You are on page 1of 281

Performance Based Evaluation of

Buildings

Nonlinear Pushover and Time History


Analysis
Reference Manual

Trevor E Kelly, S.E.


Holmes Consulting Group Ltd

Revision 5: October, 2001

Holmes Consulting Group Ltd


Level 1
11 Aurora Terrace
P O Box 942
Wellington
New Zealand

Telephone 64 4 471 2292


Facsimile 64 4 471 2336
www.holmesgroup.com

The Holmes Group of Companies


Company
Holmes Culley
Holmes Consulting Group
Holmes Fire & Safety
Optimx
Holmes Composites

Offices In
San Francisco, CA
New Zealand (Auckland, Wellington,
Christchurch, Queenstown)
New Zealand (Auckland, Wellington,
Christchurch)
Australia (Sydney)
New Zealand (Wellington)
San Diego, CA

Services
Structural Engineering
Structural Engineering
Fire Engineering
Safety Engineering
Risk Assessment
Structural Composites

Copyright 2001. This material must not be copied, reproduced or otherwise used without
the express, written permission of Holmes Consulting Group.

2001

DISCLAIMER
The information contained in these Design Guidelines has been prepared by Holmes Consulting
Group Limited (Holmes) as standard Design Guidelines and all due care and attention has been taken
in the preparation of the information therein. The particular requirements of a project may require
amendments or modifications to the Design Guidelines.
Neither Holmes nor any of its agents, employees or directors are responsible in contract or tort or in
any other way for any inaccuracy in, omission from or defect contained in the Design Guidelines and
any person using the Design Guidelines waives any right that may arise now or in the future against
Holmes or any of its agents, employees or directors.

FOREWORD

This reference manual represents the current status of the Holmes Consulting Group procedures for
the implementation of performance based evaluation of existing buildings. The procedures are
continually evolving and so features may have been added and/or changes made which are not
documented in the manual.
Performance based evaluation is a complex procedure which requires a high level of technical skill
and engineering judgment to obtain valid results. The tools described in this Manual are just tools
and do not substitute for skill and judgment.
All projects must be performed in accordance with HCG Quality Assurance requirements. In
particular, the final product must be reviewed by a Senior Engineer who has experience in the
seismic response and performance of buildings.

CONTENTS

1
1.1
1.2

INTRODUCTION

CODE COMPLIANCE ISSUES


OBJECTIVES OF THE HCG PROCEDURE

1
2

PERFORMANCE BASED DESIGN

2.1
NEW ZEALAND PRACTICE
2.2
UNITED STATES PRACTICE
2.3
ANALYSIS TYPES
2.3.1
LINEAR STATIC PROCEDURE
2.3.2
LINEAR DYNAMIC PROCEDURE
2.3.3
NON-LINEAR STATIC PROCEDURE
2.3.4
NON-LINEAR DYNAMIC PROCEDURE
2.4
IMPLICATIONS OF NDP PROCEDURE
2.5
PERFORMANCE OBJECTIVES
2.5.1
SETTING OBJECTIVES
2.5.2
ACHIEVING PERFORMANCE OBJECTIVES

3
4
4
5
5
5
6
8
8
8
11

13

HOLMES CONSULTING GROUP SOFTWARE

3.1
MODELA SPREADSHEET AND PROGRAM
3.2
ANSR-L PROGRAM
3.3
ANSROUT SPREADSHEET AND PROCESSA PROGRAM
3.4
COMPUTER SET-UP
3.5
AUTOCAD PROGRAM
3.6
COMPUTER RESOURCE DEMANDS
3.6.1
MEMORY REQUIREMENTS
3.6.2
CPU TYPE
3.6.3
DISK SPACE
3.6.4
EXECUTION TIMES

13
13
14
14
15
15
15
16
16
16

18

EARTHQUAKE LOADS

4.1
DEFINITION OF LOADS
4.1.1
GROUND MOTION AMPLITUDE
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

18
18

4.1.2
AMPLITUDE SCALING OF EARTHQUAKE RECORDS
4.1.3
FREQUENCY SCALING OF EARTHQUAKE RECORDS
4.2
NEW ZEALAND PROJECTS
4.2.1
NZS4203 REQUIREMENTS FOR NEW BUILDINGS
4.2.2
NZNSEE REQUIREMENTS FOR EXISTING BUILDINGS
4.2.3
SCALING OF EARTHQUAKE RECORDS
4.2.4
EARTHQUAKE RECORD SELECTION
4.3
UNITED STATES PROJECTS
4.3.1
UBC REQUIREMENTS FOR NEW BUILDINGS
4.3.2
FEMA-273 REQUIREMENTS FOR EXISTING BUILDINGS
4.3.3
SCALING OF EARTHQUAKE RECORDS
4.3.4
EARTHQUAKE RECORD SELECTION
4.4
USER-SELECTED RECORDS

18
19
20
20
20
21
21
26
26
26
27
28
29

30

MODELING TECHNIQUES
TECHNIQUES

5.1
FINITE ELEMENT MODELING
5.2
FLEXURAL ELEMENTS
5.3
STRUCTURAL WALL ELEMENTS
5.3.1
FLEXURAL MODELING USING GAPS
5.3.2
FLEXURAL MODELING USING EQUIVALENT SHEAR STRENGTH
5.3.3
FLEXURAL MODELING USING FRAME ELEMENTS
5.3.4
OTHER YIELDING MECHANISMS IN SHEAR WALLS
5.4
COUPLING BEAMS
5.5
WALL BOUNDARY ELEMENTS
5.6
FACE LOADED WALLS
5.7
RESTRAINTS
5.8
FLOORS
5.9
MASS AND WEIGHT IN BUILDING MODELS.
5.9.1
DISTRIBUTION OF WEIGHT
5.9.2
MASS DEGREES OF FREEDOM
5.9.3
RECOMMENDED MASS OPTIONS
5.9.4
ECCENTRICITY OF MASS
5.10
BEAM LOADS
5.11
FOUNDATION UPLIFT
5.12
BASE ISOLATORS AND DAMPERS
5.13
MULTIPLE BUILDINGS
5.14
STRENGTHENING MEASURES
5.14.1
NEW LATERAL LOAD RESISTING ELEMENTS
5.14.2
STRENGTHENING EXISTING ELEMENTS
5.15
GEOMETRIC STIFFNESS (P-) EFFECTS

30
32
32
34
35
35
35
35
36
36
36
37
38
38
39
39
40
41
42
43
43
43
43
44
44

46

MATERIAL
MATERIAL AND ELEMENT MODELS

6.1
DEGRADING STRENGTH & STIFFNESS OF FLEXURAL ELEMENTS
6.1.1
FORMULATION OF DEGRADATION

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

46
46

ii

6.1.2
IMPLEMENTATION OF DEGRADATION COEFFICIENTS
6.2
DEGRADING STRENGTH AND STIFFNESS OF PANEL SECTIONS
6.3
REINFORCED CONCRETE WALL MODELING
6.3.1
SHEAR STRENGTH ENVELOPE
6.3.2
SHEAR DEGRADATION COEFFICIENTS
6.3.3
FLEXURE DOMINATED SHEAR WALLS
6.3.3.1
EFFECT OF GRID REFINEMENT
6.3.3.2
EFFECT OF AXIAL LOAD
6.3.3.3
WALLS WITH BOUNDARY ELEMENTS
6.3.3.4
EFFECTIVE GAP LENGTH
6.3.3.5
RECOMMENDATIONS FOR MODELING FLEXURAL YIELD
6.4
REINFORCED CONCRETE FRAME MODELING
6.4.1
BEAM ELEMENT HYSTERESIS
6.5
STEEL BUILDING MODELING
6.6
UNREINFORCED MASONRY WALL (URM) BUILDINGS
6.6.1
NZNSEE RED BOOK STRENGTHS
6.6.2
FEMA STRENGTHS
6.6.3
TESTING OF MASONRY
6.6.4
CYCLIC LOADING CHARACTERISTICS
6.6.5
ROCKING STRENGTH
6.6.6
URM EVALUATION STRATEGY
6.6.7
URM STRENGTHENING USING FRP
6.7
TIMBER MODELING
6.7.1
TIMBER FLOORS
6.8
FIBER REINFORCED PLASTICS (COMPOSITES)
6.8.1
TEST DATA FOR FRP STRENGTHENING

48
53
55
55
58
62
63
67
68
69
70
71
71
74
75
75
77
79
79
80
81
82
82
82
84
85

89

MODEL PREPARATION : MODELA

7.1
UNITS
7.2
FILES WRITTEN BY MODELA
7.3
ANSR-L RUN FILE IDENTIFICATION
7.4
CONSTANTS WORKSHEET
7.5
INTERACTION SHEET
7.5.1
JOB TITLE
7.5.2
PROBLEM SIZE AND PROGRAM LOCATION
7.5.3
TIME HISTORY ANALYSIS
7.5.4
FORCE PUSHOVER ANALYSIS
7.5.5
DISPLACEMENT PUSHOVER
7.5.6
MASS LOCATIONS FOR TIME HISTORY ANALYSIS
7.5.7
SPECIAL FUNCTIONS
7.5.7.1
CLEAR OLD DATA
7.5.7.2
CONVERT ETABS
7.5.7.3
FIND COMMON SIZE
7.5.7.4
READ COMMON SIZE
7.5.7.5
HIDE / UNHIDE SHEETS
7.6
ANALYSIS DATA
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

89
90
90
91
91
91
92
93
93
93
94
94
94
94
95
95
96
97

iii

7.6.1
NZS 4203 SEISMIC DATA
7.6.2
SITE SPECIFIC SEISMIC DATA
7.6.3
INCREMENTAL EARTHQUAKE
7.6.4
MASSES AND WEIGHT
7.6.4.1
INCLUDE ELEMENT WEIGHT
7.6.4.2
INCLUDE ELEMENT MASS
7.6.4.3
INCLUDE UDL AS MASS
7.6.4.4
DISTRIBUTE DIAPHRAGM MASS AS WEIGHT
7.6.4.5
LUMP ELEMENT MASS TO DIAPHRAGM
7.6.4.6
ASSIGN MISSING MASS HORIZONTALLY
7.6.4.7
ASSIGN MISSING MASS VERTICALLY
7.6.5
TIME STEP
7.6.6
MAXIMUM DISPLACEMENT
7.6.7
DAMPING
7.6.8
COORDINATE SORT ORDER
7.6.9
FORCE TOLERANCE
7.6.10
NODE TOLERANCE
7.6.11
OUT OF PLANE WALLS
7.6.12
TIME HISTORY OUTPUT
7.6.13
SPECIAL FEATURES
7.6.14
NON-STANDARD DAMPING
7.6.15
UBC FACTORS
7.7
ELEVATIONS AND NODES
7.7.1
COLUMN LINES
7.7.2
ELEVATIONS
7.7.3
TIME HISTORY OUTPUT
7.8
MODEL PLOT
7.9
MATERIALS AND SECTIONS
7.9.1
MASONRY
7.9.2
CONCRETE
7.9.3
STEEL
7.9.4
TIMBER
7.9.5
OTHER MATERIALS
7.9.6
PANEL SECTIONS
7.9.7
BEAM/COLUMN/BRACE SECTIONS
7.9.7.1
MODELING CONSIDERATIONS FOR FLEXURAL ELEMENTS
7.9.7.2
FLEXURAL ELEMENT STIFFNESS PROPERTIES
7.9.7.3
FLEXURAL ELEMENT STRENGTH PROPERTIES
7.9.7.4
FLEXURAL ELEMENT DEGRADATION
7.9.8
BASE ISOLATORS
7.9.9
DAMPERS
7.10
PLOTTING ELEMENT HYSTERESIS
7.11
ELEMENT TYPES
7.11.1
ELEMENT OUTPUT CODE
7.11.2
WALLS
7.11.3
COLUMNS
7.11.4
BEAMS

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

98
99
99
100
100
101
101
101
101
102
102
103
103
103
106
106
106
106
107
108
108
109
109
109
110
111
112
114
115
116
117
117
119
120
121
121
122
123
126
127
128
129
130
130
131
132
133

iv

7.11.5
BRACES
7.11.6
ISOLATORS AND DAMPERS
7.11.7
GAPS
7.12
FLOORS
7.12.1
RIGID FLOORS
7.12.2
FLEXIBLE FLOORS
7.13
CONNECTIVITY
7.14
MASSES AND LOADS
7.15
DEGREES OF FREEDOM
7.15.1
CONSTRAINED DEGREES OF FREEDOM
7.15.2
EQUAL DEGREES OF FREEDOM
7.16
CALCULATION OF FLOOR MASS AND WEIGHTS
7.17
AUTOCAD DXF FILES
7.17.1
IMPORTING THE DXF FILES
7.17.2
SETTING VIEWPOINT
7.17.3
DISPLAY, PRINT AND COPY OPTIONS
7.17.4
DISPLAYING SELECTED ELEMENTS
7.17.5
WALL COLUMN ELEMENTS

134
135
135
138
138
139
140
142
143
143
144
145
146
149
150
150
151
151

152

TIME HISTORY ANALYSIS


ANALYSIS

8.1
COORDINATE SYSTEM
8.2
RUNNING ANSR-L
8.3
GETTING MEMORY REQUIREMENTS
8.4
MEASURING EXECUTION TIMES
8.5
MINIMIZING MEMORY REQUIREMENTS AND EXECUTION TIMES
8.5.1
GRID DEFINITION
8.5.2
FLEXIBLE FLOORS
8.5.3
COORDINATE NUMBERING
8.5.4
PLACEMENT OF DIAPHRAGM NODES
8.5.5
DEGREE OF FREEDOM ORDER
8.6
MODES FOR RUNNING ANSR-L
8.6.1
BATCH MODE
8.6.2
INTERACTIVE MODE
8.6.2.1
UNDEFORMED SHAPE
8.6.2.2
ANIMATION MENU SETTINGS
8.6.2.3
ANIMATION DURING ANALYSIS
8.6.3
AUTOMATIC INTERACTIVE MODE
8.6.4
SAVE AND RESTART OF ANALYSES
8.7
SETTING INTERACTIVE SCREENS IN INPUT FILE
8.7.1
OPTIONS FOR ANIMATION
8.7.2
OPTIONS FOR NODAL COMPONENTS
8.7.3
OPTIONS FOR ELEMENT COMPONENTS
8.8
SAVING AS AVI MOVIE
8.9
ANALYSIS DOES NOT START
8.9.1
FILE OPENING ERROR
8.9.2
ERROR DURING READ INPUT PHASE
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

152
153
154
155
155
155
156
156
157
157
157
158
158
159
161
163
164
165
165
167
168
168
169
170
170
170

8.9.3
ALLOWABLE DIMENSIONS EXCEEDED
8.9.4
ERROR DURING EQUATION SOLUTION
8.10
ANALYSIS TERMINATION
8.10.1
DISPLACEMENT EXCEEDED
8.10.2
CONVERGENCE NOT ACHIEVED
8.11
OUTPUT FILES
8.12
NUMERICAL STABILITY OF SOLUTION
8.13
ACCUMULATION OF ENVELOPES IN ANSR-L

171
171
172
173
173
174
175
177

179

NSP (PUSHOVER ANALYSIS)


ANALYSIS)

9.1
NSP OR FORCE PUSHOVER
9.1.1
LOAD VECTORS
9.1.2
OBTAINING PERIODS AND MODE SHAPES
9.1.3
DEFINING FORCE PUSHOVER
9.1.4
WRITE ANSR PUSHOVER MODEL
9.1.5
RESULTS FOR PUSHOVER ANALYSIS
9.1.6
CALCULATING TARGET DISPLACEMENTS
9.1.7
EVALUATING NSP PERFORMANCE
9.2
DISPLACEMENT PUSHOVER
9.2.1
COLUMN LINE
9.2.2
LEVEL NUMBER
9.2.3
NUMBER OF CYCLES
9.2.4
LOAD TYPE
9.2.5
LOAD INCREMENT
9.2.6
LOAD DIRECTION
9.2.7
CYCLE 1 AMPLITUDE
9.2.8
NUMBER OF CYCLES PER AMPLITUDE
9.2.9
PRODUCING DISPLACEMENT PUSHOVER CURVE
9.3
SPECIAL CONSIDERATIONS FOR NSP
9.3.1
EXAMPLE APPLICATION
9.3.2
RECOMMENDATIONS
9.4
DAMPING DECAY

179
180
182
182
183
183
184
188
190
190
191
191
191
191
191
192
192
192
194
194
197
197

10 RESULTS PROCESSING
PROCESSING : ANSROUT

199

10.1
INTERFACE
10.1.1
PROGRAM LOCATION.
10.1.2
SETTING FILE NAMES
10.1.3
CLEAR ALL RESULTS
10.1.4
GET GAP FORCES
10.1.5
GET ENVELOPE TIMES
10.1.6
SELECTING OUTPUT
10.1.6.1 ELEMENT OUTPUT
10.1.6.2 NODAL OUTPUT
10.1.7
SPECIAL FEATURES

199
200
200
201
201
201
201
202
202
202

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

vi

10.1.8
READ RESULTS
10.1.9
IMPORT RESULTS
10.2
COORDINATES
10.3
TIME HISTORIES OF RESULTS
10.3.1
RAW DATA
10.3.1.1 NODAL OUTPUT
10.3.1.2 ELEMENT OUTPUT
10.3.2
DISPLACEMENTS
10.3.3
ACCELERATIONS
10.3.4
DRIFTS
10.3.5
ELEMENTS
10.3.6
FORCE PUSHOVER AND DISPLACEMENT PUSHOVER
10.4
SUMMARY SHEET
10.5
AREAS
10.6
DISPLACEMENTS
10.7
ENVELOPES
10.8
SHEARS
10.9
ELEMENTS
10.9.1
TRUSSES
10.9.2
BEAMS
10.9.3
COLUMNS
10.9.4
WALLS
10.9.5
FLOORS
10.9.6
GAPS
10.10
MODELPLOT
10.11
CONSTANTS SHEET
10.12
AUTOCAD DXF FILES
10.13
CRACK PATTERN FILES
10.14
OUTPUT FILES

203
203
204
204
205
205
205
206
206
206
207
208
208
209
209
210
210
210
211
211
212
213
213
214
215
218
219
220
221

11 ACCEPTANCE CRITERIA
CRITERIA

223

11.1
DRIFTS
11.2
CONCRETE COLUMNS AND BEAMS
11.2.1
FEMA-273 GUIDELINES
11.2.2
NZNSEE GUIDELINES
11.2.2.1 CURVATURE DUCTILITY
11.2.2.2 SHEAR STRENGTH
11.3
CONCRETE BEAM COLUMN JOINTS
11.4
CONCRETE WALLS
11.5
UNREINFORCED MASONRY WALLS
11.5.1
OUT-OF-PLANE WALLS
11.6
REINFORCED MASONRY WALLS
11.7
WALLS STRENGTHENED WITH FRP
11.7.1
ANCHORAGE OF FRP
11.8
STEEL COLUMNS AND BEAMS
11.9
TIMBER FLOORS

223
223
224
225
226
226
227
228
229
230
231
231
232
233
233

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

vii

11.10
11.11

FOUNDATIONS
NON-STRUCTURAL COMPONENTS AND CONTENTS

234
235

12 EVALUATING RESPON
RESPONSE
SE

236

12.1
PERIOD OF RESPONSE
12.2
DISPLACEMENTS AND DRIFTS
12.3
ACCELERATIONS
12.3.1
VALIDATION OF THE MODEL.
12.3.2
LOADS ON COMPONENTS.
12.4
OVERALL BUILDING HYSTERESIS
12.5
COLUMNS
12.5.1
CALCULATING CURVATURE DUCTILITY CAPACITY
12.5.2
YIELD CURVATURE
12.5.3
SPREADSHEET CALCULATIONS
12.6
BEAMS
12.7
BEAM-COLUMN JOINTS
12.8
WALLS
12.9
FOUNDATION S

236
238
238
238
240
240
241
241
243
243
244
245
247
247

13 FUTURE DEVELOPMENTS
DEVELOPMENTS

249

14 PROJECT LIST

250

15 TROUBLESHOOTING

260

15.1
15.2
15.3
15.4
15.5

260
260
261
261
261

MODELA
ANSR-L
PROCESSA
PUSHOVER ANALYSIS
AUTOCAD DRAWINGS

16 BIBLIOGRAPHY

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

262

viii

LIST OF FIGURES

FIGURE 2-1 NONLINEAR STATIC PROCEDURE (NSP)............................................................................... 6


FIGURE 2-2 NONLINEAR DYNAMIC PROCEDURE (NDP) ........................................................................... 7
FIGURE 3-1 REDUCTIONS IN COMPUTER EXECUTION TIME ..................................................................... 17
FIGURE 4-1 EARTHQUAKE RECORDS USED FOR ROCK SITES ................................................................... 23
FIGURE 4-2 EARTHQUAKE RECORDS USED FOR INTERMEDIATE SITES ......................................................... 24
FIGURE 4-3 EARTHQUAKE RECORDS USED FOR SOFT SITES ..................................................................... 25
FIGURE 4-4 UBC DESIGN RESPONSE SPECTRUM.................................................................................. 26
FIGURE 4-5 FEMA-273 GENERAL RESPONSE SPECTRUM ...................................................................... 27
FIGURE 4-6 EXAMPLE SCALED RECORDS ............................................................................................. 29
FIGURE 5-1 PLANE STRESS ELEMENT GRIDS ......................................................................................... 33
FIGURE 5-2 MODELING ACCURACY .................................................................................................. 34
FIGURE 5-3 GEOMETRIC STIFFNESS EFFECTS ...................................................................................... 45
FIGURE 6-1 PARAMETERS FOR SPECIFYING STRENGTH DEGRADATION ......................................................... 47
FIGURE 6-2 DEGRADING STIFFNESS RELATIONSHIP ................................................................................ 47
FIGURE 6-3 NO STIFFNESS DEGRADATION ; NO STRENGTH DEGRADATION ............................................... 50
FIGURE 6-4 MODERATE STIFFNESS DEGRADATION (=0.5) NO STRENGTH DEGRADATION .......................... 50
FIGURE 6-5 EXTREME STIFFNESS DEGRADATION (=1.0) NO STRENGTH DEGRADATION .............................. 51
FIGURE 6-6 MODERATE STIFFNESS DEGRADATION (=0.5) WITH STRENGTH DEGRADATION ........................ 51
FIGURE 6-7 MODERATE STIFFNESS DEGRADATION (=0.5) WITH STRENGTH DEGRADATION ........................ 52
FIGURE 6-8 MODERATE STIFFNESS DEGRADATION (=0.5) WITH RAPID DEGRADATION .............................. 52
FIGURE 6-9 NON-SYMMETRICAL STIFFNESS DEGRADATION AND STRENGTH DEGRADATION............................ 53
FIGURE 6-10 SHEAR STRESS-STRAIN ENVELOPE CURVE........................................................................... 54
FIGURE 6-11 UNLOADING CHARACTERISTICS ...................................................................................... 55
FIGURE 6-12 HYSTERESIS OF SHEAR DOMINATED WALL ....................................................................... 56
FIGURE 6-13 HYSTERESIS OF FLEXURE DOMINATED WALL ...................................................................... 57
FIGURE 6-14 HYSTERESIS OF SQUAT WALL......................................................................................... 57
FIGURE 6-15 ENVELOPE FOR CONCRETE WALLS .................................................................................. 58
FIGURE 6-16 HYSTERESIS : NO DEGRADATION COEFFICIENTS ................................................................ 59
FIGURE 6-17 =0.5 =0.0 XKK = 2.67 DKODE=0.4.......................................................................... 60
FIGURE 6-18 =0.5 =0.0 XKK = 2.67 DKODE=0.75........................................................................ 60
FIGURE 6-19 =0.5 =0.0 XKK = 2.67 DKODE=0.10........................................................................ 61
FIGURE 6-20 =0.5 =0.0 XKK = 1.00 DKODE=0.10........................................................................ 61
FIGURE 6-21 REINFORCING BAR UNDER INCREASING CYCLIC DISPLACEMENT............................................. 63
FIGURE 6-22 12 HORIZONTAL X 19 VERTICAL ................................................................................... 65
FIGURE 6-23 12 HORIZONTAL X 6 VERTICAL ..................................................................................... 65
FIGURE 6-24 6 HORIZONTAL X 6 VERTICAL ..................................................................................... 66
FIGURE 6-25 1 HORIZONTAL X 6 VERTICAL ...................................................................................... 66
FIGURE 6-26 1500 KN AXIAL LOAD (1.9 MPA) ................................................................................ 67

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

ix

FIGURE 6-27 ZERO AXIAL LOAD ..................................................................................................... 67


FIGURE 6-28 SHEAR PANELS IN WEB ELASTIC; GAP EFFECTIVE LENGTH 50 MM ......................................... 68
FIGURE 6-29 SHEAR PANELS IN WEB CRACKED; GAP EFFECTIVE LENGTH 50 MM ...................................... 68
FIGURE 6-30 SHEAR PANELS IN WEB ELASTIC; GAP EFFECTIVE LENGTH 100 MM ........................................ 69
FIGURE 6-31 SHEAR PANELS IN WEB ELASTIC; GAP EFFECTIVE LENGTH 150 MM ........................................ 69
FIGURE 6-32 EFFECTIVE GAP LENGTH .............................................................................................. 70
FIGURE 6-33 EXAMPLE STRUCTURE................................................................................................... 72
FIGURE 6-34 PLASTIC UNLOADING ................................................................................................. 73
FIGURE 6-35 ELASTIC UNLOADING.................................................................................................. 73
FIGURE 6-36 COMBINED ELEMENTS IN PARALLEL .................................................................................. 73
FIGURE 6-37 TOP DISPLACEMENT FOR EARTHQUAKE LOADS .................................................................. 74
FIGURE 6-38 FREE VIBRATION DAMPING ........................................................................................... 74
FIGURE 6-39 IMPLEMENTATION OF NZNSEE STRENGTH (ZERO AXIAL STRESS)............................................ 76
FIGURE 6-40 EFFECT OF AXIAL STRESS .............................................................................................. 77
FIGURE 6-41 COMPARISON NZNSEE AND FEMA : NO SURCHARGE ..................................................... 78
FIGURE 6-42 COMPARISON NZNSEE AND FEMA : 25 M SURCHARGE................................................... 78
FIGURE 6-43 AUCKLAND CUSTOMHOUSE IN-PLACE SHEAR TESTS ........................................................... 79
FIGURE 6-44 MASONRY CYCLIC BEHAVIOUR ..................................................................................... 80
FIGURE 6-45 URM DEGRADING HYSTERESIS...................................................................................... 81
FIGURE 6-46 TIMBER FLOOR ELEMENT PROPERTIES .............................................................................. 83
FIGURE 6-47 TIMBER FLOOR HYSTERESIS ............................................................................................ 83
FIGURE 6-48 EFFECT OF FRP STRENGTHENING ON URM WALL.............................................................. 85
FIGURE 6-49 EFFECTIVE SHEAR STRESS.............................................................................................. 86
FIGURE 6-50 EXTRACTED FRP STRENGTH .......................................................................................... 87
FIGURE 7-1 INTERACTION SHEET ..................................................................................................... 92
FIGURE 7-2 ANSR MEMORY REQUIREMENTS ..................................................................................... 96
FIGURE 7-3 DIALOG FOR SETTING UP SHEETS .................................................................................... 97
FIGURE 7-4 ANALYSIS DATA SHEET ................................................................................................... 98
FIGURE 7-5 RETURN PERIOD VERSUS AMPLITUDE................................................................................ 100
FIGURE 7-6 RAYLEIGH DAMPING ................................................................................................... 105
FIGURE 7-7 COLUMN LINE INPUT .................................................................................................. 110
FIGURE 7-8 ELEVATION INPUT....................................................................................................... 111
FIGURE 7-9 TIME HISTORY OUTPUT SPECIFICATION ........................................................................... 112
FIGURE 7-10 DIALOG BOX FOR MODEL PLOT ................................................................................... 113
FIGURE 7-11 EXAMPLE OF MODEL PLOT .......................................................................................... 114
FIGURE 7-12 MASONRY INPUT ..................................................................................................... 115
FIGURE 7-13 CONCRETE INPUT ..................................................................................................... 116
FIGURE 7-14 STEEL INPUT ............................................................................................................. 117
FIGURE 7-15 TIMBER INPUT........................................................................................................... 118
FIGURE 7-16 TIMBER DIAPHRAGMS ................................................................................................. 118
FIGURE 7-17 OTHER MATERIALS INPUT ........................................................................................... 119
FIGURE 7-18 PLATE ELEMENT INPUT ................................................................................................ 121
FIGURE 7-19 (A) : FLEXURAL ELEMENT SECTION STIFFNESS .................................................................. 124
FIGURE 7-20 FLEXURAL ELEMENT DEGRADATION ................................................................................ 126
FIGURE 7-21 BASE ISOLATOR ELEMENTS ......................................................................................... 128
FIGURE 7-22 DAMPER ELEMENTS .................................................................................................... 129
FIGURE 7-23 WALL ELEMENT INPUT ................................................................................................ 132
FIGURE 7-24 COLUMN ELEMENT INPUT ........................................................................................... 133
FIGURE 7-25 BEAM ELEMENT INPUT ................................................................................................ 134

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

FIGURE 7-26 BRACE ELEMENT INPUT ............................................................................................... 135


FIGURE 7-27 GAP ELEMENT INPUT ................................................................................................ 137
FIGURE 7-28 RIGID FLOOR INPUT ................................................................................................. 139
FIGURE 7-29 FLEXIBLE FLOOR ELEMENT INPUT .................................................................................. 139
FIGURE 7-30 FLEXIBLE FLOOR ELEMENT MESH ................................................................................... 140
FIGURE 7-31 FLOOR DIAPHRAGM CONNECTIVITY INPUT ..................................................................... 141
FIGURE 7-32 ADDED MASS AND LOAD INPUT.................................................................................... 143
FIGURE 7-33 USER SPECIFIED RESTRAINED DEGREE OF FREEDOM INPUT ................................................. 144
FIGURE 7-34 USER SPECIFIED EQUAL DEGREE OF FREEDOM INPUT ........................................................ 145
FIGURE 7-35 FLOOR WEIGHTS ...................................................................................................... 145
FIGURE 7-36 DISPLAY OF FLOOR WEIGHTS ...................................................................................... 146
FIGURE 7-37 AUTOCAD LINE DRAWING .......................................................................................... 147
FIGURE 7-38 AUTOCAD THREE DIMENSIONAL DRAWING ................................................................... 148
FIGURE 7-39 AUTOCAD DRAWINGS USING NON-RECTANGULAR SHAPES ................................................ 149
FIGURE 8-1 GLOBAL AND LOCAL COORDINATE SYSTEMS ..................................................................... 153
FIGURE 8-2 ANSR-L OPENING SCREEN .......................................................................................... 159
FIGURE 8-3 UNDEFORMED SHAPE .................................................................................................. 161
FIGURE 8-4 ANIMATION SELECTION ................................................................................................ 162
FIGURE 8-5 SCREEN DISPLAY DURING ANALYSIS ................................................................................. 164
FIGURE 8-6 EFFECT OF SOLUTION PARAMETERS ................................................................................. 177
FIGURE 9-1 FORCE PUSHOVER PROCEDURE ...................................................................................... 182
FIGURE 9-2 DEFINITION OF FORCE PUSHOVER .................................................................................. 182
FIGURE 9-3 NSP WORKBOOK CONTROL SHEET............................................................................... 186
FIGURE 9-4 FEMA-273 NSP METHOD 1........................................................................................ 187
FIGURE 9-5 FEMA-273 NSP METHOD 2........................................................................................ 187
FIGURE 9-6 EXAMPLE NSP METHOD 1 ............................................................................................ 188
FIGURE 9-7 NSP LOG OF RESULTS ................................................................................................ 189
FIGURE 9-8 NDP SETS................................................................................................................. 190
FIGURE 9-9 PUSHOVER DIALOG BOX ............................................................................................. 190
FIGURE 9-10 DISPLACEMENT PUSHOVER CURVE FOR SHEAR WALL BUILDING ............................................ 192
FIGURE 9-11 DISPLACEMENT PUSHOVER CURVE FOR DEGRADING BEAM FRAME ........................................ 193
FIGURE 9-12 DISPLACEMENT PUSHOVER CURVE FOR DEGRADING BEAM AND COLUMN FRAME .................... 193
FIGURE 9-13 DISPLACEMENT PUSHOVER FORCE ................................................................................ 194
FIGURE 9-14 FORCE PUSHOVER CURVES ......................................................................................... 195
FIGURE 9-15 FORCE IMBALANCE FOR TOLERANCE 1% W .................................................................... 196
FIGURE 9-16 FORCE IMBALANCE FOR TOLERANCE 0.1% W ................................................................. 196
FIGURE 9-17 PUSHOVER DISPLACEMENT WITH DECAY ........................................................................ 198
FIGURE 10-1 INTERFACE SHEET ...................................................................................................... 200
FIGURE 10-2 MODEL PLOT DIALOG ............................................................................................... 216
FIGURE 10-3 MODEL PLOT FOR FRAME .......................................................................................... 217
FIGURE 10-4 MODEL PLOT FOR WALL ............................................................................................. 217
FIGURE 10-5 DATA ENTERED ON CONSTANTS SHEET .......................................................................... 218
FIGURE 10-6 (A) AUTOCADLT RELEASE 2.0 ..................................................................................... 219
FIGURE 10-7 STRAIN / PLASTIC ROTATION CODED DXF FILE ................................................................ 220
FIGURE 10-8 CRACK PATTERN DXF FILE ........................................................................................... 221
FIGURE 11-1 FRP EFFECT ON URM WALL SPECIMEN .......................................................................... 232

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

xi

FIGURE 11-2 ANCHORAGE OF FRP ................................................................................................ 233


FIGURE 12-1 ROOF ACCELERATION TIME HISTORY ............................................................................ 237
FIGURE 12-2 PORTION OF ROOF ACCELERATION TIME HISTORY ........................................................... 237
FIGURE 12-3 ACCELERATION RESPONSE SPECTRUM (UNDAMPED) .......................................................... 238
FIGURE 12-4 PROFILE OF FLOOR ACCELERATIONS ............................................................................. 239
FIGURE 12-5 HYSTERESIS FOR SHEAR WALL BUILDING ......................................................................... 241
FIGURE 12-6 TYPICAL MOMENT CURVATURE CURVE ........................................................................... 242
FIGURE 14-1WAIKATO POLYTECH BLOCK F...................................................................................... 252
FIGURE 14-2 CANTERBURY UNIVERSITY STUDENT UNION .................................................................... 252
FIGURE 14-3 INVERCARGILL CIVIC BUILDING ..................................................................................... 252
FIGURE 14-4 ARMAGH STREET TOWER ............................................................................................ 253
FIGURE 14-5 CHRISTCHURCH DRAINAGE BOARD BUILDING ................................................................. 253
FIGURE 14-7 BASTIA HILL WATER TOWER ......................................................................................... 253
FIGURE 14-6 THE BEEHIVE ............................................................................................................ 253
FIGURE 14-8 BEAUMONT STREET ................................................................................................... 254
FIGURE 14-9 SYMONDS STREET ..................................................................................................... 254
FIGURE 14-10 CLOCK TOWER ...................................................................................................... 254
FIGURE 14-11 LORNE STREET ........................................................................................................ 255
FIGURE 14-12 OLD BOYS SCHOOL ............................................................................................... 255
FIGURE 14-15 AUCKLAND MUSEUM ............................................................................................... 255
FIGURE 14-13 THE GATEWAY ....................................................................................................... 256
FIGURE 14-14 PANAMA STREET ..................................................................................................... 256
FIGURE 14-16 NZI AND SBI BUILDINGS .......................................................................................... 256
FIGURE 14-17 RADISSON HUNTLEY ................................................................................................ 256
FIGURE 14-18 AUCKLAND RAILWAY STATION .................................................................................... 257
FIGURE 14-19 QUEEN STREET BUILDINGS ........................................................................................ 257
FIGURE 14-20 CHRISTCHURCH CATHEDRAL ..................................................................................... 258
FIGURE 14-21 GALBRAITH BUILDING .............................................................................................. 258
FIGURE 14-22 HIGHT LIBRARY ....................................................................................................... 258
FIGURE 14-23 RIVERSIDE CASINO .................................................................................................. 259
FIGURE 14-24 COCOA BUILDING .................................................................................................. 259

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

xii

LIST OF TABLES

TABLE 2-1 SEISMIC HAZARD LEVELS ................................................................................................... 10


TABLE 2-2 BUILDING PERFORMANCE LEVELS ....................................................................................... 10
TABLE 2-3 DAMAGE TO STRUCTURAL ELEMENTS.................................................................................... 11
TABLE 3-1 EXAMPLE EXECUTION TIMES 300 MHZ PENTIUM II ................................................................ 17
TABLE 4-1 EARTHQUAKE SCALE FACTORS ........................................................................................... 22
TABLE 4-2 RECORDS AT SOIL SITES > 10 KM FROM SOURCES ............................................................... 28
TABLE 4-3 RECORDS AT SOIL SITES NEAR SOURCES.............................................................................. 28
TABLE 5-1 ACCURACY OF PLANE STRESS ELEMENTS ............................................................................... 33
TABLE 5-2 MASS OPTIONS ............................................................................................................. 39
TABLE 5-3 RECOMMENDED MASS OPTIONS ........................................................................................ 40
TABLE 5-4 CALCULATION OF MASS ECCENTRICITIES .............................................................................. 41
TABLE 6-1 EXAMPLE DEGRADATION COEFFICIENTS ............................................................................... 49
TABLE 6-2 VARIATIONS OF GRID REFINEMENT ..................................................................................... 64
TABLE 6-3 TYFO MATERIAL PROPERTIES ............................................................................................. 84
TABLE 6-4 STRENGTH PROVIDED BY FRP OVERLAY ............................................................................... 87
TABLE 7-1 FILES CREATED BY MODELA. ........................................................................................... 90
TABLE 8-1 KEYS ACTIVE FOR UNDEFORMED SHAPE ............................................................................. 160
TABLE 8-2 KEYS ACTIVE FOR DEFORMED SHAPE ................................................................................. 163
TABLE 8-3 ANIMATION OPTIONS.................................................................................................... 167
TABLE 8-4 NODAL COMPONENT OPTIONS ...................................................................................... 168
TABLE 8-5 (A) : ELEMENT COMPONENT OPTIONS ............................................................................ 168
TABLE 8-6 VARIATIONS IN SOLUTION PARAMETERS .............................................................................. 175
TABLE 8-7 EFFECT OF SOLUTION PARAMETERS ................................................................................... 176
TABLE 10-1 TRUSS COMPONENTS .................................................................................................. 211
TABLE 10-2 BEAM COMPONENTS................................................................................................... 211
TABLE 10-3 COLUMN COMPONENTS.............................................................................................. 212
TABLE 10-4 WALL COMPONENTS ................................................................................................... 213
TABLE 10-5 FLOOR COMPONENTS ................................................................................................. 214
TABLE 10-6 GAP COMPONENTS .................................................................................................... 215
TABLE 10-7 ANSROUT FILES ....................................................................................................... 222
TABLE 11-1 CONCRETE BEAM ACCEPTANCE CRITERIA ......................................................................... 224
TABLE 11-2 CONCRETE COLUMN ACCEPTANCE CRITERIA .................................................................... 225
TABLE 11-3 FLEXURAL WALL ACCEPTANCE CRITERIA ............................................................................ 228
TABLE 11-4 CRITERIA FOR WALLS CONTROLLED BY SHEAR .................................................................... 229

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

xiii

TABLE 11-5 NZNSEE LIMITS FOR FACE LOADED URM WALLS ............................................................. 230
TABLE 11-6 FEMA-273 LIMITS FOR FACE LOADED URM WALLS ........................................................... 230
TABLE 11-7 ANCHORAGE OF FRP .................................................................................................. 232
TABLE 11-8 PRESUMPTIVE ULTIMATE FOUNDATION PRESSURES .............................................................. 234
TABLE 12-1COLUMN DUCTILITY ..................................................................................................... 242
TABLE 12-2 EXAMPLE BEAM-COLUMN JOINT EVALUATION .................................................................... 246
TABLE 13-1 CURRENT TASK LIST ..................................................................................................... 249
TABLE 14-1 PROJECT LIST ............................................................................................................. 250

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

xiv

1 INTRODUCTION

The Holmes Consulting Group philosophy of innovation in structural design has by necessity
required innovation in structural analysis. Projects such as Union House base isolation, Parliament
Strengthening, Coopers Lybrand and the Museum of New Zealand, Te Papa Tongarewa, were
designed using non-linear analysis techniques as part of the evaluation procedure. Since 1991, this
type of analysis has been performed in-house using proprietary software.
In the last few years (1996 onwards), performance based design has become favored for the
evaluation and strengthening of existing buildings to resist earthquake loads. This type of design
uses similar analysis techniques to those developed by HCG and so the in-house software has been
adapted to incorporate the requirements of performance based evaluation. This manual describes
the implementation of these procedures:

1.1

CODE COMPLIANCE ISSUES

The analysis procedures described in these guidelines have been under development since the early
1990s. The material model formulations and acceptance criteria at that time were based on
engineering mechanics and experimental evidence as there were no prescriptive codes.
Since the initial development, the NZNSEE (1995, 1996) published guidelines for the evaluation of
earthquake risk buildings. Although these did not have explicit requirements for modeling for nonlinear analysis, acceptance criteria could be developed generally in conformance with these guidelines.
In 1997 FEMA published the NEHRP Guidelines for the Seismic Rehabilitation of Buildings (FEMA-273)
and it is planned to issue this as a pre-standard, FEMA-356. These guidelines provide much more
explicit procedure for modeling structures and define two non-linear procedures, the NSP and the
NDP. These guidelines also provide numerical acceptance criteria for components as a function of
the required performance level.
FEMA-273 has been the main source document for the continued development of these procedures
in the last few years. For some projects, generally in the U.S., the evaluations are fully compliant to
this document. For NZ projects, we use FEMA-273 wherever the NSNSEE guidelines do not
provide specific analysis procedures of acceptance criteria.
The basis of FEMA-273 material models for non-linear analysis is the definition of backbone curves,
which are envelopes of strength versus elastic and inelastic deformation. Some material models in
our procedures have been updated to directly incorporate FEMA-273 requirements (for example,
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

URM walls). For other materials (concrete walls, frames) the default procedures are those originally
defined based on experimental data. If you want a fully FEMA-273 compliant evaluation for these
latter types of element, you will need to define the material model yourself using the facilities built in
to the input spreadsheet.
As FEMA-273 acceptance becomes more widespread, and it is issued as a pre-standard, we will
implement the backbone curves directly into the evaluation procedure.

1.2

OBJECTIVES OF THE HCG PROCEDURE

All structural analysis techniques involve some approximations to true behavior and have limitations
in predicting the exact response to earthquakes. Non-linear time history analysis is the type of
analysis that closest matches the true response of a structure. However, this type of analysis has well
documented drawbacks, in particular the selection of appropriate time history input and accurately
modeling the hysteretic behavior of different element types.
The HCG non-linear analysis procedure has been developed to achieve two main objectives:
1. To be as accurate as, or more accurate than, any alternative method of analysis and/or
evaluation.
2. To identify seismic deficiencies correctly, and in the correct order of vulnerability.
The procedure is based on macro-modeling techniques, where the force-deflection function for a
complete element is used to define the response characteristics. For most building structures, the
size limit imposed by desktop computers provides ample scope for model refinement using this
modeling method. In general, we can use models of the same refinement as would be used for linear
elastic analysis.
This equivalent model discretization, the same as for linear elastic analysis, ensures that for an elastic
level of response the procedure will remain as accurate as less sophisticated methods of analysis. It
also allows for the eventual migration of this procedure into the standard method of analysis for the
design of new buildings as well as for the evaluation of existing buildings.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

2 PERFORMANCE BASED DESIGN

Performance based design (PBD) concepts for seismic rehabilitation are well developed in the U.S
and will increasingly form the basis for future strengthening projects in that country, particularly as
embodied in FEMA-273 (Guidelines which are to be issued as Pre-standard FEMA-356). These
concepts differ from code approaches by providing criteria to achieve specific Building Performance
Levels (Collapse Prevention, Life Safety, Immediate Occupancy and Operational) rather than setting
specific component design rules.
Prescriptive codes such as NZS4203 and the associated material codes in NZ, and the UBC in the
U.S, provide design forces and detailing requirements. Provided these requirements are met, it is
assumed that the seismic performance will be adequate. This implicitly designs to a Life Safety level
of performance. With performance based design, the actual damage to the building is evaluated at
particular levels of ground shaking and the extent of damage measured against prescribed acceptance
criteria. This approach provides detailed information to the owner on expected damage for
particular levels of earthquake. This information then forms a logical and rational basis for decision
making based on the cost effectiveness of particular strengthening schemes.

2.1

NEW ZEALAND PRACTICE

Building strengthening requirements in NZ are moving towards performance based design, although
the procedure is not formally implemented or even known by this name. The NZNSEE Guidelines
for Earthquake Risk Buildings (Red Book and Green Book) do adjust the risk factors and seismic
performance factors, implicitly varying the performance objectives. However, the objectives are
almost exclusively directed at life safety and collapse prevention rather than continuing operations.
The procedures in this manual were originally developed using NZNSEE Guidelines but as more
detailed development is performed they are moving more towards U.S. practice. This is because
there is generally no conflict between the two and the FEMA Guidelines provide more detailed rules
for this type of analysis.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

2.2

UNITED STATES PRACTICE

Performance based design has been fully described in the U.S. in guidelines published by FEMA and
ATC. These documents do not have the force of codes but do provide details of best practice for
the evaluation and strengthening of existing buildings. These are continuing to be expanded as PBD
becomes more widespread.

2.3

ANALYSIS TYPES

There is a hierarchy of four levels of structural analysis appropriate for the evaluation of existing
buildings (FEMA). Each higher level procedure provides a more accurate model of the actual
performance of a building subjected to earthquake loads, but requires greater effort in terms of data
preparation time and computational effort.
1. The Linear Static Procedure (LSP) is suitable only for regular buildings, which respond primarily
within the elastic range.
2. The Linear Dynamic Procedure (LDP) is able to model irregular buildings but is also suitable for
buildings which respond primarily within the elastic range.
3. The Non-linear Static Procedure (NSP) can evaluate buildings loaded beyond the elastic range
but is unable to fully capture the dynamics of response, especially higher mode effects.
4. The Non-linear Dynamic Procedure (NDP) is the most complete form of analysis, modeling
both dynamic effects and inelastic response. However, it is sensitive to modeling and ground
motion assumptions.
Linear Static Procedures
LSP
Linear Dynamic Procedure
LDP
Non-linear Static Procedure
NSP
Non-linear Dynamic Procedure
NDP
The most advanced form of analysis, NDP, is unique among the four methods in that it provides
realistic indications of demands on individual components loaded significantly beyond the elastic
range. Because of this capability, acceptance criteria for NDP have less inherent margin and
generally require less remedial work than the more simplified methods.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

The cost of obtaining the benefits of the most efficient and cost effective remedial work is the
increased complexity of analysis and evaluation to ensure that the sensitivity of the procedure to
modeling and ground motion is properly addressed.
The HCG procedure for the evaluation of existing buildings has been developed to implement the
two highest levels of analysis, NSP and NDP. It provides a practical, efficient and rigorous method
of implementing sophisticated non-linear static and dynamic procedures to a wide range of existing
buildings.

2.3.1

LINEAR STATIC PROCEDURE

The LSP is the equivalent static load procedure, as specified in NZS4203 and the seismic design
codes of most countries. A set of static loads is calculated based on the fundamental period of the
structure and the seismic conditions at the site (such as zone, importance of structure, soil type). The
loads are distributed up the height of the structure in a manner consistent with first mode response.
Higher mode effects are approximated by an additional fraction of the load applied at roof level in
most codes (e.g. NZS4203 and UBC).
This type of analysis is most commonly performed either by manual calculations or by using the
ETABS computer program, or similar, in NZ and the US.

2.3.2

LINEAR DYNAMIC PROCEDURE

The LDP is the response spectrum method of analysis, specified in NZS4203 and most other seismic
codes. This is the most common method of analysis for the design of new buildings, which are
unsuited for the LSP due to their size or irregularities in configuration. As computer procedures in
design offices become more sophisticated, the LDP procedure is becoming the standard design
analysis technique even for buildings for which codes permit the LSP.
In our company, the HCG spreadsheets DUCTILEIN and DUCTILEOUT have been developed to
implement the LDP using ETABS as the analysis program.

2.3.3

NON-LINEAR STATIC PROCEDURE

There is no equivalent of the NSP in either NZS4203 or UBC as it is mainly applicable to existing
buildings rather than new buildings, although a current movement toward displacement based design
methods may lead to this procedure being used for new buildings.
The procedure is sometimes called a Pushover Analysis as it involves the application of a set of loads,
or displacements, applied incrementally in one direction until failure occurs (pushover). The loads
are generally proportional to the first mode shape of the structure, although there are variations to
attempt to account for higher mode effects.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

The NSP results in a capacity curve from which equivalent viscous damping and an effective period
can be calculated. The capacity can be compared to the demand spectrum, adjusted for the
computed damping ratio, to determine whether the building is satisfactory or requires strengthening.
There are some specialized programs available for performing NSP although so far these are limited
to 2D models comprising flexural elements only. Linear elastic analysis programs (ETABS,
SAP2000) can be used to construct the pushover curve in a piece-wise linear fashion. Hinges are
inserted in the model as limit states are attained in successive elements. This can be a time
consuming process. Some programs (e.g. SAP2000) are currently being enhanced to automate this
process which will make the process more efficient. These versions of linear elastic programs, which
are suitable for NSP, will be available in the near future. Beta versions of SAP2000 for pushover
analysis have been released.
The HCG procedure described in this manual incorporates procedures for both displacement and
force based pushover analyses. These are currently under development so should be used with
caution. Figure 2-1 shows the modules used in the implementation of the NSP using the HCG
procedures.

FIGURE 2-1 NONLINEAR STATIC PROCEDURE (NSP)

MODELA
Excel Spreadsheet
input preparation

ZAP / SAP2000
Modal
Analysis

MODELA
Import Periods and
Mode Shapes

2.3.4

AutoCad DXF file


for Model Checking
Zap/SAP2000
Input Files
Periods and
Participation Factors
Mode Shapes

Calculate PF and
Factors
Data pasted to Capacity
Spectrum
Spreadsheet

NON-LINEAR DYNAMIC PROCEDURE

The NDP is the time history method of analysis, as described in NZS4203, UBC and other codes.
The basis for non-linear analysis is well known, an application of Newtons law that force equals
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

mass times acceleration. The finite element method of modeling structures as discrete elements is
also well known, and dates back to the first computers. A building is divided into degrees of
freedom and a set of simultaneous equations formed to solve for the displaced shape of the building
under a given set of forces.
For an earthquake time history analysis, the set of applied forces changes continuously as the ground
acceleration changes. If the building is being damaged then the stiffness continuously changes as
well. The response is calculated at each time step, typically every 1/100th of a second or less. At
each step the loads are changed, and the stiffness also changed if necessary. For a typical building,
this may require solving a set of several thousand simultaneous equations up to 3000 times. This is
where computer speed becomes essential.
This manual describes the implementation of the NDP for the evaluation of existing structures using
the HCG procedure. Figure 2-2 shows the modules used in the implementation of the NDP using
the HCG procedures.

FIGURE 2-2 NONLINEAR DYNAMIC PROCEDURE (NDP)

MODELA
Excel Spreadsheet
input preparation

AutoCad DXF file


for Model Checking
Multiple ANSR-L
input files

ANSR-L

Results Envelopes

Non-Linear Time
History Analysis

Node and Element


Time Histories
Acceleration Time
History for Component
Models

PROCESSA

Element and Nodal


Envelopes
Evaluation of Drifts
and Element Actions

Process results, import


to Excel
Spreadsheet

Plots of Element
Results
Write AutoCad DXF Files
to Display Cracking and
Hinging

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

2.4

IMPLICATIONS OF NDP PROCEDURE

As discussed above, guidelines for performance based design specify a hierarchy of four levels of
analysis. However, only the two non-linear options explicitly model the performance of the
components under damaging levels of earthquake.
Of the two, only the dynamic procedures fully account for the dynamic characteristics of the
structure. As earthquakes are a dynamic phenomenon, these dynamic characteristics have a large
influence on the magnitude and distribution of damage.
It is apparent that the optimum procedure will consider both non-linearity and dynamic effects, and
so the NDP is the preferred method of analysis. This method of analysis has traditionally had a
number of disadvantages, which have restricted its use. The HCG procedure has attempted to
overcome these:

Computationally intensive. Use the fastest office computers and ensure that there is sufficient
memory for the complete solution to be performed in RAM.

Input difficult to prepare. Use a spreadsheet format, based on Excel.

Results difficult to interpret. Use post-processors and spreadsheet formulas.

Appropriate time histories difficult to select. The problem of using seismic loads appropriate for the
site applies to all methods of analysis even if masked by a smoothed response spectrum. The
procedure is based on the use of multiple time histories, selected based on site considerations.

Element hysteretic behavior difficult to specify. This also applies for all methods of analysis, although it
is ignored for linear procedures. The HCG procedure uses models based on available
experimental data and continually updated.

The HCG software suite for performance based evaluation and the design of strengthening is based
on the use of ANSR-L as the analysis engine. ANSR-L is a general-purpose non-linear analysis
computer program. The other modules used in the procedures are basically data manipulation
routines for input and output, which are used to assemble the model for ANSR-L and to interpret
the results. These modules are based on Microsoft Excel as the user interface. The HCG drawing
program AutoCad is used as the presentation tool.

2.5

2.5.1

PERFORMANCE OBJECTIVES

SETTING OBJECTIVES

A performance level, such as Collapse Prevention, is not meaningful in itself unless you also specify
the level of earthquake at which this performance will be achieved. It is obviously easier to avoid
collapse for a distant M5 earthquake than it is for an M7 close by.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

The best way to describe performance objectives is to use the terminology of FEMA-273. These
guidelines define a Rehabilitation Objective as achieving a specific Building Performance Level at a particular
Earthquake Hazard Level.

Examples of different hazard levels are shown in Table 2-1. Each successive level has a lesser
probability of occurring during the life of the building. Most codes (NZS4203, UBC) specify
design loads with a 10% probability of occurrence in a 50 year design life. This is termed the
Basic Safety Earthquake 1 (BSE-1) in FEMA. Some codes, for example UBC Seismic Isolation
provisions, also specify a higher level earthquake with a much lower probability of occurring in
the 50-year earthquake, usually only 2%. In FEMA this is termed the Basic Safety Earthquake 2
(BSE-2).

