
Concrete Box-Girder Bridges. Part 1: Experimental Findings and Pre-Test Analysis. Matias A. Hube and Khalid M.
Mosalam. January 2009.
PEER 2008/102 Modeling of Unreinforced Masonry Infill Walls Considering In-Plane and Out-of-Plane Interaction. Stephen
Kadysiewski and Khalid M. Mosalam. January 2009.
PEER 2008/101 Seismic Performance Objectives for Tall Buildings. William T. Holmes, Charles Kircher, William Petak, and Nabih
Youssef. August 2008.
PEER 2007/101 Generalized Hybrid Simulation Framework for Structural Systems Subjected to Seismic Loading. Tarek Elkhoraibi
and Khalid M. Mosalam. July 2007.
PEER 2007/100 Seismic Evaluation of Reinforced Concrete Buildings Including Effects of Masonry Infill Walls. Alidad Hashemi
and Khalid M. Mosalam. July 2007.

Bracing S

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bracing S

Uploaded by

Copyright:

Available Formats

PACIFIC EARTHQUAKE ENGINEERING

EPS critical equivalent-plastic-strain criterion

Figure 5.1 Illustration of kinematic hardening parameters

is the clear distance between webs, and the depth d

is the clear distance between Ilanges.

is the clear distance between

is the clear distance between Ilanges. Based on experimental data, Fell

, m = 0.5, = 0.9, and

You might also like