Four levels of building performance are defined, as listed in Table 2-2 in increasing level of
damage. The first performance level, Immediate Occupancy (IO) requires that the building
remain essentially functional during and immediately after the earthquake. The last performance
level, Collapse Prevention (CP) will result in a building on the point of collapse and probably
unable to be economically repaired. The Life Safety level (LS) is the level usually implicit in
codes and may also result in a building which is not economic to repair.

The Rehabilitation Objectives are formed of combinations of earthquake hazard and building
performance. FEMA sets a desirable goal for rehabilitation, a Basic Safety Objective (BSO) which
comprises two targets:
1) Life Safety building performance at BSE-1, the 475 year earthquake.
2) Collapse Prevention performance at BSE-2, the 2,500 year earthquake.
Buildings properly designed and constructed to codes such as NZS4203 and the UBC may be
deemed by code enforcement authorities to meet the BSO.
Depending on the function of the structure, other objectives may be set. For example, an enhanced
objective may set higher building performance levels at BSE-1 and BSE-2 for critical facilities such as
hospitals of civil defense headquarters. Conversely, for some buildings more limited rehabilitation
objectives may be set, with inferior performance to the BSO. These are often set for older buildings
where the costs of meeting the BSO are prohibitive but it is better to achieve some improvement in
seismic performance than do nothing. In summary,
1) For most projects, the Basic Safety Objective is the default objective, equivalent performance to
a new building.
2) Enhanced objectives may be set after discussion with the owner or may be imposed by
authorities for certain classifications of facilities.
3) Limited objectives may be set for buildings such as historic unreinforced masonry buildings
where achieving the BSO would destroy the architectural fabric or where costs would be
excessive. These are usually only set after discussion with regulatory authorities. In NZ, these
are implicit in NZNSEE Guidelines by reducing the R factor.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

TABLE 2-1 SEISMIC HAZARD LEVELS

Probability of
Exceedance
50% in 50 years
20% in 50 years
10% in 50 years
2% in 50 years

Mean
Return
Period
(Years
72
225
474
2,475

FEMA
273

Codes

BSE-1
BSE-2

Design basis earthquake (DBE)


Maximum Capable Earthquake (MCE)
or Maximum Credible Earthquake
or Maximum Considered Earthquake

TABLE 2-2 BUILDING PERFORMANCE LEVELS

Performance
Level
Operational (O)

Immediate Occupancy (IO)

Life Safety (LS)

Collapse Prevention (CP)

Structural
Performance
Very light damage.
No permanent drift.
Substantially original strength
and stiffness.
Light damage.
No permanent drift.
Substantially original strength
and stiffness.
Minor cracking.
Elevators can be restarted.
Fire protection operable.
Moderate damage.
Some permanent drift.
Residual strength & stiffness in
all stories.
Gravity elements function.
Building may be beyond
economical repair.
Severe damage.
Large permanent drifts.
Little residual strength or
stiffness.
Gravity elements function.
Some exits blocked.
Building near collapse.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

Non-Structural
Performance
Negligible damage. Power and
other utilities available.
Equipment & contents secure
but may not operate due to
mechanical / utility failure.

Falling hazard mitigated but


extensive systems damage.

Extensive damage.

10

2.5.2

ACHIEVING PERFORMANCE OBJECTIVES

Table 2-3, extracted from FEMA-273, provides a general description of the overall performance of
buildings at different performance levels. The building types in Table 2-3 are those we most often
evaluate, see FEMA for damage descriptions for other structural types.
The definitions in Table 2-3 are qualitative rather than quantitative. Material sections of FEMA
provide quantitative acceptance criteria to ensure that the performance level is achieved. For
example, there are limits on beam and column plastic rotations for each of the IO, LS and CP levels
of performance. These limits depend on element details and also on the method of analysis used to
obtain the element demands.
The HCG Performance Based Evaluation procedures are based on the NSP and NDP methods of
analysis and results are evaluated in terms of the acceptance criteria for these procedures, as
discussed in Chapter 11 of this manual.

TABLE 2-3 DAMAGE TO STRUCTURAL ELEMENTS

Element

Collapse Prevention
CP

Life Safety
LS

Concrete Frames

Extensive cracking and


hinge formation in
ductile elements.
Limited cracking /
splice failure in
nonductile columns.

Extensive damage to
beams. Cover
spanning and shear
cracking (< 3 mm).
Minor spanning in
nonductile columns.
Joint cracks < 3 mm.
2% transient drift.
1% permanent drift.

4% transient or
permanent drift.
Concrete Walls

Major flexural and


shear cracks. Extensive
crushing and buckling
of reinforcing. Failure
around openings.
Boundary element
damage. Coupling
beams virtually
disintegrated.
2% transient or
permanent drift.

Boundary element
distress, including
some buckling of
reinforcement. Some
sliding at joints. Some
crushing and flexural
cracking. Coupling
beam shear and
flexural cracks.
Concrete generally
remains in place.
1% transient drift
0.5% permanent drift

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

Immediate
Occupancy
IO
Minor hairline
cracking.
Limited yielding at a
few locations.
No crushing (strain <
0.003).
1% transient drift.
Negligible permanent
drift.
Minor cracking of
walls, < 1.5 mm.
Coupling beam
cracking < 3 mm
width.

0.5% transient drift


Negligible permanent
drift
11

Element

Collapse Prevention
CP

Unreinforced
Masonry Walls

Extensive cracking.
Noticeable in-plane
and out-of-plane
offsets.
1% transient or
permanent drift
Crushing; extensive
cracking. Damage at
openings and corners.
Some fallen units.
1.5% transient or
permanent drift
Some connection
failures but no
elements
dislodged.

Reinforced
Masonry Walls

Precast Concrete
Connections

Life Safety
LS

Immediate
Occupancy
IO
Extensive cracking.
Minor cracking of
Noticeable in-plane
veneers (< 3 mm). No
and minor out-of-plane observable out-ofoffsets.
plane offsets.
0.6% transient drift
0.3% transient drift
0.6% permanent drift
0.3% permanent drift
Extensive cracking (<
Minor cracking (< 3
6 mm) throughout
mm width).
wall.
No out-of-plane
Some isolated crushing. offsets.
0.6% transient drift
0.2% transient drift
0.6% permanent drift
0.2% permanent drift
Local crushing and
Minor working at
spalling at connections cracks. Crack widths <
but no gross failure.
1.5 mm at connections.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

12

3 HOLMES CONSULTING GROUP SOFTWARE

3.1

MODELA SPREADSHEET AND PROGRAM

The complete set of input data is contained in the spreadsheet MODELA.XLS. This is a Microsoft
Excel workbook, which contains a number of pre-formatted sheets for specific data items.
Although the spreadsheet contains a number of Visual Basic routines, it is basically an input file for
the non-linear analysis procedure. A macro reads data on the sheets, writes it into a disk file and
executes a program, MODELA.EXE. This program has been written in-house in the Fortran 90
computer language and is maintained in-house.
The program MODELA reads in the description of the structure as entered in the spreadsheet and
re-formats this into the ANSR-L input file format. It assigns node numbers, degrees of freedom,
masses, element groups etc. Information about the numbering used is written into a disk file,
identified by the suffix TRC, for later use by the post-processor program.
In most cases, MODELA will be used to write multiple input files, one for each earthquake,
earthquake component orientation and eccentricity. A typical series of analyses will comprise 3
earthquakes x 2 orientations x 3 eccentricities = 18 separate runs. MODELA writes separate input
files for each analysis and also writes a batch file to submit the full sequence. For a displacement
pushover analysis, a single ANSR-L input file is written.
For a force pushover analysis, an input file is written for either the ZAP or SAP2000 linear elastic
analysis programs. These programs are used to extract mode shapes, which are imported to assemble
a force vector for the ANSR-L input file. ANSR-L is used to solve for the capacity curve.

3.2

ANSR-L PROGRAM

ANSR-L is the program which performs the detailed non-linear analysis. This program was
originally written in Fortran 66 for CDC mainframe computers and was converted to Fortran 77 for
minicomputers in the 1980s. The HCG version has been further converted by re-building using the
Lahey Fortran 90 compiler (hence the L suffix) to run under either DOS or WINDOWS. The
program is maintained in-house.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

13

3.3

ANSROUT SPREADSHEET AND PROCESSA PROGRAM

ANSROUT.XLS is a Microsoft Excel spreadsheet which imports results from one or more ANSR-L
analyses. The spreadsheet contains Visual Basic modules but the major computational effort in
reading the output files and extracting summaries is performed by the program PROCESSA.EXE
which is activated by a macro in the ANSROUT spreadsheet. Like MODELA, the PROCESSA
program has been written in-house in Fortran 90 and is maintained in-house.

3.4

COMPUTER SET-UP

The preferred set-up is to create two directories on the local C drive, named HCGPROGS and
QUAKES respectively.
The directory C:\HCGPROGS should contain the current versions of all the required programs,
spreadsheets and utilities:
MODELA.XLS
MODELA.EXE
ANSRL.EXE
ANSROUT.XLS
PROCESSA.EXE
TNT.EXE
LF90.EER
EATA.BAT
It is preferable to set a DOS path to this directory (see PATH command in your AUTOEXEC.BAT
file). This has two major advantages,
1.

The program ANSRL can be run in interactive mode without typing in the full path.

2.

The error file, LF90.EER, does not need to be copied to the job directory if an error occurs in
any of the executables.

The C:\QUAKES directory should be copied in its entirety. This is the source of the accelerogram
data files read by ANSRL. The input files reference the pathname and filename for the appropriate
earthquake files and so the sub-directory structure must be maintained.
When element time histories are requested, a large number of files may be opened simultaneously by
ANSR-L and also during the processing phase by PROCESSA. This may require a change to your
CONFIG.SYS file, which should reside in your C: root directory. This file should contain a line:
FILES=30
If it does not contain this line, insert it. If the number is less than 30, increase it to 30. If you still get
this error, increase the number of files. This change will not take effect until you next boot your
computer.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

14

3.5

AUTOCAD PROGRAM

The MODELA and PROCESSA programs write DXF files, which can be imported into drawing
programs. At HCG the standard drawing programs used are AutoCad and AutoCad-LT. The
geometry checking is most effective if AutoCad-LT is loaded on your local hard drive.
During output processing a series of stress or strain coded DXF files which can be displayed directly
in AutoCad LT are created. Another series of files, .CRA, provide crack patterns for walls. These
files are imported into AutoCad via a Lisp program, which is only available on the full version of
AutoCad.

3.6

COMPUTER RESOURCE DEMANDS

Non-linear dynamic analysis, as performed by ANSRL, is very demanding on computer resources.


Of critical importance are the amount of memory, the CPU speed and the disk space available.

3.6.1

MEMORY REQUIREMENTS

Linear elastic programs such as ETABS and SAP2000 will solve large problems on machines with
limited memory by forming the stiffness matrix in blocks and keeping most of the blocks stored on
disk while the equation reduction is being performed. You will notice when running ETABS, for
example, that forming and condensing the stiffness blocks can take a large portion of the execution
time.
Stiffness matrix blocking is practical when the equations are solved once, as in a linear elastic analysis.
It is not practical when the equations must be solved several thousand times, as can occur in a nonlinear analysis. Therefore, practical non-linear analysis on desktop computers requires sufficient
memory to keep two copies of the stiffness matrix, plus the element information, within memory.
When yielding occurs the original stiffness matrix is changed, rather than completely re-formed,
which is why two copies are kept.
For the size of problems typically being solved at HCG with this procedure, 64 MB is the minimum
practical memory and 128 MB, or 256 MB, is better. As CPU speeds increase, larger problems are
solved and more memory is required. The 300 MHz and faster machines are best with 128 MB or
more.
Extra memory has the added advantage of being able to multi-task while the ANSR analyses are
running. Once the runs are started, entering <Alt><Tab> returns you to Windows where you can
run other applications. However, to do this successfully you need lots of memory. Note also that
this may corrupt the video memory, so that animated plot displays may have degraded when you
return to the DOS window.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

15

3.6.2

CPU TYPE

For the CPU, it is simply a case of the faster the better. The current top of the line machine (2001) is
a 1.5 GHz Pentium IV. Analysis should usually be on at least a Pentium III. For a given CPU chip,
the execution time will be approximately inversely proportional to the clock speed.

3.6.3

DISK SPACE

Each ANSR analysis may take from 2 MB to 40 MB of disk space, depending on the size of the
model and the options selected. The higher number is where extra space is required to obtain time
histories of displacements and/or element forces and also to save for re-start. As there may be 18 or
more analyses run as a series, disk space may approach 1 GB.
If disk space is critical, time history output files can be processed and discarded, one or more runs at
a time. Only the ECH and D01 output files for the complete set of runs need be retained. However,
generally the size of disks on our machines is such that disk space is no longer critical.

3.6.4

EXECUTION TIMES

Non-linear analysis is computationally intensive and execution times may be significant even with the
fastest computers. The time is difficult to predict, as it is a function of:
1.

The size of the problem, as measured by the number of degrees of freedom (dof).

2.

The connectivity of the elements, which determines the bandwidth of the stiffness matrix.
For buildings with rigid diaphragms, a building with 50 column lines times 50 stories will
execute faster than a building with 250 column lines and 10 stories.

3.

The extent of non-linearity. The equations must be reduced each time a stiffness term
changes. Only the lower triangular portion of the stiffness matrix needs to be reduced,
starting from the location in the matrix of the first term which changes.

There are techniques which can minimize these effects, such as the order of numbering of the nodes,
the sequence of numbering diaphragm nodes and whether degrees of freedom are numbered from
the top or the bottom. These are discussed in more detail in Chapter 8.
Table 3-1 lists execution times for a range of structures, which have been evaluated on a 300 MHz
Pentium II with 64 MB RAM. For a given structure, these times may vary by 50% depending on
the earthquake and orientation. Figure 3-1 illustrates how the times have changed with advances in
computer technology the fastest current office desktop computers would solve the problems in
Table 3-1 approximately 6 times faster now.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

16

TABLE 3-1 EXAMPLE EXECUTION TIMES 300 MHZ PENTIUM II

Number
Of Degrees
Of Freedom
76
741
2446
2700
4226
5521
8398
12702

Band
Width
26
78
202
2676
1862
2340
212
12693

Execution
Time (for
2000 Steps)
39 secs.
1.2 mins
6.3 mins.
1 hours
10 hours
12+ hours
3 hours
36+ hours

Project Description
Plane stress cantilever
18 Story frame
40 Story frame (Cooper Lybrand)
Palmerston North Hospital
3 x 7 Story Frame/Wall Buildings
Christchurch Cathedral
17 Story shear wall
Auckland Railway Station

FIGURE 3-1 REDUCTIONS IN COMPUTER EXECUTION TIME

2001 Pentium IV 1500 1.02

2000 Pentium III 800 2.33

1998 Pentium II 300

6.27

1995 Pentium I 90

60

420

1992 486 33 Mhz PC

1985 Prime SuperMini

1800
0

400

800

1200

1600

2000

RUN TIME (MINUTES)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

17

4 EARTHQUAKE LOADS

The form of the earthquake loads depends on the type of analysis - static lateral loads for the LSP,
response spectra for the LDP, force vectors and response spectra for the NSP and time histories of
accelerations for the NDP.
Codes typically specify the design response spectra as a function of parameters such as seismic zone,
soil type, risk factor etc. The design forces are based on the forces generated by this spectrum, with
either the forces or the spectrum reduced to account for the ductility of the system.
When the analysis type is non-linear, the ductility is accounted for as part of the analysis. The
earthquake loads input are the elastic level of load and are not reduced for ductility.
4.1

4.1.1

DEFINITION OF LOADS

GROUND MOTION AMPLITUDE

The level of seismic load is usually specified by applicable codes, for new buildings, or guidelines, for
existing buildings. For some important structures the seismic load may be developed from a sitespecific study. The codes, guidelines or seismic study provide design response spectra and may also
provide some guidance in selecting appropriate time histories to represent these spectra when the
NDP is used.
For the NDP, time histories appropriate for the response spectrum must be derived. These are most
commonly based on records from actual earthquakes, which are then scaled using either amplitude or
frequency scaling.

4.1.2

AMPLITUDE SCALING OF EARTHQUAKE RECORDS

Amplitude scaling of earthquake records involves applying a constant scale factor to the measured
accelerations. As the same factor is applied to all accelerations, the effect is to scale the response
spectrum produced by the accelerogram up or down by the same factor at all frequencies. The
shape of the response spectrum is not changed.
The difficulty with this form of scaling is that there is no numerical procedure to select the scaling
factor. Over the years, different procedures have been proposed to select the scaling factor, for
example, one proposed by Housner based on the area under the velocity spectrum. However, none
of these procedures have found widespread acceptance.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

18

For a non-linear analysis, the response of a structure will be over a range of periods and so judgment
is required to select a scaling factor appropriate for the full range. Additionally, there may be
significant contribution from higher modes and so it is preferable that the excitation at these periods
also be appropriate. This is difficult to achieve with the spectral shapes of actual recorded
earthquakes.
The 1997 UBC provides specific procedures for scaling records as a function of the natural period of
the structure. These are used for U.S projects and are being evaluated for possible use for NZ
projects.
4.1.3

FREQUENCY SCALING OF EARTHQUAKE RECORDS

Frequency scaling of records modifies the energy content of the record at various frequencies so as
to change the shape of the response spectrum produced by the record. The procedure used is
iterative:
1.

Select a starting time history record, usually an actual earthquake record although artificial
traces are sometimes used.

2.

Calculate the response spectrum of the record.

3.

Convert the time history to the frequency domain using a Fast Fourier Transform (FFT).

4.

Scale the FFT amplitude coefficients at each frequency by the ratio of the ordinate of the
target response spectrum to the spectral ordinate produced by the record.

5.

Transform the record back to the time domain by using an inverse FFT.

6.

Calculate the response spectrum of the modified record.

7.

Calculate the error function. If greater than the convergence criteria, return to step 3 and
repeat the procedure.

The error function is generally set so that the average error in the spectral ordinates is less than 5%.
This procedure does not have a theoretical basis other than the empirical result that the spectrum
produced by the modified record will be similar to the target spectrum. For a linear elastic time
history analysis the correct acceleration will be obtained at each mode, within the limits of the error
function, and so multiple records each scaled by this procedure will produce a similar response.
For non-linear analysis, the frequency modified time histories may produce widely different
responses, even though the records produce response spectra which are very similar. Variations of
50% or more in maximum displacements are not unusual from records which are scaled to the same
spectra.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

19

4.2
4.2.1

NEW ZEALAND PROJECTS


NZS4203 REQUIREMENTS FOR NEW BUILDINGS

For an ultimate limit state, the seismic loads for the numerical integration time history method of
analysis, equivalent to the NDP, are defined as a design spectrum, C(T) where:
For determination of minimum strength requirements
C(T)=Sm1Ch(T,1)SpRZLu
For determination of inelastic effects and capacity actions:
C(T)= Ch(T,1)RZLu
The value of Sm1 is a function of ductility, and is inversely proportional to the ductility factor for
periods greater than 0.7 seconds.
The first equation, for minimum strength requirements, permits a linear elastic time history analysis
to be performed to obtain design actions instead of the equivalent static analysis or response
spectrum analysis. In this case, the main difference between time history analysis and the other
methods is that the factor Sm2 does not need to be considered for the time history method.
When the second equation is used, to obtain ductility demands, the elastic value of C(T) is used,
without modification by Sm1. This is expected, as the analysis incorporates ductility directly.
However, the other difference is that the Sp factor is effectively set to 1.0, rather than 0.67 as is used
for other methods of analysis. This requires that an earthquake 50% higher than the design
earthquake be used to calculate ductility demand.
This manipulation of the level of earthquake by way of the Sp factor is not applied in other cases
where it would seem to be appropriate, for example for structures designed for an elastic level of
load. The use of Sp = 1 to assess ductility demands causes difficulties in defining loads for existing
buildings.
4.2.2

NZNSEE REQUIREMENTS FOR EXISTING BUILDINGS

In the seismic evaluation of existing buildings in NZ, the relevant sections of the building Act are:
1. Section 38: Alterations to existing buildings. Essentially, this section of the act requires that a
building will comply no less after the alteration than before. This precludes weakening, or the
addition of significant mass without a proportional increase in strength. However, no structure
need be stronger than required by code, so a structure with excess reserves of strength may be
weakened.
2. Section 46: Change of Use of buildings, etc. This requires that when a buildings use is changed, it
shall comply with the structural requirements of the code, as nearly as is practically possible to
the same extent as if it were a new building.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

20

Most buildings to which this procedure is applied fall within the second option above, a change of
use. Guidelines as to how to comply with the requirement listed are available from two NZNSEE
documents, both of which have been issued in draft form:
[1]

Earthquake Risk Buildings: Recommendations and Guidelines for Classifying, Interim Securing and
Strengthening, New Zealand National Society for Earthquake Engineering, December, 1985.
Technical update issued in draft form in February, 1995 (known as the Red Book).

[2]

The Assessment and Improvement of the Structural Performance of Earthquake Risk Buildings, Draft
for General Release, New Zealand National Society for Earthquake Engineering,
December, June, 1996 (known as the Green Book).

For a new building, the level of earthquake resistance would use Sp = 0.67 and R = 1.0. The
NZNSEE Red Book recommends that, for strengthening of existing buildings, seismic loads be
based on Sp = 0.60 and R = 0.8 (high occupancy). This is 72% of the NZS4203 requirement for new
buildings. The NZNSEE Green Book recommends that the requirement be 85% of that for new
buildings.

4.2.3

SCALING OF EARTHQUAKE RECORDS

The requirements of NZS4203 are that:


1.

Analysis shall use at least three different earthquake records

2.

Scaling shall be such that over the period range of interest the 5% damped spectrum of the
earthquake record does not differ significantly from the design spectrum.

3.

Records shall have a duration of at least 15 seconds of strong ground shaking.

These requirements provide wide scope for the selection and scaling of earthquake records.
Earthquake records can be scaled using either frequency scaling or amplitude scaling. Both methods
are used in the HCG procedure.
4.2.4

EARTHQUAKE RECORD SELECTION

The earthquake record selection adopted for the HCG procedure for NZ projects is to use one
frequency scaled record and two amplitude scaled records. Table 4-1 lists the three earthquake
records used for each soil type and the scale factors applied to the records. Figures 4-1 to 4-3 plot
the 5% damped spectra of the records used.
Codes such as NZS4203 do not provide guidance as to the orientation of the earthquake along the
principal axes of the building. The HCG procedure implements the following steps:

The dominant record of the earthquake is scaled to match the NZS4203 spectrum. This is
applied along the global X axis of the model.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

21

The orthogonal record is scaled using a scale factor of 60% times the factor applied to the
dominant record. This record is applied along the global Z axis of the model.

The records are switched to apply the dominant record along Z and the lesser record along X
and the analysis repeated.

Although ANSR-L can apply acceleration records along the three axes of the model simultaneously,
the procedure to date has not applied any vertical acceleration concurrent with the horizontal
accelerations.

TABLE 4-1 EARTHQUAKE SCALE FACTORS

Record
Stiff Soil
Spectrum Compatible
El Centro, 1940
Edgecumbe, 1987, Matahina Dam
Intermediate Soil
Spectrum Compatible
El Centro, 1940
Edgecumbe, 1987, Matahina Dam
Soft Soil
Spectrum Compatible
El Centro, 1940
Loma Prieta, 1989, Oakland Wharf

Factor

File

1.00
1.00
1.05

4203rock.th
elcentro.eq
d2.ans

1.00
1.00
1.10

4203int.th
elcentro.eq
d2.ans

1.00
1.50
1.75
(1.5 T1 > 1.5 s
2.0 T1 < 0.3s

4203soft.th
elcentro.eq
file132

Figures 4-1 to 4-3 illustrate the difficulties in setting a scale factor for actual earthquake records. The
first record for all 3 sites, which is a frequency scaled El Centro 1940 record, provides a close match
to the NZS4203 spectrum. However, the second record for all 3 sites, which is the amplitude scaled
El Centro, provides a match for only a partial period range. At other periods, the spectral
acceleration is either too high or too low relative to NZS4203.
It is obvious that the specification of appropriate time history records for NZ projects is an area
where additional development work is required. It may be necessary to allow for factors which are a
function of the natural period of the structure as in the 1997 UBC.
The default scaling factors used for NZS4203 may be over-ridden by specifying user-selected
records. To implement this, the earthquake records listed in Table 4-1 are selected as site specific
records and then alternate scale factors are specified. see later in this chapter.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

22

FIGURE 4-1 EARTHQUAKE RECORDS USED FOR ROCK SITES

0.800

ACCELERATION (g)

0.700

El Centro N-S Seed


El Centro E-W Seed
NZS4203

0.600
0.500
0.400
0.300
0.200
0.100
0.000
0.000

0.250

0.500

0.750

1.000

1.250

1.500

1.750

2.000

PERIOD (Seconds)

1.000
0.900
El Centro N-S
El Centro E-W
NZS4203

ACCELERATION (g)

0.800
0.700
0.600
0.500
0.400
0.300
0.200
0.100
0.000
0.000

0.250

0.500

0.750

1.000

1.250

1.500

1.750

2.000

PERIOD (Seconds)

1.000
Matahina Dam N83E

ACCELERATION (g)

0.800

Matahina Dam N07W


NZS4203

0.600

0.400

0.200

0.000
0.000

0.250

0.500

0.750

1.000

1.250

1.500

1.750

2.000

PERIOD (Seconds)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

23

FIGURE 4-2 EARTHQUAKE RECORDS USED FOR INTERMEDIATE SITES

0.900
El Centro N-S Seed
El Centro E-W Seed
NZS4203

0.800

ACCELERATION (g)

0.700
0.600
0.500
0.400
0.300
0.200
0.100
0.000
0.000

0.250

0.500

0.750

1.000

1.250

1.500

1.750

2.000

PERIOD (Seconds)

1.000
0.900

El Centro N-S
El Centro E-W
NZS4203

ACCELERATION (g)

0.800
0.700
0.600
0.500
0.400
0.300
0.200
0.100
0.000
0.000

0.250

0.500

0.750

1.000

1.250

1.500

1.750

2.000

PERIOD (Seconds)

1.000
Matahina Dam N83E

ACCELERATION (g)

0.800

Matahina Dam N07W


NZS4203

0.600

0.400

0.200

0.000
0.000

0.250

0.500

0.750

1.000

1.250

1.500

1.750

2.000

PERIOD (Seconds)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

24

FIGURE 4-3 EARTHQUAKE RECORDS USED FOR SOFT SITES

1.2

ACCELERATION (g)

1.0
El Centro N-S Seed
El Centro E-W Seed
NZS4203

0.8

0.6

0.4

0.2

0.0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

PERIOD (Seconds)
1.6
1.4
El Centro N-S
El Centro E-W
NZS4203

ACCELERATION (g)

1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

PERIOD (Seconds)
1.2
NZS4203
Oakland 2 Storey 290 deg
Oakland 2 Storey 200 deg

ACCELERATION (g)

1.0

0.8

0.6

0.4

0.2

0.0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

PERIOD (Seconds)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

25

4.3

UNITED STATES PROJECTS

Projects located in the U.S. will generally be evaluated according to the requirements of the UBC (for
new buildings, but used by many jurisdictions for the strengthening of existing buildings) or FEMA273, which are guidelines for the rehabilitation of existing buildings.
4.3.1

UBC REQUIREMENTS FOR NEW BUILDINGS

The UBC defines a response spectrum, as shown in Figure 4-4, which is a function of the
acceleration seismic coefficient, Ca, and the velocity seismic coefficient, Cv. Each of these
coefficients is a function of the seismic zone and soil profile type. For Seismic Zone 4 the
coefficients are also a function of the near fault factors, Na and Nv, which are a function of the
proximity to known faults, and the magnitudes and slip rate.
FIGURE 4-4 UBC DESIGN RESPONSE SPECTRUM

CONTROL PERIODS
Ts=Cv/2.5Ca
T0=0.2Ts

SPECTRAL ACCELERATION (g's)

2.5Ca

Cv/T
Ca

T0

4.3.2

Ts

PERIOD (SECONDS)

FEMA-273 REQUIREMENTS FOR EXISTING BUILDINGS

FEMA-273 defines a general response spectrum, as shown in Figure 4-5. The basis for the spectrum
is the short period acceleration parameter, Ss, and the one-second parameter, S1. These are extracted
from ground motion hazard maps for the BSE-2 level. The parameters may be adjusted for other
return periods of earthquake using supplied equations.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

26

The basic acceleration parameters Ss and S1 are adjusted for the site class to obtain the spectral
acceleration coefficients SXS and SX1 which define the spectrum in Figure 4-5.

FIGURE 4-5 FEMA-273 GENERAL RESPONSE SPECTRUM

SPECTRAL RESPONSE ACCELERATION

S =
a

S
B

XS
S

S =
a

S
BT
X1

0.2 T0

4.3.3

T0

PERIOD

SCALING OF EARTHQUAKE RECORDS

The UBC and FEMA-273 Guidelines are more explicit than NZS4203 and generally follow the same
requirements. They each require a minimum of three pairs of time history components. If seven of
more pairs are used then the average results can be used for design else maximum values must be
used. The records are required to have appropriate magnitudes, fault distances and source
mechanisms for the site. Simulated time histories are permitted.
Both the UBC and FEMA-273 provide the same explicit method of scaling records:
For each pair of horizontal components, the square root of the sum of the squares (SRSS) of the 5% damped
spectrum shall be constructed. The motions shall be scaled such that the average value of the SRSS spectra does
not fall below 1.4 times the 5% damped spectrum of the design earthquake for periods between 0.2T and 1.5T
(where T is the fundamental period of the building).
To me, this requirement is unclear as to whether the average value of the SRSS is the average over all
periods for each record or the average at each period over all records. Consensus seems to be for the
latter (based on a BRANZ study group). If so, then the scaling factor for any particular record
would depend on the other records selected for the data set.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

27

4.3.4

EARTHQUAKE RECORD SELECTION

UBC and FEMA-273 doe not provide specific earthquake records but another source, ATC-40,
provides 10 records identified as suitable candidates for sites distant from faults (Table 4-2) and 10
records for sites near to the fault (Table 4-3).
These records are available on the
C:\QUAKES\FARFAULT and C:\QUAKES\NEARFA directories respectively. Each directory
contains a spreadsheet, ACCEL.XLS, which contains the 5% damped spectrum for each component
of each record and has functions to compute scaling factors.
Each record has been formatted for use as input to ANSR-L. To use them, use the options for UserSelected earthquakes in MODELA. The file names for each record, in ANSR-L format, are given in
the final columns of Tables 4-2 and 4-3.

TABLE 4-2 RECORDS AT SOIL SITES > 10 KM FROM SOURCES

No.
1
2
3
4
5
6
7
8
9
10

M
7.1
6.5
6.6
6.6
7.1
7.1
7.5
7.5
6.7
6.7

Year
1949
1954
1971
1971
1989
1989
1992
1992
1994
1994

Earthquake
Western Washington
Eureka, CA
San Fernando, CA
San Fernando, CA
Loma Prieta, CA
Loma Prieta, CA
Landers, CA
Landers, CA
Northridge, CA
Northridge, CA

Station
Station 325
Station 022
Station 241
Station 241
Hollister, Sth & Pine
Gilroy #2
Yermo
Joshua Tree
Moorpark
Century City LACC N

File
wwash.1
eureka.9
sf241.2
sf458.10
holliste.5
gilroy#2.3
yermo.4
joshua.6
moorpark.8
lacc_nor.7

TABLE 4-3 RECORDS AT SOIL SITES NEAR SOURCES

No.
1
2
3
4
5
6
7
8
9
10

M
6.5
6.5
7.1
7.1
6.9
6.7
6.7
6.7
6.7
6.7

Year
1949
1954
1971
1971
1989
1989
1992
1992
1994
1994

Earthquake
Imperial Valley, CA
Imperial Valley, CA
Loma Prieta, CA
Loma Prieta, CA
Cape Mendocino, CA
Northridge, CA
Northridge, CA
Northridge, CA
Northridge, CA
Northridge, CA

Station
El Centro Array 6
El Centro Array 7
Corralitos
Capitola
Petrolia
Newhall Fire Station
Sylmar Hospital
Sylmar Converter Stat.
Sylmar Converter St E
Rinaldi Treatment Plant

File
ecarr6.8
ecarr7.9
corralit.5
capitola.6
petrolia.10
newhall.7
sylmarh.4
sylmarc.2
sylmare.1
rinaldi.3

The spreadsheets ACCEL have functions to generate UBC and FEMA spectra for specified ground
motion parameters and plot scaled components for the records listed in Table 4-2 and 4-3, as shown
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

28

in Figure 4-6 for one of the near fault records (sylmare.1). For each project, either 3 or 7 time
histories will be selected from the earthquakes listed and appropriate scale factors selected.
The data set for records > 10 km from the source has an additional 26 data sets from earthquake
records prior to 1971. These can also be used to select appropriate starting time histories. See the
In-Structure Damping and Energy Dissipation Design Guidelines for a list of these additional records.

FIGURE 4-6 EXAMPLE SCALED RECORDS

1994 Northridge, California Sylmar Converter Sta. East


1.60
Design Spectrum
Component 1
Component 2
Lower Period Limit
Upper Period Limit

ACCELERATION (g)

1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

PERIOD (Seconds)

4.4

USER-SELECTED RECORDS

The MODELA spreadsheet provides for user-selected earthquake records and scaling factors to be
used. This option may be used for NZ projects and should be used to select records for US projects,
using the record names listed above.
The NZS4203 compatible earthquakes are stored sub-directories ROCK, INTERM and SOFT in the
directory C:\QUAKES\NZCOMP. The files are as listed in Table 4-1. To use factors other than
the default factors listed in Table 4-1, and plotted in Figures 4-1 to 4-3, these files can be selected as
user-defined files. This allows any scale factor to be applied see the input description for site
specific records in Chapter 7 for more information.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

29

5 MODELING TECHNIQUES

The basis of the HCG evaluation procedure is a non-linear analysis, either time history (NDP) or
pushover (NSP). This requires the development of a finite element model of the structure to be
evaluated. As the full evaluation typically involves 16 or more analyses, each of at least 2000 time
steps, computer resources are an important consideration in development of the model. The
refinement of the model is a compromise between accurate representation of the building and
acceptable execution times.
This section provides an overview of developing the building model and describes the concepts of
the finite element modeling. The principles in this chapter are common to both linear and non-linear
finite element models in terms of developing model geometry. The material non-linearity requires
special considerations beyond those required for linear elastic models. Details of the material and
element properties to incorporate this material non-linearity are provided in Chapter 6.

5.1

FINITE ELEMENT MODELING

The basis of the finite element model is the grid definition, in the two horizontal directions and the
vertical direction. The model is defined by the column lines, located on the plan view of the
building, and the elevations, defined as horizontal levels on an elevation of the building.
Column line locations are selected so as to locate in plan:

Building columns.

Beam ends.

Corners or edges of wall panels.

Sides of opening in walls, although small openings (600 mm or less) are usually neglected.

Internal nodes of beams, for example where hinging occurs other than at the ends of the beam.
(This can result in a large increase is complexity of the model, so use with care - see later
discussion).

Peaks in gable walls or in openings.

Any special features in the plan of the building.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

30

The total number of column lines generally ranges from 20 or so to 300 or more. Columns do not
have to be numbered sequentially and there may be gaps. For example, if you are analyzing multiple
buildings the column lines may range from 1 to 50 for Building 1 and from 101 to 151 for Building 2.
The elevations are selected to model:

The point of fixity of the structural elements (generally all at ground level but may be over
several different levels for sloping ground).

Each diaphragm level.

Tops of door openings in walls.

Tops and bottoms of window openings in walls.

Peaks in gable walls or in wall openings.

Any special features in the vertical direction.

The elevations are numbered sequentially from 1 at the base. Gaps in the numbering are not permitted. The
number of elevations may range from 2 to 100 or more. For frame structures, the number of
elevations is generally equal to one plus the number of floors. For wall structures, there will often be
4 times as many elevations as floors (each diaphragm plus one level to define the top of door
openings, plus one level each for the top and bottom of window openings).
The maximum number of nodes will be the number of column lines times the number of elevations.
The maximum number of potential degrees of freedom is the number of nodes times 6, although on
average it will be about one-half this number. The execution time is heavily dependent on the
number of degrees of freedom.
The complexity of the model will be a trade-off with the size of the model. For a small low rise
building with only perimeter walls the wall openings will be able to be modeled accurately in both the
horizontal and vertical directions as the number of nodes will be relatively small (less than 500). As
the size and height of the building increases, or the number of interior walls increases, more
approximations will be necessary to locate door and window openings etc.
Use judgment in deciding how closely you want to represent the geometry. Keep in mind the
variability of other aspects of the analysis such as material strengths, element hysteresis properties
and the input ground motion. There is no point in developing a complex model accurate to the
nearest millimeter when there is considerable variation in these other parameters.
Put some thought into the number of levels used as it is easy to end up with a large number of
elevations in wall buildings to account for variations in door and window openings. It is often
acceptable to round these openings to within 500 mm or so vertically rather than add new levels.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

31

5.2

FLEXURAL ELEMENTS

Flexural elements are used to model beams, columns, trusses and base isolators. Flexural elements
are assumed to be prismatic and yielding can occur at concentrated plastic hinges at element ends
only. The ANSR-L program has the capability of modeling non-prismatic flexural elements but the
procedures for incorporating this into the HCG procedures have not yet been developed.
Usually, flexural elements have a node at each end of the element only. Exceptions may be where
plastic hinging occurs within the span, where concentrated load and/or masses are lumped within the
span or where elements are haunched. In these cases, you may need to add column lines or
elevations to split the beam or column into several elements. This will add to the problem size and
the execution times.
If the beam reinforcing is such that plastic hinging may occur within the span rather than the ends,
the additional nodes needed to include this may make the model overly complex. In some cases it
may be possible to extrapolate the internal yield moments to equivalent yield moments at the end of
the beam and model the beam as a single element. Note that if this is done then the plastic rotations
will be under-estimated. Output rotations should be factored by the ratio of the model yielding span
length to the actual yielding span length.

5.3

STRUCTURAL WALL ELEMENTS

Structural walls can be modeled as flexural elements but are more commonly modeled as assemblages
of plane elements. Plane elements are 4 node plane stress elements used to model walls or flexible
floors. The elements are generally rectangular but can be triangular or a general four sided shape.
The elements have in-plane stiffness only.
The grid of plane elements must be selected such that corners of adjacent elements meet at a
common point to form a continuous structure. That is, the corner of an element should not be
located at the mid-point of an adjacent element (see Figure 5-1).
The situation shown in Figure 5-1 is generally obvious for panels in the same plane but
incompatibilities often occur between grids at corners. When MODELA is used to generate the
model, these incompatibilities only occur when walls are generated at a level increment greater than
unity. If every wall is generated up the height with an increment of 1 then this type of
incompatibility will not exist. Care needs to be taken when panels span more than 1 level.
The plane stress element is intended to model shear deformations. Use of this type of element is less
accurate when deformations are dominated by flexure. Table 5-1 and Figure 5-2 demonstrate the
accuracy of the element as the wall becomes dominated more by flexure (the ratio of width to height
decreases). Table 5-1 shows that:

As with all finite element modeling, results converge from the stiffer side. That is, ANSR-L
displacements are always less than the theoretical values.

As the shear deformations become more dominant, the results become more accurate.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

32

Increasing the refinement of the grid increases the accuracy of the solution for flexuredominated aspect ratios.

FIGURE 5-1 PLANE STRESS ELEMENT GRIDS

Element should be divided to


same grid as adjacent elements.

The results in Table 5-1, and Figure 5-2, show that grid refinement is more important for flexure
dominated wall panels than for pure shear panels. For walls in which the width is equal to, or greater
than, the height then the results are within 5% of theoretical regardless of the grid used. However, as
the width becomes smaller relative to the height the grid refinement becomes more important for
accuracy.

TABLE 5-1 ACCURACY OF PLANE STRESS ELEMENTS

Width, W
25
5
Height, L
5
5
Aspect Ratio, W/H
5
1
Shear Deformation
95% 42%
Flexural Deformation 5% 58%
ANSR Displacement/ Theoretical
5 x 5 Grid
97% 95%
10 x 10 Grid
95% 98%

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

5
5
5
5
10
15
20
25
0.5 0.33 0.25 0.2
15% 7% 4% 3%
85% 93% 96% 97%
90% 82% 74% 66%
97% 96% 93% 90%

33

FIGURE 5-2 MODELING ACCURACY


100%

ASNSR-II Displacement / Theoretical

95%

90%
85%
80%

5 x 5 Grid

75%

10 x 10 Grid

70%
65%
60%
0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

5.00

Aspect Ratio, W/H

These conclusions are based on elastic wall response. The plane stress elements are pure shear
elements and the yield function is based on the shear stress. As structural shear walls become taller
relative to the width, the wall yield strength is more likely to be governed by flexure rather than shear.
When flexural yield is included in the wall model the grid spacing becomes less critical. This is
discussed in more detail in Chapter 6.
The element yield function is based on a piece-wise linear shear stress - shear strain envelope. A
change in stiffness occurs when the shear strain reaches a point where the slope of the envelope
curve changes. At this stage, the total stiffness of the element is changed. This provides correct
modeling for shear yielding but not for flexural yielding. In many structural wall panels, flexural
yielding may dominate. Three methods are available to incorporate flexural yielding into models
comprising wall panels:

5.3.1

FLEXURAL MODELING USING GAPS

Gap elements can be placed at the corners of the element. Where the wall is reinforced, truss
elements representing the reinforcing bars are used to connect the two sides of the gap. The
properties of these elements can be automatically generated from the specified wall thickness and
reinforcing ratio.
This method is the most elegant, and also the most general as it does not require the user to predefine the yield type. However, it has the disadvantage of producing a complex model, especially if
all wall panel elements are modeled this way. Usually, this full flexural model is best restricted to
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

34

main shear walls in the structure, which will have a major influence on results. Further details on the
yielding characteristics of this type of model are provided in Section 6.3 of this manual.

5.3.2

FLEXURAL MODELING USING EQUIVALENT SHEAR STRENGTH

The shear yield strength is modified so that the change in stiffness occurs when flexural yield
would occur in the wall panel. This is often the most efficient procedure when plane elements are
used to represent beams such as spandrel beams. For example, a panel element representing a
coupling beam may have the shear capacity set at the Mu/Lcl where Mu is the sum of end moment
capacities and Lcl is the clear length of the beam.
This method is preferred for elements such as the piers and coupling beams of pierced shear walls,
which have a large number of separate elements. The method is less complex than adding gap
elements but does require the user to define the yield type, which the program will use whether or
not it is correct.

5.3.3

FLEXURAL MODELING USING FRAME ELEMENTS

The wall may be modeled as flexural elements, using column elements for piers and beam elements
for spandrels and coupling beams. This method may be appropriate in situations where inelastic
shear deformations are unlikely to occur and flexural yield will dominate.
If you use this method it is not advisable to mix and match as the degrees of freedom for shear
and flexural elements are different. Therefore, it is not usually advisable to model piers as wall
elements and coupling beams as frame elements (or vice versa).

5.3.4

OTHER YIELDING MECHANISMS IN SHEAR WALLS

The shear yield envelope is constructed from the code defined shear strength provided by the
concrete and by the reinforcing. Other yielding mechanisms, such as construction joint slip or
compression toe crushing, are not included. Research is needed to determine whether some wall
types require a modification to the shear envelope to include these effects.

5.4

COUPLING BEAMS

Depending on the aspect ratio, coupling beams will act as either shear or flexural elements. If walls
are modeled as plane stress elements then coupling beams should be also modeled using this type of
element regardless of mode of behavior. This is because the plane stress elements do not have any
in-plane rotational stiffness. If a flexural element connects two wall panels then it will be effectively
pin ended. In fact, as rotations are not restrained about any axes there will be no torsional restraint
to the flexural member and ANSR-L will probably give a zero divide error.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

35

Methods of refining this type of model are under investigation. As discussed above, flexural
response may be represented be setting the shear capacity to Mu/Lcl where Mu is the sum of end
moment capacities and Lcl is the clear length of the beam.
If coupling beams are modeled as flexural, rather than plane stress, elements then be aware that
NZS3101 requires low effective stiffness factors to be used for some coupling beam aspect ratios.

5.5

WALL BOUNDARY ELEMENTS

Shear walls with boundary elements are modeled as a combination of plane elements plus column
elements at each end of the wall. Boundary elements are generally used in flexure dominated walls
with gap elements placed at each end of the wall to permit flexural yielding.
When column elements are specified at the location of the gap then the column element is continued
through the gap. With this procedure, gap opening requires tension yielding in the boundary
element. This is believed to provide a realistic representation of actual behavior but further research
is required. This is discussed further in Section 6.3.

5.6

FACE LOADED WALLS

The ANSR-L model is currently developed to evaluate only the in-plane response of walls. For outof-plane response, the degrees of freedom (dof) may be fixed, slaved to the dof above or restrained
by a vertical flexural element, depending on a user-selected option and the configuration of the wall
panels.
Elements are available in ANSR-L to model face loaded walls but have not been implemented
through the MODELA program. Reinforced walls can be modeled by column elements (reinforced
both faces) or a special centrally reinforced column element. These elements may be added to
MODELA if projects arise where face loading of reinforced walls is important.
Face loaded unreinforced walls require a special model comprising wall elements plus gaps at
potential crack locations. This model is likely to remain too complex to include in the main model
for some time. However, procedures are available to evaluate unreinforced masonry walls by using a
component model, which is analyzed using acceleration time histories at points of support. These
accelerations are saved from the main model.
5.7

RESTRAINTS

Each node of the model has 6 degrees of freedom, 3 translational plus 3 rotational. To avoid
singularities in the stiffness matrix, each of these degrees of freedom must be restrained. Restraint is
provided in one of two ways:
1.

The degree of freedom is fixed. This is done, for example, at the base of the structure.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

36

2.

Elements framing into the node restrain the degrees of freedom.

The restraint provided by the element depends on the element type. A flexural element will restrain
all 6 degrees of freedom at a node, unless the ends are pinned in which case the rotational degrees of
freedom will not be restrained.
Four node plane elements have an in-plane stiffness only. This type of element restrains translational
degrees of freedom parallel to the plane of the element only. It does not restrain the out-of-plane
degree of freedom or any of the 3 rotational degrees of freedom.
The aim of the MODELA program is to assemble the elements into an input file and assign
restraints based on element connectivity to ensure that there are no zero values on the diagonals of
the stiffness matrix. However, there will inevitably be some situations where the program will not
provide a complete set of constraints.
One known example where singularities can occur is where beam or column elements span between
plane stress elements only. The program assumes that the rotational degrees of freedom are
restrained by the flexural element. However, the panel elements do not restrain torsion in the
flexural element. The solution is generally to place a flexural edge element along the plane stress
element at right angles to the flexural element which does not have torsional restraint.

5.8

FLOORS

Floors may be modeled as rigid diaphragms, flexible elements or a combination of both.


Rigid diaphragms are modeled by a combination of direct slaving of degrees of freedom and explicit
rigid links to connect the nodes to the diaphragm center of mass. The ANSR-L program performs
the slaving at the element level and only flexural elements can be slaved. For other element types
(wall panels, gaps and trusses) the MODELA program inserts a rigid link element from the element
node to the center of mass location.
Flexible floors are modeled using the same type of 4 node plane stress elements used to model walls.
Material models available include concrete and timber, with degrading strength and stiffness as
appropriate. If you will be using flexible floors, try to lay out the column lines so as to form a
regular rectangular grid if possible. This makes it much simpler to specify floor panels. MODELA
can generate the panels in both directions from a series of nodes along the X and Y axes. Internal
nodes at the intersections will be generated as required. This requires that an equal number of nodes
be used along opposite sides of the building.
Full diaphragm modeling with finite elements complicates the model by a large degree. In cases
where only part of the diaphragm is flexible (such as a necked concrete floor) a more efficient
model will use two or more rigid diaphragms at a level connected by flexible floor portions.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

37

5.9

MASS AND WEIGHT IN BUILDING MODELS.

The ANSR-L program applies masses at any or all of the 6 degrees of freedom per node, three
translational plus three rotational. If the degree of freedom at which a mass is applied is fixed then
that mass is ignored.
The MODELA program assembles nodal masses from one or more of three possible sources
1.

Diaphragm masses and mass moments of inertia

2.

Mass arising from element self weight and/or distributed loads on beam elements.

3.

User specified translational or rotational mass at specific nodes.

Weight is applied as a static load case prior to the application of the earthquake loads. In ANSR-L,
all loads are nodal forces. As for mass, if the degree of freedom at which a force is applied is fixed
then that force is ignored. The MODELA program is set up to assemble weights from the same
three sources as for the masses, as listed above.
The mass distribution is important for earthquake evaluation because the inertia loads are dependent
on the masses. The importance of correct weights depends on the element types used in the analysis.
The correct gravity loads are important, for example, for columns where yield is a function of
moment and axial load, and also for gaps, where uplift will occur once earthquake axial loads exceed
gravity loads. Weights are less important for shear panels where yield is a function of shear stress.

5.9.1

DISTRIBUTION OF WEIGHT

The most common procedure for incorporating weights into the model is by using a combination of:
1.

Element self weight.

2.

Beam distributed loads

3.

Diaphragm weight distributed to column lines.

4.

Added nodal forces.

These are described in more detail in Sections 7.6.5 and 7.13 of this manual. The options used will
depend on the type of structure and the configuration. As most nodes in the structure will have
vertical degrees of freedom, the complications described below for masses due to diaphragm
modeling do not generally apply to weights.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

38

5.9.2

MASS DEGREES OF FREEDOM

The way in which ANSR-L models rigid diaphragms requires special consideration when assembling
the mass, especially for frame structures.
When rigid floors are used the master node at the center of mass of the floor is free in the X and Z
directions (these are the ANSR-L axes, X and Z are horizontal axes). Other nodes at the same level,
which are connected to the diaphragm and which are connected to beam/column elements only, have the X
and Z degrees of freedom fixed because these local degrees of freedom are slaved to the master
node. Nodes that are connected to other element types (wall panels, gaps) have the X and Z degrees
of freedom free. This is because slaving is not available for these other element types. A rigid beam
member is used to connect these nodes to the master node.
This may be important in assembling the model mass, as masses which are applied to fixed degrees
of freedom are ignored. When element mass densities are used to apply part of the mass, the
element mass is calculated and applied to nodes at the element ends. If the degrees of freedom at
these nodes are fixed then the element mass will be ignored.
The following sections recommend methods of ensuring both the weight and mass are correctly
distributed for different types of structure. Note that these are guidelines only, there will be many
exceptions where you will use different combinations. These guidelines assume that you want
element mass calculated by the program.

5.9.3

RECOMMENDED MASS OPTIONS

There are three mass options which control the way in which the mass is distributed, as listed in
Table 5-2. These options are on the Analysis Data worksheet (see Section 7.6.5).
Recommended values of these three options for various types of structure are listed in Table 5-3.
TABLE 5-2 MASS OPTIONS

1
2
3

Lump Element Mass to Diaphragm


Assign Missing Mass Horizontally
Assign Missing Mass Vertically

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

39

TABLE 5-3 RECOMMENDED MASS OPTIONS

TYPE OF STRUCTURE
Element Mass included in Diaphragm Mass
Structures with Flexible Floors (no
diaphragms)
Wall Structure with Rigid Floors
Frame Structure with Rigid Floors
Combined Wall Frame Structure with Rigid
Floors

5.9.4

Mass Option
1
2
3
9
9
9
9

9
9
9

ECCENTRICITY OF MASS

The MODELA procedure allows for the mass locations to be adjusted to incorporate accidental
eccentricity as required by building codes. This is discussed in Sections 7.5.6, 7.7.2 and 7.11.
For models with rigid diaphragms where all mass is assigned to the floor master nodes then the
procedure for applying accidental eccentricity is straight forward as the coordinates of the center of
mass are simply moved by the specified eccentricity.
When masses are distributed (element mass or user specified mass) then the procedure to incorporate
eccentricity is more complex. The method used by MODELA is to calculate factors to apply to
masses at positive and negative locations relative to the calculated center of mass. These factors are
calculated so that the centroid will be at the specified location.
The procedure is believed to be reliable but in some cases you may wish to check the calculations.
To assist in this, calculations for the final eccentricity calculated are written to a disk file named
T$0002. This text file can be opened using programs such as WordPad. Table 5-4 shows the end
portion of the file (the initial portions of the file contain intermediate calculations, which are used to
debug the procedure if results appear incorrect).
The adjustment factors listed in Table 5-4 are the factors which are applied to masses which are
respectively on the positive and negative sides of the calculated center of mass. The mass and first
moment of mass is then listed based on the adjusted center of mass. The final section of the table
lists the original center of mass (X0, the new center of mass (X1) the target eccentricity (EC) and the
actual eccentricity (E0). The last two columns should be equal if the procedure has worked correctly.
The calculations in Table 5-4 exclude the mass specified at rigid diaphragms as the coordinates of
these locations are moved directly.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

40

TABLE 5-4 CALCULATION OF MASS ECCENTRICITIES


ADJUSTMENT
LEVEL
1
2
3
4
5
6

FACTORS
Fx+
0.8
0.8
0.8
0.8
0.8
0.8

Fx1.2
1.2
1.2
1.2
1.2
1.2

Fz+
0.8
0.8
0.8
0.8
0.8
0.8

Fz1.2
1.2
1.2
1.2
1.2
1.2

MASS AND ECCENTRICITY AFTER ADJUSTMENTS


LEVEL
Mx
MXx
Xcg
Mz
1
39.0
226.7
5.8
39.0
2
75.7
431.7
5.7
75.7
3
64.9
368.6
5.7
64.9
4
64.9
368.6
5.7
64.9
5
64.9
368.6
5.7
64.9
6
32.6
184.9
5.7
32.6

MZz
-185.8
-363.7
-311.5
-311.5
-311.5
-157.6

Zcg
-4.8
-4.8
-4.8
-4.8
-4.8
-4.8

CENTRE OF MASS AFTER ADJUSTMENTS


EC IS NEW CALCULATED VALUE, E0 IS TARGET VALUE
LEVEL DIR
X0
X1
EC
E0
1 X
7.016
5.816
-1.200
-1.200
1 Z
-3.967
-4.767
-0.800
-0.800
2 X
6.899
5.699
-1.200
-1.200
2 Z
-4.001
-4.801
-0.800
-0.800
3 X
6.879
5.679
-1.200
-1.200
3 Z
-4.000
-4.800
-0.800
-0.800
4 X
6.879
5.679
-1.200
-1.200
4 Z
-4.000
-4.800
-0.800
-0.800
5 X
6.879
5.679
-1.200
-1.200
5 Z
-4.000
-4.800
-0.800
-0.800
6 X
6.875
5.675
-1.200
-1.200
6 Z
-4.037
-4.837
-0.800
-0.800

5.10 BEAM LOADS

The ANSR-L program does not provide for the application of element loads directly. All loads are
applied to degrees of freedom, which are at nodal points. Beam loads as input (Section 7.10.4) are
converted to equivalent nodal actions by the assembly program MODELA. Beam loads may be
specified for one of two reasons:
1.

To initialize beam actions due to gravity load. The gravity moments are necessary to ensure
that yield under earthquake loads occurs at the correct time. Gravity shears add to seismic
shears.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

41

2.

To provide a mechanism to apply gravity loads to vertical elements (columns and walls) when
the yield capacity of these elements is a function of axial load.

Options are available in the MODELA spreadsheet to use the beam loads to assemble both the mass
and weight of the structure (see Section 7 for details).
The procedure used by MODELA to apply beam loads is to calculate the fixed end moments and
shear forces arising from the specified uniform or varying load. These fixed end actions are then
applied as initial forces to the beam elements in the ANSR-L model plus nodal loads at each node. If
the option to include the UDL masses is selected the shear force at each end of the beam is applied
as a nodal mass to the node at that end of the beam.
The combination of fixed end moments plus nodal actions arising from these actions should provide
the correct gravity load forces in beams. This requires that a static analysis for gravity loads be
performed prior to the dynamic analysis. This is the default.
A couple of other points relating to beam loads:

It is possible to adjust the beam yield moments to account for gravity load effects. This will
ensure that positive and negative hinging occurs at the correct time.

The beam distributed load code can be set negative in which base the beam loads are used to
define nodal weights/masses but not used to initialize beam end actions (see Section 7.10.4).

5.11 FOUNDATION UPLIFT

Foundation uplift is modeled by using gap elements, which are linearly elastic in compression but
have zero stiffness in tension. The gap elements may be placed under every column line at the
foundation level or under discrete columns. Gaps may be placed at more than one level for a
stepped foundation.
The compression stiffness should be based on the modulus of subgrade reaction and the area of the
foundation element in contact with the ground. The stiffness is invariably approximate. If
foundation uplift is an important component of seismic response then a number of analyses with
different foundation stiffness may be required to assess the sensitivity of response.
The default ANSR-L gap element as used by MODELA procedures is restricted to linearly elastic
behavior in compression and the element is fixed against translation. However, the element
formulation in ANSR-L is more general in that it can be tri-linear in compression to model soil
yielding, can slide at a specified coefficient of friction and can have a preset gap opening at the start
of the analysis. These features can be incorporated into the model using a fully specified gap.
The gap used to model uplift may have tension properties specified, for example where the building
is on piles which can resist tension or where a heavy basement is not included in the model. In this
case, a truss element will cross the gap and be anchored to the base.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

42

5.12 BASE ISOLATORS AND DAMPERS

The MODELA program has been used for a limited number of base-isolated buildings and buildings
with in-structure dampers. Development of features to model this type of structure has been limited.
Procedures are in place to model lead-rubber bearings (LRBs), high damping rubber bearings (HDR)
and steel cantilevers for base isolation. Springs and dashpots can be used in combination for instructure dampers.
The ANSR-L program has the capabilities of modeling a wide range of isolators, energy dissipators,
sliders and dampers. These include LRBs, steel cantilevers, viscous dampers, high damping rubber
bearings and Teflon sliding bearings. If projects require any of these types of device then they can be
added to the MODELA procedure.

5.13 MULTIPLE BUILDINGS

Multiple buildings can be modeled as a single ANSR-L model. The buildings can be completely
independent but in that case it would generally be more efficient to model each building separately.
The usual reason for modeling multiple buildings in a single model is if there are potential points of
impact between the buildings.
Buildings can be connected by gap elements parallel to any of the global axes. The buildings can
either abut each other (zero initial gap opening) or be a specified distance apart (the gap has a
positive initial opening). The gap element will only transfer a force if the combined displacements of
the buildings toward each other exceed the initial gap opening.
Impact between buildings causes large forces and a small time step is often required for this type of
analysis. Time steps as small as 0.001 seconds will make for a very slow analysis.

5.14 STRENGTHENING MEASURES

The usual procedure for evaluating existing buildings is to identify weaknesses in the seismic
resistance of the building, conceptualize strengthening measures and then evaluate the response to
assess the effectiveness of the strengthening measures. The evaluation of the strengthened building
is performed by modifying the model of the as-is building.

5.14.1 NEW LATERAL LOAD RESISTING ELEMENTS

Lateral load resisting elements are incorporated in the model using the same principles as for the
existing elements. There is no prohibition on multiple elements occupying the same space so
strengthening elements can be added in parallel with existing elements, that is, connected to the same
column line numbers and levels. If walls are to be strengthened by the addition of a new wall, for
example shotcrete or a composite overlay, then a panel element with the concrete or composite
properties can be added in the same location as the wall to be strengthened.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

43

Elements added in parallel to existing elements will increase the problem size slightly but do not
increase the number of degrees of freedom or bandwidth and so are relatively efficient.

5.14.2 STRENGTHENING EXISTING ELEMENTS

Apart from traditional materials (concrete, steel), structural composites (glass, carbon or other fibers
in an epoxy matrix) have a number of applications in the earthquake strengthening of buildings.
Specific uses to date have included:
1.

Confinement for flexural elements.

2.

Shear enhancement for flexural elements.

3.

Strengthening of unreinforced masonry walls.

4.

Strengthening of reinforced concrete walls.

These uses have been implemented to some extent in the HCG non-linear analysis procedure but
there are a number of issues relating to the modeling and evaluation which are not yet finalized.
HCG are active in this field and developments should be rapid. See also Section 6.8.

5.15 GEOMETRIC STIFFNESS (P-) EFFECTS

The ANSR-L program has large displacement capabilities for some element types but this is not
generally necessary for building type structures. Rather, a geometric stiffness is included in the
column elements. The geometric stiffness is incorporated in two ways:
[1]

Adding a geometric stiffness to the element elastic or elasto-plastic stiffness. This is


changed each time the elasto-plastic stiffness changes, using the current axial force.

[2]

The resisting forces are modified to add shear forces to the nodes, based on the element
axial force and its chord rotation (/H).

The flag to include these effects is set automatically for the column elements when the HCG
evaluation procedure is used.
Experience has shown that the geometric stiffness effects are generally much less marked than would
be expected based on requirements in codes such as NZS4203. This appears to be a result of a
number of effects:

Code requirements are generally based on evaluation of single degree of freedom elastic-perfectly
plastic systems. The yielded slope is zero and becomes negative once geometric stiffness is
added. In actual structure, plastic hinges form incrementally and the overall force-deflection
curve has a positive slope even if zero strain hardening is specified for individual elements.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

44

For most actual structures, a mechanism forms for only very small intervals of time, if at all.

Further research is planned on these effects.


In some situations P- effects will be important. Figure 5-3 shows one such example, taken from
the pushover analysis of a bent of the Shell Gully Motorway bridge. The two columns in the bent
each have an axial load of 4058 KN and are 13.560 m high.
The columns were modeled with % strain hardening and the sum of column shear forces versus
displacement is as shown in the upper curve in Figure 5-3. This displays the bi-linear response
expected of the element.
However, when the pushover force is plotted it exhibits a negative stiffness and diverges from the
column shear force. Without P- effects the two curves are identical. With P- the geometric
stiffness provides a shear force of (2 x 4058) x 0.8 / 13.560 = 482 KN at a displacement of 0.8 m.
As Figure 5-3 shows, this is approximately the difference in shear forces between the two curves.
If you need to quantify geometric stiffness effects the displacement pushover is often the best way to
do this. Set the element output codes to obtain the time history of maximum base shear. When this
is plotted against the pushover force, the difference will be due to P- effects.

FIGURE 5-3 GEOMETRIC STIFFNESS EFFECTS

1600
1400

FORCE (KN)

1200
1000
800
PUSHOVER FORCE
COLUMN SHEAR FORCE

600
400
200
0
0

100

200

300

400

500

600

700

800

DISPLACEMENT (mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

45

6 MATERIAL AND ELEMENT MODELS

The basis for a non-linear analysis is simply a solution of Newtons Law, force equals mass times
acceleration. This has a well-founded theoretical basis with an exact solution. Apart from the nondeterministic nature of the earthquake motions, it is the modeling of the materials, and the assembly
of these materials into finite elements, that add approximations to the solution. The more accurate
the representation of the elements the more reliable the solution.
Material non-linearity is concerned with changes to the properties of elements with varying levels of
stress. This relates to regions where stress is no longer proportional to strain. For most materials
used in structural engineering, increased stresses, especially when applied cyclically, degrade both the
stiffness and the strength of elements. This section describes the manner in which this degradation is
implemented for the ANSR-L program. A description of the material models available is provided.
The models are generally based on experimental evidence of the performance of structural
components.

6.1

DEGRADING STRENGTH & STIFFNESS OF FLEXURAL ELEMENTS

Concrete frame members loaded beyond the cracking stage will reduce in stiffness. Once loading is
such that the reinforcing steel yields the stiffness reduction becomes more marked. As the element is
cycled at deformation levels beyond yield the strength is also likely to reduce, with the amount of
degradation depending on factors such as the amount of confining steel provided. Both the stiffness
and the strength degradation are non-recoverable.
6.1.1

FORMULATION OF DEGRADATION

The formulation for the Degrading Column Element is based on the general 3D Beam-Column
element used by the ANSR-II program. The original element uses a bilinear (or tri-linear) hysteresis
relationship to relate force and deformation. Two subroutines have been added to this element to
modify its hysteretic response in bending about the two local axes. These give the user the ability to
represent both strength and stiffness degradation. The response of the element under torque and
axial load remains governed by a bilinear relationship.
Strength degradation has been incorporated by way of a user-specified strength envelope. This
envelope allows the user to express the moment strength of the member at each end as a function of
ductility. It is defined in MODELA using the parameters shown in Figure 6-1.
The program decreases the bending strength of the member about each local axis, and at each end
independently. The strength is degraded as a function of the average enveloped maximum positive
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

46

and negative rotations about a given axis, based on rotational ductility parameters specified by the
user.
The ability to model stiffness degradation has been included using a model similar to the Takeda
degrading stiffness relationship. This is illustrated in Figure 6-2.
FIGURE 6-1 PARAMETERS FOR SPECIFYING STRENGTH DEGRADATION

Multiplier on Yield Force

F1

F2
F3

M2

M3
Rotation Ductility

FIGURE 6-2 DEGRADING STIFFNESS RELATIONSHIP

F
Fy

kp

ku
ko

kl

-F y

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

47

Prior to first yield, the initial stiffness, ko governs the force deformation relationship for the element.
As for the bilinear rule, once yield occurs, the element follows the plastic stiffness, kp. This is usually
expressed as a fraction of the initial elastic stiffness.
After first yield, the unloading stiffness, ku, is reduced as a function of maximum rotational ductility
achieved in each direction. This can be expressed as:

r+ r
k u = k o max max
2ry

+
Where rmax
, rmax
and ry are the maximum negative, maximum positive and yield rotations about a

given axis respectively. The value for is specified in MODELA for each degradation type.
When =0, no degradation in stiffness occurs upon unloading from a yield excursion. The higher
the value of , the greater the degradation. A value of 0.4 is recommended for reinforced concrete
members.
The stiffness is again modified when reloading begins to occur. Firstly, when reloading from a
positive yield excursion:

ry
ku = ko

ry + (rmax ry ) + rt

And when reloading from a negative yield excursion:

ry
ku = ko

+
ry + (rmax ry ) + rt
In these equations, rt is the rotation at the time when the moment about the given axis changes sign.
That is, at the transition between unloading and reloading. The value for is specified in MODELA
for each degradation type. When = 0, the reloading branch approaches the point (ry, Fy). With
increasing values of , the reloading stiffness is softened. When =1 is specified, a point (rmax, Fy) is
approached.

6.1.2

IMPLEMENTATION OF DEGRADATION COEFFICIENTS

For implementation, the definition of the strength and stiffness degradation parameters differs
between beam and column elements:
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

48

For beams, the strength and stiffness degradation is defined separately for positive and negative
bending.
An averaging parameter, IAVG, defines whether degradation in each direction is
independent (the default, IAVG = 0) or is based on the average of the positive and negative plastic
rotations.

For columns, the strength and stiffness degradation is defined separately for bending about each
axis, that is, for My and Mz bending. The degradation is the same for positive and negative
bending, which is a requirement as the interaction diagram for a column element is symmetrical.

Examples of the effects of varying degradation coefficients are provided in Figures 6-3 to 6-9. Table
6-1 summarizes the value of each parameter used to develop these plots.

TABLE 6-1 EXAMPLE DEGRADATION COEFFICIENTS

Figure 6-3
Figure 6-4
Figure 6-5
Figure 6-6
Figure 6-7
Figure 6-8
Figure 6-9

1
0.0
0.5
1.0
0.5
0.5
0.5
0.0

2
0.0
0.5
1.0
0.5
0.5
0.5
0.5

3
0.0
0.5
1.0
0.5
0.5
0.5
0.0

4
0.0
0.5
1.0
0.5
0.5
0.5
1.0

f1
1.0
1.0
1.0
1.0
1.0
1.0
1.0

f2
1.0
1.0
1.0
0.5
0.5
0.2
0.1

f3
1.0
1.0
1.0
0.25
0.25
0.1
0.1

m1

m2

m3

Iavg

2
2
5
5

5
5
5
10

10
10
10
100

1
0
1
0

Research is continuing as to appropriate values to use. A few guidelines pending this are:

For modern frames which are well confined, use either no stiffness degradation (Figure 6-3) or
moderate stiffness degradation (Figure 6-4) with no, or very slow, strength degradation.

For poorly confined frames, use moderate to high stiffness degradation and also moderate to
high strength degradation (Figures 6-7 and 6-8).

Use judgment between these two extremes.

For bars which do not comply with lap requirements in the hinge zone, use rapid strength
degradation.

For bottom beam bars which are not fully developed, use rapid strength degradation for positive
bending (Figure 6-9).

It has not been determined yet whether it is better to use averaged or independent degradation
for beams (Figures 6-6 and 6-7). For now, it is probably best to set this flag to 1 so that the
degradation is based on the average rotation.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

49

FIGURE 6-3 NO STIFFNESS DEGRADATION ; NO STRENGTH DEGRADATION

300

BENDING MOMENT (KN-m)

200
100
0
-0.04

-0.03

-0.02

-0.01

0.01

0.02

0.03

0.04

-100
-200
-300
-400
-500
TOTAL ROTATION (radians)

FIGURE 6-4 MODERATE STIFFNESS DEGRADATION (=0.5) NO STRENGTH DEGRADATION


300

BENDING MOMENT (KN-m)

200
100
0
-0.04

-0.03

-0.02

-0.01

0.01

0.02

0.03

0.04

-100
-200
-300
-400
-500
TOTAL ROTATION (radians)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

50

FIGURE 6-5 EXTREME STIFFNESS DEGRADATION (=1.0) NO STRENGTH DEGRADATION

300

BENDING MOMENT (KN-m)

200
100
0
-0.04

-0.03

-0.02

-0.01

0.01

0.02

0.03

0.04

-100
-200
-300
-400
-500
TOTAL ROTATION (radians)

FIGURE 6-6 MODERATE STIFFNESS DEGRADATION (=0.5) WITH STRENGTH DEGRADATION


(AVERAGED POSITIVE AND NEGATIVE)

300

BENDING MOMENT (KN-m)

200
100
0
-0.04

-0.03

-0.02

-0.01

0.01

0.02

0.03

0.04

-100
-200
-300
-400
-500
TOTAL ROTATION (radians)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

51

FIGURE 6-7 MODERATE STIFFNESS DEGRADATION (=0.5) WITH STRENGTH DEGRADATION


(INDEPENDENT POSITIVE AND NEGATIVE)

300

BENDING MOMENT (KN-m)

200
100
0
-0.04

-0.03

-0.02

-0.01

0.01

0.02

0.03

0.04

-100
-200
-300
-400
-500
TOTAL ROTATION (radians)

FIGURE 6-8 MODERATE STIFFNESS DEGRADATION (=0.5) WITH RAPID DEGRADATION

300

BENDING MOMENT (KN-m)

200
100
0
-0.04

-0.03

-0.02

-0.01

0.01

0.02

0.03

0.04

-100
-200
-300
-400
-500
TOTAL ROTATION (radians)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

52

FIGURE 6-9 NON-SYMMETRICAL STIFFNESS DEGRADATION AND STRENGTH DEGRADATION

300

BENDING MOMENT (KN-m)

200
100
0
-0.04

-0.03

-0.02

-0.01

0.01

0.02

0.03

0.04

-100
-200
-300
-400
-500
TOTAL ROTATION (radians)

6.2

DEGRADING STRENGTH AND STIFFNESS OF PANEL SECTIONS

Panel sections are the 3 or 4 node plane stress elements used to model shear walls and floors. The
strength and stiffness of this type of element is defined by a shear stress-strain envelope curve, of the
form shown in Figure 6-10. The envelope curve is defined by up to 8 separate straight line segments.
Below the critical strain, c, the shear strain at each step defines the current stiffness of the element.
At transition points between the straight line segments, the stiffness is reformed using the next
tangent shear modulus until the direction of loading is reversed. On reversal, the stiffness is based
on the secant stiffness to unload the element through the origin of the hysteresis plane. This
unloading may be modified using the and factors described below.
Beyond the critical strain, c, a transition line is defined which permits strength degradation. When
the critical strain is exceeded (that is, the transition line is reached) the stiffness is based on the shear
modulus corresponding to the final slope of the envelope curve. This provides the strength
degradation characteristics. The parameter Dkode may be used to control the maximum extent of

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

53

strength degradation. The strength will not degrade below a level equal to Dkodemax where max is the
peak shear stress.
The hysteresis is modified for brittle materials, such as unreinforced masonry, so that the strength is
not maintained when the transition strain is reached but rather the strength will reduce with increased
strain in the loading direction.

FIGURE 6-10 SHEAR STRESS-STRAIN ENVELOPE CURVE

G4

max

G3

Shear Stress

G2

G1
maxDKode

Shear Strain

Xkgc

The default unloading stiffness is the secant stiffness through the origin of the stress-strain curve.
This may be modified by assigning non-unitary values to two parameters, and . The unloading
stiffness is divided into two segments. The first parameter, , defines the strain ordinate and the
second parameter, , the stress ordinate of the first unloading segment (see Figure 6-11).
For unreinforced masonry, secant unloading stiffness through the origin is appropriate and values of
= 1 and = 1 are used. For concrete, values of = 0.5 and = 1.0 give an unloading function
which is in agreement with test hysteresis curves.
The following sections, describing models for specific materials, discuss appropriate values of these
parameters and give examples of the hysteresis loops generated by various values.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

54

FIGURE 6-11 UNLOADING CHARACTERISTICS

=1=1

Strength Envelope
SHEAR STRESS

SHEAR STRAIN

6.3

6.3.1

REINFORCED CONCRETE WALL MODELING

SHEAR STRENGTH ENVELOPE

A single model for the loading portion is usually used for all reinforced concrete panels. Variations
are accounted for by varying the degradation coefficients. Figures 6-12, 6-13 and 6-14 are examples
of experimental hysteresis loops for reinforced concrete shear wall panels which have been used to
develop strength parameters shown in Figure 6-15.
For the development of the analysis hysteresis, it has been assumed that the initial loading portion of
the envelope stress-strain curve (Figure 6-15) is independent of whether the wall is dominated by
shear or flexure:
1. The initial stiffness is determined by the shear modulus of uncracked concrete. This stiffness
applies until the concrete shear strength, vc, is reached.
2. A second (cracked) stiffness is defined by a stress increment equal to the shear strength provided
by the reinforcing, vs, reached at a total strain of 0.0045 (drift angle of 1/222).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

55

3. A third (strain hardening) stiffness is defined by assuming that the amount of strain hardening is
equal to 0.25 vs. This stress occurs at a strain of 0.0075 (drift angle of 1/133).
Degradation coefficients are then used to define the strength degradation portion of the response,
depending on the type of wall response.
Input required for this type of element is the concrete shear strength, vc, the steel shear strength, vs,
and the degradation coefficients. The strength values are calculated according to code requirements
using the appropriate code. Generally, approximations will be made to allow for variables such as
axial load, which can affect the shear strength. The HCG process makes no allowance for axial load
on concrete wall panels except as it affects flexural strength (see following sections).

FIGURE 6-12 HYSTERESIS OF SHEAR DOMINATED WALL

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

56

FIGURE 6-13 HYSTERESIS OF FLEXURE DOMINATED WALL

FIGURE 6-14 HYSTERESIS OF SQUAT WALL

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

57

FIGURE 6-15 ENVELOPE FOR CONCRETE WALLS

2000

Concrete + Overstrength Steel vc+1.25vs


Concrete + steel, vc+vs

1500

SHEAR STRESS (KPa)

1000
Residual Strength
= Dkode x Vmax

Concrete shear strength, vc


500

G Concrete

0
-0.03

-0.02

-0.01

0.01

-500
-1000

0.02

0.03

Drift 1/50
Drift 1/133
Drift 1/222

-1500
-2000
SHEAR STRAIN

6.3.2

SHEAR DEGRADATION COEFFICIENTS

The strength parameters vc and vs define the loading portion of the strength hysteresis. The
degradation parameters define the manner in which the strength and stiffness degrade under
reversing loads.
If no degradation coefficients are specified then the type of hysteresis shown in Figure 6-16 will be
used. Under loading, the stress reaches a peak value and then reduces with increasing loading. This
is appropriate for unreinforced panels. Reinforced panels maintain the load for loading continuing in
the same direction and so coefficients should always be specified.
Pending further research, it is recommended that the stiffness degradation coefficient alpha be set at
0.5 and beta be set at 0.0. This provides a reasonable unloading match to the experimental hysteresis
curves.
The strength degradation coefficient xkk determines how rapidly the strength decays and the
coefficient Dkode the maximum amount of strength degradation. It is recommended that a value of
xkk = 2.67 be used unless the wall is particularly brittle. A value of Dkode = 0.4 is recommended for
shear dominated walls and 0.75 for flexure dominated walls.
Figures 6-17 and 6-18 compare these two recommended values for shear dominated and flexure
dominated walls respectively. The only difference in these two hysteresis definitions is that Dkode is
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

58

increased from 0.4 to 0.75 as the mode changes from shear to flexure. This effects the total amount
of degradation from the peak force.
For walls which are known to be heavily shear dominated and of a very squat aspect ratio then lower
values of Dkode may be appropriate. For example, in Figure 6-14 the peak strength of 35 tons reduces
to 10 tons so an appropriate values of Dkode would be 10/35 = 0.28.
As the values of Dkode reduces, the influence of the degradation slope coefficient xkk becomes
relatively more important. At higher values of Dkode the rate of strength reduction is dominated
mainly by the asymptote to the residual strength.
Figures 6-19 and 6-20 show the effect of reducing the value of xkk from 2.67 (the usual value) to 1.0
(the minimum value possible). These are for a very low residual strength, Dkode = 0.10, to make the
differences more noticeable. Even with these extreme values, the importance of the variable xkk is
relatively small.

FIGURE 6-16 HYSTERESIS : NO DEGRADATION COEFFICIENTS

2500
2000

SHEAR STRESS (KPa)

1500
1000
500

-0.10

-0.08

-0.06

-0.04

0
0.00

-0.02

0.02

0.04

0.06

0.08

0.10

-500
-1000
-1500
-2000
SHEAR STRAIN (mm/mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

59

FIGURE 6-17 =0.5 =0.0 XKK = 2.67 DKODE=0.4

2500
2000
1500

SHEAR STRESS (KPa)

1000
500

-0.10

-0.08

-0.06

-0.04

0
0.00
-500

-0.02

0.02

0.04

0.06

0.08

0.10

0.08

0.10

-1000
-1500
-2000
-2500
SHEAR STRAIN (mm/mm)

FIGURE 6-18 =0.5 =0.0 XKK = 2.67 DKODE=0.75

2500
2000
1500

SHEAR STRESS (KPa)

1000
500

-0.10

-0.08

-0.06

-0.04

0
0.00
-500

-0.02

0.02

0.04

0.06

-1000
-1500
-2000
-2500
SHEAR STRAIN (mm/mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

60

FIGURE 6-19 =0.5 =0.0 XKK = 2.67 DKODE=0.10

2500
2000
1500

SHEAR STRESS (KPa)

1000
500

-0.10

-0.08

-0.06

-0.04

0
0.00
-500

-0.02

0.02

0.04

0.06

0.08

0.10

0.08

0.10

-1000
-1500
-2000
-2500
SHEAR STRAIN (mm/mm)

FIGURE 6-20 =0.5 =0.0 XKK = 1.00 DKODE=0.10

2500
2000
1500

SHEAR STRESS (KPa)

1000
500

-0.10

-0.08

-0.06

-0.04

0
0.00
-500

-0.02

0.02

0.04

0.06

-1000
-1500
-2000
-2500
SHEAR STRAIN (mm/mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

61

6.3.3

FLEXURE DOMINATED SHEAR WALLS

Chapter 5 described three methods for modeling the flexural component of shear walls. Two of these
methods (adjusting the shear strength to reflect moment strength and modeling the component as a
frame element) are suitable for modeling local components such as wall piers or coupling beams in
pierced shear walls. When cantilever or coupled walls form a major portion of the lateral load
resisting system then the more general modeling procedure, using gap elements to model flexural
yield, should be used.
The flexural yield model procedure requires that the location of the plastic hinge be pre-defined.
This is not usually a problem as the method is most often used for cantilever or coupled walls where
yielding occurs at the base of the wall. The procedure to model a flexural hinge at the base of a wall
is:

Model the wall with shear elements as described in Chapter 5.

If the wall has boundary elements, model these as column elements at each end of the wall.

Specify gap elements at all column lines along the horizontal plane where the plastic hinge is to
occur.

Define the reinforcing area and concrete area at each gap location. This step can be automated
by defining the reinforcing content of the wall panels. In this case, MODELA calculates the area
of wall tributary to each column line. This gives the area of concrete directly. The area of
reinforcing at each gap element is obtained as the specified reinforcing content times the
concrete area.

The MODELA program will insert an additional elevation 50 mm above the level at which the gaps
are specified. The added node at the top of the gap will be constrained to have equal displacements
to the node below, to which it is connected, for all degrees of freedom except the vertical.
Therefore, the gap can open vertically only. The gap element used is actually a gap- friction element
and so could be used to model other failure modes such as sliding shear. However, development
work is required before this capability of the element can be used.
At each intersection of a panel node and the gap plane, a pair of elements will be inserted. The first
is a gap element, which can resist compression loads only. The stiffness is based on the concrete area
and E value for concrete. A truss element is added in parallel, with stiffness based on the reinforcing
area and the E value for steel. The truss element will be bi-linear (or elastic perfectly plastic if a zero
strain hardening ratio is specified for steel) with equal yield stress in tension and compression.
As the gap element provides a much higher compression strength than tension strength, the forcedeflection response of the reinforcing bar element is asymmetric, as shown in Figure 6-21.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

62

FIGURE 6-21 REINFORCING BAR UNDER INCREASING CYCLIC DISPLACEMENT

80.00

REINFORCING BAR FORCE (KN)

60.00
40.00
20.00

-0.50

0.00
0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

-20.00
-40.00
-60.00
-80.00
REINFORCING BAR EXTENSION (mm)

If a wall boundary member (column element) crosses the wall base plastic hinge, the column
representing the boundary element also crosses the gap. The column is modeled as a truss with
stiffness corresponding to that of the column but the strength in both tension and compression is set
equal to the tension strength specified for the column element. The compression area of the column
is added to that of the gap element.
The formulation described in this section is based on an engineering mechanics derivation of the
manner in which the flexural hinge forms in a reinforced concrete shear wall. Default parameters
are set which are believed to provide a reasonable representation. The discussion in the following
sub-sections describes the effects of some of these parameters. However, further work remains to
be done in refining the model for shear walls, especially in the correlation of hysteresis shapes with
experimental data.
6.3.3.1

Effect of Grid Refinement

Figures 6-22 to 6-25 illustrate the effect of grid refinement on the force-displacement properties of a
20 m high x 4 m wide shear wall, assumed 6 stories high. The wall is a straight segment without
boundary elements, with a 200 mm wall thickness and a reinforcing ratio of 0.4%. Automatic
generation of areas was used for the plastic hinge model at the base of the wall. The wall has an
axial load of 380 KN, corresponding to a stress of 0.50 MPa.
The grid refinement used to generate each of these four plots is as listed in Table 6-2. The first plot
assumes that each level is subdivided into three elements, the last three assume that each story is
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

63

represented by a single panel. The last three figures illustrate the effect of reducing the horizontal
grid refinement.
The plots are generated by applying an increasing displacement at the top of the model and recording
the force required at each step (displacement pushover).
TABLE 6-2 VARIATIONS OF GRID REFINEMENT

Figure 6-22
Figure 6-23
Figure 6-24
Figure 6-25

Number of
Horizontal
Elements
12
12
6
1

Number of
Vertical
Elements
19
6
6
6

The theoretical ultimate force was calculated by dividing the ultimate moment at the base of the wall
by the wall height. The ultimate moment is as defined by the concrete code, calculated by the
CONCDES spreadsheet using the same material strengths as used to develop the model. The
capacity reduction factor, , was 1.0 for this calculation.
The first three plots are quite similar, with the effect of the increased grid refinement mainly being to
give a more rounded form to the reloading portion of the curve. When the wall is represented as a
single panel along its length (Figure 6-25) all the reinforcing and concrete area is concentrated into
single elements at each end of the wall and the loading and reloading curves form straight lines.
The differences between these plots suggest that refined grid spacing is much less important for a
flexural wall than for a wall which responds primarily in shear. Even a single element across the wall
width, and a single panel for each story height, captures the main characteristics of response.
It is of interest to note that the use of multiple elements across the wall captures the gradual
softening of the wall under loading and unloading even though the individual elements have simple
linear elastic and elastic-perfectly plastic yield functions. The gradual softening appears to be a
function of successive reinforcing bars yielding as the top displacement increases, a geometric effect,
rather than a complex yield function of the materials.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

64

FIGURE 6-22 12 HORIZONTAL X 19 VERTICAL


150.00
Theoretical Capacity

SHEAR FORCE (KN)

100.00

50.00

-40.00

-30.00

-20.00

-10.00

0.00
0.00

10.00

20.00

30.00

40.00

30.00

40.00

-50.00

-100.00

Theoretical Capacity

-150.00
DISPLACEMENT (mm)

FIGURE 6-23 12 HORIZONTAL X 6 VERTICAL

150.00
Theoretical Capacity

SHEAR FORCE (KN)

100.00

50.00

-40.00

-30.00

-20.00

-10.00

0.00
0.00

10.00

20.00

-50.00

-100.00

Theoretical Capacity

-150.00
DISPLACEMENT (mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

65

FIGURE 6-24

6 HORIZONTAL X 6 VERTICAL
150.00

Theoretical Capacity

SHEAR FORCE (KN)

100.00

50.00

-40.00

-30.00

-20.00

-10.00

0.00
0.00

10.00

20.00

30.00

40.00

30.00

40.00

-50.00

-100.00

Theoretical Capacity

-150.00
DISPLACEMENT (mm)

FIGURE 6-25 1 HORIZONTAL X 6 VERTICAL


150.00
Theoretical Capacity

SHEAR FORCE (KN)

100.00

50.00

-40.00

-30.00

-20.00

-10.00

0.00
0.00

10.00

20.00

-50.00

-100.00

Theoretical Capacity

-150.00
DISPLACEMENT (mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

66

6.3.3.2

Effect of Axial Load

The shape of the hysteresis loop produced by a flexural hinge is heavily influenced by the level of
axial load on the wall. Compare Figure 6-22 (moderate axial load) with a heavily loaded wall (Figure
6-26) and a wall with no load (Figure 6-27). The area of the loop (and by implication the damping)
reduces with increased axial load. This effect requires further investigation.
FIGURE 6-26 1500 KN AXIAL LOAD (1.9 MPA)
250.00

Theoretical Capacity

200.00
150.00

SHEAR FORCE (KN)

100.00
50.00

-40.00

-30.00

-20.00

-10.00

0.00
0.00
-50.00

10.00

20.00

30.00

40.00

20.00

30.00

40.00

-100.00
-150.00
-200.00
-250.00
DISPLACEMENT (mm)

FIGURE 6-27 ZERO AXIAL LOAD


100.00
Theoretical Capacity
80.00
60.00

SHEAR FORCE (KN)

40.00
20.00

-40.00

-30.00

-20.00

-10.00

0.00
0.00
-20.00

10.00

-40.00
-60.00
-80.00
-100.00
DISPLACEMENT (mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

67

6.3.3.3

Walls with Boundary Elements

Walls with boundary elements provide similar hysteresis shapes to those for straight walls but there is
less waisting and so the area of the hysteresis loop is higher. Figures 6-28 and 6-29 are two
examples, the first where the web shear strength is set very high and the second where the web
cracks at a moderate shear strength. Web cracking seems to reduce the strength so as to be closer to
the theoretical value. When web cracking occurs the assumption that plane sections remain plane at
the plastic hinge no longer applies and the gap openings will no longer be proportional to the
elements distance from the neutral axis.
FIGURE 6-28 SHEAR PANELS IN WEB ELASTIC; GAP EFFECTIVE LENGTH 50 MM
400.00
300.00
Theoretical Capacity

SHEAR FORCE (KN)

200.00
100.00

-50.00

-40.00

-30.00

-20.00

-10.00

0.00
0.00

10.00

20.00

30.00

40.00

50.00

-100.00
-200.00
-300.00
-400.00
DISPLACEMENT (mm)

FIGURE 6-29 SHEAR PANELS IN WEB CRACKED; GAP EFFECTIVE LENGTH 50 MM


400.00

300.00

Theoretical Capacity
200.00

100.00

-50.00

-40.00

-30.00

-20.00

0.00
0.00

-10.00

10.00

20.00

30.00

40.00

50.00

-100.00

-200.00

-300.00

-400.00

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

68

6.3.3.4

Effective Gap Length

The plots in the preceding sections have all been based on an effective gap length of 50 mm. The
analysis of the wall with boundary members (Figure 6-29) was repeated with gap lengths of 100 mm
and 150 mm respectively. The effect is shown in Figures 6-30 and 6-31. For this particular
configuration the gap length of 150 mm provides a predicted capacity closer to the theoretical
capacity than the shorter gap lengths.
FIGURE 6-30 SHEAR PANELS IN WEB ELASTIC; GAP EFFECTIVE LENGTH 100 MM
400.00
300.00
Theoretical Capacity

SHEAR FORCE (KN)

200.00
100.00

-50.00

-40.00

-30.00

-20.00

0.00
0.00

-10.00

10.00

20.00

30.00

40.00

50.00

-100.00
-200.00
-300.00
-400.00
DISPLACEMENT (mm)

FIGURE 6-31 SHEAR PANELS IN WEB ELASTIC; GAP EFFECTIVE LENGTH 150 MM

400.00
300.00
Theoretical Capacity

SHEAR FORCE (KN)

200.00
100.00

-50.00

-40.00

-30.00

-20.00

-10.00

0.00
0.00

10.00

20.00

30.00

40.00

50.00

-100.00
-200.00
-300.00
-400.00
DISPLACEMENT (mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

69

6.3.3.5

Recommendations for Modeling Flexural Yield

The sensitivity studies suggest that the properties of the flexural hinge are not very sensitive to the
refinement of the grid used. The variation between 12 x 19 grids and 1 x 6 grids is well within the
variation expected from other parameters in the evaluation process such as the concrete strength.
Therefore, in most cases a grid selected to model geometric features will be sufficient to capture
flexural yielding response.
The discretization process used for the walls provides a strength which may be higher or lower than
the theoretical capacity, depending on such aspects as vertical load level, presence or absence of shear
cracking and effective gap length used. Sufficient sensitivity analyses to quantify these effects have
not yet been completed.
The gap length in this type of model represents the effective length of reinforcing bar which is
assumed to be strained at ultimate moment. The wall model in Figures 6-28 to 6-31 was evaluated
using gap lengths from 50 mm to 450 mm and the maximum force compared to the theoretical
maximum force, as shown in Figure 6-32. For this model, an effective gap length of 250 mm
provided the best representation of moment capacity.
This can be viewed as an effective
development length in the reinforcing bar.
Although this study is restricted in scope, these results have been used to set the default gap length at
250 mm. With further study, this parameter will probably prove to be a function of specific
characteristics of the wall and load conditions.

FIGURE 6-32 EFFECTIVE GAP LENGTH

MAXIMUM FORCE / THEORETICAL FORCE

1.600
1.400
1.200
1.000
0.800
0.600
0.400
0.200
0.000
0.050

0.100

0.150

0.250

0.350

0.450

GAP LENGTH IN MODEL (m)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

70

The plots given in the preceding sections may assist in defining properties for a specific project.
However, pending further development, you may want to develop project specific properties if
performance is very dependent on wall flexural yielding. The examples here were all generated using
the displacement pushover procedure implemented in MODELA. Similar plots could be generated
for specific walls, using the appropriate grid spacing, reinforcing ratios and axial loads for a particular
project.
6.4

REINFORCED CONCRETE FRAME MODELING

Reinforced concrete frame buildings are modeled as assemblages of frame elements representing the
beams and columns. The elements may be modeled with a stable bi-linear yield function or may
incorporate stiffness and/or strength degradation. Guidelines for these elements are provided in
Section 6.1.
Gravity load effects are usually important for frame structures as they affect:
1.

Secondary (P-) moments.

2.

The initiation of positive and negative moment yielding in beam elements.

3.

The moment strength of column elements, as represented by the moment axial load
interaction diagram.

Procedures to ensure that these gravity load effects are incorporated correctly are discussed in
Section 5.9, 7.6 and 7.13 of this manual.
6.4.1

BEAM ELEMENT HYSTERESIS

The beam element in ANSR-II has been modified to unload either plastically (producing the familiar
rhomboidal or degrading hysteresis shape) or elastically, where the unloading path follows the
loading path. This is intended for development purposes to evaluate types of prestressed concrete
frames but may have more general use.
The two types of beam element can be modeled in parallel to produce a pinched type hysteresis
loop. This is achieved by placing two beam elements spanning between the same two column lines
at the same level. One has an inelastic hysteresis, the other elastic unloading.
The two elements should sum to the required stiffness. For a rectangular beam, put two beams, each
of half the actual width. The total moment capacity will be split between the strengths of the two
elements.
To specify this type of hysteresis, a code has been added as the final column of the Beams sheet (after
the output code). This code specifies both the type of unloading and the geometric stiffness code
(whether to include P- effects). This code can have one of four values:

plastic unloading, no geometric stiffness (default).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

71

1
-1
-2

plastic unloading, include geometric stiffness


elastic unloading, no geometric stiffness
elastic unloading, no geometric stiffness

The example structure shown in Figure 6-33 has been used to generate displacement pushover curves
for three cases, using options of plastic unloading (the default), elastic unloading and two beam
elements in parallel (one plastic unloading and the second elastic unloading). The structure has
pinned base columns and all beams have the same properties so the overall force-displacement
function mirrors that of the individual beams.
Figures 6-34, 6-35 and 6-36 show the form of
hysteresis produced for each option. Note that even
though the beams are specified as bi-linear with zero
strain hardening (elastic-perfectly plastic) there is a
positive second slope to the plots. This is because
the plastic hinges in the beams form sequentially
rather than instantaneously, producing a rounded
transition from the elastic stiffness to the yielded
stiffness.

FIGURE 6-33 EXAMPLE STRUCTURE

Only limited validation studies have been performed


on these element shapes.
The response to
earthquake accelerations (Figure 6-37) and damping
decay (Figure 6-38) illustrate differences between the
element types:

The hysteretic element produces the smallest


earthquake displacement, the elastic unloading
case the largest.

The hysteretic option shows a permanent set


(residual displacement) for earthquake loads.
Neither of the other options show this effect.

The hysteretic element shows the most rapid


decay for free vibration damping but it is
initially over-damped and so does not return to its undeformed configuration.

The elastic unloading and pinched hysteresis options both vibrate approximately about zero
displacement. The pinched shape has higher damping (more rapid decay) than the elastic
unloading shape.
NOTE: If you use the parallel component model to generate the pinched hysteresis
shape you will need to sum the results of the two elements to obtain the actions on the
element. They will be reported separately in the ANSROUT workbook. If use of this
type of element becomes widespread we may automate the summation of actions for
beams occupying the same space.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

72

FIGURE 6-34 PLASTIC UNLOADING


400.00
300.00
200.00

FORCE (KN)

100.00

-100.00

-80.00

-60.00

-40.00

0.00
0.00

-20.00

20.00

40.00

60.00

80.00

100.00

60.00

80.00

100.00

80.00

100.00

-100.00
-200.00
-300.00
-400.00
DISPLACEMENT (mm)

FIGURE 6-35 ELASTIC UNLOADING


400.00
300.00
200.00

FORCE (KN)

100.00

-100.00

-80.00

-60.00

-40.00

0.00
0.00

-20.00

20.00

40.00

-100.00
-200.00
-300.00
-400.00
DISPLACEMENT (mm)

FIGURE 6-36 COMBINED ELEMENTS IN PARALLEL


500.00
400.00
300.00

FORCE (KN)

200.00
100.00

-100.00

-80.00

-60.00

-40.00

0.00
-20.00
0.00
-100.00

20.00

40.00

60.00

-200.00
-300.00
-400.00
-500.00
DISPLACEMENT (mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

73

FIGURE 6-37 TOP DISPLACEMENT FOR EARTHQUAKE LOADS


150
HYSTERETIC
ELASTIC
PINCHED

DISPLACEMENT (mm)

100

50

-50

-100
0

10

12

14

16

18

20

TIME (Seconds)

FIGURE 6-38 FREE VIBRATION DAMPING


120
100

DISPLACEMENT (mm)

80
60
HYSTERETIC
ELASTIC
PINCHED

40
20
0
-20
-40
-60
20

22

24

26

28

30

32

34

TIME (Seconds)

6.5

STEEL BUILDING MODELING

No special procedures are implemented for steel frame structures as they will usually be modeled in a
similar fashion to reinforced concrete frames.
The truss elements included in MODELA can be used to model braced structures. These elements
have a pin code which can be used to specify special hysteresis forms, for example, buckling in
compression for tension-only braces. See Sections 7.9 and 7.10 for details.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

74

6.6

UNREINFORCED MASONRY WALL (URM) BUILDINGS

Although unreinforced masonry (URM) is a brittle material, and does not have ductility as defined
for more ductile materials, it does have some non-linearity and the capacity to re-distribute load to
other elements to a limited extent before failure occurs. The procedure for modeling URM has been
developed to reflect this behavior.
Masonry wall modeling uses the degrading panel elements as described earlier in this section. The
hysteresis properties may be based on NZNSEE, FEMA or user defined parameters.

6.6.1

NZNSEE RED BOOK STRENGTHS

The default of the HCG procedure, if G and vc are omitted, or zero, is that the masonry properties
will be based on NZNSEE Red Book values. If G is omitted or blank but vc is non-zero then the
NZNSEE envelope shape will be used but the strength levels will be based on the specified shear
strength rather than NZNSEE values. Figure 6-39 shows the implementation of the NZNSEE
strength.
There are a few points to note about the envelope curve in Figure 6-39:
1. The stiffness and shear strain limits are extracted from a number of experimental studies on
unreinforced brick walls.
2. The first crack stress level (at a shear strain if 0.0005) is set to the NZNSEE allowable stress
value of 230 KPa. The peak stress before strength degradation at failure is 29% higher, at 297
KPa.
3. The strengths do not include a capacity reduction factor, . This is because non-linear time
history analysis is typically performed using probable, not dependable, strengths (NZS4203).
The hysteresis in Figure 6-39 is for a wall panel with zero axial stress. The program MODELA
approximates the effects of axial stress by implementing the following procedure:
1. Divide the maximum height of the structure into 20 equal segments, h.
2. For each masonry element group, define 20 material types. For each material type, n, increase
the cracking strength such that vc = vc + nh (0.9w) where is the coefficient of friction of the
masonry (default value 0.30), w is the density of the brick (default 18 KN/m3) and the 0.9 is a
capacity reduction factor.
3. For each masonry wall element, examine every other wall element to determine the height of
contiguous wall above the particular element, h*. Assign a material number to the element of the
integer value of h*/h.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

75

This procedure is conservative in that it accounts for the self-weight of the wall only, with no
allowance for floor tributary weights. It also assumes walls above are of equal thickness. Some
account of these factors can be made by modifying the weight density specified for the material.

FIGURE 6-39 IMPLEMENTATION OF NZNSEE STRENGTH (ZERO AXIAL STRESS)

350

300

SHEAR STRESS (KPa)

250

200

150

100
ANSR-II Probable Strength
50

NZNSEE Red Book Dependable Strength

0
0

0.0005

0.001

0.0015

0.002

0.0025

0.003

SHEAR STRAIN (mm/mm)

For walls under axial stress, the strength is increased at all strain levels as shown in Figure 6-40. This
provides a residual strength equal to the friction force for high strains.
Depending on the type of modeling being performed, and regulatory requirements, the input
properties can be used to vary the manner in which NZNSEE requirements are implemented. A
number of permutations may be used:
1.

The cracking strength is required to equal the NZNSEE value, including the factor. Set the
cracking strength, vc, to 230 x . Note that the ultimate strength will be 1.29 times this value.

2.

The ultimate strength is required to equal the NZNSEE value, including the factor. Set the
cracking strength, vc, to (230 x ) / 1.29.

3.

The ultimate strength is required to equal the NZNSEE value, excluding the factor. Set the
cracking strength, vc, to (230 / 1.29).

4.

The friction coefficient is required to include the factor. Set the value of on the input sheet to
0.3 x 0.8 / 0.9 = 0.27.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

76

FIGURE 6-40 EFFECT OF AXIAL STRESS

450
400

SHEAR STRESS (KPa)

350
300
250
200
150

No Surcharge

100

24 m Surcharge

50
0
0

0.0005

0.001

0.0015

0.002

0.0025

0.003

SHEAR STRAIN (mm/mm)

6.6.2

FEMA STRENGTHS

The FEMA guidelines provide more specific requirements for modeling URM than the NZNSEE
requirements and also relate more closely to the quality of the masonry as they provide properties for
good, fair and poor quality materials. MODELA will generate the properties for each of these
quality grades when a shear modulus of 1, -2 or 3 respectively is entered into the spreadsheet.
Figures 6-41 and 6-42 compare the hysteresis shapes of the NZNSEE and the FEMA codes.
Stresses for both codes are based on allowable stresses when material testing is not performed. The
FEMA values are for good quality materials.
When there is no, or very little, surcharge the NZNSEE allowable strengths are much higher than
the FEMA values. As the surcharge increases the FEMA values become a larger proportion of the
NZNSEE values. This is because the FEMA coefficient of friction is 0.50, compared to 0.30 for
NZNSEE.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

77

FIGURE 6-41 COMPARISON NZNSEE AND FEMA : NO SURCHARGE

350

300

NZNSEE No Surcharge
FEMA No Surcharge

SHEAR STRESS (KPa)

250

200

150

100

50

0
0

0.0005

0.001

0.0015

0.002

0.0025

0.003

SHEAR STRAIN (mm/mm)

FIGURE 6-42 COMPARISON NZNSEE AND FEMA : 25 M SURCHARGE

450
400

SHEAR STRESS (KPa)

350
300
250
200
NSNSEE 25 m Surcharge
150

FEMA 25 m Surcharge

100
50
0
0

0.0005

0.001

0.0015

0.002

0.0025

0.003

SHEAR STRAIN (mm/mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

78

6.6.3

TESTING OF MASONRY

The default masonry strengths for both NZNSEE and FEMA are based on materials for which
testing is not performed. Generally, testing will permit a higher allowable strength to be used. For
example, the Auckland Customhouse is built of brick, of quality typical of construction around the
end of the 19th century. The test strength of 413 KPa was 80% higher than the NZNSEE default of
230 KPa (Figure 6-43).

FIGURE 6-43 AUCKLAND CUSTOMHOUSE IN-PLACE SHEAR TESTS


900
800

SHEAR STRENGTH (KPa)

700
80% GREATER THAN 413 KPa
(MEAN 534 KPa)

600
500
400
300
200
100
0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

The trade off in costs between strengthening required using the default strength values and the cost
of testing should be assessed on a job by job basis. The NZNSEE requirements for testing can
result in a large number of test locations, which can be expensive and damaging. It is often best to
perform a preliminary assessment using default values and then assess the likely benefits to be
obtained from testing if the strength is increased by say 50% to 80%, as is likely if bricks are tested.

6.6.4

CYCLIC LOADING CHARACTERISTICS

For unreinforced masonry it is generally recommended that all degradation coefficients be left blank.
This results in the default masonry cyclic loading model shown in Figure 6-44 being used. However,
see also Section 6.6.6 for exceptions to this.
As shear strains applied to the model increase, the shear stress follows the envelope curve specified
for the material. If the strain exceeds the strain corresponding to peak stress then the stress will
reduce for increasing strain. In this condition the stiffness is negative and the structure will be in an
unstable equilibrium if the total negative stiffness exceeds the positive stiffness provided by elements
at the same level which have not yet reached the degrading strength portion of the curve.
Displacements will be unbounded, indicating collapse.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

79

At any stress strain point, unloading occurs through the origin, that is, along the secant stiffness.
This feature supplies the stiffness degradation, which is irreversible.

FIGURE 6-44 MASONRY CYCLIC BEHAVIOUR

400
300

SHEAR STRESS (KPa)

200
100

-0.0020

-0.0015

-0.0010

-0.0005

0
0.0000

0.0005

0.0010

0.0015

0.0020

0.0025

0.0030

-100
-200
-300
-400
SHEAR STRAIN (mm/mm)

6.6.5

ROCKING STRENGTH

The material strengths for masonry are based purely on in-plane shear. If rocking can occur in
masonry piers there are three methods which may be used to incorporate this effect:
1. Include gap elements at each corner of the pier and pre-load these with gravity loads (this may
require a separate analysis see Section 7.5). This is the most accurate form of modeling rocking
but has a significant penalty in model complexity and may require a short time step.
2. Adjust the shear strength so that the maximum shear force in the element corresponds to the
force required to initiate rocking. This requires manual calculations to determine the shear
strength. If piers are modeled this way, the hysteresis should be specified as an elasticperfectly
plastic material with no strength or stiffness degradation (entered as a User Specified Material).
3. Model rocking piers as pin-ended column elements. If piers are very slender and likely to rock at
a low level of load then their contribution to total lateral strength may be so low that they can be
ignored. A pin-ended column will provide gravity load support only.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

80

For complicated models more than one of these techniques may be used in the same model.
6.6.6

URM EVALUATION STRATEGY

The default material model for URM has a strength which rapidly degrades under increasing
displacement beyond the strain at which ultimate strength is reached. This is achieved by setting all
degradation coefficients blank, as discussed above.
With this material model, degradation of a particular element results in large unbalanced forces which
are distributed to adjacent elements, failing these in a domino effect. Consequently, by the time the
analysis terminates failure may be very extensive. It is difficult to determine from these results just
what the extent of failure would have been had the elements which failed first been stronger. In fact,
unless the model is simple enough that the time history analysis can be tracked by using the screen
animation option it may not be possible to determine which elements failed first.
A strategy to help define the hierarchy of vulnerabilities is to set the element so that strength reduces
under reversing loads rather than increasing loads by defining non-zero degradation coefficients. If
the coefficients are set = 1.0, = 0.0, XKK = 1.0 and DCODE = 0.3 then a hysteresis similar to Figure
6-45 will be generated.

FIGURE 6-45 URM DEGRADING HYSTERESIS

800
600

SHEAR FORCE

400
200

-0.02000

-0.01500

-0.01000

-0.00500

0
0.00000

0.00500

0.01000

0.01500

0.02000

-200
-400
-600
-800
SHEAR STRAIN

The difference between Figure 6-44 and Figure 6-45 is that the latter curve maintains strength until
unloading. This avoids the sudden unbalanced loads.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

81

The solution using this type of hysteresis may be non-conservative in that in reality the strength may
not be maintained at a constant level for continuing loading in the same direction if the strain
increment is large. However, if the element is to be strengthened then this does not matter as the
analysis is an interim step to locate areas needing strengthening rather than being a final evaluation of
the building.
For subsequent analyses the elements which exceeded the strain at ultimate strength (0.0015 for the
NZNSEE URM model, 0.003 or 0.004 for FEMA) are strengthened (see below) and the extent of
further damage assessed. Strengthening is continued in an incremental process until the strain does
not exceed the ultimate strain in any unstrengthened elements. The processor program PROCESSA
produces data to assist in this type of evaluation see Section 10.4.

6.6.7

URM STRENGTHENING USING FRP

Holmes Composites are involved in developing procedures for URM strengthening using Fiber
Reinforced Plastics (FRP). Based on research to date, it is recommended that the FRP be modeled
as a linear elastic panel overlaid over the masonry panels with excessive strain.
The FRP wall elements are evaluated against an allowable strain, usually 0.004. If the strain exceeds
this value then additional thickness needs to be added. See Section 6.8 for details of these materials.

6.7

TIMBER MODELING

To date, timber as a material has been used for modeling horizontal diaphragms only. The material
has been used to define the properties of the plane stress elements used to model flexible floor
diaphragms. Properties may be user-specified or may be based on published values from FEMA for
different types of construction. FEMA properties for various types of construction are built in to
MODELA see Section 7.9.
As a check on the FEMA values, an early NZ timber code (NZS3603:1981) was used. This code
provides properties for single, straight sheathed, unchorded diaphragms (Code = -2 in Modela). The
code stiffness is based on the board stiffness plus the contribution from the nail couples. Figure 646 shows that the ANSR-L model provides a reasonable match to the calculated values, with the
initial stiffness decreasing to a value approximating the stiffness of the boards alone.

6.7.1

TIMBER FLOORS

The FEMA Guidelines provide an envelope strength curve which reaches a peak strength and then
loses strength. The displacement level at which the strength degrades is a function of the length to
width ratio of the diaphragm. This is difficult to define when floors are modeled as finite elements.
For timber floors, the deformation acceptance criteria usually depend on the allowable deformations
for other elements, such as URM walls connected to the floors.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

82

A strength loss at a pre-defined strain level results in numerical stability for analysis. As tests
generally show that timber panels lose stiffness but maintain a residual strength, it is recommended
that floors be modeled as elastic-perfectly plastic after the ultimate strength based on FEMA is
reached. This provides a hysteresis of the form shown in Figure 6-47.
FIGURE 6-46 TIMBER FLOOR ELEMENT PROPERTIES
700

600

NZS3603:1981 Nail Stiffness


Sum of Individual Timber Boards

SHEAR STRESS (KPa)

Model used in ANSR-II


500

400

300

200

100

0
0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035

0.040

0.045

0.050

SHEA R STRA IN (mm/mm)

FIGURE 6-47 TIMBER FLOOR HYSTERESIS

8
6

SHEAR FORCE (KN/m)

4
2

-0.0200

-0.0150

-0.0100

-0.0050

0
0.0000

0.0050

0.0100

0.0150

0.0200

-2
-4
-6
-8
SHEAR STRAIN (mm/mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

83

6.8

FIBER REINFORCED PLASTICS (COMPOSITES)

Fiber Reinforced Plastic (FRP) is the generic name for structural composites. It is used in existing
structures to provide confinement and shear strengthening in flexural members and as an overlay to
improve shear and/or flexural strength of concrete or masonry walls.
This material is formed of glass or carbon fibers applied in an epoxy matrix. The FRP is applied in
relatively thin layers (typically 1.3 mm). The required strength is developed by building up the
thickness from multiple layers. The material is opaque and so does not form an invisible repair. It
has a smooth finish and can be painted.
HCG are taking an active role in the development of this material through our joint ownership of
Holmes Composites, a San Diego company. Design procedures are well developed for flexural
element strengthening, particularly confinement which was the initial application of FRP in structural
engineering. For wall strengthening test programs are underway and there are likely to be rapid
developments in procedures for design and analysis. Recommendations provided in this manual are
likely to be subject to revisions so check with the Development Division if you wish to use this
material.
Properties of the materials most commonly used by Holmes Composites are listed in Table 6-3.

TABLE 6-3 TYFO MATERIAL PROPERTIES

Composite
Material
Ultimate Tension Strength
Perpendicular Strength
Elastic Modulus
Ultimate Elongation
Allowable Elongation
Thickness

SEH-51 and
Tyfo S Epoxy
(Glass Fabric)

SCH-41 and
Tyfo S Epoxy
(Carbon Fabric)

420 MPa
38 MPa
21,000 MPa
0.020
0.004
1.3 mm

630 MPa
38 MPa
63,000 MPa
0.010
0.004
1.04 mm

WEB and
Tyfo Hi-Clear
Epoxy
(Glass Fabric)
224 MPa
224 MPa
14,000 MPa
0.016
0.004
0.36 mm

The most commonly used material is SEH-51, a glass fiber. This is not as strong as the carbon fiber
but has greater ultimate elongation and is much cheaper. The WEB is used where a similar stiffness
and strength is required in both directions.
Analyses performed to date have modeled the FRP SEH-51 as a wall element with a thickness equal
to 1.3 mm times the number of layers used. The material is a user-defined type with E=21,000,000
KPa and Poissons ratio = 0.2. The material strain properties are set with G1=0.0 and S1=0.01.
This retains linear elastic behavior to a shear strain of 0.01. Pending further research data, the shear
strain is limited to 0.004. If the analysis shows FRP panels with shear strains exceeding this value
then the thickness needs to be increased.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

84

6.8.1

TEST DATA FOR FRP STRENGTHENING

There is a limited amount of test data on the use of FRP for strengthening URM and concrete walls.
As an interim measure, pending further research, extracted properties from a test series performed by
Reinhorn et al. have been used to develop what is believed to be a conservative approach to
modeling FRP strengthening.
Wall specimens tested were 70 high (1.78m), 72 long (1.83m) and constructed of 2 wythes of 4
brick (102 mm). Walls were tested both with and without FRP overlays. Figure 6-48 shows the
strength envelope obtained from the bare wall and the strengthened wall.

FIGURE 6-48 EFFECT OF FRP STRENGTHENING ON URM WALL

300
Bare Wall
Extrapolated Bare Wall
Wrapped Wall

250

FORCE (KN)

200

150

100

50

0
0

10

15

20

25

30

35

40

45

DISPLACEMENT (mm)

From the wall dimensions the effective shear deformation was calculated and stress strain curves
constructed for the bare wall and the strengthened wall, as shown in Figure 6-49. The bare wall
resisted a peak shear stress of 331 KPa at a shear strain of 0.0024. This compares reasonable well
with the strength curve used for URM analysis based on NZNSEE strengths, which provides a peak
stress of 298 KPa at a shear strain of 0.0015. This curve is also shown on Figure 6-49.
When the wall was strengthened with FRP the envelope strength increased from 331 KPa to 717
KPa, an increase of 130%. The strain at which the peak stress occurred increased from 0.0024 to
0.0065, an increase of 170%.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

85

The test on the bare wall was terminated at a shear strain of 0.004. To derive the strength of the
composite wrap it was assumed that the strength at this strain was maintained to the maximum shear
strain in the strengthened wall, 0.0065. This extrapolation is shown as a dotted line in Figures 6-48
and 6-49.
FIGURE 6-49 EFFECTIVE SHEAR STRESS

800
700

Bare Wall
Extrapolated Bare Wall
Wrapped Wall
ANSR NZNSEE Model

SHEAR STRESS (KPa)

600
500
400
300
200
100
0
0.0000

0.0020

0.0040

0.0060

0.0080

EFFECTIVE SHEAR STRAIN (mm/mm)

The difference between the bare wall strength envelope and the strengthened wall was used to
develop an implied strength of the FRP overlay, as shown in Figure 6-50. This shows an
approximately linear elastic strength up to a peak of 179 KN at a strain of 0.0065. Beyond this strain
the FRP delaminated and the strength reduced to a level similar to that of the bare wall.
The derived overlay strength shown in Figure 6-50 has been correlated with the known FRP
properties as shown in Table 6-4.
The FRP used was applied to both faces of the test wall. One side was strengthened with 2 bands of
SEH-51, each 152 mm wide x 1.3 mm thick. The other side was overlain with a 0.36 mm thick layer
of WEB.
Using the properties of the materials listed in Table 6-4, and assuming a full overlay of each of the
two materials, the calculated shear forces at a strain of 0.0065 are 135 KN in the SEH-51 and 25 KN
in the WEB, providing a total force of 160 KN. This is 10% less than the measured strength
enhancement of 179 KN.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

86

FIGURE 6-50 EXTRACTED FRP STRENGTH

SHEAR FORCE PROVIDED BY WRAP (KN)

200

160
Wrap Alone
Average Stiffness
120

80

40

0
0.0000

0.0010

0.0020

0.0030

0.0040

0.0050

0.0060

0.0070

0.0080

SHEAR STRAIN (mm/mm)

TABLE 6-4 STRENGTH PROVIDED BY FRP OVERLAY

SEH-51

WEB

Total

0.0013
0.00036
0.00166 Thickness, t (m)
21,000,000 14,000,000
Elastic Modulus, E (KPa)
0.2
0.2
Poissons Ration
8,750,000 5,833,333
Shear Modulus,G (KPa)
0.0065
0.0065
Shear Strain,
56875
37917
Shear Stress v = G (KPa)
1.8288
1.8288
Length, L (m)
135
25
160 Shear Force V = tLv (KN)
179

Measured Shear Force, KN

There are still outstanding issues related to URM strengthening using FRP and further testing is
underway. Pending further data becoming available, it is believed that a conservative procedure to
use is as follows:

Add a new panel at the same location as the URM panels requiring strengthening.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

87

Set the thickness as the layer thickness (Table 6-3) times the number of layers to be used. The
number of layers will generally be determined by a trial and error process.

Set the material type as isotropic linear elastic using the E value from Table 6-4 and a Poissons
ratio of 0.20.

Evaluate results using a strain limit of 0.004 in the FRP elements.

The ANSROUT results processing spreadsheet has been set to assist with this evaluation see
Section 10.4. To use the processed results, name the walls which are strengthened with a string
which includes the characters TYFO (not case sensitive). For example, use TYFO3 for a 3-layer
overlay.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

88

7 MODEL PREPARATION : MODELA

Model preparation is performed entirely on an Excel spreadsheet, MODELA.XLS. The most up to


date version of the spreadsheet should be obtained from the HCG server at the start of a project and
saved into the job directory.
Your local computer should be configured as discussed in Section 3.
The MODELA spreadsheet has a macro which shells the DOS program MODELA.EXE. This
program in turn generates the ANSR-L input files, AutoCad DXF files for geometry checks and a
TRC file which is read by the post-processor program (Section 10).
Data can obviously be entered into the spreadsheet in any order. However, the model can also be
generated at various stages of development. For complicated models, it may be most effective to
generate the model piece by piece, checking geometry at each stage. For example, the model with
walls only may be generated and the geometry checked in AutoCad before the beams and columns
are added.
The sheets generally contain blocks of contiguous data stretching across and down from the cell
containing the name. Cells containing data are generally filled with a light yellow color.
Extra rows can be added within the data blocks but extra columns should not be added
to the spreadsheet within the block of input cells.
Columns can be added outside the range of the input data cells (the light yellow shaded cells) for
storage of data used for intermediate calculations not required by MODELA.

7.1

UNITS

The analysis procedure based on the MODELA spreadsheet is set for S.I. metric units only. The
length unit is meters (m) and the force unit KiloNewton (KN). Stresses are consistent with this, that
is, in KiloPascals (KPa). Be careful when entering strength values which are usually represented as
MPa. For example, concrete strength will be 30,000 KPa. Masses are in tonnes, that is, KN weight
divided by 9.81.
As the ANSR-L input files are set up using specific material properties and scaling factors it is not
permissible to use other consistent sets of units. At this stage, if you have a model in kips and feet
units, for example, use Excel spreadsheet formulas to convert the quantities to S.I. units.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

89

7.2

FILES WRITTEN BY MODELA

The output from MODELA is a series of files as listed in Table 7-1 (see Section 7.3 for details of the
identification characters defining the files).

TABLE 7-1 FILES CREATED BY MODELA.

jjrrqie.ANS

.
.
jjZ0 or jjS0

A series of input files for the ANSR-L program. Number of files equal to 2 x
(number of earthquakes) x (number of eccentricities) x (number of earthquake
increments).

jjrr.BAT

A batch file to run the series of ANSR-L, ZAP or SAP2000 files listed above.

jjrr.TRC

A file containing the correlation between the data on the spreadsheet and the
ANSR-L file as generated. Read by the post-processor program, can also be
used find the location of ANSR-L node numbers (see Section 10.2).

jjrr.DXF

An AutoCad DXF file which can be imported to display the geometry. This file
represents elements as line elements and will include diaphragm connectivity if
requested (see Analysis Data). The drawing produced by this file is used to
generate the crack patterns (see Section 10.13).

jjrrF.DXF

An AutoCad DXF file of geometry with all elements drawn fully three
dimensional, that is, beam and column elements drawn to correct breadth and
width, wall elements drawn to correct thickness.

T$0000
T$0001
T$0002

A series of temporary files which may or may not be created. These are used to
write out intermediate data if problems are suspected with the model as
generated. They can be deleted.

7.3

ZAP or SAP2000 input files for modal analysis

ANSR-L RUN FILE IDENTIFICATION

The MODELA program generates a series of ANSR-L files of the form

jjrrqie.ANS
where
jj

is the run number defined on the Interaction sheet

rr

is the revision number defined on the Interaction sheet

is the earthquake number, from 1 to 2 times the number

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

90

of earthquakes (1 is earthquake 1, oriented X, 2 is earthquake


1 oriented Y, 3 is earthquake 2 oriented X etc).
i

is the incremental earthquake, an alphabetic series from A


to the number of increments. The series starts from the highest
factor. For example, if you have specified incremental
earthquakes from 0.50 to 1.00 at 0.25, A will be 1.0, B 0.75
and C 0.50.

is the eccentricity, numbered from 0 to the number requested on


the interaction sheet (maximum 5).

The software packages used as part of the HCG procedure are not all fully compatible with long file
names and so the total name should be kept to not more than 8 characters. This requires that the
total length of the Run Number plus the Revision Number be 5 characters or less.
7.4

CONSTANTS WORKSHEET

The Constants sheet of the MODELA workbook contains data used by the macros and dialog boxes.
There are special fields for the analysis of special purpose Zap masonry wall evaluation models but
these should not be used without taking advice as they are not documented in this manual.

7.5

INTERACTION

SHEET

The interaction sheet, shown in Figure 7-1, contains push buttons to control the macros to write the
files. The title and directory information is set on this sheet when a job is first initiated. This sheet is
not then used until after the job data have been entered on the remaining workbook sheets.

7.5.1

JOB TITLE

The first step in creating the job specific worksheet is to enter the job title, up to 80 alphanumeric
characters, and set the job directory. This can be entered by typing directly into the field or by
activating the Set Directory button and then selecting the directory in the dialog box which will appear.
The Run Number and Revision Number are used to identify the ANSR-L runs. It is preferable to keep
these as brief as possible, usually only one character each. This is to ensure that file names are less
than 8 characters (see Section 7.2).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

91

7.5.2

PROBLEM SIZE AND PROGRAM LOCATION

The bottom box on this sheet has data cells for the size of Common, the location of the Fortran
executable programs and a code for Master Nodes. See the discussion on Special Functions later in
this section for appropriate values for Common and Master Nodes.
The path to the MODELA and PROCESSA programs will depend on the directory structure you
have set up to contain the files. The program name can be followed by a D token to set the
program to run in debug mode. With this option, MODELA.EXE will write additional information
into the .TRC file. This option is only used if an error occurs during the Write ANSR Files step or if
the model appears to be incorrect.
If you have used the D option, remove the token once you have found and fixed the problem. The
.TRC file written with the D option is not compatible with the post-processor used by ANSROUT.
FIGURE 7-1 INTERACTION SHEET
COMPUTER PROGRAM MODELA
ANSR-II Nonlinear Time History Analysis Models
Job Title:
Run Number:
Revision:

Cooper Lybrand As-Built


1 Directory : C:\jobs\COOPERL

Time History
Analysis
Write ANSR Files
Standard
Write ANSR File to
Get
Gap Loads
Write ANSR File
Reading
Gap Loads

Force Pushover
Analysis

Displacement
Pushover

Write ZAP
Model for
Periods

Set
Pushover
Parameters

Read Periods

Write
ANSR
File

Import Periods
Write ANSR
Model with
Pushover Loads

Mass Locations for Time


History Analysis
Centre of Mass Analysis
+X +Y Eccentricity

C:\PROGRAMS\MODELLER\MODELA
1

Code for Master Nodes : 0, 1 or 2 (see Manual)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

Clear Old Data

-X -Y Eccentricity
+X -Y Eccentricity
-X +Y Eccentricity

Convert ETABS

Options for Force


Pushover Analysis
2
1
0.5
1000
0.5

Problem Size and Program Locations


3,500,000 Size of Common (increase for large jobs - see Manual)
C:\PROGRAMS\MODELLER\PROCESSA
Location of programs

Special
Functions

0.6

X Mode Number
Z Mode Number
Base V/W
Number of Steps
X Period
Z Period

Apply Fy

Apply Fx

Apply Fx

Apply Fy

Apply Fz

Apply Fz

Apply Mx

Apply Mx

Apply My

Apply My

Apply Mz

Apply Mz

Find COMMON
Size
Read COMMON
Size
Hide/Unhide
Sheets

92

7.5.3

TIME HISTORY ANALYSIS

Generally, the type of analysis performed will be Standard. For this type of analysis, the gravity loads
are applied as a set of static loads prior to the application of seismic loads. The displaced shape and
elements forces form the initial conditions for the earthquake accelerations which are then applied as
dynamic loads.
For some structural types, applying the gravity loads will give rise to a displaced shape and/or
element stress conditions which are not compatible with the actual structure. This is especially so for
unreinforced masonry structures, in which shear cracking may occur in lintels etc when the gravity
loads are applied.
The second and third time history analysis options are intended to provide a more reasonable
evaluation for this type of structure where the effect of the gravity load is to provide the initial
compression force on the gap elements representing flexural cracking. The procedure is as follows:
1. Activate the option button Write ANSR File to Get Gap Loads.
2. Run the batch file created. This will analyze a single ANSR-L run file, which applies the gravity
load and then stops after step 1 of the earthquake analysis, before inertia loads are applied.
3. Process this file using the Get Gap Loads option of the ANSROUT spreadsheet (see Section 10).
This will read the compression forces due to gravity in all the gap elements and write them to a
disk file.
4. Return to the MODELA spreadsheet and activate the Write ANSR Files Reading Gap Loads
option button. This will write a set of ANSR-L files with no static load applied but the gap loads
applied as initial forces on the gap elements.
Only gap elements are initialized. This procedure should not be used for models
containing column elements, where the yield moment is a function of axial load. Static
loads are required to correctly initialize the columns axial forces.
7.5.4

FORCE PUSHOVER ANALYSIS

The Force Pushover method is an implementation of the FEMA-273 and ATC-40 procedure. Force
vectors proportional to the first mode shape, proportional to mass and/or in a triangular distribution
are applied incrementally to obtain a force-displacement capacity curve. This is then compared to a
demand curve, adjusted for damping.
Section 9.1 of this Manual provides a full description of this procedure and the manner in which it is
implemented.

7.5.5

DISPLACEMENT PUSHOVER

The Displacement Pushover is a single step procedure to develop the cyclic performance characteristics
of the structure being evaluated. A cyclic displacement is applied incrementally at a specified node
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

93

point and the force required to displace the building is recorded at each step. This allows a global
hysteresis curve for the entire structure to be extracted. It may be seen as the analytical equivalent of
applying a displacement forcing function to an experimental specimen in the laboratory.
Although the displacement pushover does not follow published procedures and so cannot be used
directly for evaluation, it does provide very valuable information relating to the building being
analyzed and can be useful in identifying anomalies in the model.
Section 9.2 of this Manual provides a full description of this procedure and the manner in which it is
implemented.

7.5.6

MASS LOCATIONS FOR TIME HISTORY ANALYSIS

The program permits up to 5 center of mass locations, as listed on the check box on this sheet. Two
types of eccentricity are incorporated in MODELA, eccentricity of the diaphragm mass and
eccentricity of the distributed element mass. Either or both may be included within a single model.
Diaphragm mass is based on eccentricities associated with each diaphragm. The coordinates of the
master node representing the centroid of the diaphragm are moved by the eccentricities specified for
the particular diaphragm. Each diaphragm has an eccentricity specified (in meters) and so different
diaphragms, even if at the same level, can have different eccentricities.
The second type of eccentricity is applied to masses distributed at each level of the model. The
masses arising from element self weight, plus any user specified masses, at a particular level are
factored so that the effective mass centroid is moved by the eccentricity specified
In neither case are changes made to the mass moment of inertia. For the diaphragm eccentricity, this
is consistent with the assumption made for other types of analysis (for example, response spectrum).
For the distributed mass the mass moment of inertia arises from the spatial distributions of the mass
points and so the mass moment of inertia is effectively changed.

7.5.7

SPECIAL FUNCTIONS

7.5.7.1

Clear Old Data

This option button is intended to clear the spreadsheet at the start of a new project. It
has not been extensively tested so use with caution.
7.5.7.2

Convert ETABS

This option button will prompt you to select an ETABS .EKO file for the model you want to
generate. A macro will then scan the .EKO file and enter the appropriate data into the sheets of the

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

94

workbook. You will need to enter strength data separately as the ETABS model does not contain
this information.
Although every effort is made to keep this option current, you should carefully check any models
generated using this option in case data are entered in the wrong fields or omitted.

7.5.7.3

Find Common Size

The ANSR-L memory management is based on a large common block termed blank common, which
is an amount of memory reserved to store all the structural data, stiffness and mass matrices, and
element data. Pending upgrades to the program, the size of this memory block must be specified by
the user.
The amount of memory required is dominated by the storage required by the stiffness matrix. This is
in turn determined as the product of the number of degrees of freedom (dof) times the average
bandwidth (difference in dof numbers for nodes common to an element). The number of dofs is set
by the complexity of the structure but the bandwidth also depends on the order in which the nodes
are numbered.
This option will write a series of 18 ANSR input files with the Check Data mode set so that the
structure is input but no analysis is performed. At the end of the input echo the memory storage
requirements are written to the output file (See Section 8.3). These 18 files consider all permutations
of coordinate sort order and master node codes.
Determining common size this way is time consuming and so should only be used for moderate and
large sized structural models. If the analysis can be completed within the default blank common size
of 2,000,000 then it is probably not worth worrying too much about optimal settings.
The procedure is to activate this button and run the batch file created, the same procedure as is used
for standard ANSR analyses. The next button is then used to read the memory requirements and
import them to the bottom of the Interaction sheet.
Note that the process of writing the 18 ANSR files will take longer than would be require to write the
ANSR files for 18 earthquake/eccentricity options. This is because the coordinates must be reordered and regenerated for each file, rather than generated once as for earthquake / eccentricity
options.
7.5.7.4

Read Common Size

The Read Common Size button will scan the 18 ANSR .ECH files and write the memory required for
each option starting at line 40 of the Interaction sheet. Figure 7-2 shows an example. For this
example, the memory requirement ranges from 3.6M to 10.8M, depending on the order of sorting
coordinates and the master node code, DIAPEND. For this problem, it is optimum to set the sort
order as XYZ and the master node code as either 1 or 2.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

95

FIGURE 7-2 ANSR MEMORY REQUIREMENTS


COMMON

ORDER

6,647,409

XYZ

DIAPEND
0

3,684,009

XYZ

3,684,009

XYZ

6,649,001

XZY

3,734,825

XZY

3,734,825

XZY

6,761,817

YXZ

6,808,185

YXZ

6,744,033

YXZ

6,915,705

YZX

6,968,949

YZX

6,904,797

YZX

9,945,781

ZXY

7,886,209

ZXY

7,886,209

ZXY

10,825,345

ZYX

8,510,953

ZYX

8,510,953

ZYX

The smallest memory size does not necessarily mean the shortest run time but the two do seem to be
closely correlated. The run time can also be influenced by whether nodes are numbered from the
bottom or top (see Section 7.6.14 in the Analysis Data section). If this option is changed then the
common sizes should be re-computed.

7.5.7.5

Hide / Unhide Sheets

The complete MODELA workbook contains 22 separate sheets, all of which may or may not be
used for a specific project. This option allows you to hide and unhide sheets to tailor the project to
the options you will be using. Note that this is purely a convenience tool, it does not delete sheets
and sheets may be hidden and/or unhidden as you work on the project. When this button is
activated the dialog box shown in Figure 7-3 will be displayed. Check the sheets which you want
visible in the workbook.
In addition to the sheets listed in Figure 7-3, there are a number of dialog sheets, which are normally
hidden. If you need to show these sheets they can be unhidden from the Excel Format menu (Sheet
option). Normally, these dialog sheets will not be changed.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

96

FIGURE 7-3 DIALOG FOR SETTING UP SHEETS

7.6

ANALYSIS DATA

The Analysis Data sheet (Figure 7-4) sets a wide range of parameters which define how the model and
the analyses are set up.
Always check the options on this sheet carefully as the settings will often be left over
from a previous project and may not be suitable for your project.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

97

FIGURE 7-4 ANALYSIS DATA SHEET


ANALYSIS CONTROL DATA
Seismic Loading

SITE SPECIFIC

Soil Type

INTERMEDIATE

R Factor
Z Factor
Sp Factor
Lu Factor

Site Specific Time Histories (Set Seismic Load to Site Specific)


Number of Time Histories
1
Number

1.00
1.20
1.00
1.00

X LOAD
Number
Factor
1
1.00

Y LOAD
Number
Factor
0
0.00

Duration
(secs)
20.00 C:\QUAKES

Z LOAD
Number Factor
0
0.00

Masses & Weight


Include Element Weight
Include Element Mass
Include UDL as Mass
Distribute Diaphragm Mass as Weight
Lump Element Mass to Diaphragm

1.55

Assign Missing Mass Horizontally

Incremental Earthquake

0.10
1.00
0.10

Assign Missing Mass Vertically

Coordinate Sort Order


First
Second
Third
X
Y
Z
Gravitational Constant
9.81
Force Tolerance
0.05
(fraction of Weight - recommend .01 to 0.10)
Node Tolerance (m)
0.005
TIME HISTORY OUTPUT

Time Step
0.010
First
Period
3.00

Maximum Displacement
1
m

seconds
Damping
0.05

Minimum Level
Maximum Level
Increment

Second
Period
0.250

Damping
0.05

Diaphragm Nodes

Special Features

User Specified Nodes


Element Time Histories

Number DOFs from Top

NON-STANDARD DAMPING
Type
K or M

Column
(for M)
Material
(for K)

Level
(M only)

OUT OF PLANE WALLS


SAVE each run for re-start
Slave DOFs active

Direction
(M only)
ANSR

Damping
Ratio

Period

Axes

Fix Degree of Freedom


Slave to Upper / Lower DOF

Include Connectors in DXF

Model Spanning Column

ZAP Masonry Model


Use -X in batch file

Fix/Spanning as Needed

Ignore Gravity Loads


Use SAP instead of ZAP

UBC FACTORS
Z
0.400
I
1.000
R
4.500
Nv
1.000
Ca
0.440
Cv
0.640

7.6.1

NZS 4203 SEISMIC DATA

The seismic data for the analyses are currently automated for the New Zealand code NZS4203 only.
For other codes, and for situations where site specific time histories are to be used, acceleration time
history files must be defined directly.
For NZS4203 loading (set in the Seismic Loading edit box) you need to set the soil type and the
amplitude factors R, Z, Sp and Lu. See Section 4 for a discussion of these factors.
The Number of Time Histories cell should be set to 3, the minimum number to comply with the code.
The program is currently set up for a maximum of 3 time histories. Less than 3 time histories may
be used if you are checking out the model but any final analyses should use 3. To use more time
histories you will need to use the Specific Time Histories option.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

98

7.6.2

SITE SPECIFIC SEISMIC DATA

When you wish to use site specific earthquake records you will need disk files of the acceleration time
histories for the earthquakes. The files must be in the correct format for ANSR-L. This format is
relatively flexible, requiring only that the data be in columns in a text file. However, the top of the
file must contain control data in a specific format. Consult the ANSR-L manual for details of these
lines. Each acceleration file may contain multiple components, typically 2 translational plus the
vertical component.
NOTE : ANSR files written with MODELA currently read only the first two records.
Typically, these will be the two horizontal components. Also, each file must contain
two records as a minimum.
For each record to be used (up to the Number of Time Histories set) enter the component number and
scale factor for that component for each of the global ANSR-L axes. The duration to be used for the
analysis should also be set. This will depend on the structure and record. You can use the Select
button to pick file names from a dialog box.
Generally, each record will be applied twice, the first with the dominant component along the global
X axis and the second with the dominant component along the Z axis. The component numbers
and factors will then be transposed for the X and Z axes for the next analysis.
Acceleration time histories are available in-house for a large number of recorded earthquakes in New
Zealand and overseas and many more are available over the Internet. A research project is planned
to format and categorize the acceleration time histories.

7.6.3

INCREMENTAL EARTHQUAKE

The analysis is performed for the full amplitude of the earthquake motions specified, either NZS4203
or site specific. There is an option to apply the earthquake records incrementally by specifying the
maximum amplitude, the maximum amplitude and the increment. These values are all relative to the
full amplitude.
This option is useful in constructing a fragility curve for a building, in the form of maximum base
shear versus displacement. At each point on the curve the extent of damage to the curve can be
mapped. Curves such as shown in Figure 7-5 can be used to relate the earthquake amplitude to its
return period.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

99

FIGURE 7-5 RETURN PERIOD VERSUS AMPLITUDE

1.4
1.2

RISK FACTOR, R

1.0
0.8
0.6
0.4
0.2
0.0
1

10

100

1000

RETURN PERIOD (YEARS)

7.6.4

MASSES AND WEIGHT

The Masses and Weight options define how the ANSR-L masses and weights are assembled. These
two quantities are independent in the ANSR-L model, even though in reality they relate to the same
physical property of the building.
The mass distribution is important for earthquake analyses as the inertia loads are dependent on the
masses. The weights are applied as a vertical static load prior to the start of the earthquake analysis.
The importance of correct weights depends on the element types used in the analysis. The correct
gravity loads are important, for example, to incorporate P- effects in frame structure, for columns
where yield is a function of moment and axial load, and also for gaps, where uplift will occur once
earthquake axial loads exceed gravity loads. Weights are less important for shear panels where yield
is a function of shear stress only.
The options set in this section define the procedures used to calculate and model mass and weight.
See previous discussion in Section 5.9 for recommendations relating to which of the options to use.

7.6.4.1

Include Element Weight

Element weights are generated using the material weight density and the element area or thickness.
If this option is selected then the weight of each element is distributed equally to each node of the
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

100

element and applied as part of the static gravity loads. A weight density must be specified for the
materials or the calculated weights will be zero.

7.6.4.2

Include Element Mass

Element masses are generated similarly to weights, from mass density and the element area or
thickness. If this option is selected then the mass of each element is distributed equally to each node
of the element and applied as a nodal translational mass in each of the three directions. Rotational
masses are not calculated. Masses are cumulative to masses already specified for the joint. A mass
density must be specified for the materials or the calculated masses will be zero.

7.6.4.3

Include UDL as Mass

Uniformly distributed loads may be applied to beam elements (see later in this Section). These loads
initialize element end moments and shears so that positive and negative yield will correctly include
gravity load effects. The fixed end forces are also applied to nodes at each end of the beam and part
of the gravity load case.
If this box is checked then the total load due to the UDL will be calculated and one half of the
weight, divided by gravity, will be applied as translational masses to each beam node. These are
cumulative.

7.6.4.4

Distribute Diaphragm Mass as Weight

For a regular building a quick and reasonably accurate method of specifying gravity loads is to
distribute the diaphragm mass times gravity over each node at a level. When this option is checked
the total weight is assumed distributed equally over all nodes at the same level.
The mass is converted to weight by factoring by the acceleration due to gravity and then distributed
over all nodes at the same elevation as the diaphragm which are connected to vertical elements
(columns and walls).

7.6.4.5

Lump Element Mass to Diaphragm

When this box is checked it has an effect only if the Include Element Weight box is also checked. As
element weights are calculated they are added to the diaphragm mass rather than applied directly at
the element end nodes. The element contribution is also added to the rotational moment of inertia
of the floor. For elements between floors, mass is lumped to the nearest floor.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

101

The element mass is assigned to the diaphragm node to which the element is connected, which is the
default of diaphragm 1 unless you have specified differently (Section 7.12). There are some
conditions where this procedure may not be satisfactory:
1.

For mass at nodes between floors, the node is not specified as connected to any diaphragm
and so the default of diaphragm 1 is used. In some situations, the mass may be better
distributed to another diaphragm.

2.

For mass at a floor which has been specified as connected to diaphragm 0 (effectively, a
disconnected node) then it will be assigned to Diaphragm 1.

See Section 5.9 for further discussion.

7.6.4.6

Assign Missing Mass Horizontally

Mass which is assigned to degrees of freedom which are fixed is ignored. Fixed degrees of freedom
may occur at slaved nodes for frame members and at the wall out-of-plane degrees of freedom. Mass
may be assigned to these nodes from element densities or from user-specified mass.
When this option is checked, the program calculates the total mass which is assigned to fixed degrees
of freedom and then distributes it equally to those nodes at the same elevation which have a free
degree of freedom in the particular direction. This procedure is performed separately for each of the
three translational directions.
This horizontal distribution of mass is generally preferred over the vertical distribution as it retains
the correct overturning moment due to inertia loads.
If this option is used then, on the completion of the mass assignment, MODELA executes the
option to Lump Element Mass to the Diaphragm for any masses which could not be distributed
horizontally. This will occur when all nodes at a level are slaved to a diaphragm and there are no
translational degrees of freedom other than the master node.

7.6.4.7

Assign Missing Mass Vertically

This mass assignment is for similar reasons to those listed above. However, rather than assign
masses to the nodes on the same level, the mass at the fixed degree of freedom is moved to the
nearest free degree of freedom on the same column line as the fixed degree of freedom.
This type of distribution may be most suitable for the out-of-plane nodes of shear walls where face
loads are distributed vertically to the floors above and below.
If this option is used then, on the completion of the mass assignment, MODELA executes the
option to Lump Element Mass to the Diaphragm for any masses which could not be distributed
horizontally. This will occur when there are no free nodes on the same column line above or below a
particular node.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

102

7.6.5

TIME STEP

The time step is the interval over which the acceleration is assumed constant for the direct
integration solution of the equations of motion. A general rule of thumb for the time step is to use a
value not greater than 1/100th of the fundamental period of response. For example, a time step of
0.01 seconds will be suitable for structures with a period of response of 1 second or more. This will
capture modes with a period of 10 times the time step or more accurately (an 0.01 second time step
includes modes of periods of 0.10 seconds or more) and of 5 times the time step with less accuracy
(0.05 seconds).
Generally a time step of not greater than 0.01 seconds is used and this is suitable for most structures
other than very stiff wall type structures. The ANSR-L program does not solve for the natural
periods. These can be obtained by performing a Zap/Sap analysis (see Section 9.1) or indirectly by
running a single time history, processing displacements and plotting displacements at the roof level.
An estimate of the natural period can be made by measuring the time interval between zero
displacement crossings on the displacement plot (see also Section 12.1).
Models with elements which stiffen, such as gap elements, may need a finer time step than that
estimated above to ensure that unbalanced loads do not become too large. The time step for a
model with many gap elements may need to be found by trial and error. The definitive check is to
continue reducing the time step until the difference in peak responses from subsequent analyses is
less than a specified tolerance, generally about 1% to 2%. In practice, it is more likely that the
maximum time step is the value at which numerical instability does not occur. That is, the longest
step for which the convergence cannot be achieved message is not received.

7.6.6

MAXIMUM DISPLACEMENT

The Maximum Displacement provides a means of stopping the analysis once the building becomes
unstable. The ANSR-L program will continue solving the equations of motion until displacements
exceed the maximum value which can be represented in double precision arithmetic (10308). In many
cases, a lot of computer time will be wasted continuing an analysis long after the structure would
have failed.
The analysis will stop when the displacement at any nodes exceeds the value entered in this cell.
The displacement exceeded termination is sometimes a local effect due to a fictitious or real local instability
see Section 8.10 for strategies to deal with local instability.

7.6.7

DAMPING

These cells provide for two periods and two damping values to be entered. The periods are in
seconds and the damping is a fraction of critical damping. These parameters are used to calculate the
viscous (velocity dependent) damping for the analysis. This damping is in addition to the hysteretic
damping implicit in the force deformation functions for the yielding elements.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

103

The damping fraction is generally assumed to be 0.05 times critical. This is the value in codes such as
NZS4203 and UBC. The response spectra in these codes are based on 5% damping. This value is
appropriate for most materials stressed near the yield level.
Damping less than 5% may be applicable for materials with low inherent damping, such as welded
steel or prestressed structures, if they are used for bare frames. However, for most buildings nonstructural components such as cladding and partitions will generally increase damping so that 5% is
appropriate even with these materials.
The viscous damping specified in ANSR-L is Rayleigh damping where the damping matrix, [C], is
constructed from the mass matrix [M] and stiffness matrix [K] as

[C] = [M] + [K]


where and are user-specified coefficients. These two constants may be calculated for any two
periods of response, T1 and T2, which have associated viscous damping ratios, 1 and 2 as:

4 (T1 1 T2
(T22 T12 )

T1T2 (T2 1 T1
(T22 T11 )

The use of two constants, and , allows damping to be specified exactly at two periods. At all
periods between these two periods the damping will be less than specified and for periods outside the
range of these two periods the damping will be greater than specified. At any period, T, the viscous
damping can be calculated as:

aT
+
4
T

The worksheet Damping calculates this formula for the damping factors you specify in this section
and plots the damping versus period function.
Figure 7-6 shows the total damping, plus the components from mass and stiffness respectively, for
an example structure where and have been set to provide damping of 5% at periods of 0.05
seconds and 1.0 seconds. The mass damping increases with increasing periods whereas the stiffness
damping decreases with increasing period. The correct damping is applied at the two periods
selected. The minimum damping applied is approximately 2.1% at a period of 0.2 seconds.
As shown in Figure 7-6, the damping increases rapidly for periods less than, or greater than, the
specified values. The longer period should allow for period lengthening due to yielding. The shorter
period should be set approximately equal to the shortest period for which the response can be
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

104

captured by the time step, say 5 to 10 times the time step or 0.05 to 0.10 seconds for the usual time
step of 0.01 seconds.
If the range of periods selected is too wide then the higher modes may be under damped. For
example, for frames and wall structures the second mode period is generally in the range of 1/3 to
1/5 the first mode period, or from 0.2 to 0.33 for the structure with damping shown in Figure 7-6.
The mode in this range will be under damped and a more accurate response would probably be
obtained by increasing the lower period specified for damping from 0.05 to 0.10 seconds.
The coefficients are generally applied as scalar quantities, that is, the same coefficient is applied to all
components of the mass and stiffness matrices. ANSR-L does have the capability of varying the two
coefficients so that different damping is generated by different mass points and elements. It is quite
complex to solve for the vector values of {} and {} for these situations although Modela
procedures do provide some facilities for this see Section 7.6.14. If you need to use non-standard
damping, be careful to ensure that values are appropriate to your model.

FIGURE 7-6 RAYLEIGH DAMPING

10%
Mass Damping
Stiffness Damping
Total Damping
Target Damping

9%
8%
7%
6%
5%
4%
3%
2%
1%
0%
0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

105

7.6.8

COORDINATE SORT ORDER

The coordinate sort order sets the order of numbering for the axes. It should be set so as to
minimize the memory requirements and execution times. The larger the model the more important
this becomes. For most structures, Z (the vertical axis) will always be first. See Section 8.5.3 for a
discussion of this item.

7.6.9

FORCE TOLERANCE

The force tolerance is used to specify the maximum unbalanced force, as a fraction of the total
weight of the structure, during the ANSR-L dynamic analyses. If the unbalanced force at any time
step exceeds this value then iterations within the time step, up to a maximum of 5, will be performed
in an attempt to reduce the force below the tolerance. If convergence cannot be achieved after the
specified number of iterations the program will terminate with an error message at the end of the
.D01 file.
The force tolerance is generally set between 0.01 and 0.10 of the weight, with the larger value for
structures with a large potential number of plastic hinges and/or stiffening elements such as gaps. In
these models the total overshoot in a time step may become large without necessarily indicating
numerical stability. If the convergence not achieved message is received for a value of 0.10 then the time
step is probably too large.
See Sections 8.10.2 and 8.12 for further discussion.

7.6.10 NODE TOLERANCE

This value is used by MODELA when generating the ANSR-L models. If the location of a nodal
point is within this tolerance of another node it is assumed that the nodes are in the same location.
This may be useful is some circumstances, for example if node locations are generated using
equations in EXCEL and there are rounding errors when the same node is generated from two
directions.

7.6.11 OUT OF PLANE WALLS

The 4 node plane stress elements have in-plane translational stiffness only. If the elements (1) span
from floor to floor, (2) have plane stress elements connecting at right angles or (3) are connected to
flexural elements then the out-of-plane degrees of freedom will be supported. If none of these
conditions apply then the out-of-plane stiffness must be restrained some other way. This section
provides options available to apply this restraint:
1. Fix the degree of freedom. This procedure can be used provided the element is parallel to one
of the global ANSR-L axes. If it is not parallel, fixity along any global axis will provide some
degree of in-plane restraint. For structures with a rectangular grid layout this is the simplest and
most efficient method. An early disadvantage was that the animated on-screen displays would
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

106

draw the deformed shape with zero displacement at these nodes, which distorted the picture.
ANSR-L has now been modified to plot the displacement as the average of those above and
below the fixed node, alleviating this disadvantage.
2. Slave the degree of freedom to the upper or lower degree of freedom. If this option is selected
then MODELA will select the nearest node on the same column line which has the particular
degree of freedom restrained by one of the three conditions above. The displacement at the outof-plane node in the unrestrained direction will then be slaved (set to be equal to) the
displacement of the restrained node. This has the advantage of improving the appearance of the
animated shape. If has the disadvantage of increasing the memory requirements of the model,
and probably also increasing the execution times. As for 1. above, it is only suitable for elements
parallel to the global axes. With the modification to option 1. above, this option is now
obsolete.
3. Model a spanning column. With this option, MODELA inserts columns spanning between the
unrestrained node and the nearest nodes above and below which are restrained in the out-ofplane direction. The column will be linear elastic, with properties based on the wall thickness
and the panel length. Ends of the column will be pinned. This is the most complete model of
the actual situation and ensures accurate animated displays. However, it also adds the most
complexity to the model. This increased complexity will be noticeable in memory requirements
and execution time. However, it is the only method suitable for panels which are not parallel to
the global axes.
4. Fixed or spanning column as required. This is a combination of 1. and 3. whereby the degree of
freedom is fixed if the panel is parallel to the global axes and a spanning column is added if this
condition is not met. This is the most efficient solution for models not based on a rectangular
grid and is generally recommended.
If you will be running your programs in batch mode, use the most efficient solution, Option 1 if a
rectangular grid else use Option 4. If you want to observe the animated display, or will be making
an .AVI file, use Option 3. In the latter case, after you have completed the animation it is usually
more efficient to re-generate the ANSR-L files using a more efficient option to get the response for
the remaining analyses in batch mode.
The best option for analysis efficiency and accuracy is Option 4, which fixes the degree
of freedom where possible and uses spanning columns only where necessary. Option 3
may be needed to display animations or make AVI files. Although every effort has
been made to ensure that the model with this option will be equivalent to the
recommended option, the model is more complex and so more prone to errors. Do
not use these options for the analyses which you use to perform the final evaluation of
the structure.
7.6.12 TIME HISTORY OUTPUT

Time histories of nodal and element responses may be generated. These check boxes determine
which time histories are selected for a particular series of analysis.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

107

Diaphragm nodes, if selected, will be output at all floor center of mass nodes as defined in the Floors
sheet unless some elevations are suppressed (see Section 7.7). If the User Specified Nodes box is
checked, the nodes output are those specified on the Elevations & Nodes sheet (see Section 7.7).
Elements time histories will be output only for elements selected on the appropriate element sheet.

7.6.13 SPECIAL FEATURES

This series of check box will be used as additional features are added to the procedure. So far, there
are a number of options:
1. SAVE each run for re-start. This option will configure the ANSR-L input file so that the data is
saved to a disk file for re-start if the <Esc> key is pressed during any analysis. This option may
be useful for very large models. See Section 8.6.
2. Number the degrees of freedom from the top of the model (default is bottom up). This option
may improve memory requirements and/or execution times for some models. See Section 8.5.
3. Slave degrees of freedom active. This option is obsolete.
4. Include connectors in DXF. This option will draw lines from the center of mass of each floor
node to the nodes which are connected to the particular floor node. This is used to debug the
model. See Section 7.16.
5. ZAP Masonry model. This is a special purpose type of analysis which is not documented in this
Manual as it has limited application.
6. Use X in batch file. This option will set the ANSR run mode to display the animated deformed
shape. See Section 8.6.
7. Ignore gravity loads. Use this option for models where gravity may be causing undesirable
effects and gravity loads are not important to the analysis. For example, masonry buildings
modeled without gap elements are not influenced by gravity loads. However, gravity
deformations may cause unexpected effects such as cracking of lintels before any earthquake
loads are applied.
8. Use SAP instead of ZAP. This option is only used for Force Pushover. This develops an input file
for SAP2000, the preferred option if you have this program installed.
Other options will be added to this group as they are implemented.

7.6.14 NON-STANDARD DAMPING

The ANSR program has more flexibility in specifying damping that is reflected in MODELA input.
In particular, the mass damping may be varied between different degrees of freedom and the
stiffness damping factors may be varied between different element groups. This section allows the
generality to be accessed.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

108

Use of these features requires caution. The non-standard damping option on this sheet permits
damping to be varied at different nodes and damping to be varied with different materials. The
most common use of these factors is for water retaining structures, where the mass representing the
water sloshing mode has lower damping than the remainder of the structure (water damping is only
%, compared to 5% normally assumed for structural materials).
7.6.15 UBC FACTORS

Preliminary development has been started to incorporate UBC evaluation as an alternative to


NZS4203. As this is completed, this Manual will be revised to detail the procedure. The UBC
factors specified on this sheet are used to define a response spectrum which is written to the ZAP or
SAP2000 files for the Force Pushover option.

7.7

ELEVATIONS AND NODES

The Elevations & Nodes worksheet defines the geometry of the model, as a series of column line
locations located on the horizontal plane and a number of elevations located on the vertical plans.
NOTE: If a node is placed within the length tolerance (default 5 mm) of a floor master
node then the floor node will be moved by the length tolerance in the X and Y
directions. This is to avoid the possibility of a zero length element being specified to
connect a slave node to a master node. If this change may be significant for your
structure, set the length tolerance smaller so the node will be moved by a lesser amount.
This sheet contains column of data which is used to define the geometry, as described below. There
are additional columns on the sheet which can be used to input nodal increments, conversions from
other units etc. You can add columns as required but do not add columns within the blocks of
columns with specified data (for example, do not add a column between X and Y coordinates but
you can add as many columns as you like to the right of Support Level.

7.7.1

COLUMN LINES

The column lines are entered in the labeled columns and rows on the sheet, as shown in Figure 7-7.
Data required to describe the columns are the column number and the X and Y coordinates of the
column line. Column numbers must be a valid integer number. The numbers can be in any order
and need not be consecutive. If column numbers are repeated, the last definition will be used.
Optional data items are the column angle in degrees (default 0) and support level (default 1). The
column angle is used to define the stiffness of column sections relative to the global axes. For
example, if a rectangular column is used with different orientations at different column lines then
some columns will be at 0 and others at 90. A single section definition can then be used. An
alternative method to model this would be to set all columns to the default of 0 and define two
different sections, with B and D reversed.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

109

FIGURE 7-7 COLUMN LINE INPUT

COLUMN LINE COORDINATES


Note : Set support level -ve to use pinned support.
COLUMN LINE COORDINATES
45
X
Y
ANGLE
Support
COLUMN
LINE
(degrees)
Level
1
0.000
0.000
2
7.500
0.000
3
15.000
0.000
4
22.500
0.000

The support level is the level at which the column line is assumed to be fixed. The default is level 1.
The node at the support level is assumed fully fixed (all 6 degrees of freedom). If the support level is
negative then the column line at the absolute value of this level is assumed to be pinned (restrained in
the three translational directions but free in the three rotational directions).
Additional columns are provided adjacent to the column line coordinates. These may be used for
intermediate steps, for example incrementing column coordinates or converting units. Extra
worksheet columns can be added if required between the Support Level column and the Level Number
column.

7.7.2

ELEVATIONS

Required data for elevations is the Level Number and the Elevation. The level number must start at 1
and increment by 1. The elevation must be a valid number and the elevation for a level must be a
larger number than for any lower-numbered elevation. Elevations must differ by a distance at least
equal to the value specified for the node tolerance (Section 7.6.10).
Optional data for elevations is the Name, the X Eccentricity and the Y Eccentricity. The level name, if
supplied, will be used to identify output quantities such as floor displacements. Although unlimited
in length, The names are generally 3 to 8 character descriptors, for example, L02 or LEVEL2. Level
names should not contain embedded blanks.
The X and Y eccentricities are accidental eccentricities to be applied to building masses to
incorporate torsional effects. The eccentricities specified in this section apply only to the local
masses, that is, masses from element densities and user specified nodal masses. Eccentricities of rigid
diaphragm masses are specified on the Floors sheet.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

110

FIGURE 7-8 ELEVATION INPUT

ELEVATIONS
e pinned support.
ELEVATIONS
LEVEL
NAME
ELEVATION
NUMBER
1
GROUND
0.000
2
1st
4.570
3
2nd
8.220
4
3rd
11.870
5
4th
15.520
6
5th
19.170
7
6th
22.820
8
7th
26.470
9
8th
30.120
10
9th
33.770
11
10th
37.420

X
Eccentricity

Y
Eccentricity

>0 to
Omit
Element
Time History

This portion of the input is also used to define whether to suppress story shear output for any
particular levels. This is only relevant if an element output code of 99 is used to obtain story shear
summations (see sheets for each element type). If you have a large number of elevations then the
story shear output file can become large. Suppressing some levels reduces the amount of data stored.
Note that the elements are assigned to story levels based on the level at the lowest node of the
element. If you want story shears in the Roof story then the code would need to be 0 at the Fourth
level to obtain summation of elements spanning from Fourth to Roof.
7.7.3

TIME HISTORY OUTPUT

The time history output cells are optional and are specified if time histories are to be saved at nodes
other than diaphragm master nodes (This box must be checked on the Analysis Data sheet see
Section 7.6.12).
If data are supplied, each row must contain valid column and level numbers and the axes about
which the time histories are required. The axes are specified as single characters referring to the
ANSR-L axes (that is, Y is vertical). For example, the string XZ in the Axes column would produce
output along both horizontal axes for the specified column and level. See Figure 7-9, which specifies
time history output at column line 15 in both horizontal directions for levels 4, 7 and 10.
There is no limit on the number locations for which time histories can be obtained but for a large
number of nodes the output file can become very large. Processing time may also become
substantial. Generally, output should be obtained for no more than about 20 nodes in total, that is,
the sum of diaphragm nodes plus user selected nodes. In special cases, more may be requested but
usually for only a limited number of analyses.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

111

FIGURE 7-9 TIME HISTORY OUTPUT SPECIFICATION


TIME HISTORY OUTPUT
LEVEL
AXES
COLUMN
NUMBER NUMBER
15
4
XZ
15
7
XZ
15
10
XZ

7.8

MODEL PLOT

This Model Plot worksheet is used to obtain a plan view plot of the model geometry, optionally
including the column and/or element numbers. The plot uses Excel drawing tools and is relatively
primitive. It is intended as a quick intermediate step. The AutoCad drawings (Section 6.16) are the
main tool for final checks of model geometry.
The worksheet has a single button, labeled UPDATE, in the upper left corner. If this button is
activated then the dialog box shown in Figure 7-10 will be displayed. An example of the type of plot
produced is shown in Figure 7-11.
Paper size is generally set to A4, but you can use A3 if you have a printer capable of printing this size
and have a complex model.
For any plan view of a model with a moderate or large number of column lines it is difficult to get
column and/or element numbers displayed so that the values are legible. Use trial and error to set
the font size and also the horizontal and vertical offsets. These offsets are multiples of the font size
and determined where the column or element number is written relative to the position of the node
point or the line representing the element.
The highest and lowest levels to include in the model are set by the dialog box. These may be set so
as to plot a single level or multiple levels. The range set depends on the purpose of your plot and
complexity of the model. Only elements within this level range will be displayed.
For each element type options are available to omit or plot the element. Each element type can be
plotted as a line, or as a line with the ID number or type number also shown on the plot. Select the
appropriate option boxes for the type of elements used in your model.
After the element selections there are a number of general check boxes. These are used to plot node
locations, with or without identification numbers. (If the Plot Node Locations box is not checked then
only selected element types will be included in the plot). Options are also available to show

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

112

diaphragm nodes and the connectivity of diaphragms. These last two options are best used with
display of a single level at a time.

FIGURE 7-10 DIALOG BOX FOR MODEL PLOT

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

113

FIGURE 7-11 EXAMPLE OF MODEL PLOT

446

447

448

449

450

444

440

420

441

421 422

401 402

7.9

403

423

404405

424425 426

427

406 407
408

453

442 452

428
429

430
431 432

409 410

411

433

412

434

435

413
414 415

445

443

436 437 438

416 417

451

418

439

419

MATERIALS AND SECTIONS

This section is the most important to the process of the non-linear analysis as it describes the
materials, sections sizes and the post-yield performance of the elements, including any stiffness
and/or strength degradation. See Chapter 6 for a detailed description of the characteristics of the
material and section models.
The materials characteristics are constantly being refined, modified and extended. The descriptions
in this manual will be updated as this process continues.
All materials are stored together in the MODELA program (that is, masonry, concrete, steel etc) and
so material numbers must be unique. If masonry materials are numbered 1 to 3 then no concrete,
steel, timber or other materials can have one of these three numbers. Gaps in the number sequence
are permitted.
Similarly to materials, plate and beam/column/brace section numbers must be unique and gaps are
permitted. Use a numbering system which is convenient for your project, for example, plate
sections numbered from 1,2 3 etc and beam/column/brace sections numbered from 101, 102, 103
etc.
You can add extra rows for any material or section type as required and you can also delete existing
rows if you have only a few types. However, always leave at least the first row for each material and
section type.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

114

7.9.1

MASONRY

Masonry input data is as shown in Figure 7-12. The meaning of the values in each column is as
follows:

ID number is a unique material identification, a valid integer number not used for any other
material.

fm is the masonry compressive strength, in KPa. If not specified the default is 10,000 KPa.

G is the shear modulus. Leave blank to use the NZNSEE Red Book properties for masonry.
Set to 1, -2 or 3 to use FEMA properties corresponding to good, fair or poor masonry
respectively.

is Poissons ratio, set to 0.25 if omitted.

vc is the cracking shear strength of the masonry. If omitted, cracking strength from NZNSEE or
FEMA as appropriate will be used.

is the coefficient of friction, multiplied by the dead load stress to obtain the increase in shear
strength due to gravity.

w is the weight density of the masonry, default 0.0.

m is the mass density of the masonry, default 0.0.

alpha, beta, xkk and dkode are degradation coefficients, as defined above. If omitted, a default
masonry hysteresis shape will be used. This is recommended for unreinforced masonry.

FIGURE 7-12 MASONRY INPUT

MASONRY
ID

f'm

vc

Alpha

Beta

Xkk

Dkode

The defined masonry properties are intended for unreinforced masonry. Reinforced masonry
generally behaves more like reinforced concrete than unreinforced masonry and is usually modeled as
a concrete material with appropriate modifications to properties.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

115

7.9.2

CONCRETE

Figure 7-13 shows the input fields for concrete elements. If the material is to be used for flexural
elements only, then only the concrete strength fc need be supplied. The other concrete properties
will be calculated from this. However, if the concrete is to be used for panel elements (walls) then
strengths and degradation coefficients need to be supplied.

FIGURE 7-13 CONCRETE INPUT

CONCRETE
ID
1

f'c
30000

vc
1000

vs
500

w
24

m
2.45

Alpha

Beta

Xkk

Dkode

The specific data items listed in Figure 7-13 are:

The ID number is a unique material identification, a valid integer number not used for any other
material.

fc is the concrete compressive strength, in KPa. If not specified the default is 25,000 KPa.

E is the elastic modulus. Leave blank to use the NZS3101 calculated value, which is 6900 +
3320fc.

is Poissons ratio, set to 0.20 if omitted.

vc is the cracking shear strength of the concrete. Default is 500 KPa.

vs is the shear strength provided by the reinforcing. Default is 500 KPa.

w is the weight density of the concrete, default 0.0.

m is the mass density of the concrete, default 0.0.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

116

alpha, beta, xkk and dkode are degradation coefficients, as defined above. These are only required if
the material is to be used for panel elements. If omitted, a default hysteresis shape will be used.
This is not recommended for concrete so values should always be included here.

7.9.3

STEEL

Steel input parameters are as shown in Figure 7-14. As for the other material types, the steel
materials must each have a unique identifier not shared by any other material.
As steel properties are usually constant, the only input parameter required is the yield stress, Fy,
which defaults to 430,000 KPa. The elastic modulus defaults to 200,000 MPa and Poissons ratio to
0.20. Default values for the weight and mass density are zero. The Strain Hardening Ratio for steel
(Strain H.R in Figure 7-14) is used only if steel reinforcing is used to model wall flexural yielding
(with gaps at each end of the wall see Section 5.3).
A zero SHR is appropriate for mild steel for relatively small elongation (up to about 10 times yield
strain). For elements strained beyond this, a positive SHR in the range of 0.025 to 0.05 may be
appropriate. As peak strains cannot be determined prior to the analysis, some iteration may be need
if this parameter is important.
The input fields provide for degradation coefficients for steel the same as for concrete. It is unlikely
that these would be used.

FIGURE 7-14 STEEL INPUT

STEEL
ID
2

7.9.4

Fy
430000

Strain H.R.

Alpha

Beta

Xkk

Dkode

TIMBER

Timber input parameters are as shown in Figure 7-15. As for the other material types, the timber
materials must have unique identifiers not shared by any other material. Timber does not have
default values of E or except that, if E is specified as an integer less than zero, FEMA properties
will be used. If vc is blank or zero then the material will be linear elastic.
If a shear capacity, vc, is specified the timber will be modeled as a bi-linear degrading strength and
stiffness material. The strength will degrade to 10% of the yield strength, vc, at a strain equal to 10
times the yield strain.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

117

FIGURE 7-15 TIMBER INPUT

TIMBER
ID
3
4
5

E
-1001
-1001
-1001

vc

w
0.5
2.5
2.5

m
0.05
0.25
0.25

Alpha
0.50
0.50
0.50

Beta
0.00
0.00
0.00

Xkk
2.67
2.67
2.67

Dkode
0.40
0.40
0.40

If a negative integer is entered for the elastic modulus, E, then the properties of the timber will be
selected from the FEMA diaphragm options listed in Figure 7-16.

FIGURE 7-16 TIMBER DIAPHRAGMS


TIMBER DIAPHRAGMS (Uses NEHRP Values)
Type

Chorded Unchorded

Single Straight sheathed


Double Straight Sheathed
Single Diagonal
Diagonal + Straight
Double Diagonal
Structural Panel (Blocked)

-1
-3
-5
-7
-9
-11

-2
-4
-6
-8
-10
-12

Structural Panel (Unblocked)

-13

-14

For Types -1 to -10 strength is calculated from NEHRP.


For Types -11 to -14 enter shear strength as vc.
Values assume a diaphragm effective thickness of 25 mm.
Structural Panels
Nail Size (mm)
Spacing (mm)
q (KN/m)
Ultimate Strength (KPa)
(for 25 mm thickness)

4.00
100
8.04
515

Maximum Nail Sizes


Thickness <=7.5
2.80
<=12.5
3.3
15+
4.00

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

118

As discussed in Section 6-7, it is recommended that timber floors be modeled with an elasticperfectly plastic hysteresis rather than degrading strength. To activate this option, subtract 1000
from the value of E (for example, for diaphragm Type 1, enter 1001). An option is also available
to model the floor as linear elastic by subtracting 100 from the value of E but this will not limit
maximum forces and so is not generally recommended.
The FEMA values for a 25 mm thickness can be adjusted to NZNSEE recommended values for
floors by adjusting the thickness. The FEMA strength for type 1 corresponds to a stress of 71
KPa. The effective model thickness can be calculated by dividing the NSNSEE strength in KN/m
by 71. For example,

Strength of 6 KN/m is specified by the NZNSEE Red Book for floors with straight tongue and
groove sheathing or for roofs with straight sheathing and roofing applied directly to the sheathing. The
appropriate thickness would be 6/71 = 0.085 m.

Strength of 4 KN/m is specified for timber stud walls with wood or metal lathe and plaster. The
appropriate thickness would be 4/71 = 0.056 m.

In summary, the recommended input for timber floors is as shown in Figure 7-15. This will provide
a model which will maintain strength but in which the stiffness will degrade. The floor acceptance
criteria are then based on limitations for elements supported by the floor.

7.9.5

OTHER MATERIALS

The Other Materials option is intended to model plate elements with non-defined properties. For
example, composite overlays have been modeled using this material type. Figure 7-17 shows the
spreadsheet input fields.

FIGURE 7-17 OTHER MATERIALS INPUT


OTHER
ID

E1 E2 E3 P1 P2 P3 GAB Angle Fy
Alpha Beta Xkk Dkode G1 S1 G2 G2 G3 S3 G4
S4 G5 S5 G6 S6 G7 S7

The material specified here is intended to include both flexural and panel type elements. For the
former type, only E plus either G or need be specified. For panel elements, a series of stiffness
values (G1, G2, G3 etc) and strains (S1, S2, S3 etc) are used to define the envelope stress-strain
curve. The type of stiffness and strength degradation is defined by the factors alpha, beta, Xkk and
Dkode , as described above for masonry and concrete. See Chapter 6 for further details.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

119

The default for 4 node elements is to use the special purpose degrading strength and stiffness
element described in Chapter 6. This is ANSR-L element Type 8. The original ANSR-II 2D Plane
element can be used by specifying a negative value for Dkode. This element has four material types:
1.

Isotropic linear elastic material. Specify E and only.

2.

Orthotropic linear elastic material. Specify three modulus of elasticity values EA, EB, EC and
three Poissons ratios, AB, AC and BC.

3.

Elastic-perfectly plastic material with von Mises yield criterion. Specify E, and y.

4.

Mroz material model this is undocumented and superseded by Type 8.

If you need to use these finite element models then consult the ANSR-II User Manual or the
Development Division for advice.

7.9.6

PANEL SECTIONS

Panel sections are used to model walls or flexible floors using triangular or quadrilateral elements.
The properties of the individual panels are specified using the spreadsheet fields listed in Figure 7-18.
The specific data items listed in Figure 7-18 are:

ID number is a unique section identification, a valid integer number not used for any other
section type.

The material identification. This must correspond to a material ID specified on this sheet.

The reinforcing ratio, for concrete panels only. This is the vertical reinforcing ratio. At this
revision, is used only to calculate properties for flexural reinforcing.
If the horizontal
reinforcing varies, different panel types need to be specified for each value of vs, which is defined
as part of the concrete material definition.

The panel thickness. This defines the element stiffness.

The thickness for weight. If blank, assumed to be zero. Usually this will be equal to the
thickness but sometimes extra will be allowed to include the weight of non-structural finishes etc.

The reduction factor, which is applied to the element thickness and so reduces the stiffness for
both shear and flexure. Any value applied here is not applied to the weight, which is always
based on the unreduced thickness. The default reduction factor is 1.0.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

120

FIGURE 7-18 PLATE ELEMENT INPUT

PLATE SECTIONS
ID

7.9.7

MATERIAL REINFORCING THICKNESS THICKNESS REDUCTION


ID
RATIO
FOR
FACTOR
WEIGHT

BEAM/COLUMN/BRACE SECTIONS

Section properties for all flexural members are entered in two portions of the spreadsheet, as shown
in Figure 7-19. The first portion defines the section dimensions and stiffness properties, the second
portion the strength properties. A single section size may have several different strengths (for
example, different reinforcing content where used in different locations). Flexural members in the
model are defined by the section type and the strength type at each end.

7.9.7.1

Modeling Considerations for Flexural Elements

Stiffness
The current model developed using MODELA uses a bi-linear stiffness for the flexural elements.
An actual reinforced concrete element will have three stiffness values, representing the uncracked,
cracked and yielded states. The uncracked state generally applies only for the first cycle and to
relatively small amplitudes and so is neglected for the analysis. Therefore, the stiffness values
specified correspond to the cracked stiffness and the yielded stiffness.
The cracked stiffness can be based on code requirements (for example, the Commentary to
NZS3101) or calculated (using moment curvature spreadsheets or formulas see Priestley, 1998).
The NZS3101 values for columns are a function of axial load. As the axial load may vary during
earthquake motions, it is recommended that stiffness values be based on the gravity load stress in the
column.
The evaluation phase (Section 12) requires that the effective stiffness be consistent with the stiffness
in the moment curvature spreadsheet used to calculate the ultimate curvature.
Development work is continuing in this area. Pending resolution of this, the following effective
stiffness values are recommended for flexural elements:

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

121

For all beams, use 0.4Ig and base properties on a rectangular section (neglect T or L beam action.
Use between 0% and 1% strain hardening.

For columns, use 0.3Ig to 0.6Ig with the higher factor applying to well confined sections.

Use 0% to 1% strain hardening for columns with relatively low axial loads, 0% for moderate to
high axial loads.

These are guidelines based on current knowledge and may be revised as further research data come
to hand regarding effective stiffness.
Strength
Generally, time history analyses performed to obtain maximum inelastic displacements and ductility
demands are based on probable element strengths, which are calculated based on specified minimum
material strengths and a capacity reduction factor of unity (= 1.0). This is the recommendation
from the NZS4203 Commentary.
In some circumstances it may be desirable to check the effect on response of overstrength in some
elements. For example, beam overstrength will influence the extent and distribution of column
hinging. In frame structures, it is common to perform two series of analyses, the first using probable
strengths throughout to obtain deformations and the second using beam overstrength to obtain
maximum demands on columns (or column overstrength for a column hinging mechanism).
Shear demand is a function of member flexural strength and generally overstrength capacities are
appropriate for determining maximum shear forces. However, in the current procedure the shear
capacity is not incorporated into the non-linearity. The shear capacity is checked against demand
using envelope results from the analyses. This permits the effects of flexural overstrength to be
incorporated into the evaluation of the response against acceptance criteria (see Section 12).

7.9.7.2

Flexural Element Stiffness Properties

Data required to define the dimensional properties (Figure 7-19 (a)) are as follows:

ID number is a unique section identification, a valid integer number not used for any other
section type.

The material identification. This must correspond to a material ID specified on this sheet.

The section type, which may be one of a number of pre-defined shapes or may be O (signifying
Other). The currently defined shapes are listed at the end of this section of the worksheet.

Primary dimensions of the section, in the X direction and Y direction. For columns, these
dimensions are parallel to the X and Y axes used to define coordinate locations. If a column
angle has been specified the dimensions are rotated by this angle. For beams, the X dimension is
the depth below the floor level and the Y dimension is the beam width.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

122

The next dimension field in used only for beam sections which have an upstand above the floor
diaphragm level. In this case, the total beam depth is the sum of the data items in Columns E
and G. The definition of beams in these two segments does not affect the stiffness but does
affect the display in the AutoCad drawing and the rigid end zones of columns framing into the
beams.

Two further dimension fields, Tfl and Tw, are used to define section shapes such as RHS, T etc,
as shown in the legend below the data input fields.

If type Other is specified, stiffness properties A, Ixx, Iyy, J, Avx and Avy are input. These are related
to the X and Y axes as for dimensions above. If the type is one of the pre-defined section types
then these fields can be used to over-ride the calculated values. In this case, one or more
parameters can be entered. Only the non-zero values will be used to over-ride calculated values.
For example, if the area, A, is entered but the next four fields are blank, the calculated area will
be replaced by the value entered but the calculated Ixx, Iyy etc will be used.

Reduction factors may be specified for both the area and the moment of inertia. The default for
these values is 1.0, that is, no reduction is applied to properties based on the gross section (or the
properties specified, if any).
NOTE : If you are including degradation of beams and/or columns in your model then
you need to specify the member dimensions, even if you are using type Other and
inputting properties. This is because the plastic hinge length is specified as a factor to
apply to the element dimension.

The section properties may be selected from defined sections by double clicking from the Type cell
for a particular section number. The menu shown in Figure 7-19 (b) will be displayed with the
choices shown. If you select type Other then you have the choice of Manual Input (existing values will
be cleared) or picking a steel section from the HCG database of steel sections. If you select a steel
section the properties will be written to the Other cells and the section names will be written to
Column A.
7.9.7.3

Flexural Element Strength Properties

The flexural element strengths are used for braces, beams and columns. The number of fields used
depends on the type of element (Figure 7-19 (c)):

A brace (pin ended member) requires axial (compression and tension) strength only.

A beam requires positive and negative moment strengths only. The positive moment is the
sagging moment (bottom steel yielding) and the negative moment the hogging moment (top steel
yielding).

A column requires both axial and moment strengths. For a column, the moment strengths are
the balanced moments about the X and Y axes respectively.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

123

FIGURE 7-19 (A) : FLEXURAL ELEMENT SECTION STIFFNESS


BEAM COLUMN BRACE SECTIONS
ID MATERIAL
ID
1
2
3
4
11
12
13
14
15

1
1
1
1
1
1
1
1
1

TYPE X Dimension Y Dimension Beam


(See
(Beam D
(Beam
Depth
Above
Notes)
Below)
Width)
R
0.700
0.700
R
0.600
0.600
R
0.600
0.600
R
0.500
0.500
R
0.600
0.500
R
0.600
0.500
R
0.600
0.400
R
0.600
0.400
R
0.600
0.400

Required for Type OTHER


Area Ixx Iyy J Avx Avy Reduction Factors Steel
A
Area
Inertia Section
Tfl Tw
1.0
1.0
1.0

0.4
0.4
0.4

1.0
1.0
1.0

0.4
0.4
0.4

NOTES
1

Valid Types are:


CIRCLE Circular (D only required)
RECT
Rectangular (D and B required)
PIPE
Pipe (D and Tw required)
BOX
RHS (B, D, Tw and Tfl required)
I-SECT (B, D, Tw and Tfl required)
L-SECT (B, D, Tw and Tfl required)
T-SECT (B, D, Tw and Tfl required)
C-SECT (B, D, Tw and Tfl required)
sectname Valid section name from Database sheet
OTHER Any Other type - enter A, Ix, Iy, J, Avx, Avy (Avx and Avy are optional)

(B) DEFINED STIFFNESS PROPERTIES


OTHER
RECT
CIRCLE
PIPE
BOX
I-SECT
C-SECT
T-SECT
L-SECT

'USER' Library:
Manual Input
125TFB
100TFB
760UB244
760UB220
760UB197
760UB173
760UB147
690UB140
690UB125

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

OK
Cancel

124

(C) : FLEXURAL ELEMENT SECTION STRENGTH

BEAM/COLUMN/BRACE STRENGTHS

ID

Compression
Strength

Tension
Strength

1
2
3
4
11
12
13
14
15

12487
9670
9099
7112
10000
10000
10000
10000
10000

2334
2334
1621
2118
10000
10000
10000
10000
10000

X Axis
Axis for
Columns
+ve
Moment
Strength
1521
1047
933
672
587
489
480
309
309

Y Axis
Axis for
Columns
-ve
Moment
Strength
1521
1047
933
672
667
541
560
480
480

Strain Buckling Stiffness


Hardening Code Degrading
Ratio
>0 Yes

For columns, the yield moments about each axis and the axial load are coupled. The interaction
diagram was calculated based on the nominal material strengths.. The interaction between bending
moments and axial load is defined by:

My

M
yu

Mz
+
M
Zu

F Fo
+

Fu

= 1.0

where
1
Fo = ( Fut Fuc )
2

Fu =

1
( F + Fuc )
2 ut

My, Mz and F denote bending moments about the element y and z axes and the axial force
respectively. Subscript u denotes ultimate. Fut and Fuc are axial ultimate strengths in tension and
compression.
The flexural elements are currently modeled with a bi-linear force-deformation relationship. (The
ANSR-L elements can model tri-linear relationships but appropriate input has not yet been
developed for this). The elements can be strain hardening after yield, in which case a strain
hardening ratio is input. This ratio is the factor by which the initial, elastic st6iffness is multiplied to
obtain the yielded stiffness. The default is zero.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

125

For brace elements (pin ended), the ANSR-L program provides a buckling code which determines
the post-yield behavior of the elements. A code of 1 indicates that the member buckles at the
specified compressive strength.
NOTE : The Buckling Code in the strength section of the spreadsheet is no longer used
by Modela. The options for buckling code are set on an element basis in the Braces
sheet.
The stiffness degrading code can be set to a positive integer if the element is to have a degrading
strength and/or stiffness. The value refers to the set of degradation coefficients as defined in the
following sub-section.

7.9.7.4

Flexural Element Degradation

Procedures for incorporating flexural element strength and/or stiffness degradation are described in
Chapter 6.
Figure 7-20 lists input items to describe the flexural element degradation. The ID numbers of the
degradation types are referenced to the beam/column/brace section specification above.

FIGURE 7-20 FLEXURAL ELEMENT DEGRADATION


DEGRADATION TYPES
ID

Modulus Icr/Ig
of Rupture
(x f'c)

Plastic Strength Degradation Coefficients


Stiffness Degradation Coefficients
IaveH Hinge
(leave blank for defaults)
(My columns, M+ beams)
alphabeta+
beta- (use 0) Length F1
F2
F3
M1
M2
M3
or alpha z or beta y or beta z
Length

alpha+
or alpha y

Strength Degradation Coefficients


(Mz columns, M- beams)
F1
F2
F3
M1
M2 M3

The ID number must be unique for each degradation type.

The first two fields after the ID (modulus of rupture and ratio of cracked to gross moment of
inertia) are for future developments and are not currently used.

The and factors are defined differently for beams and columns. See Chapter 6 for a full
description.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

126

The plastic hinge length is a factor to apply to the depth of the element to obtain the actual
plastic hinge length.

See Chapter 6 for a description of the strength degradation coefficients.

7.9.8

BASE ISOLATORS

The HCG procedure is intended to be suitable for the evaluation of isolated buildings although so far
only a few buildings have been analyzed using the procedure in its current format.
The ANSR-L program has been used for four types of base isolator:
1. Lead rubber bearings. These are modeled as bi-linear cantilever column elements.
2. High damping rubber bearings. These are usually modeled as equivalent hysteretic damped
elements, similarly to lead rubber bearing.
3. Teflon sliders. These are based on a special purpose element developed during the Te Papa
development phase, using results of tests at the University of Auckland. The coefficient of
friction of the element reflects the dependence on pressure and velocity determined by the tests.
4. Yielding steel cantilevers. These elements are modeled as bi-linear cantilever column elements
using properties as described by Skinner, Robinson and McVerrry.
Figure 7-21 shows the input cells for the isolator elements. These are generally self-explanatory.
Advice on modeling base isolated structures should be obtained from the R&D division as there are
special considerations for damping, acceleration time histories and other aspects of modeling.
Two changes have been made to the input to reflect special requirements of some projects:
1.

The code for Teflon Type can be used with LRB isolators to indicate whether a bi-linear or trilinear stiffness formulation is used. If Type is LRB and Teflon Type is an integer greater than
zero then a tri-linear curve will be used, else bi-linear.

2.

The width dimension, B, is used to define the vertical stiffness if the isolator type is LRB.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

127

FIGURE 7-21 BASE ISOLATOR ELEMENTS


ISOLATOR
ID

TYPE
(See
Notes)

Lead Rubber or High Damping rubber


Ku
Kr
Fy
(KN/m)
(KN/m)
(KN)

(Set this value > 0


for lead rubber to
use a trilinear model,
else bi-linear

NOTES
1

2
3

Teflon
Type

Kvert
For LRB
Yielding Steel Cantilevers (dimensions in m)
Rectangular
Both Types
Circular
m
B
T
D
Dr
L
R

Valid Types are:


L
Lead-Rubber Bearing
H
High Damping Rubber
T
Teflon Slider
S
Steel Cantilever
Teflon Types are UFP, 15GF, 25GF or UFT. Use UFP generally.
m is the estimated strain in the cantilever (between 0.01 and 0.07). Use 0.07 if unsure.

Dr

R
L
T

B or D

7.9.9

DAMPERS

Procedures to design and evaluate the effects of added damping are underway as a separate research
project (see In-Structure Damping and Energy Dissipation Design Guidelines). As part of this development,
element properties for springs or viscous dampers can be specified as shown in Figure 7-22. These
elements are oriented in a similar manner to braces and base isolators.
The properties required for the two types of damper are:
1.

A spring damper, described by the stiffness and the yield force.

2.

A viscous damper, described by the damping coefficient, C, and the exponent, . If the
device as a velocity limit, beyond which the damping force is constant, this can be entered
as VMAX.

The spring damper, equivalent to a truss element with a specified yield force and zero strain
hardening, and could also be included in the model as a brace. For details of the viscous damper
force function see the damper design guidelines.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

128

FIGURE 7-22 DAMPER ELEMENTS

DAMPING ELEMENTS
ID

101

TYPE
(See
Notes)
D

Keff
C

Fy
alpha

Vmax

1.0

These values for spring components


These values for dashpot components

Valid Types are


K
Spring element
D
Dashpot element
Damper sections are combined with isolators and other sections so must not have
the same ID as any other section.

7.10 PLOTTING ELEMENT HYSTERESIS

Often, you will want to check the actual hysteresis shape represented by the material and element
properties you have specified in the sections above. The hysteresis plots for plane stress elements in
Chapter 6 of this manual were obtained by a displacement pushover analysis of a single story
structure. The model used a series of wall elements, each spanning between the same column lines
and same level. Each element represented a different section (and material) type.
The key to obtaining these hysteresis loops is to use an output code of 2 for the panels (the last
column on the Walls sheet). A time history tabulation of element stresses and strains will then be
output to the .D01 output file. Steps in obtaining an Excel plot are:
1.

Run a displacement pushover analysis and edit the .D01 file. Delete the top couple of
hundred lines, down to where the tabulated values start at time 0.02 seconds. Then find the
string Envelope near the bottom of the file and delete all the lines after the end of the table
(which will end at 20.00 seconds). Save the file with the top and bottom lines stripped off.

2.

Import the file to Excel using the fixed width option. The first two columns identify the
group number and element number. If the material is changed, the elements will be in
different groups and the element number column will be all 1s. The next column is time.
The following two columns are the r and s coordinates of the output point, always zero so
ignore them. The next column is the yield code, an integer indicating which slope on the

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

129

envelope the response lies on. Following this are three stresses, S11, S22 and S12 and three
strains, 11, 22 and 12. For this type of element, the shear stresses and strains define the
response.
3.

If you chart S12 against 12 you will get the hysteresis. (You will need to change the order of
the columns in the formula bar on the chart as the hysteresis will be rotated 90 degrees from
the normal orientation).

If you have obtained multiple element time histories from a single analysis, use the Excel filter
function to plot each hysteresis, one at a time.
The ANSROUT processing also provides a method for plotting the hysteresis of total story force.
See Section 10 for more details.

7.11 ELEMENT TYPES

All sheets for each element type generally follow a similar format. The element type is identified and
its connectivity is specified, that is, the columns lines and/or levels which the element connects to.
Other information needed to fully describe the element is also specified where needed, such as pin
ends for columns and gravity loads for beams. The following sub-sections describe the specific of
input for each type.

7.11.1 ELEMENT OUTPUT CODE

For each element, the final spreadsheet column is headed Output Code. Envelope forces and
deformations are always output for every element. This code specifies what element time history
output is required, if any.

A code of 1 indicates that the time history of element output is to be written into the .D01
output files as the analysis progresses. This creates a very large output file and should be used
with caution.

A code of 2 will write a time history of output but without headings at each step. This reduces
the size of the D01 file and makes it simpler to import the file to Excel. However, it is only
efficient for a limited number of element time histories.

A negative code indicates that the element time history output is to be written to a separate
output file. The output is more compressed and efficient using this option. If the element has a
unique code number then output for that element will be output separately. If more than 1
element has the same output code then the forces for the group of elements defined by the same
output code will be summed and output as a single result.

An output code of 99 will automate the story shear procedure. Each element which has this
code will be generated with an output code in ANSR-L equal to the negative of the level number
at the bottom of the element. This option is the simplest way of obtaining plots of story shear

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

130

versus time but may produce large output files, and take a long time to process, if there is a large
number of levels. See Section 7.7.2 for suppressing output for some stories to reduce the size of
this file.
The summation option (negative code) can be used to obtain output for a series of panels at a level
which form a wall, for example, or all elements at a level to obtain the total story shear. Such
grouping is generally only meaningful for elements which are at the same level. To simplify
identification during processing (Section 10) it is recommended that the output code be the negative
of the level number, consistently for all element types. For example, all columns and wall panels
between Levels 4 and 5 could have an Output Code = -4 to obtain the total story shear at that level.

7.11.2 WALLS

Figure 7-23 is a portion of the wall element input spreadsheet. The input fields are:
1. A wall name which is used to identify layers in the AutoCad model. Use up to 8 characters,
without any embedded blanks. Typically, wall names are a single character identifying the
elevation (A, B etc) or a direction descriptor (North, South, etc).
2. The section type refers to one of the plate sections defined on the Materials and Sections
worksheet. An exception is for openings in walls, where the section type may be identified as 0.
An opening specified for the same region as is also defined as a wall portion will remove that
wall segment. This is often a convenient way to specify a large area of solid wall with only a few
discrete openings.
3. The Element code is left blank for a rectangular element. Set to T for a triangular, or any nonrectangular, element.
4. The bottom level is the level at which the base of the wall segment is located.
5. The top level is the level at which the top of the segment is located.
6. The level increment allows levels to be skipped in generating the wall up the height. This should
be used with caution as a value greater than 1 may pose a danger of incompatibilities with
intersecting walls.
7. The column lines connected by the wall segment. Up to 12 column lines can be defined per row.
If there are more than 12, start the next row with the 12th value and continue from there.
8. An output code for the walls. See Section 7.11.1.
When the Element cell is set to T, MODELA will read four pairs of coordinates, {Ci,Li} where Ci is a
valid column line number and Li a valid level number. The four pairs should describe the perimeter
of an element reading clockwise from top left. If the 4th pair of coordinates is omitted, it will be
assumed equal to the first pair and so describe a triangular element.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

131

FIGURE 7-23 WALL ELEMENT INPUT


WALL PANELS
NOTES
1
2
3
3

WALL
NAME

Set Element to blank unless you are specifying a 3 or 4 node non-rectangular element - set to "T"
For type = T, enter C1 L1 C2 L2 C3 L3 (optionally C4 L4) where C1,L1 are column and level
intersection at Node 1 etc. Order : Top Left, Top Right, Bottom Right, Bottom Left
Elements will be generated from Bottom Level to Top Level from C1 to C2, C2 to C3 etc.
for example, bottom level 1 top level 2 will specify 1 element up the height, from 1 to 2
Wall name will generally be a character or number identifying the elevation. Use up to 8 characters.
Walls in AutoCAD are identified by the first character of this name. Recommend A to Z

SECTION Element Bottom


TYPE
Level
(0 = Hole)

Top
Level

NOTE: Set Output Code to -99 to sum storey shears


Level Col 1 Col 2 Col 3 Col 4 Col 5
Col 12 OUTPUT
Incr
CODE

7.11.3 COLUMNS

Column input is as listed in Figure 7-24. Each column is identified by a valid section type, a strength
type at the top and a strength type at the bottom. These must each be valid identifiers from the
Materials and Sections sheet.
As for walls, column elements span from a bottom level to a top level at a specified increment.
Columns typically span from diaphragm level to diaphragm level. For frame structures there is
typically one level per diaphragm and so the level increment will be 1.
For wall / frame structures there are often more levels in the model than there are diaphragms to
include door and window openings in walls. Extra column nodes at these intermediate levels add
significantly to the problem size as each internal node has six degrees of freedom. Therefore, if the
column does not intersect with a beam or wall element at the intermediate levels, the level increment
should be set greater than one.
Column pins may be included by setting the pin code to 1 (pin at bottom), 2 (pin at top) or 3 (pin at
both ends). If a pin code of 3 is specified the column will be modeled as a truss member because this
is computationally more efficient than the beam/column element. A pin is defined as a moment
release about the three element local axes (two bending moments plus a torsional release).
Up to 15 columns may be listed on a row. All columns on the same row must be of the same section
and strength type, span between the same levels and have the same pin end condition.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

132

FIGURE 7-24 COLUMN ELEMENT INPUT


COLUMN ELEMENTS
NOTES
1
2
3
4
5

Specify one or more columns per row. Columns on a row must have the same section and span
between the same levels. One column is inserted from Level at Bottom to Level at Top.
The code for PIN is 0 (no pin) 1 (pin at bottom) 2 (pin at top) or 3 (both ends).
A column pinned top and bottom is better modelled as a brace.
Columns of same type will be generated from Bottom Level to Top Level at increment listed.
Set Output Code to -99 to sum storey shears

SECTION Strength Strength Level Level Level Pin Col Col Col Col Col Col
TYPE
Type
Type
at
at
Incr
Line Line Line Line Line Line
Bottom
Top
Bottom Top
1
2
3
4
5
6
1
1
1
1
2
1
1
2
3
4
5
6
1
1
1
2
4
1
1
2
3
4
5
6
1
1
1
2
4
1
16
17
2
1
1
4
38
1
1
2
3
4
5
6
2
1
1
4
38
1
16
17

Default 0 P Delta
1 No P-Delta
Col OUTPUT KGEOM
Line CODE
CODE
15
15
15

The default MODELA procedure is to set the geometric stiffness flag on for columns and off for
other elements. The element sheets now have an input column added (after the output code) where
you can specify this. You might use this, for example, to quantify P- effects by repeating an analysis
with and without the flag on.
As P- effects are generally used for columns, the default of 0 is to include geometric stiffness, a
value of 1 will suppress it. Note that this is reversed for the beams.

7.11.4 BEAMS

The beam input, shown in Figure 7-25 generally follows a similar format to the column input except
that:
1. Distributed beam loads may be specified. The load types may be uniform or varying. The effect
of beam loads is to initialize beam end moments and shear forces and these loads may also be
applied to the structure as nodal weights and/or masses depending on the options set on the
Analysis Data worksheet (see Section 7.6). See Section 5.10 for a discussion of beam loads.
2. Beams span between two column lines. Multiple column lines, up to 14, which together form a
bent may be specified across the row. Beam elements will be generated from C1-C2, C2-C3 etc.
The beam span load option allows for uniform and varying loads. The program takes account of
beams which cantilever or are part of a bay which has been sub-divided into two or more elements.
For beams, the default is Geometric Code is 0, which specifies no geometric stiffness. As geometric
stiffness only effects members with axial load and chord rotation this flag generally has no effect for
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

133

beams. However, this flag is also used to specify the hysteresis type for beams (see Section 6.4).
Values of 1 and 2 will activate the elastic unloading option, compared to the default of hysteretic
unloading.

FIGURE 7-25 BEAM ELEMENT INPUT


BEAM ELEMENTS
NOTES
1
2
3
4
5

Specify one line of beams per row. Beams on a row must be of same section type. Beams will be generated
from C1-C2, C2-C3, C3-C4 etc. Beams may be incremented up levels if they are the same type.
The code for PIN is 0 (no pin) 1 (pin at bottom) 2 (pin at top) or 3 (both ends).
A beam pinned top and bottom is better modelled as a brace.
Load Type 1 = uniform, 2 = varying, w at centre, 3 = varying w at Left, 4 = varying w at right
Set Output Code to -99 to sum storey shears

SECTION Strength Strength Bottom Top Level Pin Distd Load Col Col Col Col
TYPE
Type
Type
Level Level Incr
Load Type Line Line Line Line
End I
End J
Top
w
1
2
3
4
3
1
1
2
4
1
1
2
3
4
3
1
1
2
4
1
14
15
16
17
4
1
1
5
38
1
1
2
3
4
4
1
1
5
38
1
14
15
16
17

No P-Delta (Default)
0
P-Delta
1
No P-Delta Elastic Unloading
-1
P-Delta Elastic Unloading
-2
Col
Col OUTPUT KGEOM
Line
Line
CODE CODE
5
14
5
1
5
1

NOTE : The distributed load type can be set negative if the loads are to be used to
assemble nodal loads but not to initialize beam end actions. For example, a value of 1
will apply a nodal load of wl/2 at the nodes at each end of the beam but will not apply
fixed end moments to the beam.
7.11.5 BRACES

A brace as defined by MODELA is basically any flexural element which is not horizontal or vertical
and so is not parallel to one of the global axes. The element is most commonly used to model
diagonal brace type elements but may be used to connect any two pairs of columns/level
coordinates.
The brace input, shown in Figure 7-26, generally follows a similar format to the column input except
that:
1. The code for the pin end has additional values to allow for special types of truss elements. These
allow for ductile yielding plus various forms of buckling and brittle failure. Contact the R&D
division if you want details on these special types of truss.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

134

2. A brace element is specified by a bottom level and a top level and a bottom column line and top
column line. Usually both the levels and the columns will differ, producing a diagonal brace up
the height of the building. In special circumstances, the levels may be the same (horizontal
member) or the columns may be the same (vertical member). In no cases will both the levels
and columns be the same as the element would be zero length.
FIGURE 7-26 BRACE ELEMENT INPUT
BRACE ELEMENTS
NOTES
1
2
3
4

5
Section
Type

Specify up to 7 braces per row (same section type and spanning between same levels).
Use level increment > 1 for braces which span levels
The code for PIN is 0 (no pin) 1 (pin at bottom) 2 (pin at top) or 3 (pin both ends).
PIN is also used to specify elements to be modelled as trusses as follows:
PIN=10 Yielding truss element
PIN=11 Elastic buckling in compression
PIN=12 Elastic bucking in tension
PIN=13 Elastic in compression, brittle failure in tension
Set Output Code to -99 to sum storey shears
Strength Strength Bottom Top Level Pin Bot Top Bot Top Bot Top
Type
Type
Level Level Incr
Col Col Col Col Col Col
End I
End J
Line Line Line Line Line Line
1
1
2
2
3
3

Bot Top Output


Col Col Code
Line Line
7
7

7.11.6 ISOLATORS AND DAMPERS

Isolators are modeled identically to braces, as described in the preceding sub-section. As for braces,
isolators can connect any pair of level/column coordinates and can be vertical, horizontal or obtuse.
Most commonly isolators are modeled as vertical elements at the base of the structure with a pin at
top. Yielding cantilever type elements may be horizontal (for example, Union House).

7.11.7 GAPS

Gap elements are most commonly used to model one of the types of compression only
configurations encountered in structures:
1. Uplift of foundation elements (vertical gap, with an opening parallel to the horizontal plane).
2. The compression block component of flexural yielding at the base of a cantilever wall (vertical
gap, with an opening parallel to the horizontal plane). This gap is used in parallel with a
reinforcing bar tension element.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

135

3. Pounding between buildings (horizontal gap, with an opening parallel to the vertical plane).
4. Horizontal soil springs above base level, for example where retaining walls bear against
surrounding soil.
The input cells in Figure 7-27 are used for all types of gap, and for any combination of the three.
1. For vertical gaps (modeling foundation uplift or wall yielding) specify column line 1, column line
2 and the column increment. Gap elements will be generated from C1 to C2 at the specified
increment. If C2 and the increment are omitted a single gap element will be inserted.
Note that if the gap direction is defined in Column L then the column for C2 should be
blank if the gap is grounded. The cells for Column Increment are used to define the
second level if the gap spans levels leave this blank too if the gap is grounded or is at
a single level.
2. For a horizontal gap (pounding) set the cell in Column L to 1, 2 or 3 to define the direction of
the gap (respectively parallel to the X, Y or Z axes). The gap element then connects C1 to C2.
If C2 is zero or blank, the gap element connects C1 to the ground. The default is for the gap to
be oriented such that tension is for movement between the two nodes in the positive direction of
the axis. To change the direction of the gap, add 10 to the code, that is, use 11, 12, or 13 for X,
Y or Z axes respectively.
3. The level number at which the gap is to be inserted is entered in Column D. Levels cannot be
generated.
4. The compression area and compression material are required for all gap types. It is
recommended that these values be set as small as possible to avoid excessive unbalanced loads.
5. The tension area and tension material are required only for gaps used to model flexural yielding.
These are intended to represent the reinforcing bars crossing the plane of the gap.
6. The gap effective lengths used to calculate strain in the reinforcing bar. Pending further study, a
value of 0.05 to 0.45 metres is recommended. Default is 0.25 (250 mm).
NOTE : The gap dimension in the ANSR-L model is always 50 mm. The length
specified here is used to calculate the stiffness of elements crossing the gap. If you
specify a length of 500 mm the AutoCad plot will look the same. However, the
stiffness of the gap, reinforcing bar elements and any boundary elements will be onehalf of what they would be with the default 250 mm gap length.
7. If the generate areas cell contains a Y then the compression and tension areas will be
automatically calculated from the length and thickness of wall panels incident on the specified
column line. For this option to work correctly, the reinforcing ratio must be specified for the
wall panel type.
8. Output code is as for other elements.
9. Column L defines the gap direction (see above).
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

136

10. An initial opening may be set for the general gap. This is generally used to define the initial
separation between two buildings separated by a gap. Note that pounding will occur once this
separation is exceeded, regardless of the difference in coordinates between C1 and C2. If this
value is negative then it is interpreted as an initial compression load in the gap.

FIGURE 7-27 GAP ELEMENT INPUT


GAP ELEMENTS
NOTES
1

Specify column lines and levels at which gap elements are required. Columns but not levels
may be generated.
For gaps used to model foundation uplift, set tension area to zero. If compression
area is also zero, very stiff subsoil conditions will be used.
Gaps are also used to model flexural yielding at the bottom of wall panels. For these gaps the
compression area is effective area of concrete and tension area the effective area of reinforcing
tributary to that node. An effective length is used to calculate strain in the reinforcing steel.
Pending further study, use 0.05 to 0.15 metres for this.
If Generate Areas is set to Y the tributary areas for walls incident to this Column Line at
this level will be used to calculate compression and tension areas (input reinforcing ratio
for the wall type for this option to work correctly).
A general gap will span from C1 to C2. For this type, an initial gap opening (in m) may be specified.
If the gap is grounded, set C2 to zero and specify direction in Column L (1 = X, 2 = Y, 3 = Z)
Column L must be non-zero for all general gaps. If the gap spans levels, put the second level in the
Column Increment column (Column C)
Set Output Code to -99 to sum storey shears

Column
Line
1

Column Column Level Compression Compression Tension Tension Effective Generate


Line
Incr.
Number
Area
Material
Area
Material Length
Areas
2
(Y/N)

2
3

Set to
>0
General

Initial
Load
(-ve)
or
Initial Coefficient Sticking
Opening
of
Factor
(+ve)
Friction
SF

Output
Code

Travel
Factor
CB

The final three input columns in the Gap sheet refer to properties of the element normal to the gap
plane. These define the coefficient of friction of the sliding surface and sticking and travel factors,
which define the reduction of the friction force with distance traveled. Check with the R&D division
if you need to model special sliding surfaces.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

137

7.12 FLOORS

The evaluation procedure includes both rigid and flexible diaphragms (or no diaphragms). Input for
both types is provided on the worksheet Floors.

7.12.1 RIGID FLOORS

Figure 7-28 provides the input for rigid diaphragms:


1. The first level at which these properties apply.
2. The last level at which these properties apply, if there are more than one level with the same
mass and centroid properties (default = first level).
3. The increment between the first and last level (default = 1)
4. The diaphragm number for this level. Diaphragms at each level start at one and are in sequence.
(default = 1)
5. The X coordinate of the center of mass.
6. The Y coordinate of the center of mass.
7. The translational mass.
8. The rotational mass.
9. The X eccentricity of the mass for this diaphragm.
10. The Y eccentricity of the mass for this diaphragm.
11. A flag to indicate whether to suppress time history output of the center of mass node. This is
only meaningful if you have requested time history output at these locations (Section 7.6.13).
Note that if the center of mass coordinates for any floors coincide with a column node (within the
distance tolerance, default of 5 mm) then the coordinates will be adjusted by a distance equal to the
distance tolerance in both the X and Y directions to ensure that there will be no zero length
connector elements. If this is an unacceptable distance, make the distance tolerance smaller.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

138

FIGURE 7-28 RIGID FLOOR INPUT

RIGID FLOORS
NOTES:
1 Enter data for rigid floors here. If floors are flexible, model as elements below in
lower portion of this sheet.
2 Floor levels can be incremented.
3 For each floor, enter coordinates at centre of mass, mass and mass moment of inertia
and eccentricities if required.
4 Enter a non-zero value for Suppress Output to omit nodal time history
Level
1
1

Level Level Diap.


2
Incr Number
38
1
1

X
C.M
0.000

Y
C.M
0.000

MASS

MMI

100

1000

Ex

Ey

Suppress
Output

7.12.2 FLEXIBLE FLOORS

Flexible diaphragms are modeled as finite element grids using the same type of plane stress element
as is used for wall elements. The form of input depends on the manner in which the floor is
modeled.
Figure 7-29 shows the spreadsheet cells describing the flexible floor elements. The first three
columns describe the first level, last level (default = first level) and level increment (default = 1).
Floors can be generated up height if they have the same properties and geometry.

FIGURE 7-29 FLEXIBLE FLOOR ELEMENT INPUT


FLEXIBLE FLOORS
NOTE : Set Section type to -SECTION for a 4 node element
Then enter 4 node numbers as Cx1, Cx2, Cx3, Cx4
Level Level Level Section
Column Lines in X Direction
Column Lines in Y Direction
1
2
Incr Type Cx1 Cx2 Cx3 Cx4 Cx5 Cx6 Cx6 Cx8 Cx9 Cx10 Cy1 Cy2 Cy3 Cy4 Cy5 Cy6 Cy7 Cy8 Cy9 Cy10

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

139

The section type is a valid identifier for one of the plate sections specified on the Materials & Sections
sheet.
If the finite element layout is on a rectangular grid, then the mesh can be generated by specifying a
series of column line numbers along the X and the Y directions respectively. The edge column lines
must have been specified. Internal column lines (at the intersection of lines drawn parallel to the X
and Y axes) will be inserted by MODELA if they do not exist.
For example, to generate the finite element mesh shown in Figure 7-29 the column lines in the X
direction would be 1, 2, 3, 4, 5, 6 and 7. The column lines in the Y direction would be 21, 22, 23, 24
and 25. Coordinates would be specified for all these column lines. The column lines at the
intersections would be automatically generated if they did not exist.
If the finite element layout is not on a rectangular or regular grid then the elements can be entered
one by one, defining the four column lines which describe the perimeter of the element. The four
column nodes are listed in the cells headed Cx1, Cx12, Cx3 and Cx4. This condition is indicated by
entering the section type as the negative value of the actual type.

FIGURE 7-30 FLEXIBLE FLOOR ELEMENT MESH

C25
C24
C23
C22
C21

C1

C2

C3

C4

C5

C6

C7

7.13 CONNECTIVITY

The Connectivity worksheet defines which columns lines are connected to which diaphragm. If your
structure has flexible diaphragms, or no diaphragms, then all data cells on this sheet should be empty.
For flexible diaphragms, connectivity is defined by whether elements are incident to nodes which are
connected to floor finite elements.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

140

For rigid diaphragms, the default is to assume that all column lines are connected to diaphragm 1 at
levels which contain rigid diaphragms. Therefore, there are only two conditions which require data
on this sheet to be entered:
1. If there are multiple diaphragms at one or more levels.
2. If there are column lines which are not connected to the diaphragm at one or more levels.
If either condition applies, then the following information is supplied on the cells shown in Figure 731 (Note that as all columns are connected to diaphragm 1, only those to which this assumption does
not apply need be specified):
1. The first column, last column and column increment to which this specification applies. (default
last column = first column, default column increment = 1)
2. The first level, last level and level increment to which this specification applies. (default last level
= first level, default level increment = 1)
3. The diaphragm number to which these nodes are connected. A value of 0 indicates a
disconnected node, otherwise enter a diaphragm number which is a valid identification number
specified for the particular level(s).
A diaphragm specification will over-ride any previous specification for the same node. For example,
if most columns at a particular level are disconnected, then the first entry will be from 1 to the
number of columns with a diaphragm number of 0. Subsequent lines will list only the column lines
connected to the diaphragm.
If there are flexible floor elements as well as rigid floors then additional column lines may be
generated to locate internal floor nodes (see preceding section). In this case, to connect (or
disconnect) all columns at a level, enter the second column as a very high number to ensure that all
column are included.

FIGURE 7-31 FLOOR DIAPHRAGM CONNECTIVITY INPUT

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

141

FLOOR CONNECTIVITY
NOTES:
1
2
3
4

Column
1

All elements are assumed connected to a floor if a rigid floor is specified at a level.
If columns are disconnected or multi-diaphragms are used, specify on this sheet.
Only columns not connected to diaphragm 1 need be specified.
Note that extra nodes may be generated for floor panels.
To ensure all nodes are disconnected, set Column 2 to a high number.
Column
2

Column
Incr

Level
1

Level
2

Level
Incr

Diaph
Number

7.14 MASSES AND LOADS

The structure masses and weights are generally assembled from element densities and floor
specifications. This sheet permits additional masses to be applied to the model.
Weights are applied as a static load case. Results from this analysis form the initial conditions for the
dynamic analysis. This sheet also allows weights to be added to this load case.
If masses or loads are to be added, then the following information is supplied on the cells shown in
Figure 7-32:
1. The first column, last column and column increment to which this specification applies. (default
last column = first column, default column increment = 1). If the level number is negative (see
2. below) then the load is applied at the diaphragm node and the first column is interpreted as
the diaphragm number.
2. The first level, last level and level increment to which this specification applies. (default last level
= first level, default level increment = 1). If the level is negative, the mass/load is applied at the
diaphragm master node at this level and the column number indicates which diaphragm master
node the mass/load is to be applied at.
3. The load case number. The maximum value in this cell sets the number of static load cases
applied to the structure prior to the dynamic analysis. Usually, only a single static case,
representing gravity, is applied but other loads cases may be defined. Each load case will write
an output file (S0n, where n is the load case number). If you are just adding mass, set the load
case to 1.
4. The loads, which may be in any of the three translational and three rotational directions, in the
model coordinate system where X and Y are horizontal. Weights are specified as negative Z
loads (these are converted by MODELA to -Y in the ANSR-L coordinate system see Section
8.1).
5. The masses, which may be different in each of the three translational directions.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

142

FIGURE 7-32 ADDED MASS AND LOAD INPUT


ADDED MASSES AND LOADS
NOTES:
1
Masses and loads are assembled from floor properties and elements densities.
If added masses or loads are required, enter on this sheet.
Loads are signed. Vertically downwards loads are negative Z.
2
3
Set level number to -ve to use master node. For this option, Column number = diaphragm number
Column Column Column Level Level Level Load Load Load Load Moment Moment Moment Mass Mass Mass
1
2
Incr
1
2
Incr Case
X
Y
Z
X
Y
Z
X
Y
Z

7.15 DEGREES OF FREEDOM

The MODELA program will assemble the degrees of freedom (DOFs) at each node depending on
whether the node is a support node and the element types incident on the node. For the majority of
structures, the DOFs will be correct and no data will be required on this sheet. In fact, unless you
have a very unusual configuration, if the program does not assemble the DOFs correctly, notify the
R&D division.
If DOFs are to be modified then this can be done either by specifying restraints directly or by linking
the degrees of freedom at specified nodes.
7.15.1 CONSTRAINED DEGREES OF FREEDOM

If the constraints generated for degrees of freedom are to be modified, then the following
information is supplied on the cells shown in Figure 7-33:
1.

The first column, last column and column increment to which this specification applies.
(default last column = first column, default column increment = 1). If the level number is
negative (see 2. below) then first column is set to the diaphragm number.

2.

The level to which this specification applies. Level generation is not permitted. If negative, the
mass/load is applied at the diaphragm master node at this level and the column number
indicates which diaphragm master node the degree of freedom code is to be applied at.

3.

The degree of freedom code. Valid codes are 0 or blank (the degree of freedom is free), 1 (the
degree of freedom is fixed) or 1 (the degree of freedom assembled by MODELA is retained.
Degrees of freedom are in the ANSR-L coordinate system (see Section 7.1)
The 1 option is included to allow for situations where only some of the 6 possible DOFs are to be
subject to user over-ride.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

143

An example of the use of user specified DOFs would be a structure where there were several
basement levels. Below ground, the structure is restrained from horizontal translation but column
shortening and joint rotation could occur. For the levels above the base of the model, but below or
at ground level, the DOFs for all column lines would be set as:
1 -1 1 -1 -1 -1
which uses the MODELA DOFs in all directions except the horizontal translations.
Use this option with care as undesirable supports may be introduced into the model.

FIGURE 7-33 USER SPECIFIED RESTRAINED DEGREE OF FREEDOM INPUT

CONSTRAINED DEGREES OF FREEDOM


NOTES:
1
Over-ride any program specified DOFs with values given here.
2
Set DOF to -1 to use whatever is set by MODELA
3
Set level number to -ve to use master node.
For this option, Column number = diaphragm number
Column
1

Column
2

Column
Incr

Level

DOF
X

DOF
Y

DOF
Z

DOF
XX

DOF
YY

DOF
ZZ

7.15.2 EQUAL DEGREES OF FREEDOM

In some situations, it may be good modeling practice to enforce equal displacements between
particular nodes in one or more directions. To achieve this, the cells from the DOFs sheet shown in
Figure 7-34 are used.
Nodes are connected using a six digit code indicating respectively X, Y and Z translations and XX,
YY and ZZ rotations. A 1 in a particular location indicates equal displacements. For example, a
code of 101010 indicates that the X, Z and YY displacements are set to be equal. The connected
nodes are described as their column/level identification. Nodes need not be on the same level.
Up to 10 nodes can be specified on a single row. If more than 10 are constrained to be equal, repeat
the first column on subsequent rows.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

144

As for constrained degrees of freedom, use equal degrees of freedom with care as undesirable
supports may be introduced to the model.

FIGURE 7-34 USER SPECIFIED EQUAL DEGREE OF FREEDOM INPUT

EQUAL DEGREES OF FREEDOM


NOTES:
1
Code is 6 digit, relating to global axes e.g. 110000 for equal X and Y
2
Each node is described as a column/level pair
3
For more than 10 nodes, repeat C1 on subsequent lines
Equal
DOF
Code

C1
Column Level

C2
Column Level

C3
Column Level

C10
Column Level

7.16 CALCULATION OF FLOOR MASS AND WEIGHTS

The Weights worksheet is a duplicate of the same sheet in the DUCTILEIN workbook, where floor
mass properties are calculated from a series of column lines describing the perimeter., as shown in
Figure 7-35.
For levels at which you want the weight and mass calculated, enter the level, diaphragm number and
load in KPa. In the yellow cells, enter a sequence of column numbers to define the perimeter of the
floor. These may be entered in either clockwise or counter-clockwise order.

FIGURE 7-35 FLOOR WEIGHTS


CALCULATION OF FLOOR PROPERTIES
1
2
3

Level
8

Enter node numbers describing perimeter (clockwise or anticlockwise) in columns J, K etc.


Double Click on LEVEL Cell to Display Floor As Defined
NOTE : This sheet is for calculations only, values are not linked to the Floors sheet.
Copy the Values to the Floors sheet manually.

Diaphragm Load
Area
1
6.706 1035.69

Weight
6945.37

Xcg
0.00

Ycg
-14.75

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

Mass
707.99

MMI
Columns describing Perimeter of Floor Diaphragm (C
121358
1
2
3
4
5
6
7
8

145

NOTE: This sheet is provided to help determine diaphragm properties but is NOT
linked to the Floors sheet. You will need to copy properties calculated in this sheet to
the Floors sheet to use the calculated values.
As you enter column numbers the calculations for area, weight, mass, coordinates of the centroid and
mass moment of inertia will be displayed.
If you double click in Column A on the row for a particular floor then the floor will be displayed as
shown in Figure 7-36.
FIGURE 7-36 DISPLAY OF FLOOR WEIGHTS

FLOOR DISPLAY
Weight

6.706 KPa
Results:

Area
Xcg
Ycg
Weight
Mass
Mass Moment of Inertia

1036
0.001
-14.746
6945
708.0
121358.1

0.000
-25.000 -20.000 -15.000 -10.000 -5.000 0.000
-5.000

5.000

10.000 15.000 20.000 25.000

-10.000
-15.000
-20.000
-25.000
-30.000
-35.000
-40.000

7.17 AUTOCAD DXF FILES

The MODELA executable writes two AutoCad compatible DXF files (Drawing Exchange Files).
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

146

1. An AutoCad DXF file which can be imported to display the geometry. This file represents all
elements as line elements, drawn on their centerline, and will include diaphragm connectivity if
requested (see Section 7.6, Analysis Data). The drawing produced by this file may also be used to
generate the crack patterns (see Section 10). This file will have the specified job name plus the
.DXF suffix (see Figure 7-37 for an example line drawing file)
2. An AutoCad DXF file of geometry but with all elements drawn fully three dimensional, that is,
beam and column elements drawn to correct breadth and width and wall elements drawn to
correct thickness. This file will have the specified job name plus F (for Full model) plus the
.DXF suffix (see Figure 7-38 for an example three-dimensional drawing file).
To import these files you will need either the full AutoCad program or AutoCad LT loaded on you
computer. Alternatively, you can transfer the DXF file to a computer which does have AutoCad.
The full three-dimensional drawing will draw non-rectangular shapes, as shown in Figure 7-39. This
has not been tested for all possible shapes and orientations and so there may be some unpredictable
drawings produced. In some cases, the orientation may not be what you want as the analysis model
accuracy does not depend on this (the input stiffness properties are used). You can modify some
aspects, for example, a C column can be turned around by specifying a column angle of 180.
FIGURE 7-37 AUTOCAD LINE DRAWING

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

147

FIGURE 7-38 AUTOCAD THREE DIMENSIONAL DRAWING

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

148

FIGURE 7-39 AUTOCAD DRAWINGS USING NON-RECTANGULAR SHAPES

7.17.1 IMPORTING THE DXF FILES

The procedure for importing a DXF file once it has been created by MODELA is:
1. Open AutoCad LT.
2. If the dialog box for a new drawing appears, ensure that the option for Use Prototype is NOT
checked. Once you have checked this option, you can also un-check the option Show this Dialog
at Startup and ignore this step in future.
3. If the dialog box for a new drawing does not appear, and you are not sure about whether a
prototype is being used, select New from the File Menu and follow the step above.
4. From the File Menu, select the option to Import/Export and, within this option, choose DXF In.
You will then be prompted to pick the file.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

149

7.17.2 SETTING VIEWPOINT

Once the file has been imported, it will be displayed as a plan view. Two methods to change the
viewpoint are:
1. Type the command VP <Enter> then type the vectors for each axis (for example, 1,1,1 <Enter>
to give an isometric view0.
2. Select the View Menu, option 3D Viewpoint, choose Axes and dynamically rotate the axes to the
viewpoint you want.
You may need to experiment to find the clearest view. Switching the vectors from 1 to 1 will view
from the opposite direction.

7.17.3 DISPLAY, PRINT AND COPY OPTIONS

For a complex model, the View Menu provides options which will make the view clearer:
1. Select Hide to remove lines which would be hidden in the selected view (the first view in Figure
7-38 has been displayed using this option).
2. Select Shade and then 256 Color Edge Highlight. If the model is complex, the 256 Color
option may be better (the seconds view in Figure 7-38 has been displayed using the 256 color
option).
The Zoom command, also on the View Menu, allows display of more detail.
The Print command on the File Menu does not permit printing of the Shaded drawing but does
permit hidden line removal.
To copy the AutoCad drawing to other applications (such as Excel or Word) you can use the <Print
Screen> key to copy the entire screen to the clipboard or use one of the AutoCad options on the
Edit Menu:
1. Copy Image, which will copy the display, including shading. Resolution will be as for a bitmap.
With this option, select a window to copy using the mouse. Press <Enter> to copy to the
clipboard.
2. Copy Vectors. This provides a better quality copy but does not include shading. If hidden lines
have been removed on the screen then they will be removed on the Copy Vector copy. When
you select this option, select objects by drawing a box around them and then press <Enter> to
copy them to the clipboard.
Once in the clipboard, they can be entered into other applications using the Edit / Copy command.
The lines drawings in Figures 7-37 to 7-38 have been copied as vectors, the shaded drawings as
images.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

150

7.17.4 DISPLAYING SELECTED ELEMENTS

For a complicated model, checking is often best performed by examining drawings showing only part
of the model. One option is to Zoom to part of the drawing. Another option is to display only
selected element types or sections. The element types have been assigned to separate layers to
facilitate this.
Layers can be turned on or off, or colors changed, by selecting Layer Control on the Settings
Menu. Layers for all element types, some of which may not be used in the particular model, will be
included in the list.

Beams and Columns are identified by a layer name comprising BEAM or COLUMN plus a section
number. Walls are identified by the user-specified name from the input worksheet. Connectors (if
specified) are identified as CONNECTOR plus level number plus diaphragm number. For example,
CONNECTOR102 would connect nodes to the center of mass of Diaphragm 2 at Level 10.

7.17.5 WALL COLUMN ELEMENTS

If some walls in the model are not parallel to the global axes then wall columns may be included in
the model to restrain out-of-plane stiffness. These elements will be included in the DXF files and
may confuse the drawing. The simplest option is to use the Layer Control option as discussed above
and turn off the layer named WALLCOLUMNS.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

151

8 TIME HISTORY ANALYSIS

The analysis engine on which the HCG procedure is based, ANSR-L, is a hybrid of the ANSR-II and
ANSR-III programs, which were developed at the University of California, Berkeley in the 1970s
and 1980s. The theoretical basis and details of solution techniques are provided in the EERC
reports listed in the bibliography. These should be consulted for technical details of the program.
This section describes the operational aspects of the program.

8.1

COORDINATE SYSTEM

The model development in MODELA follows the convention where the horizontal axes are X and
Y and the vertical axis is Z. This is for consistency with other HCG software packages such as
ETABS and AutoCad.
The ANSR-L program requires that the Y axis be vertical if the slaving option is to be used for
beam-column elements. To enable this option, the MODELA program rotates the axes so that
Xa = Xm
Ya = Zm
Za = -Ym
Subscript a refers to the ANSR-L coordinate system, m the MODELA coordinate system. Take this
into account when checking the coordinates generated for the ANSR-L model.
Figure 8-1 shows the ANSR-L global coordinate system and the local coordinate systems defined
using the default orientation. The element orientation is defined be the use of a K node, which lies
in the local x-y plane and defines the definition of the y-axis. The default location of the K node is:

For beam elements, the K node is placed parallel to the Y (vertical) axis. Therefore, the major
axis moment, Mz, acts about a horizontal axis normal to the beam centerline axis. The local y
axis is parallel the to global Y axis.

For column elements the K node is placed parallel to the global Z (horizontal) axis. The local y
axis is parallel to the global Z axis and the local z axis is parallel to the global X axis.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

152

The column K node location shown in Figure 8-1 is for a specified column angle of 0 degrees (see
Elevations and Nodes, Section 7.7). Any column angle specified on this sheet will rotate the location of
the K node in the counter-clockwise direction. For example, if a column line has an angle of 90
specified then the K node for the columns will be parallel to the global X axis. The local y axis will
be parallel to the global X axis and the local z axis parallel to the global Z axis.

FIGURE 8-1 GLOBAL AND LOCAL COORDINATE SYSTEMS

y
K

Y
x

z
I

K Node
for
Beams

BEAM LOCAL AXES


PRIMARY MOMENT
x
X
K Node
for
Columns

Z
ANSR-L GLOBAL AXES

y
COLUMN LOCAL AXES

8.2

RUNNING ANSR-L

The MODELA program will write a series of ANSR-L input files (each with a .ANS suffix) plus a
batch file to run the analyses. The analyses can be started by
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

153

1.

Double clicking on the batch file in Explorer.

2.

Shelling to DOS mode, changing to the job directory and typing the name of the batch file.

3.

Restarting the computer in DOS mode, changing to the job directory and typing the name of
the batch file.

The third option will only be used for a very large model, where the memory required by ANSR-L is
close to the limit of physical memory in the computer.
This procedure starts the full set of analyses running, that is, each earthquake, orientation and
eccentricity. Often, it is necessary to start a single analysis to determine memory requirements
and/or to debug the model. This section describes procedures for this.

8.3

GETTING MEMORY REQUIREMENTS

The ANSR-L memory management is based on a large common block, an amount of memory which
is reserved to store all the structural data, stiffness and mass matrices, and element data. Earlier
versions of Fortran required that the size of such common blocks be defined at compile time. This
required that the program be re-compiled to change the capacity of the program. Later Fortran
standards, which ANSR-L is now compiled under, permit the size of these blocks to be allocated
dynamically at run time.
Eventually we will modify ANSR-L to compute and assign the memory required automatically but
for now you need to specify this. The program is compiled with a default common size of 2 MB,
which is suitable for small to medium size problems. For larger problems, you need to specify the
amount of memory to reserve on the Interaction sheet of the MODELA workbook. The procedure to
determine how much memory is required is as follows:
1.

Generate the ANSR files from MODELA.

2.

Open any one of the .ANS files with a text editor.

3.

Change the first word of the first line from START to CHECK. This sets the program to run
in data check mode only.

4.

Run this input file in the Interactive mode (see Section 8.6). When you get the undeformed
shape plotted on the screen, terminate with <Esc>.

5.

Open the echo file. ECH with a text editor and go to the bottom of the file.

6.

A few lines from the bottom will be a line Storage Needed if Element Data in Core. The value at
the end of this line is the memory allocation that must be entered on the CONSTANTS sheet
(round it up to the nearest 100,000).

This value is the number of 4 byte long integer numbers required. The actual memory requirement,
in bytes, is this value multiplied by 4. For example, a value of 4,000,000 on this line would require
computer memory of 16 MB to run.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

154

In theory, ANSR-L can run in virtual memory if memory requirements exceed the physical memory
in the computer. In practice, unless the entire problem is retained in main memory the execution
times will be too slow to be practical. The original ANSR-II program was developed with an out-ofcore equation solver and the ability to store element data in disk files. As disk operations are so slow
relative to memory operations, these capabilities are no longer supported in the ANSR-L program.
The input files set flags so that the analysis will not start unless all information can be held in
memory.
If the memory required is more than is physically present in your computer (find out by entering
MEM in DOS mode) then you must either try to reduce memory usage (see next section), borrow a
computer with more memory or upgrade the memory in your computer.
Note that the MODELA spreadsheet provides a procedure for automating this process see Section
7.5 of this Manual.

8.4

MEASURING EXECUTION TIMES

The execution time for each analysis is listed at the bottom of the .D01 file. If you are checking a
number of alternatives to see the effect on run times then it will be more efficient just to run a few
steps rather than the full analysis. Generally 100 to 200 steps will be enough to compare alternative
numbering systems etc.
The default .ANS files are set for a 20 second duration and so the number of steps is set as 20
divided by the time step, e.g. 2000 for a 0.01 second time step. You will find this number 6 lines
from the bottom of the .ANS file as the first field on the line. Change this with a text editor to run a
shortened analysis.

8.5

MINIMIZING MEMORY REQUIREMENTS AND EXECUTION TIMES

For large problems, there are several factors to consider to minimize memory requirements. In
general, memory requirements are reduced by reducing the amount of storage required for the
stiffness matrix. Reductions in this storage generally also reduce execution times.
The following sub-sections describe some methods of reducing memory requirements and execution
times. Not all methods will have a positive effect on a given problem and there is no sure way to
assess the effect other than to try different options and measure the effect on memory requirements
and run time. For small jobs this will not be warranted but for large models it will be worthwhile
assessing sensitivities to these parameters and select the best values for the particular project.

8.5.1

GRID DEFINITION

The main user variable available to minimize memory and execution time is the size of the problem.
Avoid any unnecessary column lines and elevations and use features such as internal beam or column
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

155

nodes only when absolutely necessary. Rationalize the modeling of openings. If the levels of
window and/or openings differ by a relatively small amount, say less than about 500 mm, use a single
elevation at the average point.
If walls are subdivided within stories to allow for openings, columns do not need to have internal
nodes at these levels and should span from floor level to floor level unless connected to another
element at the internal nodes.
If some walls have a large number of intermediate levels to define openings, other walls that are not
interconnected may not need the same number of intermediate openings.
Watch for
incompatibilities as described in Section 5.1 though.

8.5.2

FLEXIBLE FLOORS

Flexible floors modeled using finite elements can make a large increase to the problem size. This
type of modeling may be unavoidable (for timber floors, for example). In cases where only part of
the diaphragm is flexible (a necked concrete floor) a more efficient model will use two or more
rigid diaphragms at a level connected by flexible floor portions.

8.5.3

COORDINATE NUMBERING

The degree of freedom numbers are assigned by ANSR-L in the order the nodes are input to the
program. The program does not have a bandwidth optimization module and so the nodes should be
numbered so as to keep the average bandwidth as small as possible. (The bandwidth is related to the
maximum node number difference between nodal points of each element).
MODELA sorts and numbers the nodes using a wavefront method based on a user-specified
coordinate axis order. The order is specified in the Analysis Data worksheet where the three axes are
listed in the order in which they are to be numbered. The axis listed first changes most slowly, the
axis listed last most rapidly. For example, if the order is Y, X, Z then Y will be held constant and the
nodes numbered in order of increasing Z for a constant X coordinate. (The coordinate axis
convention is that used by MODELA, with X and Y the horizontal axes).
The general rule is to number the axis with the smallest number of nodes along it the fastest. For
example, if a building had one bay in the X direction and six bays in the Y direction then the
numbering order would be Y, X.
If a building has rigid floors then it is generally most effective to number the nodes by floor because
all nodes at a floor will be connected through the floor master node. Therefore, rigid floor buildings
will be numbered Z, X, Y or Z, Y, X.
For medium and large projects is may be advisable to test several different numbering schemes. The
scheme which provides the minimum total memory requirement will generally be the most effective
(see Section 7.3 for the procedure to assess memory requirements). See Section 7.5 for an automated
procedure for this.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

156

8.5.4

PLACEMENT OF DIAPHRAGM NODES

When rigid diaphragms are used, a master node is added at the center of mass of the floor. The
default is to locate this node in the sort order defined for all other nodes. Depending on the type of
structure, there may be advantages to locating these nodes differently.
The Interaction worksheet has a diaphragm node flag (see Section 7.5). This is generally 0 but may
also be 1 or 2, in which case the diaphragm nodes are numbered either after all nodes at a level (Code
1) or after all nodes (Code 2).
It is not well defined under what conditions a 1 or 2 will be more effective than the 0 and, as for the
other parameters, a number of sensitivity runs to determine the effect on total memory requirements
may be justified. The automated procedure described in Section 7.5 can also be used to determine
an appropriate value for this parameter.

8.5.5

DEGREE OF FREEDOM ORDER

The order in which the degrees of freedom are numbered is generally starting from 1 with node
number 1. This may be reversed, with the first degree of freedom being the last node and the dof
numbers increasing as the node numbers decrease.
The degree of freedom order can make a difference in execution time. This is because the stiffness
matrix is not completely reduced when a yield event occurs. Only the coefficients at dofs greater
than the highest numbered dof at which a change occurs are reduced. For certain types of building
this can make a big difference.
For example, if a base isolated building is numbered from the base up then most stiffness changes
will be at the first few dofs, those connected to the ground. Therefore, at each change almost the
entire matrix will need to be reduced. If the nodes are numbered from the top down then the dof
numbers where changes most occur will be close to the highest numbered dofs and the number of
reductions will be much less. This can more than halve run time.
For large jobs, it may be worth trying the model numbered both ways see Section 8.4 for measuring
the execution times.

8.6

MODES FOR RUNNING ANSR-L

The batch file written by MODELA is set to run all the ANSR files in batch mode. The program
can also be run in a number of different modes directly from a DOS screen. The four options
available are:
ANSR [-W] filename
ANSR B filename
ANSR X [-W] filename
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

Interactive
Background (Batch mode)
Automatic interactive
157

ANSR R filename

Restore saved analysis

The B mode is most useful where there are multiple analyses to be performed. The interactive
mode is useful to debug the model and observe the deformed shape of the structure during the
analysis.
The interactive mode waits for user input as each stage, and at the end of the analysis. For this
reason, do not set the first run in a batch file as interactive and leave the file to run unattended as it
will wait for user input at the completion of the first analysis. Use the X mode in this situation.
The optional W token for the interactive modes can be used to reverse the screen colors during
animation. The default is to animate on a white background. If the W token is used the
background color will be black.
8.6.1

BATCH MODE

When run in the batch, or background, mode ANSR-L does a data check and then continues directly
into the analysis, provided there are no fatal errors, with no intervention from the user.
As the analysis progresses the screen will show progress as the integration step number and the time.
The progress screen also lists the number of stiffness changes so far, the applied loads for the current
step and the unbalanced load for the current step.
The stiffness changes indicate the total number of times that the stiffness matrix has been changed,
which generally corresponds to the number of steps at which a change in yield state in at least one
element has occurred. For some analyses, where iterations within a time step are required, there may
be more stiffness changes than time steps. This is rare. The number of stiffness changes is a good
indication of the degree of non-linearity for a particular analysis.
The applied load value is the Euclidian norm of the inertia loads, that is, the square root of the sum
of the squares of the inertia loads at all nodes. This provides some measure of the effective
acceleration at a particular time step. The unbalanced load is a similar Euclidan norm of the
unbalanced loads at each time step, generally due to yield overshoot. The unbalanced load is applied
as a corrective load at successive steps and so will tend to reach large values during extensive yielding
and then reduce when response is largely elastic.
The only active key when ANSR-L is running is batch mode is the <Esc> key, which is used to
terminate the analysis. When this key is entered, the program will write the envelope of results for
the portion of the analysis completed and terminate. If the SAVE option is activated, the .SAV file
will be written and the analysis can be re-started if required.

8.6.2

INTERACTIVE MODE

The interactive mode, the default mode when there is no token on the command line, differs from
the background mode in that it displays the undeformed shape, the animated shape as the analysis
progresses and time history of nodal and/or element results if requested. At each stage, the program
accepts user input to define what is shown.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

158

When ANSR is first activated, the screen shown in Figure 8-2 appears with the name of the input file
as provided on the command line. This screen will indicate whether or not the input file exists in the
directory. If not, you can type in the correct name. When the name is correctly entered, press the
<Esc> key to continue to read the input file and display the undeformed shape.

FIGURE 8-2 ANSR-L OPENING SCREEN

8.6.2.1

Undeformed Shape

The undeformed shape will be displayed, as shown in Figure 8-3, once the input file has been read in
with no fatal errors. When this screen is displayed, a number of keys are active as listed in Table 8-1.
Notes for Table 8-1:
1.

Options N, :, and V toggle the action on and off with successive keystrokes.

2.

The model orientation and size, as adjusted and as displayed on the screen at this stage, will
form the basis for the animated shapes.

3.

Option N will display node numbers unless only 1 element group is displayed in which case it
will display element numbers.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

159

4.

Option : toggles filling of panel elements on and off. The fill state when this screen is exited
will define the fill state for the animated display. Fill may not work correctly when elements
are deformed during animation.

5.

See Section 8.8 for a discussion of the use of the <F1> key.

This screen is generally used to check element connectivity etc. If the model appears satisfactory,
entering M will continue to the next menu. Otherwise, pressing the <Esc> key will terminate the
analysis.
TABLE 8-1 KEYS ACTIVE FOR UNDEFORMED SHAPE

KEY

F
B
L
R
U
D
N
:
V
<Esc>
1
2..n
.nn
X
Y
Z
<F1>
C
Any other key (for example, M)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

ACTION
Rotate
Zoom Forward
Zoom Backward
Pan Left
Pan Right
Pan Up
Pan Down
Number Nodes/Elements
Fill Panel Elements
Make diaphragm ties visible
Terminate analysis
Display Group 1 elements only
Display Group n elements only
As above when Group > 9
Draw with X axis vertical
Draw with Y axis vertical
Draw with Z axis vertical
Dump screen to bitmap file
Reset defaults
Continue with analysis

160

FIGURE 8-3 UNDEFORMED SHAPE

8.6.2.2

Animation Menu Settings

The animation menu will be displayed after the undeformed shape display (Figure 8-4). This menu is
used to set options for displaying the animated shapes during the analysis plus any time history or
element component plots.
The default in the interactive mode is to display an animation of the deformed shape with a
maximum deflection scale factor set to 1% of the building dimension along the vertical axis
(normally, 1% of the building height). This menu allows the default animation to be changed and
also allows up to 6 plots of node or element results to be displayed on the screen and updated as the
analysis progresses.
NOTE: Be careful in entering data in the fields in this menu as the data entry is not a
full text editor. You need to type in your values and then press <Enter>. Check that
the field displays what you want. If not, type it in again.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

161

The values set on this menu are:


1.

Display code, on the first line. This is for development, ignore it for now.

2.

Display animated shape. Toggle for available choices (see Section 7.7.1 for discussion).

3.

Maximum displacement. The deformed shape is scaled such that 10% of the screen width
equals the maximum displacement set here. You may need to experiment to determine a
satisfactory scale factor.

4.

Plots of node or element results. Toggle the first field on this line to obtain choices of nodal
results, element results or energy plot. (The energy plot is under development, try it at your
own risk). The remaining fields on each line depend on the choice selected for the first field.
See Section 7.7.2 and 7.7.3 for details.

Once you have selected the animation parameters and any results plots, press the <Esc> key to
continue with the analysis.

FIGURE 8-4 ANIMATION SELECTION

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

162

8.6.2.3

Animation During Analysis

Once the analysis commences, a screen display as shown in Figure 8-5 will be displayed and updated
each time step as the analysis progresses. While this screen is displayed the keys listed in Table 8-2
are active.

TABLE 8-2 KEYS ACTIVE FOR DEFORMED SHAPE

KEY
<Esc>
:
S
C
H
D
U
<F1>

ACTION
Terminate analysis
Fill Panel Elements
Step through analysis
Return to Continuous Mode
Halve display increment
Double display increment
Toggle undeformed shape
Dump screen to bitmap file

Notes:
1.

The animation is originally in continuous mode. If you press S, the mode switches to step-bystep mode, where the program will progress one time step each time you press the S key. It
will continue in this model until you press the C key.

2.

The screen display is updated each step. For simple models the time taken to update the
display may be a significant portion of the total time. Each time D is pressed the display
increment will be doubled, that is, every 2 steps, 4 steps, 8 steps etc. up to the maximum
number of time steps. The step can be halved by pressing the H key.

3.

See Section 8.8 for a discussion of the use of the <F1> key.

The analysis will continue for the number of time steps specified unless the <Esc> key is pressed, in
which the case program will write the envelope of results for the portion of the analysis completed
and terminate. If the SAVE option is activated, the .SAV file will be written and the analysis can be
re-started if required.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

163

FIGURE 8-5 SCREEN DISPLAY DURING ANALYSIS

8.6.3

AUTOMATIC INTERACTIVE MODE

The automatic interactive mode, -X, is similar to the interactive mode in that it displays the deformed
shape as the analysis progresses but this mode does not wait for any user input. The animated shape
will use default values unless you have modified the top of the input file to set parameters (Section
7.7).
This mode is useful if you want to observe the first of a series of analyses, to ensure that the response
is reasonable, but then want the remainder of the analyses to continue unattended. This is achieved
by changing the first line in the batch file from B to X.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

164

8.6.4

SAVE AND RESTART OF ANALYSES

The MODELA spreadsheet provides an option to SAVE analyses for restart (see Section 7.6). When
this option is activated, the input files for ANSR-L contain the codes for saving analysis results for a
re-start.
To use this option, the analysis can be stopped at any time by pressing the <Esc> key. When this is
done, the program writes the complete memory contents into a file named aaaaaa.SAV. Provided
this file has not been deleted, the analysis can be continued at any time by entering ANSRL R
aaaaaa.
When an analysis is re-started, screen displays are not restored. Therefore, the re-start analysis
continues as though the B flag had been entered for the initial analysis, regardless of whether the
analysis was interactive or background.
Generally, the analyses will be running from a batch file with a sequence of input files. When the
<Esc> key is pressed, the current run will terminate but the next run will then start. There is no
elegant way to stop this pattern, you need to enter the <Ctrl><Alt><Del> re-boot sequence. When
this is done, if you have shelled from Windows you can terminate the task, or if in DOS the
computer will re-boot. The batch file is set to delete the .SAV files, as they are typically large.
However, to permit this re-boot sequence to work, the file delete is not done immediately after the
analysis, but rather after the one subsequent to this. This may seem confusing, look in the batch file
and you will see how this is accomplished!
If you have re-booted, it is advisable to clean up the directory. There may be a series of files with
strange names, for example, BABBaaaa. Delete these. You should also delete any output files
which may have been created by the run which was aborted.
To re-start from you batch file, edit the batch file with for example DOS EDIT or WordPad. Delete
all the lines up to the file which you stopped using the <Esc> key. On this line, change the B to
R. Re-submit the batch file.

8.7

SETTING INTERACTIVE SCREENS IN INPUT FILE

As described above, for interactive mode ANSR-L has a menu to set screen displays during analysis.
The menu operation is quite primitive and time consuming to set each time the analysis is run if a
number of time history plots are required.
An alternative method, which has not yet been implemented via MODELA, is to add a series of lines
to the top of the input file to define the screen plot requests.
In time, MODELA will be modified to include these lines. In the meantime, you can activate this
option by creating a file with the screen definition lines in it. Then, copy this file to the top of your
ANSR-L input file and run it in the interactive mode. (It is best to have the data in a separate file so
that you can copy it in each time you create a new input file through MODELA).
If present in the ANSR-L input file, the following lines must by present in their entirety at the very
top (before the START line) of the ANSR-L input file:
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

165

*MENU
1
1
0.2
0.0
-1
1.0
1
2
2
0
0
0
START

0.0
56
4
4
0
0
0

1
5
5
0
0
0

1
2
3
0
0
0

0
0
0
0
0
0

0
0
0
0
0
0

0.2
0.03
400
0.0
0.0
0.0

0.0
0.0
0.0
0.0
0.0
0.0

This line always required


IPINC (1)
Animation (2)
Maximum Displacement (3)
Angles and (4)
Undeformed shape (5)
Factor to scale model (6)
6 Lines of optional time
history plots. Always insert
all 6 lines, even if no time
history plots are requested.
Use zeroes for plots not
required. (7).

NOTES:
1.

This first code, IPINC, is not currently used. Set to 1.

2.

The options for animation of the structures deformed shape are as listed in Section 8.7.1.
Normal values are 2 for none, 1 for display, 4 for an animated display plus the display of a bar
chart of number of hinges. (Option 4 is useful only for frame structures).

3.

Estimate the maximum displacement anywhere in the structure. This sets the scale for
displaced shape animation.

4.

The angles and define the orientation of the 3D structure. Use 0.0, 0.0 for a plane (XY)
structure. You can override these values using the arrow keys when the undeformed shape is
on the screen. Usual 3D values are 30.0 20.0.

5.

Set this flag to 0 to have the undeformed shape displayed over the deformed shape, -1 to
suppress it. During animated display, the U key will toggle this option on and off.

6.

Scale factor. Use 1.0 for this in most cases. Some models will only fill part of the screen and
this can be used to increase the size of the model.

7.

Up to 6 plots of either nodal components or element components can be displayed during


solution. Use these lines to define any plots required. The first key sets the type of plot. A
value of 1 indicates nodal components, 2 indicates element components. This remaining
values depend on the type of plot, as follows:
Nodal Responses (Field 1 = 1).
Plots of nodal responses versus time.
Field 2 =
Field 3 =
Field 4 =

Node number.
Direction for response (1 = X, 2 = Y, 3 = Z)
Type of response (see Section 7.7.2)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

166

Field 5 =
Field 6 =
Field 7 =

Second node number


only required for relative displacements (If Field 4 = 4).
0 (not used for nodal responses).
Maximum scaled value on plot axis.

Element Components (Field 1 = 2).


Plots of element components versus time or versus other components (for example, Column
P versus M).
Field 2 =
Group Number
Field 3 =
Element Number
Field 4 =
Type of response (see Section 7.7.3)
For line elements (2D or 3D beams or columns)
add the number of components to this value to plot the response
at End J.
Field 5 =
Other component to plot (set to 0 to plot versus time).
Field 6 =
0 (not used for element responses).
Field 7 =
Maximum scaled value for first component.
Field 8 =
Maximum scaled value for second component. (Set to 0.0
for plots versus time).
For animated displays, the user must provide estimates of maximum values for scaling before an
analysis. Often a preliminary analysis is required to get estimates of maximum values.

8.7.1

OPTIONS FOR ANIMATION

The animation options, which are the same as those toggled from the menu, are listed in Table 8-3.
The basic options are to show no animation (2, 6), the displaced shape (1, 5), the hinge patterns (3, 7)
or both hinging and displaced shape (4, 8).

TABLE 8-3 ANIMATION OPTIONS

Value
1
2
3
4
5
6
7
8

Option
YES
NO
HINGING
H + DIS
YES + SAVE
NO + SAVE
HINGING + SAVE
H + DIS + SAVE

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

167

The HINGING option displays a bar chart showing the number of beam plastic hinges and column
plastic hinges at each time step. This option is generally only meaningful for frame type structures
(Figure 8-5 shows this option).
Options 5 to 8 are similar to options 1 to 4 except that they save the screen at each time step for
conversion to bitmaps and AVI movie files (see Section 8.8).

8.7.2

OPTIONS FOR NODAL COMPONENTS

The nodal component plotted can be any one of the four options listed in Table 8-4. Option 4,
relative displacement, requires a second node number to be entered. The plotted curve is then the
difference between displacements at these two nodes.

TABLE 8-4 NODAL COMPONENT OPTIONS

Value
1
2
3
4

8.7.3

Option
Displacement
Velocity
Acceleration
Displacement Relative

OPTIONS FOR ELEMENT COMPONENTS

The number of element components varies depending on the element type, as shown in Table 8-5.
The line elements (2D and 3D beam-column and degrading beam-column) each have two times as
many components as listed in Table 8-5. The values listed in Table 8-5 are for end I. To obtain the
equivalent values at end J, add the number of components for that element type to the value for End
I.
The gap element used by MODELA is the velocity dependent gap-friction for gaps, rather than the
standard gap friction. The 2D beam-column element is used for beams and the 3D beam- column
for columns. If degradation is specified for flexural elements then the 3D degrading beam-column is
used for beams and columns.

TABLE 8-5 (A) : ELEMENT COMPONENT OPTIONS

Truss
1
2
3
4

Yield Code
Axial Force
Extension
Accum. Ext +ve

3D Beam
Column
Yield Code 1
Yield Code 2
Y Bending Moment
Z Bending Moment

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

2D Beam
Column
Yield Code
Bending Moment
Shear Force
Axial Load
168

Truss
5
6
7
8
9
10

Accum. Ext ve

3D Beam
Column
Torsion T
Axial Load P
Y Plastic Rotation
Z Plastic Rotation
T Plastic Rotation
P Plastic Rotation

2D Beam
Column
Plastic Rotation
Accum +ve Pl Rot
Accum -ve Pl Rot

(B) : ELEMENT COMPONENT OPTIONS

1
2
3
4
5
6
7
8
9
10
8.8

Gap
Friction

Degrading
Shear

Velocity
Dependent
Gap-Friction

Yield Code N
Normal Deform
Normal Force
Yield Code F
X Frict Deform
Y Frict Deform
X Frict Force
Y Frict Force

Yield Code
Shear Stress
Stress 11
Stress 22
Shear Strain
Strain 11
Strain 22

Yield Code N
Normal Deform
Normal Force
Yield Code F
Velocity
Frict Coeffic
X Frict Deform
Y Frict Deform
X Frict Force
Y Frict Force

3D
Degrading
Beam
Column
Y Moment
Z Moment
Torsion
Y Shear
Z Shear
Axial Load
Y Rotation
Z Rotation
Y Pl Rot
Z Pl Rot

SAVING AS AVI MOVIE

Some development work is under way to save the screen displays for later recovery. This is to enable
the animation to be studied at closer to real time for large models. This option is activated in the
selection of animation type (see Section 8.7).
When one of the SAVE options is selected then the screen display is simplified (no logo or border)
and written to a file at each displayed time step. This is quite a slow process.
The creation of the movie file from the disk files created during the ANSR-L analysis is a two step
process:
1.

Convert the files to bit map format using the HCG program PROCBMP.EXE.

2.

Covert the bit map files to .AVI format using the Windows program AVICON32 (available
from the Auckland office).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

169

This process is still under development and the results are quite crude. The quality may be improved
with time.
Do not attempt to get bit map files if you are also getting element time histories as there will be a file
naming conflict.

8.9

ANALYSIS DOES NOT START

Although MODELA is intended to produce input files completely compatible with ANSR-L there
will always be cases where the analysis will not run. This may be due to bugs in the programs, or
complexities in the model which are not properly accounted for by MODELA.
When the analysis terminates, it may leave the screen display still in graphical mode. If
this happens, type MODE 80 <Enter>. This should return the display to normal.
Analyses with errors may terminate while reading the input file or when the stiffness matrix has been
assembled. In the latter case, you may get screen messages that there is a zero divide in Subroutine Dot.

8.9.1

FILE OPENING ERROR

When element time histories are requested, a large number of files may be opened simultaneously by
ANSR-L and also during the processing phase by PROCESSA. See Section 3.4 Computer Setup for
information on avoiding a potential error due to this.

8.9.2

ERROR DURING READ INPUT PHASE

For an error in the input file, the last file written (sort by date) will be the .ECH file, which will be
incomplete. Open this file with a text editor and go to the bottom of the file. The last line will give
an indication of where the error occurred. Hopefully, this will provide some clues as to why the
analysis terminated.
Often the last line will be an element which caused the problem. The element will be identified by
the nodes at ends I and J for a flexural member or the 4 nodes for a plane stress elements. To find
out where these nodes are, you will need to go to the .TRC file, which is written by MODELA at the
same time as the ANSR input files. This file can be read with WordPad or any other text editor. The
first lines will be the level numbers, followed by the coordinates. Each coordinate line provides the
ANSR node number, the level number, the column line number, a diaphragm flag (0 unless a center
of mass node), then the X, Y and Z coordinates. These data lines allow you to locate an ANSR
node in terms of the MODELA levels and column lines (see also Section 10.2 for an alternative
method of finding node reference column and level numbers).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

170

A more effective method of obtaining column and degree of freedom numbers is to


read results from the ANSROUT spreadsheet. This will import node and degree of
freedom information even if the analysis is not complete. See Section 10.
A common problem is that some elements have zero stiffness. This may be because a dimension or
strength is set to zero. Check the information at the start of the element group to make sure that all
relevant stiffness and strength parameters have non-zero values.
If you have not defined sufficient memory for the problem as defined then the analysis will not start.
The end of the .ECH file will provide information as to total memory required.

8.9.3

ALLOWABLE DIMENSIONS EXCEEDED

ANSR-L has fixed dimension arrays used to store element information such as stiffness, strength,
eccentricities and loads. The consequences of exceeding these dimensions are unpredictable. The
latest version of ANSR-L has been set to be rigorous in ensuring that these dimensions are not
exceeded. If the dimension is exceeded, the last lines in the .ECH file will be similar to
MAXIMUM OF 200 STIFFNESS TYPES ALLOWED
As ANSR-L is developed the dimensions will be changed so as to be allocatable rather than fixed.
However, in the meantime, if you get this message you will need to reduce the number of stiffness,
load patterns, eccentricities or whatever caused the error.
The MODELA program groups elements according to material type. For example, all columns of a
particular material number will be in a separate group. Therefore, you can reduce the number of
variations by splitting the columns into more than one group by defining two or more materials, each
with the same physical properties but a different identification number. The column section
numbers are similarly split, with the same dimensions but the different material identifications. Some
of the columns will then be defined with the alternative section types.

8.9.4

ERROR DURING EQUATION SOLUTION

If the analysis terminates because of a zero divide, the message printed on the screen should provide
a do loop range, from N to M. (This message is repeated in the last file written by ANSR, either .S01
or .D01). The zero stiffness will be one of these two degrees of freedom, N or M. To find the
location of these,
1.

Open .ECH with WordPad or some other text editor.

2.

Search for the phrase degrees of freedom.

3.

Each node will be listed together with the numbers of the degree of freedom for X, Y, Z
translation then rotation about the X, Y and Z axes.

4.

Find the degree of freedom numbers, and record the node numbers and directions.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

171

5.

Go to the .TRC file and find the level number and column line for those nodes, as described
above.

6.

Check all elements incident on these nodes to try to determine why there may be no stiffness
attached.

An alternative to Steps 1 to 5 is to import results into the ANSROUT workbook. The Coordinates
sheet in that workbook provides the degree of freedom numbers and the column line and level
equivalence. See Section 10.2.
Sometimes it can be quite difficult to identify the problem. One option is to fix the suspect degrees
of freedom (see DOF sheet in MODELA.XLS) to get the analysis running and then check the
response for odd behavior.

8.10 ANALYSIS TERMINATION

Once the analysis had started, it will terminate for one of a number of reasons:
1.

The specified earthquake record duration is reached. This is a normal termination.

2.

User terminated with <Esc> key. This is a normal termination. If the SAVE option has been
activated then the analysis can be completed later using the R option.

3.

There is a zero or negative stiffness on the diagonal (see Section 8.9).

4.

The maximum displacement is exceeded. This generally indicates that the structure has
collapsed, either globally or locally. A message will be written to the .D01 file before the
envelope values.

5.

Convergence cannot be achieved. This is because unbalanced loads exceed the limit specified.
A message will be written to the D01 file after the envelope values and before the time log.

6.

The disk is full. Free up disk space and re-submit.

7.

Any other error.

When you have submitted the files as a series through the batch file created by MODELA, it may not
be clear whether the analyses have been completed or not. The ANSROUT spreadsheet has an
option to obtain envelope times (see Section 10). This option reads the D01 output files to find the
time at which the envelope is written. If this is less than the specified length of record (generally 20
seconds) then an abnormal termination occurred. This may occur for all files or for only some of the
files.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

172

8.10.1 DISPLACEMENT EXCEEDED

The ANSROUT spreadsheet has an option to write to file the node numbers where displacements
exceed a specified amount. This helps to find out where the maximum values were exceeded.
Running the input file in interactive mode often aids in interpreting performance.
In most cases, the displacement exceeded message indicates that the structure is unstable and requires
strengthening to resist this level of earthquake load.
However, the displacement exceeded message is also sometimes a local effect which may be due to a
fictitious or real local instability.
A fictitious local instability is caused by modeling features which result in a very low stiffness at some
degree of freedom, for example a cantilever which yields under gravity load. This type of instability
can generally be fixed by changing the model.
A real local instability can occur, for example, in an unreinforced masonry building with a large
number of wall panels. Some portion of the wall may collapse, but not the complete line of
resistance, so that the stability of the structure is not endangered. In these circumstances this damage
may be acceptable. The analysis can be completed by setting the maximum displacements very large.
The failing elements will often cause large unbalanced loads so a reduced time step may be necessary
for this type of analysis.
If you are permitting local collapse in you model you will need to use alternative acceptance criteria
as maximum displacement (or drift) is usually the primary criterion. In this type of structure, it is
often advisable to plot the time history of displacements at the diaphragm nodes to ensure overall
stability.

8.10.2 CONVERGENCE NOT ACHIEVED

The analysis will terminate with this message when the unbalanced loads exceed the limit set in the
MODELA spreadsheet. This message most commonly occurs in models with gap elements or
degrading panel elements.
A step-by-step analysis program such as ANSR-L assumes that all elements remain elastic for the
duration of the time step and then checks the status at the end of the time step. If a stiffness change
would have occurred part way through the time step then there will be an unbalanced load because
of the overshoot. This is applied as a corrective load during the subsequent time step but causes a
temporary unbalance (violation of equilibrium).
If the unbalanced load exceeds the specified tolerance then the program will iterate within the time
step in an attempt to achieve equilibrium within the tolerance.
For most models very few iterations are necessary and the solution remains stable. However, some
element types provide large unbalanced loads, especially gap elements. If a gap element is in tension
and closing then it can displace a relatively large amount within a single time step. If the gap would
have closed then this displacement times the compressive stiffness gives rise to a very large
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

173

unbalance. This problem is aggravated if there are a large number of gap elements which may close
within the same time step.
There are two ways of mitigating the unbalanced load effects in gap elements:
1.

Use as small a compressive stiffness as possible for the element.

2.

Reduce the time step as necessary.

There are other solution strategies to avoid the need for fine time steps, such as variable steps or the
event-to-event solution. However, these have yet to be implemented within ANSR-L and this will be
at some future time.
The other cause of large unbalanced forces is a structure on the verge of instability. Check the
magnitude of displacements and element stresses and deformations as these may show that the
structure needs strengthening for this level of earthquake.
You may want to experiment with a larger tolerance on forces if you have analyses terminating
because of lack of convergence when the response otherwise seems satisfactory. The screen during
batch mode lists the magnitude of unbalanced forces. If the structure is unstable then this will
usually be reflected in the pattern of the unbalanced forces, which will increase with each step rather
than fluctuate up and down as for a stable structure. See Section 8.12 for further information on
numerical stability.

8.11 OUTPUT FILES

Each ANSR-L analysis will produce a series of output files. These will be one or more of:
.ECH
.S01
.S02 etc
.D01
.ONN
.SAV
.001
.002
.00n

Data echo
Static Analysis Results
If more than one load case
Dynamic Analysis Envelopes
Dynamic Analysis Node Time Histories
Save for re-start
Element Group 1 Time History
Element Group 2 Time History
Element Group n Time History

NOTE : File names .001, .002 etc will be used to store bitmaps if you are saving screen
images. Do not attempt to get element time histories and also save screen images in the
same run.
These file are usually read and imported by ANSROUT / PROCESSA. However, they can be
opened and edited using any text editor (for example, WordPad).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

174

There is a batch file, EATA.BAT, in the c:\hcgprogs directory. This file can be used to delete
unwanted output files. Enter EATA aaaaa to delete all output files for analysis aaaaa. Enter EATA
* to delete all output files for all runs in a directory. If they are valid runs, make sure you have
processed the results before doing this.

8.12 NUMERICAL STABILITY OF SOLUTION

As discussed previously, the two input parameters which control the accuracy and stability of the
solution procedure are the integration time step and the force tolerance. This latter parameter
defines the unbalanced load beyond which the program will iterate within a time step to achieve
equilibrium rather than apply a corrective load at the next time step.
The effect of these parameters have been studied on a 14 story coupled shear wall building under
earthquake loads. The building elastic period is approximately 1.0 seconds. This type of building
contains softening elements (shear cracking) and stiffening elements (concrete compression blocks at
the base flexural hinges). A total of 8 analyses were performed, as listed in Table 8-6:

Run 31A0 was the baseline analysis, using the default time step of T/100 = 0.01 seconds and a
force tolerance of 0.1 times the total weight.

Runs 32A0 and 33A0 used the same time step but decreased the force tolerance by factors of
100 and 10 respectively.

Runs 32A0 and 33A0 also used the same time step but increased the force tolerance by factors of
10 and 100 respectively.

Runs 36A0 and 37A0 used the same force tolerance as the baseline analysis but decreased the
time step by factors of 2 and 4 respectively.

The final analysis, 38A0, reduced the force tolerance by a factor of 100 and the time step by a
factor of 4. For comparative results, this is considered the most accurate analysis.
TABLE 8-6 VARIATIONS IN SOLUTION PARAMETERS

Tolerance / W
Time step

31A0 32A0 33A0 34A0 35A0 36A0 37A0 38A0


0.1 0.001 0.01
1
10
0.1
0.1
0.001
0.010 0.010 0.010 0.010 0.010 0.005 0.0025 0.0025

Each model was evaluated for 20 seconds of a spectrum compatible time history. A comparison was
made of the global response parameters (top displacement and base shear) and local response
parameters (wall shear strain, wall flexural crack at the base and concrete compression force at the
base hinge). Table 8-7 and Figure 8-6 compare the results. All have been normalized to analysis
38A0, the most accurate analysis.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

175

The recommended values of time step and tolerance (31A0) produce results with a maximum
variation of 6% from the most accurate analysis in wall strain and a smaller variation in all other
parameters. This would be considered an acceptable margin.

For this example, a smaller tolerance does not improve results (32A0 and 33A0).

A larger tolerance provides inaccurate results in the gap element compression although other
parameters are within acceptable limits (34A0 and 35A0).

A smaller time step improves results although the improvement is relatively small compared to
the almost 100% increase in runtime.

These results justify the recommended values used in this manual. However, the dynamic
characteristics of some structure may be such that other values are appropriate. If you have any
indication of numerical instability in your results, it is advisable to perform sensitivity analyses such as
this on a particular project.

TABLE 8-7 EFFECT OF SOLUTION PARAMETERS

Displacement
V/W
Crack Width
Compression
Shear Strain

31A0
2%
-1%
1%
0%
-6%

32A0 33A0 34A0 35A0 36A0 37A0 38A0


-8%
5%
2%
2%
6%
1%
0%
-1%
0%
-1%
-1% -1%
0%
0%
-1%
7%
1%
1%
4%
1%
0%
9%
0% 131% 131% -1% -2%
0%
-4% -1% -6%
-6% -1% -2%
0%

NOTE : This example used flexural yield gaps at the base. These gaps have a tension
strength provided by the reinforcing bar. Rocking walls without tension ties appear to
be more sensitive to the force tolerance and a smaller tolerance is recommended,
usually 0.01W. As for all types of complex mode, do some parametric analyses and
check equilibrium and stability of results.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

176

FIGURE 8-6 EFFECT OF SOLUTION PARAMETERS

RESPONSE NORMALISED TO 38A0

2.50
Displacement
V/W
Crack Width
Compression
Shear Strain

2.00

1.50

1.00

0.50

0.00
31A0

32A0

33A0

34A0

35A0

36A0

37A0

38A0

ANALYSIS NUMBER

8.13 ACCUMULATION OF ENVELOPES IN ANSR-L

The results from time history and pushover analysis from ANSR-L analysis which are used for most
of the evaluation are the envelope values, the peak maximum and negative values which are
accumulated through the analysis and printed at the end.
The way in which ANSR-L accumulated envelope values for elements can influence results in some
circumstances when iterations are performed within a time step (when the unbalanced loads exceed
the specified force tolerance). This is because the element envelopes are updated each time the
element response is calculated. During iterations the element results at intermediate steps may be
higher than the actual values at the end of the time step. These erroneous values will be recorded as
envelope values even though they are corrected later within the same step.
This appears to have a practical effect only in stiffening systems. For most element types the
stiffness degrades and so this is not an issue. However, for a closing gap element the stiffness
increases greatly when impact occurs. For this revision, the gap element in ANSR-L has been
modified to accumulate element envelopes only at the final iteration. This appears to have corrected
erroneously large compression forces indicated in previous versions of the program.
For technical reasons, it is preferable not to make this change to other element types unless there are
benefits to be obtained. At the time the element forces are calculated it is not known whether there
will be further iterations. Therefore, the change to the gap element required two copies of the
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

177

envelope to be maintained for every gap element so that the previous values could be reverted to if
iterations were performed. If this were done for every element type it would increase memory
requirements and execution times for uncertain benefits.
If you are suspicious about envelope results for any element component then the envelope values can
be checked against time history output for a particular element the time history of results is only
printed at the last iteration. If you do encounter any discrepancies, notify the Development Division
and we will investigate making this change to other element types.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

178

9 NSP (PUSHOVER ANALYSIS)

A pushover analysis is an implementation of the Nonlinear Static Procedure (NSP). A non-linear


model of the building is subjected to an increasing lateral loads. The procedure is fully specified in
the U.S. documents FEMA-273 and ATC-40. The implementation of these is termed the Force
Pushover in these procedures. The FEMA and ATC-40 NSP methods are an alternate to the time
history method as a means of performing a complete building evaluation.
An alternate NSP is where a static displacement is applied rather than a static force. This procedure
has the advantage is being relatively simple to apply and being able to track the capacity curve under
cyclic displacements. However, it does not currently form a complete evaluation method in terms of
any published codes of practice or evaluation guidelines. It is most often used as a supplement to a
time history analysis to provide additional information about the structure being analyzed.

9.1

NSP OR FORCE PUSHOVER

As the NSP procedure in not documented for NZ use, most of our projects have used the NDP
(time history analysis). We do not have such a large database of completed projects which use NSP
as we do for NDP. These sections provide some guidance in this procedure but if you use the NSP
on your project, consult the Development Division to ensure that you get the latest revisions to the
evaluation tools.
Make sure that you read through the sections of the FEMA-273 Guidelines and
Commentary before you attempt to apply the NSP. The procedures discussed here
automate the process but you still need a thorough knowledge of the method to ensure
valid results.
In the application of the NSP, a target displacement is calculated and all structural components
evaluated against acceptance criteria for the forces and deformations at this target displacement.
FEMA-273 provides two methods of calculating the target displacement:
1.

Method 1 is based on statistical data studies of single-degree-of-freedom yielding systems.


An effective period is defined and the elastic displacement at this period is modified by
factors accounting for inelastic actions and hysteresis shape effects. This results in a target
displacement for a specific level of earthquake.

2.

The alternative NSP procedure in the FEMA Commentary, Method 2, is based on the
ATC-40 requirements. This method is based on calculating the target displacement as the
point where the spectral capacity curve and the demand design curve intersect. This method
incorporates damping by establishing a demand design curve by reducing the 5% damped
acceleration spectrum to allow for the equivalent damping due to inelastic action. As the
damping is displacement dependent, the solution for the target displacement is iterative.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

179

Once the target displacement is obtained, both methods follow the same evaluation procedure.
FEMA does not express a preference for either method although the fact that Method 1 is in the
Guidelines and Method 2 only in the Commentary seems to imply greater acceptance of Method 1.
Our HCG spreadsheets for NSP evaluation include both methods.
Experience to date on the NSP is that Method 1 is best suited to stiff structures, such as URM
buildings. Method 2 is better suited to more flexible structures such as moment frames. For most
projects, we attempt to calculate target displacements using both methods. In some cases, one of the
methods will not provide a valid solution, which generally indicates that the other method is best
suited for that class of structure.
9.1.1

LOAD VECTORS

FEMA-273 requires that the load vectors be profiles that approximately bound the likely distribution
of inertia forces in an earthquake. It requires that at least two patterns of lateral load be considered.
1.

A uniform pattern, based on lateral forces that are proportional to the total mass at each
floor level. This is termed the uniform distribution these guidelines.

2.

A lateral load pattern based on Cvx (the FEMA-273 equivalent static load distribution) if
more than 75% of the total mass participates in the fundamental mode. This is termed the
power distribution these guidelines.
or
A lateral load pattern consistent with the story shear distribution calculated from a Response
Spectrum Analysis which includes at least 90% effective mass.

The lateral load based on Cvx, option 2. above, is calculated from the formulas:

Fx =

wx hxk
n

w h

k
i i

i =1

The power, k = 1.0 for T 1.0 and 2.0 for T 2.5. Linear interpolation is used to calculate values of
k for intermediate values of T.
The ATC-40 provides an alternative lateral load distribution based on the first mode of vibration
only. Lateral loads are applied in proportion to the product of story masses and the first mode shape
of the elastic model of the structure:

Fx =
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

w x x
V
w x x
180

The MODELA program will write 6 ANSR-L input files when the Force Pushover option is activated
as follows:
File
1
2
3
4
5
6

Direction
X
X
X
Z
Z
Z

Load
ATC-40
Uniform
Power
ATC-40
Uniform
Power

The ATC-40 load vector is not fully compliant with FEMA-273. The uniform load distribution is
always required by FEMA-273. The power distribution is the second FEMA-273 load condition, but
is only applicable when the fundamental mode provides at least 75% effective mass.
NOTE : The response spectrum based load vector has not yet been implemented. If
you need this for an NSP, contact the Development Division.
The ATC-40 pushover procedure requires that factors be used to transform the capacity curve to the
capacity spectrum, which plots Sa against Sd. Definitions used to develop these values are:
N

(w

)/ g

(w

)/ g

i i1

PF1 =

i =1
N

2
i i1

i =1

( wi i1 ) / g
i =1

1 = N
N

2
( wi ) / g ( wi i1 ) / g
i =1
i =1

Sa =

Sd =

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

V /W
1

roof
PF1 roof ,1
181

The PF1 and factors are calculated by MODELA to allow the capacity spectrum to be constructed.

9.1.2

OBTAINING PERIODS AND MODE SHAPES

Preparation of the Force Pushover input file requires a number of steps, as shown in Figure 9-1.
These steps are performed in the order shown.
1.

2.

3.

The first step is to write a linear elastic model to


extract the periods and mode shapes. The input FIGURE 9-1 FORCE PUSHOVER PROCEDURE
file may be formatted for either ZAP (an inhouse general-purpose program) or SAP2000.
Force Pushover
The latter program is preferred if available as it is
Analysis
more user-friendly and has superior on-screen
Write ZAP
graphics.
Model for
Periods

When you have written the ZAP model, run the


batch file in a similar manner to that used to run
ANSR-L. The screen prompts should be selfexplanatory. For SAP2000, import the .s2k file
which is produced by MODELA.

Read Periods
Import Periods
Write ANSR
Model with
Pushover Loads

The Read Periods option executes the program


PROCESSA which reads the SAP2000 or ZAP
output file and writes the periods, participation
factors and mode shapes to a comma delimited
output file. During processing, the factors PF1
and are also calculated.

4.

The Import Periods option opens the commadelimited file and imports it into the MODELA
workbook into the sheet named ForceVectors.

5.

The Write ANSR option reads the data and


writes the NSP files.
The data required to
define the force pushover, described below, need
to be entered before this step.

FIGURE 9-2 DEFINITION OF FORCE


PUSHOVER

Options for Force


Pushover Analysis
2
1
0.5
1000
0.5
0.6

9.1.3

DEFINING FORCE PUSHOVER

Once the Force Vectors have been imported, the force


pushover load case is defined using the portion of the
Interaction sheet shown in Figure 9-2. Definitions are
provided for pushover analyses in the two horizontal
directions, X and Z.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

X Mode Number
Z Mode Number
Base V/W
Number of Steps
X Period
Z Period

Apply Fy

Apply Fx

Apply Fx

Apply Fy

Apply Fz

Apply Fz

Apply Mx

Apply Mx

Apply My

Apply My

Apply Mz

Apply Mz

182

The cells listed are used to define:


1.

The mode numbers corresponding to the fundamental mode in the X and Z directions
respectively. Generally, these will be the modes with the numerically highest participation
factor in the particular directions (see ForceVectors sheet).

2.

The maximum base shear force to apply, as a fraction of the total weight of the structure. The
analysis may not continue to this level of load as numerical instability may occur at a lower
level of load if a mechanism forms. You will need to use judgment to set this value.

3.

The number of steps. Typically, 1000 steps provide good resolution for the capacity curve.
As noted above, if instability occurs then the analysis may terminate after a lesser number of
steps.

4.

The fundamental periods in the two horizontal directions. These are required to calculate the
power function, k, for the FEMA-273 lateral force distribution.

5.

Which force combinations to apply for each direction. Usually, for the X pushover curve only
the X forces will be applied and similarly for Z, only the Z forces. However, the modes will
generally have non-zero components in the other directions and if required these can be
applied. FEMA-273 and ATC-40 are not clear on this aspect.

There are some uncertainties regarding implementation of Force Pushover, particularly regarding
whether to apply only the primary force component, all translational components or all six
components. For now, it is recommended that only the primary force component be applied.

9.1.4

WRITE ANSR PUSHOVER MODEL

The option button to write an ANSR pushover model will write six time history files, as described
earlier. The 1st and 3rd files refer to the ATC-40 procedure and are not used for a FEMA-273
evaluation. For the latter, you can delete these two files from the batch file.
The ANSR analyses are performed using the same procedure as for the time history analysis,
described in Section 8.

9.1.5

RESULTS FOR PUSHOVER ANALYSIS

The pushover analysis produces the force versus displacement function for the complete building.
This capacity if then used to calculate the target displacement using either Method 1 or Method 2
from FEMA-273.
Once the target displacement is calculated, the structural components are
evaluated at this displacement level using the evaluation criteria for the particular components. In
most cases, target displacements will be calculated for two levels of earthquake, usually BSE-1 and
BSE-2.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

183

As the target displacement is not known prior to the analysis, the element results are stored at every
5th step, typically 200 sets for a 1000 step analysis. This increment is selected to keep output files to
a manageable size as output every steps produces extremely large files.
The pushover loads are applied as static load case two and the results from the analysis are written to
file .S02. When results are processed and the target displacements have been calculated, this file is
re-processed and the saved results at the nearest steps are imported into the output spreadsheet for
evaluation.

9.1.6

CALCULATING TARGET DISPLACEMENTS

The ANSROUT post-processing spreadsheet (see Section 10) has an option to read pushover force.
If this option is selected the PROCESSA program will write a comma delimited file for each
pushover analysis with a .DIS suffix. Each row of this file will contain six force components (Fx, Fy,
Fz, Mx, My, Mz) together with one column of displacements for each node and direction for which
output was requested.
The procedure for calculating the NSP displacement has been automated in the spreadsheet
NSP.XLS. The Control sheet from this workbook in shown in Figure 9-3. The calculation procedure
for target displacements follows the step-by-step procedure on this sheet.
Each workbook solves for a target displacement in each direction for a particular earthquake level
and load distribution. You will typically need at least 4 workbook for a complete evaluation, two
earthquake levels and two load distributions.
1.

Select the DIS files, which are produced by the ANSROUT procedure. One file is selected
for each horizontal direction.

2.

Select the file containing the modal analysis information. This file is written as part of the
generation procedure for the load vectors (MODELA).

3.

Import the files. This macro will import the capacity curves and plot them on the control
sheet, as shown on the bottom of Figure 9-3.

4.

Set the column numbers for the forces and displacements. The locations of the columns
for displacements depends on the number of output points. Usually, the target
displacements will be calculated at roof level see FEMA.

5.

Define the spectra. Spectral shapes can be as defined by either ATC-40 or FEMA. This
macro will prompt for the appropriate ground motion parameters.

6.

Define which spectrum is to be used for Method 2, the ATC-40 method. The FEMA
spectrum is always used for Method 1.

7.

Set the performance level and framing type consult FEMA for appropriate values.

8.

Analyze, for either or both of FEMA 273 (Method 1) or ATC-40 (Method 2).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

184

The Analyze option will solve for target displacements in the X and Z directions for the procedures
selected. The process will attempt to solve for the full spectrum but, if the capacity of the structure
is such that this cannot be achieved, the process will calculate the fraction of full load which it can
resist.
Figures 9-4 and 9-5 show examples of the plots produced by the procedure. Note that the
acceleration and displacements for the two methods are not directly comparable. This is because
Method 2 converts the capacity curve to equivalent spectral values using the modal factors and
PF1 described in Section 9.1.1.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

185

FIGURE 9-3 NSP WORKBOOK CONTROL SHEET

Non-linear Static Procedure for Performance Based Design


X
Z

Choose .DIS Files

D:\90000\NSP\53.EIG

Choose .EIG File

Import Files

Pushover Curves

X
Z

Select Modes

X
Z

Force Column
2
Fx
4
Fz
Mode

Direction
1
3

1
2
1
313

Define Spectra

Choose Spectra
2
(Only required for ATC-40 Method)
FEMA Factors

LS
1

Displacement Column
12
C 33-X-L 12
52
C 33-Z-L 12
T
0.477
0.334

PF1
60.446
51.37

a1
0.60
0.43

phiRoof1 Weight
0.0209 5.95E+04
0.0179 5.97E+04

Fraction of Deisgn Load


Specify x-performance point
Specify y-performance point
FEMA

Performance Level
Framing Type

Use FEMA 273 Procedure

Analyse

Use ATC40 Procedure

14000
12000
10000
8000

Fx

6000

Fz

4000
2000
0
0

20

40

60

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

80

100

186

FIGURE 9-4 FEMA-273 NSP METHOD 1

Capacity Demand Spectrum: FEMA273 Procedure, X- Direction:


Fraction of full load: 0.36
Performance Point: 312

Base Shear (Fraction of Weight)

0.25

0.2

0.15

0.1
Pushover Curve
Bilinear Curve
Performance Point

0.05

0
0

10

20

30

40

50

60

70

80

90

100

180

200

Roof Displacement (mm)

FIGURE 9-5 FEMA-273 NSP METHOD 2


Capacity Demand Spectrum: ATC40 Procedure, X- Direction:
Fraction of full load: 0.31
Performance Point: 312

0.45
Pushover Curve

0.4

Bilinear Curve

Spectral Acceleration (g)

0.35

Elastic Demand Spectrum


Reduced Demand Spectrum

0.3

Performance Point

0.25
0.2
0.15
0.1
0.05
0
0

20

40

60

80

100

120

140

160

Spectral Displacement (mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

187

Figure 9-6 is an example of the NSP Method 1 calculations for the Gateway project. This provides
an example of the calculation of the factors required by the FEMA-273 Method 1 procedure.
FIGURE 9-6 EXAMPLE NSP METHOD 1

Parameter
Elastic Lateral Stiffness
Effective Lateral Stiffness
Yield Base Shear / Weight
Strength Ratio
Factor to relate spectral displ. to roof displ.
Factor to relative inelastic displ. to linear response
Modification factor for hysteresis shape
Factor for dynamic P- effects
Response spectrum acceleration at Te
Effective Fundamental Period
Target Displacement

Units
Ki
kip/in
Ke
kip/in
Vy/W kip/kip
R
C0
C1
C2
C3
Sa
g
secs
Te
inches
t

LS
BSE-1
88523
85974
0.294
2.134
1.300
1.387
1.255
1.000
0.816
0.193
0.672

CP
CP
85% BSE-2
BSE-2
88523 88523
85974 85974
0.294 0.294
3.905 4.594
1.300 1.300
1.416 1.416
1.450 1.450
1.000 1.000
1.494 1.758
0.193 0.193
1.451 1.707

0.400
0.350

BASE SHEAR / W

0.300
0.250
Capacity Curve

0.200

Bi-Linear Approximation
0.150

Target Displacement

0.100
0.050
0.000
0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

2.00

DISPLACEMENT (inches)

9.1.7

EVALUATING NSP PERFORMANCE

The procedure described above calculates the target displacement for each level of load. The stresses
and deformations in each component are then evaluated at this displacement level. This requires that
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

188

the results be extracted at this step of the analysis. The ANSROUT processor program writes a
summary files, with a suffix .LOG, to determine which set of results to extract. This is a commadelimited file which can be opened in Excel.
Figure 9-7 shows a portion of a LOG file. The top of the file summarises the element types in the
model, listing the number of elements in each group and the type of element. At each data set which
is saved (every 5th step) the force and displacement is tabulated. A summary is also provided of the
number of elements which exceed respectively the IO, LS and CP performance level criteria.
Portions of the file not included in Figure 9-7 provide more detailed information for each element
group.

FIGURE 9-7 NSP LOG OF RESULTS


GROUP
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

NELS
1487
439
233
143
88
15
30
325
106
60
300
188
27
126
580
203
365

TYPE
8
8
8
8
8
8
8
5
5
5
2
2
2
2
1
1
5

SET

FORCE

DISP

NAME
MASONRY WALL SHEAR PANELS
CONCRETE WALL SHEAR PANELS
USER-SPECIFIED WALL SHEAR PANELS
CONCRETE WALL SHEAR PANELS
USER-SPECIFIED WALL SHEAR PANELS
USER-SPECIFIED FLOOR SHEAR PANELS
CONCRETE FLOOR SHEAR PANELS
CONCRETE BEAMS
STEEL BEAMS
CONCRETE BEAMS
CONCRETE COLUMNS
STEEL COLUMNS
MASONRY COLUMNS
STEEL BRACES
CONNECTING GAP
REINFORCING BARS
RIGID DIAPHRAGM CONNECTORS
SUM OF GROUPS
>LS
0
0
0

>CP
0
0
0

1
2
3

35
208
382

-0.03
0.00
0.04

>IO
0
0
0

62
63
64
65
66
67
68

10629
10802
10976
11150
11324
11497
11671

3.87
4.05
4.24
4.43
4.65
4.88
5.14

0
0
0
1
2
2
2

0
0
0
0
0
0
0

0
0
0
0
0
0
0

99
100
101
102
103
104

17055
17228
17402
17576
17749
17923

18.17
18.74
19.34
20.00
20.68
21.38

74
76
81
86
87
88

17
15
13
10
9
11

9
11
13
17
20
21

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

189

The information in the LOG file is used to determine


which data set to process. For example, if the target
displacement were 20 mm for the example shown in
Figure 9-7 then the appropriate data set for that
earthquake level would be Set 102.
These sets are then specified in the ANSROUT
workbook, as shown in Figure 9-8, and the results
imported. See Chapter 10 for details.

9.2

FIGURE 9-8 NDP SETS


NSP SET NUMBERS FOR OUTPUT
NUMBERS FOR

NUMBERS FOR

CRACK FILES

FORCE OUTPUT

10
36
44

47
63
90

44

90

DISPLACEMENT PUSHOVER

With this option, a displacement trace is applied at a specified location of the model, typically at the
roof level. This allows the hysteresis of displacement versus force for the entire structure to be
generated.
The pushover displacement is applied by inserting a unit length truss element between the column
line specified for the application of the displacement and a fixed support. A displacement trace is
then applied at the end of this truss. The axial force in the truss element is the pushover force
required to produce the displacement. The truss element may not be exactly 1.0 long as the support
node is moved slightly if a userFIGURE 9-9 PUSHOVER DIALOG BOX
specified node or floor node occupies
the same coordinates.
Pushover Analysis Parameters

When
you
specify
pushover
parameters and check the box for
pushover load (see Analysis Data sheet)
then a single ANSR-L input file will be
created. This file is run in the same
manner as for any dynamic time
history analysis using an earthquake
acceleration record. The ANSROUT
post-processor can then process the
output files produced see Section 10.
When this option button is activated,
the dialog box shown in Figure 9-9 will
appear.

9.2.1

Column Line

Level Number

Number of Cycles

Master Node

Sine

Load Type

Linear

Load Increment

Load Direction

OK

0.005

Cycle 1 Amplitude

Number of Cycles per Amplitude

COLUMN LINE

The column line at which the


displacement is to be applied. A number from 1 to the maximum column line number. If the
displacement is to be applied at the center of mass of the rigid diaphragm (the usual case) then check
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

190

the Master Node box adjacent to this item. When the Master Node box is checked the program ignores
the column number specified.
9.2.2

LEVEL NUMBER

The level at which the displacement is to be applied, a number from 1 to the maximum number of
levels. If the Master Node box is checked then make sure the level you select has a rigid diaphragm
specified.

9.2.3

NUMBER OF CYCLES

The displacement is applied over 2000 time steps. These steps may incorporate anywhere from to
8 cycles of displacement. A cycle would be the initial positive loading cycle only, from zero to
maximum displacement. A cycle would be from zero to maximum displacement and back to zero.
As the number of cycles increases then stiffness and strength degradation effects will become more
apparent.

9.2.4

LOAD TYPE

The load type specifies the shape of the time function of applied displacement. The options
currently are sine, sawtooth and decay. The sine shape is recommended for most cases. The decay
shape is to generate damping decay curves see Section 9.4.

9.2.5

LOAD INCREMENT

The load increment can be constant (same amplitude for each cycle), linear (each cycle increasing
from the previous cycle by the same amount) and geometric (each cycle increasing from the previous
cycle by twice the previous increment).
Selecting Constant will produce hysteresis loops which overlay each other as none of the elements
currently implemented degrade purely due to the number of cycles. Options used are usually linear
or geometric.

9.2.6

LOAD DIRECTION

The pushover displacement can be applied parallel to either of the global horizontal axes, X or Y
(these are MODELA axes, corresponding ANSR-L axes are X and Z). Each analysis can include
only one direction.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

191

9.2.7

CYCLE 1 AMPLITUDE

The amplitude for cycle 1, together with the load increment above, will define the displacement
amplitude for every cycle. The value is a number, in meters.

9.2.8

NUMBER OF CYCLES PER AMPLITUDE

Generally only 1 cycle per amplitude will be used, that is, the displacement amplitude will be
incremented each cycle. Provision for more than one cycle at each amplitude is for future elements
which may incorporate cyclic degradation.

9.2.9

PRODUCING DISPLACEMENT PUSHOVER CURVE

Pushover displacement is produced from the ANSROUT spreadsheet (Section 10) which has a check
box for pushover displacement. When this option box is checked the pushover force and
displacements are written to a disk file with the suffix .PSH. This is a comma-delimited file which
can be imported into Excel. The columns provide time, displacement and force at each time step.
Columns B and C can be plotted to provide the overall building hysteresis. The force units are KN,
you need to divide these values by the weight in your model to obtain the normalized base shear.
The weight can be obtained by summing the masses in the ANSROUT workbook (Displacements or
Shears sheets).
Some examples of normalized pushover curves are shown in Figure 9-10 (Shear Wall Building),
Figure 9-11 (Frame with Degrading Beams, Non-Degrading Columns) and Figure 9-12 (Frame with
Degrading Beams and Columns).
FIGURE 9-10 DISPLACEMENT PUSHOVER CURVE FOR SHEAR WALL BUILDING

0.15

BASE SHEAR / WEIGHT

0.10

0.05

0.00
-500

-400

-300

-200

-100

100

200

300

400

500

-0.05

-0.10

-0.15
ROOF DISPLACEMENT(mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

192

FIGURE 9-11 DISPLACEMENT PUSHOVER CURVE FOR DEGRADING BEAM FRAME


0.15

BASE SHEAR / WEIGHT

0.10

0.05

-200.00

-150.00

-100.00

-50.00

0.00
0.00

50.00

100.00

150.00

200.00

-0.05

-0.10

-0.15
DISPLACEMENT (mm)

FIGURE 9-12 DISPLACEMENT PUSHOVER CURVE FOR DEGRADING BEAM AND COLUMN FRAME
0.15

BASE SHEAR / WEIGHT

0.10

0.05

-200.00

-150.00

-100.00

-50.00

0.00
0.00

50.00

100.00

150.00

200.00

-0.05

-0.10

-0.15
DISPLACEMENT (mm)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

193

9.3

SPECIAL CONSIDERATIONS FOR NSP

There are some issues with the NSP relating to the solution procedures, especially as they apply to
Unreinforced Masonry (URM) or other types of strength degrading system. This is because the
procedure requires the application of a monotonically increasing force vector.
For equilibrium, the sum of the internal resisting forces in the elements must equal the external
imposed loads. It follows that for a strength degrading system the internal resisting forces cannot
increase as the load, and so displacement, increases beyond peak strength. Equilibrium will not be
achieved and the analysis will terminate. The value set for the force tolerance will influence the
displacement at which the analysis terminates.
9.3.1

EXAMPLE APPLICATION

The Gateway is a 3 story URM building. The first story is dominated by a combination of rocking
piers and shear dominated elements. The hysteresis of this story was developed by applying a
cycle of displacement pushover, as shown in Figure 9-13.
Important features of the displacement pushover analysis are demonstrated in Figure 9-13:

The overall force deformation follows the FEMA backbone curve for URM.

Equilibrium is satisfied, as shown by the fact that the sum of the reactions is very close to the
pushover force.
FIGURE 9-13 DISPLACEMENT PUSHOVER FORCE
12000

10000

8000

6000
Pushover Force
Wall Forces (from VBA file

4000

2000

0
0.00

5.00

10.00

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

15.00

20.00

25.00

30.00

194

The NSP is based on a force vector monotonically increased. As Figure 9-13 shows, for
displacements exceeding approximately 5 mm, any increase in force will cause unbounded
displacements as the stiffness of the system is zero until, for displacements exceeding 12 mm, it
becomes negative.
The force pushover was applied to a force level of 1.0 times the seismic weight, W, in 1000 steps
with solution tolerances set to 10%, 1% and 0.1% of the weight, respectively. The final analysis, with
a tolerance of 0.01W, was repeated by reducing the force increment per step by a factor of 4. This
was achieved by specifying the maximum force as 0.25W. Figure 9-14 shows the pushover curves
generated for each variation.
FIGURE 9-14 FORCE PUSHOVER CURVES

12000
10000

FORCE

8000

Tolerance = 10%W
Tolerance = 1%W

6000

Tolerance = 0.1%W
Tolerance = 0.1%W DT = DT/4

4000
2000
0
0.00

5.00

10.00

15.00

20.00

25.00

30.00

DISPLACEMENT

The force pushover curve in Figure 9-14 is not expected to correlate exactly with the displacement
pushover in Figure 9-13 because the force vector applies a distributed load to the structure whereas
the displacement pushover applies a single concentrated force at 1st floor level. However, they
should show the same general characteristics.
If Figure 9-14, the loading curve up to about 5 mm displacement was largely independent of the
force tolerance. Beyond that, with the larger tolerances, 10%W and 1%W, the load was apparently
maintained for displacements well beyond the 12 mm level at which force degradation would be
expected. The pushover displacement was measured at the first story and so higher displacements
would be expected at the roof level, the control point for the force pushover procedure. However,
they would not be higher by this amount.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

195

Figures 9-15 and 9-16 plot equilibrium for the 1% and 0.1% tolerances, respectively. These figures
plot the external load (applied force) and the resisting force (reactions). In the former case, the 1%
tolerance represents a force of approximately 550 KN and so the equilibrium unbalance is permitted
to reach this value. This allows the load to be maintained to artificially large displacements. When
the tolerance is set at 55 KN, Figure 9-16, and a finer time step is used, the maximum displacements
of 17 mm are closer to what was expected from the displacement pushover.
FIGURE 9-15 FORCE IMBALANCE FOR TOLERANCE 1% W
12000

10000

FORCE

8000

6000

Applied Force Tolerance = 1%W


Reactions Tolerance = 1%W

4000

2000

0
0.00

5.00

10.00

15.00

20.00

25.00

30.00

DISPLACEMENT

FIGURE 9-16 FORCE IMBALANCE FOR TOLERANCE 0.1% W


12000

10000

FORCE

8000

6000
Applied Force Tolerance 0.1%W
Reactions Tolerance 0.1%W

4000

2000

0
0.00

5.00

10.00

15.00

20.00

25.00

30.00

DISPLACEMENT

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

196

9.3.2

RECOMMENDATIONS

For NSP evaluations, exercise caution in selecting load increments and tolerances. The example
represented here is probably the most severe type of model for the NSP as the structure has no
resistance other than URM walls. However, it is recommended that the tolerance be set at 0.001W,
or at least not greater than 0.01W, on the MODELA workbook for all projects using NSP. The
total load applied (Base V/W) should be set with as small a margin to expected ultimate strength as
possible, or the number of steps increased, to ensure as small a force increment as practical.
The NSP tolerance is much less than used for the NDP, where tolerances are generally 0.1W. The
reasons for this are:

The force tolerance for NDP includes dynamic effects from damping and stiffening elements as
well as equilibrium imbalance. The static procedure does not have these effects, all imbalance
represents lack of equilibrium.

For the NDP procedure the applied loads are cyclic. The unbalanced forces reach a peak value
and then reduce with subsequent cycles. The NSP is applied monotonically and the unbalanced
forces tend to continue to increase.

For any project using NSP, it is recommended that force equilibrium be checked by generating
reaction curves similar to those in Figures 9-15 and 9-16. Set the element output code to 99 and
process element time histories to obtain the sum of reactions to compare to applied load.

9.4

DAMPING DECAY

The displacement pushover option of MODELA has an added option for load type, termed Decay.
When this option is selected a pushover displacement will be applied for cycle, that is, the first
loading sequence, to the displacement amplitude selected. The load will then be released and the
structure allowed to vibrate freely.
The procedure is the analytical equivalent of the experimental snap-back method used to measure
damping in some types of structure.
This option provides a displacement trace of the form shown in Figure 9-17. The final portion of
the plot demonstrates the damping decay. If you import this into the spreadsheet, you can use Excel
functions to determine a series of positive and negative peaks. This will provide an indication of the
period as a function of displacement for non-linear structures.
The effective damping ratio, as a fraction of critical damping can be determined from the
logarithmic decay using the peaks m cycles apart. The displacements are extracted for cycle n and
cycle n+m and the damping calculated using the following formula:

ln

n
= 2 m
n+m

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

197

Figure 9-17 has an exponential curve plotted as well as the calculated displacement trace. This curve
has an equation of

= 0 e t
For non-linear structures the period (and so frequency, ) and damping, will be a function of
displacement. It will not be possible to develop a single exponential curve as for the elastic structure
in Figure 9-15. In this case, the logarithmic decay formula can be used to compute damping between
subsequent cycles (m=1) and the average over a number of cycles (m>1).
It would be possible to develop a spreadsheet template to automate period and damping calculations
once the displacement trace is imported. This could include fitting a best fit exponential curve.
This development is not currently planned and could be done in the future depending on our
research needs.

FIGURE 9-17 PUSHOVER DISPLACEMENT WITH DECAY

250

Measured Displacement Trace


Fitted Damping Curve
Fitted Damping Curve

200
150

DISPLACEMENT (mm)

100
50
0
0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

-50
-100
-150
-200
-250
TIME (Seconds)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

198

10

RESULTS PROCESSING : ANSROUT

Results from the ANSR-L analyses are processed via an Excel workbook, ANSROUT.XLS. The
most up to date version of the spreadsheet should be obtained from the HCG server at the start of a
project and saved into the job directory.
Your local computer should be configured as discussed in Section 3. When the ANSROUT
workbook is first opened, check the program location on the Interface worksheet. Cell B27 should
contain the location (full path) of the PROCESSA.EXE file.
The ANSROUT spreadsheet has a macro which shells the DOS program PROCESSA.EXE. This
program reads the trace file (.TRC), which was written by MODELA, to obtain a description of the
model geometry and element types. It then reads results from the ANSR-L output files and writes
them to comma delimited files. These files are then imported to Excel, using macros, for further
processing.
The output processing provides a number of options for selective processing (for example, you do
not need to process nodal time histories and you do not need to obtain results for all element types).
You can later repeat the post-processing with different options, for example, if there are some large
displacements you may wish to process nodal time histories first to check the response. To permit
re-processing, you must have retained the ANSR-L output files .ECH, .D01, .ONN and any element
time histories (.00n).
The process is controlled by the Interface worksheet in the ANSROUT workbook.
When element time histories are requested, a large number of files may be opened simultaneously. If
you have a problem with this, see Section 3.4 Computer Setup for information on increasing the
number of file handles available.
10.1 INTERFACE

The Interface sheet is as shown in Figure 10-1. This sheet is the control, used to set processing
parameters and to activate the macros used to read the ANSR-L files and import them to the
spreadsheet. If is also used to activate special purpose functions such as obtaining gap forces and
envelope times.
There is also a portion of this sheet, not shown on Figure 10-1, which is used to specify output sets
for an NSP analysis see Section 9.1.7.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

199

FIGURE 10-1 INTERFACE SHEET

Processing of ANSR-II Analysis Results


Read Results

Clear All Results

Select

Get Gap Forces

Import Results

Get Envelope Times


Element Output
Envelopes Only

Maxima Only

Individual Results

Time Histories

Envelopes & Individua

Maxima & Time Historie

Truss Elements

Displacement Time Histories

3D Beam Columns

Acceleration Time Histories

2D Beam Columns

Drift Time Histories

Plane Stress Elements

Element Time Histories

Gap Elements

Pushover Force

Write .CRA File for AutoCAD


Write Displacements

Don't Colour Code Force Results

Output Files to Process


C:\jobs\Gateway\NSP\31A0
C:\jobs\Gateway\NSP\32A0

11
12
13

Write Plate DXF File

Trace Displacements >


Minimum Crack Number

File

Autofill

9
10

Pushover Displacement

Write Base Shears

Set Output File


Names
1
2
3
4
5
6
7
8

Nodal Output

Name of .TRC File


C:\jobs\Gateway\NSP\3

0.300
1

Suppress Element Group Time Histories

14
15
16
17
18

C:\HCGPROGS\PROCESSA

Location of program

19
20

10.1.1 PROGRAM LOCATION.

The path to the program PROCESSA.EXE must be entered in cell B27 of this sheet. This must
include the program name but without the suffix, for example,
C:\HCGPROGS \PROCESSA
10.1.2 SETTING FILE NAMES

The first parameter set is the name of the .TRC file, which was written by MODELA at the time the
ANSR-L input files were written. You can type the path and filename directly into the cell but it is
usually more convenient to use the Select button. This will display a File Open dialog box. Select the
directory and .TRC file for the results which you want to process.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

200

The names of the output files to be included for processing are entered on the Interface sheet. There
are three ways of entering these files:
1. Type the names directly into the appropriate cells. The name must include the full path.
2. Select the output files individually by using the Set Output File Names button. With this option,
the File Open dialog box will be displayed. This will list all ANSR-L output echo files in the
directory (.ECH). Open each file you want included in the processing; enter Cancel when done.
3. The Autofill button will automatically include each .ECH file in the directory where the .TRC file
resides.
The usual procedure is to run each variation (that is, each series of ANSR-L analyses) in its own subdirectory. With this, the third option, using Autofill, is used to select all files. Note that if you want
to exclude some files you can modify the list produced by Autofill.

10.1.3 CLEAR ALL RESULTS

This button will clear all the sheets containing results from previous analyses. The macro clears the
entire sheet so be careful with this option. If you have added formulas to cells you may want to
selectively clear sheets manually rather than use this procedure.

10.1.4 GET GAP FORCES

As discussed in Section 7.5, a separate analysis may be performed to obtain the pre-load on gap
elements for some types of model. This button will read the ANSR-L output file and write the gap
axial forces to a disk file. The file is read in by the MODELA spreadsheet. When this option is
being used, a single output file is selected.

10.1.5 GET ENVELOPE TIMES

This button is used to find out whether all analyses ended normally by examining the envelope times
in the output files. When you activate this button, each output file will be read and the envelope
time written to the Summary worksheet. Times less than the earthquake duration (generally 20
seconds) indicate that an abnormal termination occurs.

10.1.6 SELECTING OUTPUT

A range of check boxes determine the output to process for elements and nodes. They also set
special features which are available.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

201

10.1.6.1 Element Output

For element output, select envelopes, individual results or both. For large models and/or a large
series of ANSR-L analyses, the Individual Results options will produce extremely large spreadsheets.
Use this option with care.
Selective element types can be used to minimize the amount of data imported. Usually, all element
types will be checked so that, if present in the model, the results will be imported. If an element type
is not checked then the PROCESSA program will not write results into the summary sheet.
Therefore, if you change your mind you cannot simply check the box and Import Results again, you
also have to Read Results again.

10.1.6.2 Nodal Output

Nodal output can be processed to obtain maximums only, time histories only or both. Generally,
nodal output is processed to provide time histories of response. The time histories are written to
disk files. It is up to the user to perform further processing.
Options for nodal output are to obtain one or more of displacement, acceleration, drift, element or
pushover results. The last are for either force or displacement pushover, depending on which option
was analyzed. See Section 9 for more details.
Nodal output processing may be time consuming if you have a large number of files and/or a large
number of time history output node locations. Often it is preferable to process the results without
time history output to check that element results appear reasonable. The processing can then be
repeated getting time history output.
10.1.7 SPECIAL FEATURES

1. Write Base Shear. Check to read in story shears. Normally leave checked.
2. Write .CRA file for AutoCad. Check to have the processor program write an AutoCad input file
to display cracking in wall elements. Only meaningful for models containing wall elements. You
need the full version of AutoCad to import these files as there is a Lisp program required to
display the cracking.
3. Write Displacements. Check to read in story displacements. Normally leave checked.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

202

4. Write Plate DXF File. Write an AutoCad file with the wall elements assigned to layers depending
on the level of maximum shear strain. Check this box for models with wall elements. This DXF
file can be imported to AutoCad LT.
5. Dont Color Code Force Results. The cells on the sheets containing element forces are color
coded, according to the file which provided the governing force or deformation component. If
this box is checked then this step is omitted. This speeds up the Import Results step.
6. Trace Displacements. This cell contains a displacement value, in meters. During processing, a
file named T$0000 is created. If the displacement at any node exceeds the value in this cell then
the node number and displacement is written to this file. This is a way to trace local instability
in the results, manifested by excessive displacements at a limited number of nodes.
7. Minimum Crack Number. This relates to the AutoCad cracking display file. For concrete and
masonry, the elastic stiffness is crack 0, the next stiffness crack 1 etc. Usually, especially for
concrete, cracking at level 1 indicates only that the initial cracking strength is exceeded and such
cracking would not be visible. This cell can be used to enter the minimum crack number to
display. Normally, this will be set to 2 for both concrete and masonry.
8. Suppress Element Group Time Histories. This is only relevant if you are obtaining element time
histories. The default (this box unchecked) is to write the total shear at each story and also the
shear contributed by each element group. If this option is checked then only the total shears will
be written (see Section 10.3.5).

10.1.8 READ RESULTS

This button activates the DOS program PROCESSA. When pressed a macro will write the current
selections to a disk file named jjjj.PRA and then shell the executable file PROCESSA. Provided the
program runs normally, you will see the DOS progress screen displayed as the results are read from
the ANSR-L output files.
This task reads the output results and re-writes then to comma-delimited files but does not import
them to the spreadsheet until the Import Results macro is activated.

10.1.9 IMPORT RESULTS

This button activates the macro which imports the summary of results into the individual sheets.
The macro operation opens the comma de-limited files one at a time (see Section 8.11), selects the
used range and then pastes the values into the appropriate sheet in the workbook.
If the results are to be Color Coded then the file names on the Interface worksheet will be colored to
provide a key as to colors used on the individual results spreadsheets. The cells containing all
element force results will have the font color set to match the color key of the file which produced
the governing result for that particular force component.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

203

The information used to assign colors is contained on the imported sheet as extra columns. After all
results columns (plus other information, such as element dimensions and coordinates) there will be a
series of columns containing integers from 0 to the number of files processed, one column for each
force component. Once the color coding has been performed, these columns can be cleared and
used for additional processing formulas.

10.2 COORDINATES

The coordinates sheet is a data sheet to aid in interpreting results, displaying results and tracing errors
in the ANSR-L input file. For analyses which terminate normally, this sheet will not usually need to
be referred to. The ModelPlot sheet (see below) reads coordinates from this sheet. The sheet also
lists the degrees of freedom at each node to aid in finding errors in models which do not terminate
normally.
The first section of this sheet echoes the column line coordinates and the elevations as entered in the
MODELA spreadsheet.
The second section of the sheet, after the column lines, lists the nodal points in the ANSR-L model.
For each node, the reference column line and elevation as defined in MODELA are listed. This is
followed by the degree of freedom numbers assigned to the node in the 3 translational and 3
rotational directions. A zero value indicates that the particular degree of freedom is fixed.
This last section of the Coordinates sheet provides the information needed to track zero divide
problems more efficiently than searching the ECH file (see Section 8.9)

10.3 TIME HISTORIES OF RESULTS

Time histories may be specified for elements and/or nodes. These are usually processed to produce
maximum values and time histories. In some cases they will be processed to produce maximum
values only. Situations when time histories will be saved to obtain maximum values only are:

To obtain maximum floor accelerations. These are not obtained from the D01 output file
envelopes and can only be obtained from time histories.

To obtain maximum story drifts. These can be obtained approximately from the D01
displacement envelopes. However, the only way to obtain accurate drifts is to calculate the
difference in displacement between successive levels at every time step.

The PROCESSA program processes the time history results and writes them into a series of disk files
as comma delimited ASCII values (if time history output is requested). It also writes maximum
values (if requested) into a disk file, which is imported to the Envelope sheet in the results workbook
(see below).
The time history files are in a format suitable for not import into a spreadsheet but the processor
program does not actually import them. They are simply processed into files in the job directory and
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

204

it is up to the user to perform any additional processing. The time history results are processed for
individual analyses and a separate output file will be created for each analysis and each output type.

10.3.1 RAW DATA

The time history output is written to ASCII files, which can be opened with a text editor (WordPad),
imported into Excel (as fixed width files) or processed further as described in the following sections.

10.3.1.1 Nodal Output

Nodal output is contained in a series of files with a .ONN suffix. This file contains one line per
nodal output component per time step, in fixed width fields. There are 7 columns containing the
following data:
1.

Time, in seconds.

2.

Nodal displacement, in meters.

3.

Nodal velocity, in meters/second.

4.

Nodal acceleration, in meters/second/second.

5.

Nodal inertia force, in KN.

6.

ANSR-L node number.

7.

Direction (in ANSR-L coordinate system).

You can import this file to Excel and use filtering to obtain time histories instead of using
PROCESSA. In fact, you will have to do this if you want to obtain time histories of quantities not
currently processed (velocities and inertia forces).

10.3.1.2 Element Output

Element time history output, if requested, will be in disk files with the suffix .001, .002, .003 etc
where the numeral refers to the group number. The files will have a title line (or lines) plus one line
per element output group per time step. (Each output group is defined by a negative output code
value for the particular element type, as set on the element sheets in the MODELA workbook).
The first column in each will always be time and the second the group number. The remaining
columns will be output components depending on the element type and will be identified by column
headings in the first lines in the file.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

205

If the automatic code 99 was used to specify element output then the group number will be the
level number to which the lowest node of the element is connected. The group numbers will not be
sequentially if output has been suppressed at some levels.

10.3.2 DISPLACEMENTS

The processed displacements are stored in a series of files with the .DIS suffix. The time histories
will be for the diaphragm nodes and/or user selected nodes as defined in the MODELA workbook.
Data are stored in columns, with the first column containing time and subsequent columns the
displacement values. Columns have an identifier in Row 1.
The displacements are factored by 1000 by the PROCESSA program to convert meter units to
millimeters.

10.3.3 ACCELERATIONS

Processed accelerations are stored in a series of files with the .ACC suffix. The time histories will be
for the same nodes for which displacements were requested. As for displacements, data are stored
in columns, with the first column containing time and subsequent columns the acceleration values.
Columns have an identifier in Row 1.
The accelerations are factored by 1/9.81 to convert meter/second2 units to multiples of gravity (gs).

10.3.4 DRIFTS

Processed accelerations are stored in a series of files with the .DRF suffix. The drifts depend on the
form of displacement output requested:

For diaphragm nodes, the displacements between center of mass nodes at adjacent floors are
calculated and divided by the difference in elevation between the two floors. Drifts are
calculated for each diaphragm separately.

For user specified nodes, the displacement at a particular column line is subtracted from the
displacement at the next higher level for the same column line at which displacements have been
requested.

Examples of these calculations are:


1.

Diaphragm output requested. A building with a single diaphragm at each level will have a drift
calculated for every story. If it also has, say, diaphragm 2 at Levels 2 and 4 then a drift will
also be calculated for diaphragm 2 as the difference in displacements between Levels 2 and 4
divided by the elevation change between these two levels.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

206

2.

Use selected output. If, say, Column 5 has output at Levels 1, 2 and 3 then drifts will be
calculated using column 5 displacements from L1-L2 and from L2-L3. If output is requested
at only levels 1 and 3 then only one drift will be calculated, from L1-L3.

An additional drift will be calculated at each diaphragm and user specified column line by calculating
the drift between the lowest level for which output is specified and the elevation specified for Level
1. This may not be correct for stepped foundations or some other types of configuration.

10.3.5 ELEMENTS

The PROCESSA program will read any element time histories which exist (files with suffix .001, .002
etc) and write a summary of the shear forces into comma delimited disk files.
The element processing is aimed at obtaining time histories of story shears and works best when
element output has been defined to produce this (see Section 7.10). The steps followed by the
program are:
1.

Open an output file, identified by a .VBA suffix.

2.

Loop through each element group in turn and check for the existence of a file with the suffix
.00n, where n is the element group.

3.

If the particular file exists, read in shear force output for each output group number, assuming
group number refers to story number.

4.

Sum shear forces for all groups which have the same story number.

5.

Output the total shear forces for each story and also the shear from each group which
contributed to the sum. (The individual group shears can be suppressed, see Section 10.1.7)

Generally, the only groups containing time history output will be columns, braces and walls. The
program will sum the total wall shear plus total column shear about each of the global ANSR-L
horizontal axes and output these.
The .VBA file can be opened in Excel (comma delimited) and the time history of story shear plotted,
with separate plots for the columns and walls if required. If the .DIS file is also imported then story
hysteresis plots can be constructed.
The number of columns in the .VBA file may exceed the maximum allowed by Excel (256 columns).
The required number of columns will be the number of levels for which output is obtained times two
directions for the totals. In addition, there will be a column for each individual group at each level in
each direction. As discussed earlier, each material used for the walls and columns is assigned to a
separate element group. If you have a large number of materials, and/or a large number of levels,
you will exceed the column limit of 256. In this case, only the first 256 columns will be available
from Excel. The number of columns can be reduced by:

Getting story shears at a limited number of levels (See Section 7.7.2 for the procedure to
suppress particular levels).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

207

Processing only total shear forces, not the individual element groups (See check box on Interface
sheet).

10.3.6 FORCE PUSHOVER AND DISPLACEMENT PUSHOVER

These options are only relevant when you have performed a pushover analysis. When either of these
option boxes is checked, the pushover forces and displacements will be written to disk files with the
suffix .PSH for a displacement pushover analysis or the suffix .DIS for a force pushover analysis.
These are comma-delimited files, which can be imported into Excel. The columns provide time,
displacement and force at each time step. The output files can be plotted to provide the overall
building hysteresis.
As for the other time history traces, the files are created but no other action is taken. See Section 9
for further information.

10.4 SUMMARY SHEET

This sheet is earmarked for further development. It was originally developed as a summary sheet to
aid in the evaluation of unreinforced masonry buildings and its usefulness is largely confined to this
type of building.
Data blocks contained on this sheet are:

For each ANSR-L analysis file, the filename, the scaling factors and the earthquake component
to which the scaling factor is applied. Scaling factors include a 9.81 factor to convert units of
metres/second2 to units of gravity. Divide by 9.81 to obtain the factor applied to the amplitude.

For each file, the base shear coefficient in the X and Z directions. The base shear on this sheet is
calculated as the sum of the envelope shear forces in panel elements only. If the structure has
other elements at base level (e.g. columns) the shear force in these elements is not included in the
total. Currently, the only accurate means of obtaining the base shear is to save and process the
element time histories. If the model contains only walls then this will be an upper bound shear
force as it is calculated as though all maximum values occurred simultaneously. This may not
necessarily be true.

The maximum displacements in each of the 3 translational directions. These are the maximum
values anywhere in the structure.

For each analysis, the number of unstrengthened wall panels which exceed the FEMA shear
strain limits for URM in-plane walls for each performance objective. These are currently set by
the program as IO (Immediate Occupancy) < 0.001, LS (Life Safety) < 0.003 and CP (Collapse
Prevention) < .004. The number of wall panels where the strain exceeds the strength
degradation limit (0.0015) is also listed. These panels are those which require strengthening.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

208

The program examines each wall panel to determine whether it is strengthened, where
strengthening is defined to be true if the panel has an overlay of a wall type which contains the
sequence of characters TYFO (not case sensitive). A URM panel with an FRP overlay generally
has a strain limit of 0.004. The number of strengthened panels with strains which exceed this
limit is listed.

The number of FRP panels (as above, containing TYFO in the name) which exceed the strain
limit of 0.004 are listed.

For URM buildings, the usual procedure is to run a series of analyses using the incremental
earthquake option. This sheet will then provide information as to which level of earthquake causes
damage such as to require strengthening and the level of earthquake for which the FRP panel strains
are within the allowable limit.

10.5 AREAS

This sheet lists the total areas of the panel elements. Three areas are listed at each elevation in the
model:
1.

The shear area of each wall elevation in each of the two horizontal directions.

2.

The face area of each wall.

3.

The face area of floor panels.

This is provided for general information, usually to develop quantities, and is not used further by the
program.

10.6 DISPLACEMENTS

Displacements are listed for each analysis plus the envelope values from all analyses. At each
elevation defined in the model, the following information is listed:
1.

Level name.

2.

Elevation of this level.

3.

The total mass applied to nodes located on this level.

4.

The maximum X, Y and Z displacements and the ANSR-L node at which they occurred.
Displacements are in units of mm.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

209

10.7 ENVELOPES

The Envelopes sheet will contain data only if time history nodal output has been saved and processed
with the option to provide maximum values. This sheet contains maximum displacements,
accelerations and drifts for each file and also the envelope of each quantity.
The values are listed for each level. The elevation and maximum positive and negative values are
listed so that the Excel charting tools can be used to display results. See Section 10.3 for more
information on processing of time history results.

10.8 SHEARS

The Shears worksheet is similar to the Summary sheet in that it was developed for models containing
solely panel elements. The shears listed on this sheet for a particular level are actually the sum of the
shear forces in all panel elements at this level. Shear forces from other element types are not
included in the total.

10.9 ELEMENTS

Results from each element type will be imported into the sheet for that particular type. The output
imported depends on the element type but there are some items which are common to all types:

The element number, which is the ANSR-L model number.

The element location in the structure, defined by the column line (C) and the level (L). For most
element types the column and levels at ends i and j will be listed.

Element forces and deformations, a function of element type.

Following the element force components, the X, Y and Z coordinates of End I of the element
will be listed. These are in the ANSR-L coordinate system. These are provided as an aid to
filtering.

Information related to the element properties, such as shear area for panels and EI for flexural
elements. Data will be identified by the column heading.

After the final data item will be a number of columns without headings containing integer values
from 0 to the number of files processed. These are used to assign colors to the results cells.
Once the file has been imported the data in these columns can be deleted.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

210

10.9.1 TRUSSES

Truss element components are the axial force and axial deformations, as shown in Table 10-1.

TABLE 10-1 TRUSS COMPONENTS

Component
Tension
Compression
Max +ve Ext
Max ve Ext
+ve Acc
-ve Acc

Description
Maximum tension force in truss
Maximum compression force in truss
Maximum positive (tensile) extension.
Maximum negative (compressive) extension.
Maximum accumulated positive extension.
Maximum accumulated negative extension.
(See ANSR Manual for description).

The gap results may also be on this sheet. This is because the gaps are sometimes modeled as truss
elements which buckle under tension forces. This type of element is used to enable stiffness
damping to be applied, which aids numerical stability. For gap elements, the positive extension is
the gap opening and the negative extension the gap compressive deformation.
10.9.2 BEAMS

Beam output components are listed in Table 10-2. The flexural results are for positive bending
(bottom steel in tension) and negative bending (top steel in tension). Shear force values are the
maximum absolute value.
Axial loads are provided for beams although these are meaningless if the beams have a diaphragm.
Values are maximum and minimum at each end although usually the loads are the same at each end
as there are no axial loads applied within the beam span.
The beam dimensional and stiffness properties are provided as an aid to the evaluation of beam
results (see Section 11).

TABLE 10-2 BEAM COMPONENTS

Component
Mpi
Mni
Vi
Rpi
Rnj
Mpj
Mnj
Vj

Description
Maximum positive moment at End I
Maximum negative moment at End I
Maximum shear force at End I
Maximum positive plastic rotation at End I
Maximum negative plastic rotation at End I
Maximum positive moment at End J
Maximum negative moment at End J
Maximum shear force at End J

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

211

Component
Rpj
Rnj
Pimax
Pimin
Pjmax
Pjmin
D
B
E
Imaj
Imin

Description
Maximum positive plastic rotation at End J
Maximum negative plastic rotation at End J
Maximum axial load at End I (Compression ve)
Minimum axial load at End I (Compression ve)
Maximum axial load at End J (Compression ve)
Minimum axial load at End J (Compression ve)
Beam Depth
Beam Width
Youngs Modulus
Major axis moment of inertia
Minor axis moment of inertia

10.9.3 COLUMNS

Column output components are listed in Table 10-3. The flexural and shear results are for the local
Y axis (bending along the global X axis) and Z axis (bending along the global Y axis). All values are
the maximum absolute value.
Column yield occurs when the vector of the bi-axial moments and axial loads exceed the limit
imposed by the perimeter of the interaction diagram. The total plastic rotation which occurs after
yield has a component in the two moment directions and also in the axial direction. These are the
three plastic rotation values listed at each end of the column. Generally, the evaluation is based on
the flexural plastic rotations, Ry and Rz.
The column dimensional and stiffness properties are provided as an aid to the evaluation of column
results (see Section 11).

TABLE 10-3 COLUMN COMPONENTS

Component
Myi
Mzi
Vyi
Vzi
Ryi
Rzj
Rpi
Myi
Mzi
Vyi
Vzi
Ryi
Rzj
Rpi

Description
Maximum Y moment at End I (= MODELA X axis)
Maximum Z moment at End I (= MODELA Y axis)
Maximum Y shear force at End I
Maximum Z shear force at End I
Maximum Y plastic rotation at End I
Maximum Z plastic rotation at End I
Maximum axial plastic rotation at End I
Maximum Y moment at End J
Maximum Z moment at End J
Maximum Y shear force at End J
Maximum Z shear force at End J
Maximum Y plastic rotation at End J
Maximum Z plastic rotation at End J
Maximum axial plastic rotation at End J

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

212

Component
Pimax
Pimin
Pjmax
Pjmin
D
B
E
Iy
Iz

Description
Maximum axial load at End I (Compression ve)
Minimum axial load at End I (Compression ve)
Maximum axial load at End J (Compression ve)
Minimum axial load at End J (Compression ve)
Column Y dimension
Column Z dimension
Youngs Modulus
Y axis moment of inertia
Z axis moment of inertia

10.9.4 WALLS

The panel elements have three local axes representing stresses in the horizontal and vertical planes
plus the shear stress. The wall output components, shown in Table 10-4, list the maximum and
minimum values of each of the stress components and also the strain components associated with
each stress. The panel shear area is the area of the horizontal plane projection.

TABLE 10-4 WALL COMPONENTS

Component
Strains
F12+
F12S12+
S12F11+
F11S11+
S11F22+
F22S22+
S22AREA

Description
Maximum Absolute Shear Strain
Maximum Positive Shear Stress
Maximum Negative Shear Stress
Maximum Positive Shear Strain
Maximum Negative Shear Strain
Maximum Positive Vertical Stress
Maximum Negative Vertical Stress
Maximum Positive Vertical Strain
Maximum Negative Vertical Strain
Maximum Positive Horizontal Stress
Maximum Negative Horizontal Stress
Maximum Positive Horizontal Strain
Maximum Negative Horizontal Strain
Panel Shear Area

10.9.5 FLOORS

Flexible floors are modeled using the same elements as walls and so the floor output components,
Table 10-5, are the same as for walls.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

213

TABLE 10-5 FLOOR COMPONENTS

Component
Shear Strain
F12+
F12S12+
S12F11+
F11S11+
S11F22+
F22S22+
S22AREA

Description
Maximum Absolute Shear Strain
Maximum Positive Shear Stress
Maximum Negative Shear Stress
Maximum Positive Shear Strain
Maximum Negative Shear Strain
Maximum Positive Vertical Stress
Maximum Negative Vertical Stress
Maximum Positive Vertical Strain
Maximum Negative Vertical Strain
Maximum Positive Horizontal Stress
Maximum Negative Horizontal Stress
Maximum Positive Horizontal Strain
Maximum Negative Horizontal Strain
Panel Shear Area

10.9.6 GAPS

Table 10-6 lists the components output for the gap element. The ANSR-L element is a combined
gap-friction element although in most applications using this procedure the degrees-of-freedom
which would permit sliding normal to the gap are fixed. The output components provide forces and
deformations for gap opening and closing and also for the sliding component.
NOTE: Some gaps are modeled as truss elements rather than gap elements and the
output will be included on the Truss sheet. The same output is supplied.
When gap elements are used to model reinforced concrete wall cracking or anchor bolt uplift, a pair
of elements is used, the gap to model compression and a truss element to model the reinforcing or
the bolts. The Truss sheet will contain results for these elements. The extensions in the truss should
correspond to the deformations in the gap at the same location.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

214

TABLE 10-6 GAP COMPONENTS

Component
Uplift
Tension
Displacement
Compression
Friction D
Friction F

Description
Maximum gap opening displacement
Tension in gap (should be zero or ve (Compressive))
Maximum gasp closing displacement
Compression force in gap
Displacement normal to gap
Force normal to gap

10.10 MODELPLOT

The ModelPlot sheet is used to display beam, column or wall results. Plots are displayed by bents,
where a bent is defined as an assemblage of frame elements in the same vertical plane. The Update
button on this sheet activates the dialog box shown in Figure 10-2. The data entered on this box are:

Paper size and orientation (A3 or A4 paper in either landscape or portrait orientation).

Font size, for numerical values.

Horizontal and vertical offset of start of result value relative to the node. These are in multiples
of the font size.

Highest and lowest levels to include in the plot. These refer to level numbers as defined in the
MODELA workbook.

Column numbers defining the start and end of the bent. These refer to column numbers defined
in the MODELA workbook.

The beam and/or column results to display. These refer to numerical values on the plot.
Available components are displayed in drop down boxes.

Options to show member identification and whether to indicate beam and column nodes with a
filled circle on the plot.

A check box for repeat plot results. If this box is checked then a series of start and end column
lines will be read from the Constants sheet. Each plot will be automatically printed as it is
generated. This is equivalent to entering a series of Column A and Column B values on the
dialog box.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

215

The model plot facility is continually under development and may have additional features to those
shown on Figure 10-2.

FIGURE 10-2 MODEL PLOT DIALOG


Model Display Options
Paper Size

11

Font Size

22

Highest Level to Include

none

Beam results to display

Lowest Level to Include

Wall Strain

Column/Wall results to display

56

Column A on beam line

63

Landscape

Orientation

OK

A3

Column B on beam line

1.5

Horizontal Offset

0.5

Vertical Offset

Show member IDs

Cancel

Repeat plot results

Show beam nodes


Show column nodes

To use the model plot facility, copy the formulas on the BeamRes and ColRes sheets to ensure that
they cover the full range of results for your project. Then, from the dialog box shown in Figure 10-2,
choose the components for beams and columns to display. Figure 10-3 shows an example of the
plot display used to show the ratio of maximum shear to shear capacity ratio for both beams and
columns.
The last option of the column results drop down menu is the option to display Wall Strain. This
option displays wall panels between Column A and Column B and shades the panels depending on
the maximum shear strain. The shades to use for each strain increment are read from the Constants
sheet. Figure 10-4 is an example of this display.
The ModelPlot option is still under development and the features in the most recent version of
ANSROUT may not correspond exactly to what is described in this version of the manual.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

216

FIGURE 10-3 MODEL PLOT FOR FRAME

Beam Results: Shear Vu / Vcap


Column Results: Shear Vu / Vcap

5
1.53

1.66
1.04

4
2.38
3

200

207

214

221

FIGURE 10-4 MODEL PLOT FOR WALL


19
18

17
16
15

14
13
12

11
10
9

8
7
6

> 0.0100
> 0.0075

5
4

> 0.0000

20
133134

135 136

137

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

138

139 140141

142 143

217

10.11 CONSTANTS SHEET

The Constants sheet contains a mix of user and program data, mainly the latter. User information
which may be entered on this sheet is in the portion shown on Figure 10-5 only. Do not change any
other data on this sheet. User specified data is:

Plot Column Data. Pairs of column numbers may be entered to define the bents to plot for the
ModelPlots worksheet.

DXF Colors. A series of shear strains and associated color codes, and plastic rotations and
associated color codes, may be listed. The element will be assigned the particular color
depending on strain or plastic rotations.

Wall Plot Colors. A series of shear strains and associated color codes for Excel plots on sheet
ModelPlot.

Other user data may be added to this sheet. Do not alter data at any other locations on this sheet
unless it is clearly indicated that this is permitted.
FIGURE 10-5 DATA ENTERED ON CONSTANTS SHEET
DATA FOR REPEAT
PLOTTING OF
RESULTS

CHANGE DATA IN
YELLOW SHADED
CELLS ONLY

3
4
77

79
92
80

Strain Levels for DXF File


(Maximum of 6 Strains)
Strain
0.0013
0.0025
0.0075
0.0100
0.0150
0.0200

Colour
254
254
252
252
252
255

Pl. Rotation
0.00000
0.00500
0.00600
0.00500
0.00600
0.00750

Colour
254
253
252
251
250
255

Strain Levels for Workbook Plot


(Up to 10 Strains)
Strain
0.0000
0.0050
0.0075

Colour
27
17
25

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

218

10.12 AUTOCAD DXF FILES

If the check box to Write Plate DXF files is checked a series of AutoCad DXF files will be written.
There will be one file for each ANSR-L analysis which is processed. There will also be a drawing
containing envelope values over all files. The envelope DXF file will have the name of the TRC file
plus E.DXF.
The format of these files is currently evolving. They currently contain the following information:

Plate elements (walls and floors), assigned to layers, which are color coded according to
maximum absolute shear strain. The color for each strain level is as defined on the Constants
sheet.

Flexural elements (columns and beams) assigned to layers, which are color coded according to
maximum plastic rotation. . The color for each plastic rotation level is as defined on the
Constants sheet. The procedure for the element display is:
Columns are plotted as two equal portions, the lower half colored for plastic rotation at the
bottom end, the upper half colored for plastic rotation at the top end.
Beams are plotted as four equal portions, (a) End I ve plastic rotation (b) End I +ve plastic
rotation (c) End J +ve plastic rotation and (d) End J ve plastic rotation.
Element segments are plotted as a line element (centerline) if yielding does not occur,
otherwise as a three dimensional element.

The files are imported into AutoCad or LT in a similar fashion to the files produced by MODELA
(see Section 7.16).
Generally, the colors selected will be shades of gray to black so that the coded drawing can be printed
on a black and white printer. There are differences between the two releases of AutoCadLT used in
HCG offices. See Figures 10-6(a) and (b) for the color codes for gray scales for the two versions.
FIGURE 10-6 (A) AUTOCADLT RELEASE 2.0

(B) AUTOCADLT RELEASE 3.0

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

219

FIGURE 10-7 STRAIN / PLASTIC ROTATION CODED DXF FILE

10.13 CRACK PATTERN FILES

If the check box to Write CRA File for AutoCad is checked a series of ASCII disk files with a .CRA
suffix will be written. There will be one file for each ANSR-L analysis which is processed. There is
currently no provision to produce a plot for envelope values over all files.
The CRA files are imported by way of a Lisp program, crackup.lsp, which is loaded as an application
to AutoCad. The Lisp program requires two drawing files, crack.dwg and pl_hinge.dwg. If you want to
use the crack pattern display, you will need these three files loaded on the AutoCad directory. Note
that the applications are not available in the LT version of AutoCad, only the full version.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

220

FIGURE 10-8 CRACK PATTERN DXF FILE

10.14 OUTPUT FILES

A number of files are used as part of the processing of the ANSR-L results. These are listed in Table
10-7. The files containing data imported to ANSROUT may be deleted once results are imported, as
may the .PRA and .EN1 files.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

221

The other files may be used for further processing, such as plotting nodal response quantities (ACC,
DIS, DRF) or shear force time histories (VBA). These can be deleted when processing is complete.

TABLE 10-7 ANSROUT FILES

File
Type
Comment
Files used internally by MODELA and PROCESSA
ASCII
Written by MODELA.EXE
.TRC
.PRA
Written by ANSROUT.XLS macro
Time History Output Files
.DIS
Comma-delimited
Displacement time histories
.ACC
Acceleration time histories
.DRF
Drift time histories
.EN1
Envelopes for DIS,ACC and DRF
Files imported to ANSROUT.XLS
.DSP
Comma-delimited
Files read in by Import Results macro.
.SUM
These can be deleted once results
.SHE
have been imported.
.E0n
.ARE
Special Purpose Files
.VBA
Comma-delimited
Base shear time history
.PSH
Pushover force-displacement
Files used by other Applications
ASCII
Gap initial force input for MODELA.
.GAP
.CRA
Crack patterns input for AutoCad
Temporary Files with error-tracing information
T$0000
ASCII
Displacements exceeding limit
AutoCad results display files.
.DXF
ASCII
Series of DXF files (one per analysis)
with codes stresses / strains
DXF file with envelope strains from all
E.DXF
analyses.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

222

11

ACCEPTANCE CRITERIA

The main source documents for acceptance criteria are the NZNSEE and FEMA-273 guidelines.
For projects outside NZ we use the FEMA-273 requirements. For NZ projects, acceptance criteria
are generally based on the NZNSEE Green Book although these are usually supplemented with the
FEMA-273 criteria. This is because the U.S. criteria are much more fully developed than the NZ
equivalents.
The criteria in the section address the Life Safety and Collapse Prevention performance objectives as
these are generally most critical. If you have other performance objectives on your project, consult
FEMA-273.
The requirements in this section are extracted from FEMA-273 and NZNSEE but are representative
only and are not intended to reproduce all requirements in these guidelines.
You have a
responsibility to consult these source documents to obtain criteria appropriate for you project.
Some of these acceptance criteria require processing beyond that performed by the ANSROUT
spreadsheet described in the preceding chapter. Chapter 12 discusses this supplemental processing.

11.1 DRIFTS

FEMA-273 does not have an explicit restriction on drift as the criteria are aimed at maximum
element deformations. Restrictions on drift apply indirectly when the elements are assessed.
For NZ projects, the limiting drift is the value set by NZS4203 for time history analysis, 0.025 times
the story height. This is generally interpreted as the maximum drift at any column line. If your
structure is close to the drift limit then you may need to obtain time histories of displacement at
corner columns so that accurate drift ratios are obtained. If drifts are much less than the limit then
the center of mass drifts will probably be sufficient to demonstrate compliance with this requirement.

11.2 CONCRETE COLUMNS AND BEAMS

Criteria are discussed below for concrete frame elements as required by FEMA-273 and NZNSEE.
The FEMA-273 requirements are based on maximum plastic rotations, which are obtained directly
from the results of the NSP and/or NDP analyses. The NZNSEE requirements are based on
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

223

curvature ductilities, which requires processing of the plastic rotations obtained from the analyses
and requires assumptions about yield curvature.
In my opinion, criteria based on plastic rotations are more suited to these methods of analysis than
those based on curvature ductility so, if possible, base your project criteria on the FEMA-273 limits.
11.2.1 FEMA-273 GUIDELINES

FEMA-273 provides modeling parameters and numerical acceptance criteria for reinforced concrete
beams and reinforced concrete columns as a function of such parameters as reinforcing content,
transverse reinforcing and shear stress.
Tables 11-1 and 11-2 respectively provide examples of the criteria for beams and columns. See the
tables in FEMA-273 Chapter 6 for further details and for criteria for types of buildings such as posttensioned or flat slab.
1.

In these tables, linear interpolation between values is permitted.

2.

Transverse reinforcing is classified as either C (conforming) or NC (non-conforming). A


component is conforming if stirrups within the plastic hinge regions are closed and spaced
at d/3 and, in components of moderate to high ductility demand, the strength provided
by the stirrups (Vs) is at least 75% of the design shear. Otherwise, the component is nonconforming.
TABLE 11-1 CONCRETE BEAM ACCEPTANCE CRITERIA

Plastic Rotation (radians)


Beams Controlled by Flexure
Transverse
V__
IO
LS
CP
-
Reinforcing
b
df
w
c
bal
0.0
0.25
C
0.005
0.02
0.025
0.0
0.50
C
0.005
0.01
0.02
0.5
0.25
C
0.005
0.01
0.02
0.5
0.50
C
0.005
0.005
0.015
0.0
0.25
NC
0.005
0.01
0.02
0.0
0.50
NC
0.0
0.005
0.01
0.5
0.25
NC
0.005
0.01
0.01
0.5
0.50
NC
0.0
0.005
0.005
Beams Controlled by Shear
Stirrup spacing d/2
0.0
0.0
0.0
Stirrup spacing > d/2
0.0
0.0
0.0
Beams Controlled by inadequate development or splicing
Stirrup spacing d/2
0.0
0.0
0.0
Stirrup spacing > d/2
0.0
0.0
0.0
Beams Controlled by inadequate embedment into b/c joint
0.01
0.01
0.015
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

224

TABLE 11-2 CONCRETE COLUMN ACCEPTANCE CRITERIA

Plastic Rotation (radians)


Columns Controlled by Flexure
P__
Transverse
V__
Reinforcing
Agfc
bwdfc
0.1
0.25
C
0.1
0.50
C
0.4
0.25
C
0.4
0.50
C
0.1
0.25
NC
0.1
0.50
NC
0.4
0.25
NC
0.4
0.50
NC

IO

LS

CP

0.005
0.005
0.0
0.0
0.005
0.005
0.0
0.0

0.01
0.01
0.005
0.005
0.005
0.005
0.0
0.0

0.02
0.015
0.015
0.01
0.01
0.005
0.005
0.0

The processing used to determine compliance with NZNSEE criteria can also be used to check for
FEMA-273 limits by the addition of spreadsheet functions (see Chapter 12).
For most existing concrete frames we encounter, the allowable beam plastic rotations will be in the
range of 0.005 to 0.010 radians. Columns will generally have allowable plastic rotations of about
0.005. These compare with the usual assumption for modern, well detailed concrete frames to be
able to resists plastic rotations of 0.025.

11.2.2 NZNSEE GUIDELINES

The NZNSEE Guidelines for evaluating earthquake prone buildings provide two criteria for
reinforced concrete frame elements:
1. Maximum curvature ductility. This is a function of member geometry, flexural reinforcing and
transverse reinforcing.
2. Maximum shear stress. The shear capacity of the members is the sum of the strength provided
by the concrete and the shear steel. The concrete strength component is a function of the
curvature ductility.
FEMA provides alternate procedure for evaluating concrete frame elements based on maximum
plastic rotations as a function of the confining steel, shear stress and axial load levels. These are
more directly applicable to the output produced by ANSR-L and so may be developed as an alternate
to NZNSEE this is being evaluated.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

225

11.2.2.1 Curvature Ductility

The NZNSEE recommendations limit the available curvature ductility factor to u/y where u is the
ultimate curvature and y the yield curvature. For a beam, yield curvature is given by:

y =

y
d kd

For a column, this will be influenced by axial load level. The ultimate curvature for a beam or
column is given by

u =

cu
c

where c = the neutral axis depth at the ultimate curvature and cu depends on the extent of
confinement. For unconfined concrete cu = 0.003. For confined concrete a conservative value is
given by:
cu = 0.004 ( 1 + 1.1 psfyt)
where ps = ratio of volume of transverse reinforcement to volume of concrete core and fyt = yield
strength of transverse reinforcement.
To convert curvature to rotations, the plastic hinge length Lp may be approximated as 0.5h where h is
the section depth. A less conservative value is
Lp = 0.08L + 0.022fydb
where db is the diameter of longitudinal reinforcement and fy is the yield strength of longitudinal
reinforcement.

11.2.2.2 Shear Strength

The NZNSEE Green Book defines the probable shear strength of columns without plastic hinging
as
Vp = 0.85 (Vc + Vs + Vn)
The concrete shear, Vc, is calculated as
Vc = k fc 0.8 Ag where k = 0.29
The shear resisted by the steel, Vs, is
Vs = Avfytdcot30/s
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

226

and the steel resisted by compressive axial load, Vn, is


Vn = N*tan
When plastic hinging occurs, the concrete shear strength factor, k, reduces with increased curvature
ductility.

For uniaxial lateral seismic forces, the value of k = 0.29 applies to curvature ductilities up to 3,
reducing linearly to a value of 0.10 at ductilities of 7 or greater.

For biaxial forces, the 0.29 factor reduces linearly from a ductility of 1 to 0.10 at a ductility of 5.

For beam elements, the probable shear strength without plastic hinging is:
Vp = vcbwd + Avfytd / s
Where
vc = kfc.
When beam hinging occurs the concrete shear coefficient, k, is calculated is also calculated as a
function of the curvature ductility. It reduces linearly from 0.20 for ductilities less than 3 to 0.05 for
ductilities of 7 or greater.

11.3 CONCRETE BEAM COLUMN JOINTS

Neither FEMA-273 nor the NZNSEE guidelines permit inelastic deformations in concrete beam
column joints of primary structural members. For both sources the allowable shear stress is a
function of the square root of the concrete strength and a modification factor, in FEMA-273 and k
in NZNSEE.
The NZNSEE guidelines specify k factors which differ between exterior and interior joints but does
not distinguish whether the joint does or does not have beams framing in from the transverse
direction. The FEMA-273 factors allow for more configurations and in some projects evaluated
using NZNSEE it may be advisable to also assess the FEMA-273 factors to refine the k value.
The total strength of the joint is calculated based on the concrete contribution, as discussed above,
plus the shear strength provided by the joint steel, if any. This is then compared to the shear
demand on the joint (see Section 12 for procedures to evaluate this).

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

227

11.4 CONCRETE WALLS

NZNSEE guidelines do not provide evaluation criteria for reinforced concrete walls and so FEMA273 criteria are used for all projects.
For shear walls in flexure, the limiting plastic rotations are as listed in Table 11-3 and for walls
controlled by shear the limiting drifts are provided in Table 11-4. For most flexural walls on our
projects the allowable plastic rotations for flexural hinges will be in the range of 0.002 to 0.008
radians.
For low-rise buildings with pierced shear walls, most walls will be controlled be shear and Table 11-5
lists the maximum drift for these walls as 0.0075 for CP. The coupling beam segments of these walls
have a higher limit unless the shear stress is high.

TABLE 11-3 FLEXURAL WALL ACCEPTANCE CRITERIA

Plastic Rotation (radians)


Shear walls and wall segments
(As-As)fy+P
Confined
V__
IO
Boundary
twlwfc
twlwfc
0.1
0.25
Yes
0.005
0.1
0.50
Yes
0.004
0.25
0.25
Yes
0.003
0.25
0.50
Yes
0.001
0.1
0.25
No
0.002
0.1
0.50
No
0.002
0.25
0.25
No
0.001
0.25
0.50
No
0.001
Columns supporting discontinuous shear walls
Conforming transverse reinforcing
0.003
Non-conforming
0.0
Shear wall coupling walls
Longitudinal and
V__
Transverse Reinforcement
twlwfc
Conventional, conforming
0.25
0.006
0.50
0.005
Conventional, non-conforming
0.25
0.006
0.50
0.005
Diagonal Reinforcing
n.a.
0.006

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

LS

CP

0.010
0.008
0.006
0.003
0.004
0.004
0.002
0.001

0.015
0.010
0.009
0.005
0.008
0.006
0.003
0.002

0.007
0.0

0.010
0.0

0.015
0.010
0.012
0.008
0.018

0.025
0.015
0.020
0.010
0.030

228

TABLE 11-4 CRITERIA FOR WALLS CONTROLLED BY SHEAR

Shear walls and wall segments


All segments
Shear wall coupling beams
Longitudinal and
Transverse Reinforcement
Conventional, conforming
Conventional, non-conforming

Drift (%)
or Chord Rotation (radians)
IO
LS
CP

V__
twlwfc
0.25
0.50
0.25
0.50

0.40

0.60

0.75

0.006
0.004
0.006
0.004

0.012
0.008
0.008
0.006

0.015
0.010
0.010
0.007

For NZ projects, concrete walls usually are evaluated against an acceptance criterion of maximum
strain not exceeding 0.0075. This is the CP limit in Table 11-4 and is the point beyond which
significant strength degradation may occur. For coupling beams, higher allowable strains are
permitted and if necessary the evaluation may distinguish between wall panels based on their
location.
When flexural yielding is modeled using gaps and reinforcing bar elements the maximum rotation can
be calculated as
pl=pl/L
where pl is the maximum gap opening (equal to the reinforcing bar extension) and L is the length of
the wall segment. Procedures for evaluating against these criteria are discussed in Section 12.

11.5 UNREINFORCED MASONRY WALLS

For NZ projects, unreinforced masonry walls are evaluated against an acceptance criterion of
maximum strain not exceeding 0.00150. This is the point beyond which significant strength
degradation may occur.
If FEMA-273 criteria are being used, the limiting strains are 0.003 for LS and 0.004 for CP for bedjoint sliding. This is the usual failure criterion but check FEMA-273 for your particular wall
configuration.
FEMA-273 also provides limits on the rocking mode for URM walls. When gaps are used to model
rocking the maximum gap opening can be compared against these criteria. For collapse prevention
the limit is 0.004H/L. For a pier 2 m high by 1 m wide, for example, the maximum acceptable gap
opening would be 8 mm.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

229

11.5.1 OUT-OF-PLANE WALLS

Both NZNSEE and FEMA-273 provide maximum height-to-thickness rations for URM walls as a
function of seismic zones. These limits are listed in Tables 11-5 and 11-6. The values from the two
sources are identical except that FEMA-273 has allowable H/t values for the lowest seismic zones
which are not in the NZNSEE guidelines.
Provided walls are within these limits, no further evaluation is required. If walls are outside these
limits then a more detailed evaluation can be performed to determine whether strengthening is
required.
The NZNSEE Guidelines provide an energy method for evaluating walls outside the limits in Table
11-5. In our experience, this is non-conservative as it predicts survival of walls for which a detailed
non-linear analysis shows failure. We advise caution if the method is used and generally use one of
two alternative approaches:
1.

Use a more recent procedure based on statistical studies by Opus International Consultants.

2.

Perform a detailed time history of the walls using the acceleration time histories from the
diaphragm levels of the building analysis as input.

If you are evaluating to FEMA-273 and the walls exceed the H/t limits in Table 11-6, use the second
method above.
TABLE 11-5 NZNSEE LIMITS FOR FACE LOADED URM WALLS

Wall Types
Walls of one story buildings
1st story walls of buildings of more than one story
Walls in top story of buildings of more than one story
All other walls

Seismic Zone
Z > 0.8
Z 0.8
16
13
18
15
14
9
16
13

TABLE 11-6 FEMA-273 LIMITS FOR FACE LOADED URM WALLS

Wall Types
Walls of one story buildings
1st story walls of buildings of more than one story
Walls in top story of buildings of more than one story
All other walls

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

SX1
0.26g
20
20
14
20

0.24g<SX1
0.37g
16
18
14
16

0.37g<SX1
0.5g
13
15
9
13

230

11.6 REINFORCED MASONRY WALLS

The NZNSEE guidelines do not provide criteria for reinforced masonry walls and so FEMA-273
criteria are used for all projects. Check Table 7-5 in FEMA-273, where allowable drifts for each
performance level are provided as a function of the compression on the wall, the reinforcing content
and the aspect ratio of the wall.
The allowable drifts for the CP performance level range from 0.2% to 2.6% and is a function of the
factors listed above:

Allowable drifts are highest for the smallest reinforcing ratios.

Allowable drifts reduce as the axial stress on the wall increases.

Slender walls (low L/heff ratio) have higher drift limits than squat walls.

11.7 WALLS STRENGTHENED WITH FRP

The strengthening elements used for the seismic rehabilitation of existing buildings include
traditional components such as concrete walls and shotcrete overlays and steel braced frames. More
recently, Fiber Reinforced Polymers (FRPs) have been used to improve confinement of flexural
members and to increase the in-plane shear strength of wall panels.
Criteria for traditional components are provided by FEMA-273, similar to those for existing
elements. For alternative materials and methods of construction not currently included in FEMA273, the guidelines provide experimental requirements to obtain engineering properties. ICBO
AC125 provides design guidance for FRP applied to walls. Although complete test data are not yet
available for FRP applied to URM walls, a conservative approach can be developed from data
available.
Figure 11-1 shows test results for a wall panel tested at SUNY, Buffalo. The FRP increases both the
strength and the maximum shear strain of the wall. For peer reviewed projects we have performed in
the U.S., the maximum drifts for panels strengthened with FRP overlays have been restricted to two
times that for a non-strengthened wall, 0.006 for LS and 0.008 for the CP limit. Figure 11-1 indicates
that this is conservative. However, the anchorage length supplied will limit this strain and in practice
the maximum value used is 0.006, as described below.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

231

FIGURE 11-1 FRP EFFECT ON URM WALL SPECIMEN

120
Bare Wall

SHEAR STRESS (psi)

100

Wrapped Wall

80

60

40

20

0
0.0000

0.0040

0.0080

0.0120

0.0160

0.0200

SHEAR STRAIN (in/in)

11.7.1 ANCHORAGE OF FRP

The maximum stress which the FRP can resist is a function of the available anchorage length. Tests
at San Jose State University have developed the strength versus anchorage length function as shown
in Table 11-7 and Figure 11-2.
For the design strain of 0.004 normally used for FRP an anchorage length of 12 (300 mm) is
required. If anchorage of 18 (450 mm) or longer is provided then the strain can be increased to
0.006.
TABLE 11-7 ANCHORAGE OF FRP

Anchorage
Length
(inches)
4
6
12
18
24
30

Load
(lbs)
6180
8200
11220
19150
17000
18900

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

Stress Test Strain


(ksi)
(in/in)
10.1
13.4
18.3
31.3
27.8
30.9

0.0025
0.0033
0.0046
0.0078
0.0069
0.0077
232

FIGURE 11-2 ANCHORAGE OF FRP


0.009
0.008

FAILURE STRAIN (inches)

0.007
0.006
0.005
0.004
Test Strain
Design Strain

0.003
0.002
0.001
0.000
0

10

15

20

25

30

35

ANCHORAGE LENGTH (inches)

11.8 STEEL COLUMNS AND BEAMS

Few of the buildings we evaluate have steel elements as the primary lateral load resisting system. For
steel moment frames, braced frames or eccentrically braced frames, consult FEMA-273 for modeling
and evaluation procedures. For moment frames the performance is closely related to the joint
configuration.
Special modeling techniques may be required for some forms of steel structure. If the FEMA-273
requirements for you particular steel structure seem to require stiffness and/or strength functions not
currently available in our procedure, consult the Development Division.
11.9 TIMBER FLOORS

The modelling of timber floors is generally based on FEMA-273 provisions. In practice, the
acceptance criteria applied to timber floors are usually based on acceptable deformations in the
elements connected to the floor. These are most commonly URM walls.
Although FEMA-273 is not explicit on limits, we generally limit maximum timber floor deformations
to not more than 50% of the minimum thickness of the walls connected to the floor.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

233

11.10 FOUNDATIONS

The evaluation of foundations can be based on ultimate load capacity determined by one of three
methods:
Presumptive Capacities.
FEMA-273 provides capacities for spread footings and mats, which can be used when the amount of
information on foundation soil properties is limited. Table 11-8 is extracted from FEMA-273. See
that source for further details.

TABLE 11-8 PRESUMPTIVE ULTIMATE FOUNDATION PRESSURES

Class of

Materials2

Vertical
Foundation
Pressure (KPa)3

Massive crystalline bedrock


Sedimentary and foliated rock
Sandy gravel and/or gravel
Sand, silty sand, clayey sand,
silty gravel & clayey gravel
Clay, sandy clay, silty clay and
clayey silt

383
192
192
144

Lateral
Bearing
Pressure
KPa/ 300 mm
depth
below
natural grade4
115
38
19
14

967

10

Lateral Sliding1
Coefficient5 Resistance6
KPa

0.80
0.70
0.70
0.50

12.4

1.

Lateral bearing and lateral sliding resistance may be combined.

2.

For organic clays and peat a foundation investigation is required.

3.

Values are for footings having a minimum width of 300 mm and minimum depth of 300 mm into natural grade.
Except where 7. applies, increase of 20% allowed for each additional 300 mm of width or depth to a maximum
of 3 times the tabulated value.

4.

May be increased by the amount of the designated value for each additional 300 mm of depth to a maximum of
15 times the tabulated value.

5.

Coefficient applied to the dead load.

6.

Lateral sliding resistance value to be multiplied by contact area. In no case shall the lateral sliding resistance
exceed one-half the dead load.

7.

No increase for width is allowed.

Prescriptive Capacities.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

234

Capacities may be inferred from construction documents for the existing building or previous
geotechnical reports.
FEMA-273 permits the bearing pressure for a spread footing to be assumed to be twice the allowable
dead plus live load specified for design. For piles and piers, the ultimate capacity may be assumed to
be 50% greater than the dead plus live loads used for design.
In general, any footing component may be assumed to have a capacity 50% greater than the original
design working load.
Site-Specific Capacities
A detailed analysis may be performed by a geotechnical engineer to determine ultimate foundation
capacities.

11.11 NON-STRUCTURAL COMPONENTS AND CONTENTS

There are three procedures which have been used to develop design coefficients for diaphragms and
components:
1.

Use the accelerations from the NDP. For NZ evaluations the accelerations are generally
extracted from the Sp = 1.0 analysis and are factored by Sp = 0.67 to obtain design forces.
For most applications, the average maximum acceleration values are used, where maximum
accelerations are over all nodes at a level and average is over all earthquakes.

2.

Use FEMA-273 equations, (3-9) for diaphragms and (11-2) for components. The forces are
reduced by an m factor. These coefficients are used for deformation controlled
components.

3.

As for 2., but adjust these coefficients to correspond to the actual strength of the building,
as calculated from the NSP. The accelerations from equations (3-9) and (11-2) will be
limited by the building strength. The maximum coefficients may be different in the two
directions. These coefficients are not reduced my m and are used for force controlled
components.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

235

12 EVALUATING RESPONSE

The primary aim of evaluating the response of the building is to check whether the component
forces and deformations meet the criteria requirements. In doing so, however, it is important to also
evaluate other features of response, which may indicate whether the model provides a reasonable
representation of the actual building behavior.
The items included in the evaluation in this section are intended both as a check against acceptance
criteria and, as far as possible, to ensure that the response is reasonable for the particular structure
and level of earthquake load.

12.1 PERIOD OF RESPONSE

The period of response of a building provides a good indication of whether a model has reasonable
stiffness and mass. For example, if a 10 story frame building vibrates with a period of 0.10 seconds
then it is far too stiff, similarly if the period is 10 seconds then it is far too soft.
The Zap/Sap model developed to implement the Force Pushover option provides the periods of the
structure. If this option is used then these periods provide an indication of the correctness of the
model.
ANSR-L does not provide calculation of periods. An estimation of the period of response can be
made by plotting displacements and measuring the distance between zero crossings. A more accurate
assessment of period can be made by using the HCG spreadsheet SPECTRUM.XLS. This
workbook has an input sheet into which an acceleration or displacement time history can be pasted.
A response spectrum can then be generated from this time history. Peaks on the floor response
spectrum indicate the periodicity of response. A zero damped spectrum shows the peaks in most
detail.
An example of this procedure is shown in Figures 12-1 to 12-3. The roof acceleration from an ACC
file produced by ANSROUT is shown in Figure 12-1. A 2 second portion of this record, plotted in
Figure 12-2, shows an apparent period of about 0.4 with a shorter period wave of about 0.10 second
period superimposed on it.
Figure 12-3 is a 0% damped response spectrum of this acceleration time history. This shows clear
peaks at periods of 0.09 seconds and 0.42 seconds. It is reasonable to expect that these two periods
are the two fundamental modes in the direction in which the accelerations are recorded.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

236

Often the spectrum will show a broad peak or multiple peaks at the longest dominant period. For
example, Figure 12-3 shows a secondary peak at 0.46 seconds. This is an effect of the period
lengthening due to yielding. This effect is typically more pronounced on the fundamental mode than
on higher modes.
FIGURE 12-1 ROOF ACCELERATION TIME HISTORY

1.50

1.00

ACCELERATION (g)

0.50

0.00
0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

18.00

20.00

9.80

10.00

-0.50

-1.00

-1.50

-2.00
TIME (Seconds)

FIGURE 12-2 PORTION OF ROOF ACCELERATION TIME HISTORY

1.50

1.00

ACCELERATION (g)

0.50

0.00
8.00

8.20

8.40

8.60

8.80

9.00

9.20

9.40

9.60

-0.50

-1.00

-1.50

-2.00
TIME (Seconds)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

237

FIGURE 12-3 ACCELERATION RESPONSE SPECTRUM (UNDAMPED)

25.00

ACCELERATION (g)

20.00

15.00

10.00

5.00

0.00
0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

PERIOD (Seconds)

12.2 DISPLACEMENTS AND DRIFTS

For structures in which drift is critical, it is usually advisable to obtain time histories of nodal
displacements at corner column lines to ensure that the worst case is checked. ANSROUT can be
used to check envelope displacements and drifts at column lines at which time history output is
requested.
The evaluation procedure for walls is often based on the maximum drift imposed on the wall (see
later in this chapter) and there are direct limits imposed for NZ projects. In these cases, drifts
should always be calculated from the time histories of displacement and values calculated from
envelope displacements may under-estimate maximum values.

12.3 ACCELERATIONS
12.3.1 VALIDATION OF THE MODEL.

Values of the peak acceleration can indicate possible errors in the model. Spectra generated from the
accelerations can be used to check periodicity (see above).
The acceleration profile with height of the structure is generally assumed to be linearly increasing,
forming an approximately triangular distribution. The actual profile can be plotted from the values in
the Envelope sheet in the ANSROUT workbook if time histories have been processed.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

238

An example is shown in Figure 12-4. This shows an increase in acceleration at the upper levels, a
common characteristic of flexible buildings and the reason codes such as NZS4203 and the UBC
require an additional inertia force at the top of the building. Note that the distribution is not the
triangular shape assumed by codes as the accelerations at lower levels tend towards the maximum
ground acceleration, which was 0.25g for the input motion used to generate Figure 12-4.

FIGURE 12-4 PROFILE OF FLOOR ACCELERATIONS

19

17

ELEVATION (m)

15

13

11

5
-0.8

-0.6

-0.4

-0.2

0.2

0.4

0.6

0.8

FLOOR ACCELERATION (g)

The spectral acceleration can be obtained from the spectrum of the seismic input to the structure and
the approximate period of response. This acceleration occurs at the centroid of the triangle, twothirds the height for a building of uniform mass distribution. The acceleration at the top will
therefore be about 1.5 times the spectral value.
For example, assume a building on intermediate soil with a 1 second period and Z=1.2, Lu = 1 and R
= 1. For time history analysis to obtain ductility demand,
C(T)

=
=
=

Ch (T, 1) RZLu
0.50 x 1.0 x 1.2 x 1.0
0.60

For this building, the expected elastic acceleration level at the roof would be approximately 1.5 x 0.60
= 0.90. This value forms an upper bound on the acceleration expected from the non-linear analysis.
Yielding of the structure acts as a fuse to reduce floor accelerations. A perfectly elasto-plastic,
single-degree-of-freedom system would limit accelerations to the yield level. For example, such a
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

239

system with a yield level of 0.10 times the structure weight would have maximum accelerations of
0.1g. In real structures, yielding will initiate in some elements but a full yielded mechanism generally
does not form until much higher loads than those required to cause first yield. Also, higher modes
can cause high accelerations, which are less influenced by yielding than the fundamental mode.
The value of C(T) only applies for the spectrum compatible time history as the scaled, recorded
earthquake motions have spectral peaks and valleys. The actual spectral acceleration at a particular
period may vary considerable from the NZS4203 value.

12.3.2 LOADS ON COMPONENTS.

Maximum accelerations at particular nodes can be used as seismic coefficients for evaluating or
designing rigid parts or components. Spectra generated from the acceleration time histories can be
used to evaluate or design flexible parts. These spectra should be generated using damping ratios
appropriate to the component being evaluated.
Depending on the importance of the component, the acceleration time histories may be used as input
to a separate non-linear model of the component itself. This can be used for evaluating the response
of walls under face loads, for example.
There is some uncertainty as to how to use floor accelerations and/or floor spectra to evaluate
components. NZS4203 requires that parts be designed using a minimum amplification of 2 over the
maximum acceleration regardless of the rigidity of the part. It may also be necessary to perform an
analysis using member overstrength capacities to obtain upper bound accelerations.
If the acceptance criteria for your project require floor accelerations, you will need to save time
histories of nodal response for floor nodes and possibly other nodes. When these are processed, the
maximum accelerations are imported to the Envelopes sheet by ANSROUT.

12.4 OVERALL BUILDING HYSTERESIS

The force or displacement pushover analysis can be used to construct an overall strength hysteresis
curve for the building. The amplitude and shape of these curves provides a check on the correctness
of the model. Figure 12-5 is an example hysteresis curve for a shear wall structure. Features to
check on this plot are:

Is the yield shear force as a function of weight reasonable?

Is the ultimate shear force (total strain hardening) justified by the yield hierarchy (plastic hinges
forming after the initial yield) or by the material strain hardening specified?

Is the stiffness and/or strength degradation consistent with the element properties?

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

240

Figure 12-5 shows how the strength continues to increase for a model with multiple yielding
elements as yielding progresses. This contrasts with the single panel element models, which
produced the hysteresis shapes given in Chapter 6.

FIGURE 12-5 HYSTERESIS FOR SHEAR WALL BUILDING

0.15

BASE SHEAR / WEIGHT

0.10

0.05

0.00
-500

-400

-300

-200

-100

100

200

300

400

500

-0.05

-0.10

-0.15
ROOF DISPLACEMENT(mm)

12.5 COLUMNS

The acceptance criteria for columns are based on curvature ductility for the NZNSEE guidelines or
plastic rotations for FEMA-273 guidelines.
The plastic rotations are output directly for all column elements and reported on the Columns sheet
and so checking for FEMA-273 criteria is straight-forward. Plastic rotations are printed for the
three vector components of plastic deformation (X, Y and axial). There will generally be an axial
component as the direction of loading when yield occurs will have a vertical component. For
evaluation, the higher of the two plastic rotations about the X and Y axes is used to compare with
FEMA-273 limits.
For NZNSEE evaluation, use the procedures outlined below. These are implemented by copying
formulas in the ColRes sheet of the ANSROUT workbook. Tailor these sheets to suit your project.

12.5.1 CALCULATING CURVATURE DUCTILITY CAPACITY

The curvature ductility capacity can be calculated using procedures such as those incorporated in a
spreadsheet developed by Chris Allington at the University of Canterbury.
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

241

This spreadsheet calculates the moment curvature function up to a limiting value, which is the
minimum of two conditions, (1) the concrete strain reaches the maximum value permitted or (2) the
moment reduces below 80% of the peak value. The axial loads used should correspond to the loads
on the columns when maximum plastic rotation occurred.
The moment curvature spreadsheet produces curvature ductility capacities, ,. The maximum
allowable plastic rotations can be calculated as the ultimate curvature multiplied by a plastic hinge
length of 0.5 times the column depth.
Table 12-1 provides an example of capacities produced by the spreadsheet. The values in this table
are all based on confining steel of 6.3 mm diameter bars at 203 mm centers. Figure 12-6 provides an
example of the moment curvature curve produced.

TABLE 12-1COLUMN DUCTILITY

File
MC1
MC2
MC3
MC4
MC5
MC6

Depth Width No. Size Axial EIeff


pl
Of
Of Load
/ EIg
Bars Bars
760
560
8
28
255 20.9 0.257 0.036
508
508
6
22
387 14.3 0.253 0.025
660
560
8
25
332 21.3 0.287 0.033
325
480
4
25
234 22.2 0.378 0.031
325
375
4
22
110 24.8 0.324 0.035
375
400
4
22
225 27.3 0.381 0.033

FIGURE 12-6 TYPICAL MOMENT CURVATURE CURVE


Moment-Curvature
600
500
Moment (kNm )

400
300
200
100
0
-0.02

0.02

0.04

Curvature (1/m)

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

0.06

0.08

0.1

242

12.5.2 YIELD CURVATURE

The curvature spreadsheet provides the maximum allowable curvature ductility. However, output
from ANSR-L is the maximum plastic rotation. Plastic rotation can be related to the ultimate
curvature by the plastic hinge length but the yield curvature is required to compare plastic rotations
with allowable ductility.
The yield curvature of a column depends on the axial load level. If the yield moment is known, the
curvature can be calculated as My/EI. The actual curvature in the ANSR-L model is based on the EI
input to the model. However, the curvature ductility calculated by the spreadsheet is based on an
effective EI. The EI used to calculate yield curvature should be consistent with the moment
curvature spreadsheet EI.
The spreadsheet defines the yield curvature based on first yield of a reinforcing bar. This curvature
can be used to define an effective EI as a ratio of the EI based on the gross concrete section. For
the example project, the effective EI ranged from 0.26 to 0.38 of the gross value (Table 12-1). An
upper bound value of 0.4 was used in the evaluation to produce analysis results consistent with the
definition used to define curvature ductility.

12.5.3 SPREADSHEET CALCULATIONS

The ANSROUT envelope spreadsheet for the column results contains maximum moments, shear
forces, axial loads and plastic rotations for each column element. As the columns are threedimensional, moments, shears and rotations are reported about each axis of the element. The
evaluation procedure implemented in the ColRes spreadsheet for actions about each of the two axes is
based on the following:
1. A plastic hinge length of 0.5 times the column dimension is calculated.
2. The effective stiffness is calculated as EIeff = 0.4 EIgross.
3. The yield moment is assumed equal to the maximum of the moments at each end of the column.
4. The plastic curvature, pl, is calculated as the maximum plastic rotation from the analyses
divided by the plastic hinge length.
5. The yield curvature, y, is calculated as the yield moment divided by the effective stiffness.
6. The curvature ductility, , is calculated as 1 + pl / y
7. The ultimate shear stress is calculated as vu = Vmax / bd.
8. The concrete shear coefficient, k, is calculated as a function of the curvature ductility, between
0.29 and 0.10.
9. The minimum axial load is calculated.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

243

10. The concrete shear strength due to axial loads is calculated as vcn = Pmin tan . The angle is
calculated assuming a neutral axis depth of 1/3 the section dimension.
11. The concrete shear strength is calculated as vc = 0.8 k fc where the 0.8 factor converts gross
area to shear area.
12. The steel shear strength is calculated as vs = (vu vc vcn) / .
13. The required shear steel content Av/s is calculated as vs bd / fytdcot 30.
The procedure is repeated about each axis and the shear steel required compared with the steel area
provided to assess the column adequacy.

12.6 BEAMS

For FEMA-273 evaluations the beam acceptance criteria are based on maximum plastic rotations.
As for columns, these can be checked directly against plastic rotations imported to the ANSROUT
workbook. For beams, maximum plastic rotations are reported for positive and negative bending at
each end of the beam. Usually, the evaluation uses the highest of these four values.
The NZNSEE evaluation procedure for beams is generally similar to that for columns except for
simplifications caused by bending about one axis only and the absence of axial loads. The procedure
is implemented on the BeamRes sheet:
1. A plastic hinge length of 0.5 times the beam depth dimension is calculated.
2. The effective stiffness is calculated as EIeff = 0.5 EIgross.
recommendations from NZS3101.

This ratio is based on the

3. The yield moment is assumed equal to the maximum of the moments at each end of the beam.
4. The plastic curvature, pl, is calculated as the maximum plastic rotation from the analyses
divided by the plastic hinge length.
5. The yield curvature, y, is calculated as the yield moment divided by the effective stiffness.
6. The curvature ductility, , is calculated as 1 + pl / y
7. The ultimate shear stress is calculated as vu = Vmax / bd.
8. The concrete shear coefficient, k, is calculated as a function of the curvature ductility, between
0.20 and 0.05.
9. The concrete shear strength is calculated as vc = k fc
10. The required steel shear strength is calculated as vs = (vu vc) / .
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

244

11. The required shear steel content Av/s is calculated as vs bd / fytd


As for the column evaluation, this procedure is implemented in the spreadsheet using envelope
values.

12.7 BEAM-COLUMN JOINTS

Both FEMA-273 and NZNSEE guidelines set limits on the maximum shear stress in beam-column
joints. This requires calculation of the imposed shear stress and the shear capacity of the joint.
For concrete frames, the processor program PROCESSA writes a file containing information used to
evaluate the beam column joints. The file has the job name with the suffix BMC and contains the
following information for each joint:
Beams
XI, XJ, YI, YJ
Columns
Top, Bottom

B, D, M+,MB, D, VX, VY, PMAX,PMIN

In this list,

XI is the beam parallel to the X axis for which End I frames into the joint, etc.

B and D are the section dimensions.

M+ and M- are the maximum positive and negative beam moments.

VX and VY are the maximum column shears about the two axes.

PMAX and PMIN are the maximum and minimum column axial loads.

The force components are the maximum and minimum values from each time step of all the analyses
processed and so do not necessarily occur simultaneously. However, it is conservative to use this set
of forces.
The forces are written into a comma-delimited file, which can be imported to Excel. A template
spreadsheet, BeamColumns.XLS, is available for this purpose. The data from the BMC file is pasted
into this spreadsheet and Excel functions in the cells evaluate the joints. The workbook is currently
set up to evaluate for NZNSEE procedures but FEMA-273 requirements are quite similar. There
are some approximations involved and you will need to adjust these so that the evaluation is
conservative for you project.
The procedure is implemented in the following steps:

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

245

Set parameters for your building.


These include concrete strength, fc, details of tie sets in the joints and the k (or for FEMA-273)
factors to use at exterior and interior joints. Also, define a factor, f, to use to transform the beam
moments to compression and tension forces. This factor is defined such that the tension T (or
compression, C) = M/fD where M is the applied moment and D is the overall depth. This factor
expresses the effective depth between the centroid of the compressive and tension forces as a
fraction of the depth. Typical values will be in the range of 0.8 to 0.6, with lower values for high
reinforcing content. You will need to do some calculations to get a value for this and may need to
vary it between beam types.
Calculate Allowable Shear Strength.
The concrete shear strength is a function of the joint dimensions, minimum axial load, concrete shear
strength and the factor k (or ). The workbook assembles these from the input data and calculates
the maximum allowable joint stress in each direction. The information on the number of beams
framing into the joint is used to distinguish between interior and exterior joints. The strength
supplied by the tie sets, if any, is added to the concrete shear strength.
Calculate Applied Shear
The maximum joint shear from the beams is calculated by calculating for bending in the beams in the
positive and negative directions and subtracting the column shear from the higher of the two
conditions.
Calculate Shear Ratio
The ratio of the applied shear to the shear strength is calculated for each joint. A ratio greater than
one indicates a deficiency. The spreadsheet assembles the total number of joints in each direction
plus the number of deficiencies, as shown in the example in Table 12-2. In this example, the joints
are satisfactory in the X direction but the demand exceeds the capacity by an average of 9% in about
30% of the joints in the Y direction (189 of 616 joints).
TABLE 12-2 EXAMPLE BEAM-COLUMN JOINT EVALUATION

Total
Count
Max
Average

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

Ratio
X
616
0
0.00

Ratio
Y
616
189
1.27
1.09

246

12.8 WALLS

Depending on the wall segment configuration and material, acceptance criteria may be shear strain,
drift or chord rotation. The panel element formulation is based on the total shear strain and strains
are output for shear deformations plus vertical and horizontal strains. The output spreadsheet lists
maximum values of all these strains.
The procedure used to model walls using finite elements produces peak shear strains in individual
elements. If multiple elements are used within a story height and along a wall length the strains in the
elements will incorporate effects of the shear distributions along the length/height and the effects of
stress concentrations around openings.
Acceptance criteria for walls governed by shear, both concrete and masonry, in FEMA-273 are based
on maximum lateral drifts rather than shear strains at local points in the wall. Effectively, they
represent a global demand rather than local demand.
On this basis, most URM and concrete wall projects have been evaluated on the basis of the
maximum story drift. The peak drift is compared to the allowable limit for the material type.
Provided the story drift is within the particular performance limit for that earthquake, the response is
acceptable.
The strains in individual elements can be used to determine where damage will occur and will also be
used to evaluate coupling beams, assuming shear strain is equivalent to chord rotation.
Use the ModelPlot option to display walls where strains exceed the acceptance criteria for the
particular performance level and use the plots assess whether panels are walls or coupling beams.
For rocking URM walls the acceptance criteria are based on maximum drifts and the gap openings
can be used to evaluate response against these criteria. For example, the CP limit is 0.004heff/L. For
a pier 2 m high by 1 m wide the maximum acceptable gap opening would be 8 mm. You will need
to assemble pier dimensions to assess performance of rocking walls.
For flexural concrete walls the assessment of response is also based on the gap opening. As
discussed earlier, the maximum rotation is pl=pl/L where pl is the maximum gap opening (equal
to the reinforcing bar extension) and L is the length of the wall segment. For example, a if a flexural
wall 3 m long had an allowable plastic hinge rotation of 0.010 radians, the maximum acceptable gap
opening would be 0.010 x 3000 = 30 mm.
The allowable value is a function of flexural and shear stress so you will need to set up worksheet
functions to evaluate flexural walls. The stresses and dimensions are available on the Walls results
sheet and so can be used to develop procedures to assess the walls.
12.9 FOUNDATION S

Foundations generally need to be evaluated against allowable vertical loads and total base shear.
Vertical loads are typically obtained from the gap element maximum compression forces or from pile
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

247

loads if piles are modeled. If you have fixed base without gap elements then vertical loads will not be
obtained. Include gap elements if you will need to evaluate foundation bearing pressure, unless
vertical loads can be derived from element results immediately above the foundation.
To evaluate shear, you will need the total base shear force. For the NDP this can be obtained from
the element shear summation, for the NSP it will be the total applied force at the target displacement.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

248

13 FUTURE DEVELOPMENTS

A number of extensions and enhancements are planned as part of the development process of the
evaluation procedure. Table 13-1 lists those identified to date.
If you have a project for which one or more of these items will be important, please advise the
Development Division and the item will be addressed with a higher priority.
TABLE 13-1 CURRENT TASK LIST

Task
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.

Reinforced walls : calculation of vc, vs, from fc,


Automate BeamRes and ColumnRes sheets in AnsrOut
Extend AnsrOut plots for wall shear strain
Research hysteresis shapes, especially from FEMA 307
Remove dimensions for element properties in ANSR
Clear Old Data safe to use?
Acceptance Criteria. NZNSEE/FEMA implementation
Document, tidy up Capacity Spectrum spreadsheet
Rationalize ANSR element time history output code
Shear / Axial failure in beam/column elements
Degradation as a function of duration
Read variable number of components per earthquake (e.g. Vertical)
Automate time history scaling
Procedures for foundation modeling
Alternative units (e.g. U.S)
Torsion springs for beam/column joints
Use wall gap element to model sliding shear failure.
Steel bracing type element (ex PC-ANSR)
Automated FEMA backbone curves
Improved solution procedure (ex ANSR-III)
Extended procedures for Base Isolation
Extended procedures for Energy Dissipators
Automate Composite Strengthening Procedures
Face Load Models

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

249

14

PROJECT LIST

A project list is maintained, by category of building, to enable consistent procedures and reporting
techniques to be used for all evaluations that are performed using this procedure. Reports on each of
the buildings listed in Table 14-1 are available should you wish to use them as a basis for reports on
future projects.
As you complete projects using these procedures, please advise the Development Division so that
the table can be kept current.

TABLE 14-1 PROJECT LIST

Type
Unreinforced Masonry Buildings

Project
Christchurch Arts Center Clock Tower
Christchurch Arts Center Old Boys High
Bellamys Building, Christchurch
Armagh Street Tower, Christchurch
Christchurch Cathedral

St Marys Chapel, Christchurch)


Auckland War Memorial Museum
Auckland Civic Theatre

Auckland Customhouse
Auckland Railway Campus (ex Railway Station)
Nathan / Australis Buildings, Auckland
127 Lichfield St, Christchurch
Waikato Polytech Block F
Cocoa & Chocolate Building, CA
Auckland Gas Company Office & Workshop
Union Pacific Railroad Station, Salt Lake City,
25-29 Symonds St, Auckland
Concrete Frame & Wall Buildings

Invercargill Civic Administration Building


Christchurch Drainage Board Building
University of Canterbury Student Union

University of Canterbury Hight Library


Palmerston North Hospital
Middlemore Hospital Galbraith Building
Radisson Huntley, Santa Monica, CA
Christchurch Civic Office
Clear Communications, Christchurch
Beehive, Wellington
80 Anzac Ave, Auckland
Copyright 2001. This material must not be copied,
reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

250

Type

Project
153 Featherston, Wellington
Panama Chambers
Middlemore Hospital Staff Residence
Middlemore Hospital Maternity Block
Puke Ariki Building
NZI Building, Auckland
SBI Building, Auckland
Riverside Casino, Hamilton

Multiple Buildings Including Pounding

Health Link South, Christchurch


192-196 Queen St, Auckland

Base Isolation of Existing Buildings

Parliament Buildings

Base Isolation of New Buildings

Museum of New Zealand, Te Papa Rongarewa


Hutt Valley Hospital
Missouri Botanical Gardens
Whareroa Boiler

Special Structures

Kapuni Regenerator Tower


Ammonia Convertor
Refining Column Distillation III
Bastia Hill Water Tower, Wanganui
Invercargill Liquid Pitch Pipeline
Timbertank Reservoirs (37 projects)
ACI Glassworks Chimney
Shell Gully Motorway Bridge

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

Wellington Maritime Museum


Union House Added Floors

251

FIGURE 14-1WAIKATO POLYTECH BLOCK F

FIGURE 14-2 CANTERBURY UNIVERSITY STUDENT UNION

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

FIGURE 14-3 INVERCARGILL CIVIC BUILDING

252

FIGURE 14-5 CHRISTCHURCH DRAINAGE BOARD BUILDING

FIGURE 14-4 ARMAGH STREET TOWER

FIGURE 14-6 BASTIA HILL WATER TOWER

FIGURE 14-7 THE BEEHIVE

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

253

FIGURE 14-8 BEAUMONT STREET

FIGURE 14-9 SYMONDS STREET

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

FIGURE 14-10 CLOCK TOWER

254

FIGURE 14-11 LORNE STREET

FIGURE 14-12 OLD BOYS SCHOOL

FIGURE 14-13 AUCKLAND MUSEUM

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

255

FIGURE 14-15 PANAMA STREET

FIGURE 14-16 NZI AND SBI BUILDINGS

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

FIGURE 14-14 THE GATEWAY

FIGURE 14-17 RADISSON HUNTLEY

256

FIGURE 14-18 AUCKLAND RAILWAY STATION

FIGURE 14-19 QUEEN STREET BUILDINGS

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

257

FIGURE 14-20 CHRISTCHURCH CATHEDRAL

FIGURE 14-21 GALBRAITH BUILDING

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

FIGURE 14-22 HIGHT LIBRARY

258

FIGURE 14-23 RIVERSIDE CASINO

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

FIGURE 14-24 COCOA BUILDING

259

15 TROUBLESHOOTING
Nonlinear analysis is more complex than linear elastic analysis. The successful completion of an
evaluation using the NSP or NDP requires care in preparation of the model and evaluation of results.
Although the procedure documented here is intended to automate the model preparation, analysis
and processing steps, there will inevitably be problems associated with one or more of these phases.
This section provides a list of the more common problems which have been encountered.

15.1 MODELA

Check that you have cleared previous data from the workbook. Check hidden sheets, such as
Connectivity, DOFs, Masses & Loads, which are used infrequently.

Check that the material types used to define sections have been specified and that you have not
duplicated identification numbers.

Make sure the section and strength types used to define elements have been defined on the
Materials & Sections sheet and that you have not duplicated identification numbers.

If the model generation process seems very fast, it may mean that an error message has been
written into the .TRC file. Open this file with Wordpad or similar and check the bottom of the
file. You may see a message such as Cannot find section / strength type n. See above to remedy this.

The program reads input data from a file with the suffix .MDA. Sometimes, if you scan this file
with a text editor you will notice odd data, which will help you locate an error.

15.2 ANSR-L

See sections 8.9 and 8.10 for a discussion of possible errors in ANSR-L. A summary of things to
check is:

Unable to open files. Check your config.sys file (Section 3.4)

Error during input phase. Check whether a material property or section dimension is set to zero.
See bottom of .ECH file for location of error.

Error during input phase. May be too many stiffness, eccentricity or initial load definitions for
an element group. See bottom of .ECH file for location of error. See Section 8.9.3 to fix.

Analysis does not continue after completion of input phase. The memory size (blank common)
may be too small for the problem. See bottom of .ECH file for requirements.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

260

Solution terminates at first step. Usually zero stiffness somewhere. See Section 8.9.4.

Solution terminates part way through. Maximum displacement may be exceeded. This usually
indicates instability. Check that you have not set too small a limit on the Analysis Data sheet.

Solution does not converge. This message will be near the bottom of the .D01 file. Generally
indicates failure. See Section 8.10.2.

15.3 PROCESSA

Open the .TRC file with Wordpad or similar and search for a string containing ERROR. Some
error messages which do not prevent the analysis running are written to this file and prevent
PROCESSA executing.

The message ERROR in K Node in the .TRC file refers to the AutoCad drawing of the elements
and indicates a problem with element connectivity / direction. Check out the element defined
where the error is written and see whether there is an error on the MODELA spreadsheet.

15.4 PUSHOVER ANALYSIS

Make sure the node you specify for the pushover displacement has elements connected to it.

For modal analysis, some unrestrained degrees of freedom acceptable in ANSR-L may cause an
error. In particular, pinned columns used to model masonry mullions are okay for ANSR-L but
will cause an error in ZAPL. You may need to constrain these degrees of freedom.

ZAPL had a bug which prevented solution using multiple blocks. This has been fixed but may
occasionally re-surface. If solution does not work with more than one stiffness block, try
increasing the Common size.

Check also the potential problems listed for ANSR-L above.

15.5 AUTOCAD DRAWINGS

If walls do not display correctly, there may be wall columns in the model. These are added to
restrain out-of-plane degrees of freedom for walls not parallel to the global axes. Go to Layer
Control and freeze the WALLCOLUMN layer.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

261

16 BIBLIOGRAPHY
The procedure has been developed using code provisions, rehabilitation guidelines, experimental data
and analysis techniques from a variety of sources. The documents listed in this bibliography
represent the resources used in developing the methodology.

[1]

Code of Practice for General Structural Design and Design Loadings for Buildings, NZS 4203:1992,
Standards New Zealand, 1992.

[2]

UBC, 1994, Appendix to Chapter 16, Division III, International Conference of Building
Officials, Whittier, CA, 1994.

[3]

NEHRP Guidelines for the Seismic Rehabilitation of Buildings, FEMA-273, Federal


Emergency Management Agency, Washington D.C. October, 1997

[4]

ATC Seismic Evaluation and Retrofit of Concrete Buildings, Applied Technology Council,
California, November, 1996

[5]

Repair of Earthquake Damaged Concrete and Masonry Wall Buildings, FEMA-308, Federal
Emergency Management Agency, Washington D.C. May, 1999

[6]

Earthquake Risk Buildings : Recommendations and Guidelines for Classifying, Interim Securing and
Strengthening, New Zealand National Society for Earthquake Engineering, December,
1985. Technical update issued in draft form in February, 1995.

[7]

The Assessment and Improvement of the Structural Performance of Earthquake Risk Buildings,
Draft for General Release, New Zealand National Society for Earthquake Engineering,
December, June, 1996.

[8]

Park, R and Paulay, T, Reinforced Concrete Structures, John Wiley and Sons, 1986.

[9]

T Paulay and N J N Priestley, Seismic Design of Reinforced Concrete and Masonry Buildings,
John Wiley & Sons, 1992.

[10]

Concrete Structures Standard : Part 1 - The Design of Concrete Structures, NZS 3101:Part 1:1995,
Standards New Zealand, 1992.

[11]

Priestley, N J. N. Seismic Behaviour of Unreinforced Masonry Walls, Bulletin of New Zealand


National Society for Earthquake Engineering, Vol. 18, No. 2, June, 1985.

[12]

E L McDowell, K E McKee and E Sevin, Arching Action of Masonry Walls, Proc. of the
American Society of Civil Engineers, Journal of the Structural Division, Vol. 82, No.
ST2, March, 1956.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

262

[13]

K E McKee and E Sevin, Design of Masonry Walls for Blast Loading, Proc. of the American
Society of Civil Engineers, Journal of the Structural Division, Proc Paper 1511, January,
1958.

[14]

ABK Joint Venture, Methodology for Mitigation of Seismic Hazard in Existing Unreinforced
Masonry Buildings, El Segundo, CA, Agbabian Associates, ABK-TR-01 to ABK-TR-08,
1981 to 1984.

[15]

T. E. Kelly, Earthquake Resistance of Unreinforced Masonry Buildings, New Zealand National


Society for Earthquake Engineering, Technical Conference, Rotorua, New Zealand,
April 1995.

[16]

Wesley, H. Banon and R. B. Jenkins Seismic Capacities of Masonry Walls at the Big Rock Point
Nuclear Generating Plant, Structural Engineering in Nuclear Facilities, Volume 1, ASCE,
September, 1984.

[17]

T. E. Kelly, M. R. Button and R. L. Mayes Seismic Evaluation of Reinforced Masonry Walls,


Structural Engineering in Nuclear Facilities, Volume 1, ASCE, September, 1984.

[18]

Mondkar, D.P. and Powell, G.H., 1979, ANSR II Analysis of Non-linear Structural
Response User's Manual, EERC 79/17, University of California, Berkeley, July.

[19]

Habibullah, A, ETABS Three Dimensional Analysis of Building Systems USERS MANUAL,


Version 6.0, Computers and Structures Inc, Berkeley, CA, October, 1994,

[20]

Mengi Y. and McNiven H., A Mathematical Model for Predicting the Non-linear Response of
Unreinforced Masonry Walls to In-Plane Earthquake Excitations, Earthquake Engineering
Research Center, Berkeley, CA, Report No. UCB/EERC-86/07, May, 1986.

[21]

Blaikie, E. L and Spurr D. D Earthquake Vulnerability of Existing Unreinforced Masonry


Buildings, Works Consultancy Services, December, 1992.

[22]

Epperson, G and Abrams, D Evaluating Lateral Strength of Existing Unreinforced Brick Piers
in a Laboratory, Proc. 5th North American Masonry Conference, Urbana-Champaign,
1990.

[23]

Magenes, G and Calvi, G. M Cyclic Behaviour of Brick Masonry Walls, 10th World
Conference on Earthquake Engineering, Madrid, 1992.

[24]

R. I. Skinner, W. H Robinson and G H McVerry, An Introduction to Seismic Isolation,


Wiley, 1993.

[25]

M J N Priestley, Brief Comments on Elastic Flexibility of Reinforced Concrete Frames and


Significance to Seismic Design, Bulletin of the New Zealand National Society for
Earthquake Engineering, Vol. 31, No. 4, December, 1998.

[26]

Kelly, T.E. and Chambers, J.D. Analysis Procedures for Performance Based Design, 12th World
Conference on Earthquake Engineering, Auckland New Zealand, February, 2000.

[27]

Reinhorn, A.M., and Madan, A., Evaluation of TYFO-W Fiber Wrap System for In-Plane
Strengthening of Masonry Walls, Report No. 95-AMR-0002, Dept. Of Civil Engineering,
State University of New York, Buffalo, August 1995.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

263

[28]

Base Isolation of Structures - Design Guidelines, Holmes Consulting Group, Revision 0, July,
2001.

[29]

In-Structure Damping and Energy Dissipation Design Guidelines, Holmes Consulting Group,
Revision 0, July, 2001.

[30]

Opus International Consultants Methodology for the Assessment of Face Loaded Unreinforced
Masonry Walls Under Seismic Loading, Report prepared for the EQC Research Foundation,
May, 2000.

Copyright 2001. This material must not be copied,


reproduced or otherwise used without the express, written
permission of Holmes Consulting Group.

264