@Seismicisolation

@Seismicisolation
TALL BUILDING
STRUCTURES:
ANALVSIS AND DESIGN

Bryan Stafford Smith

McGill University
Montreal, Quebec
Canada

Alex Coull
University of Glasgow
Glasgow, Scotland
United Kingdom

JOHN WILEY & SONS, INC.

New York • Chichester • Brisbane • Toronto • Singapore

@Seismicisolation
@Seismicisolation
In recognition of the imponance of preserving what has been
written, it is a policy of John Wiley & Sons, Inc.. to have books
of enduring value published in the United States printed on
acid-free paper. and we exen our best effons to that end.

Copyright© 1991 by John Wiley & Sons. Inc.

All rights reserved. Published simultaneously in Canada.

Reproduction or translation of any pan of this work

beyond that permitted by Section I07 or I 08 of the
1976 United States Copyright Act without the permission
of the copyright owner is unlawful. Request� for
permission or funher information should be addressed to
the Permissions Depanment, John Wiley & Sons, Inc.

Library of Congress Cata/ogi11g in Publication Data:

S1af!"ord Smith. Bryan.
Tall building structures: analysis and design/Bryan
Stafford Smith, Alex Coull.
p. em.

Includes bibliographical references.

I. Tall buildings-Design and construction. 2. Structural
engineering. I. Coull, Alex. II. Title.
THI611.S59 1991
690-dc20 90-13007
CIP

Printed and bound in Singapore

10 9 8 7 6 5 4 3 2

@Seismicisolation
@Seismicisolation
 To Betty and Frances

@Seismicisolation
@Seismicisolation
@Seismicisolation
@Seismicisolation
TALL BUILDING STRUCTURES

@Seismicisolation
@Seismicisolation
@Seismicisolation
@Seismicisolation
 PREFACE

This book is the indirect outcome of 25 years of research on tall building structures
by the two authors. It began with their liaison in the mid-1960s at the University
of Southampton, England, and has since continued in their respective Universities,
of Surrey, McGill, Strathclyde, and Glasgow.
At the commencement of the period, the evolution of radically new structural
forms gave great stimulus to devising appropriate methods of analysis. In the suc­
ceeding quarter-century there have been great advances in the design and construc­
tion of tall buildings throughout the world, and in the associated development of
analytical techniques.
In the early days, approximate techniques were being devised for specific,
largely two-dimensional, structural forms, and the analysis of complex three-di­
mensional systems represented a formidable challenge. Since then, there have been
significant advances in both computer hardware and software: the power of com­
puters has increased dramatically, and a large number of comprehensive general
purpose analysis programs have been developed, based on the stiffness method of
analysis. In principle at least, it is now theoretically possible to analyze accurately
virtually any complex elastic structure, the only constraints being the capacity of
the available computer, time, and cost.
However, the great power of this analytical facility has to be handled judi­
ciously. Real building structures are so complex that even an elaborate computa­
tional model will be a considerable simplification, and the results from an analysis
will always be approximate, being at best only as good as the quality of the chosen
model and method of analysis. It is thus imperative to be able to devise an ana­
lytical model oJ the real structure that will represent and predict with appropriate
accuracy, and as efficiently and economically as possible, the response of the
building to the anticipated forces. Models required for the early stages of design
will often be of a different, lower level of sophistication than those for checking
the final design.
The task of structural modeling is arguably the most difficult one facing the
structural analyst, requiring critical judgment and a sound knowledge of the struc­
tural behavior of tall building components and assemblies. Also, the resulting data
from the analysis must be interpreted and appraised with discernment for use with
the real structure, in order to serve as a reasonable basis for making design deci­
sions.
The rapid advances in the past quarter-century have slowed up, and the era is
now one of consolidation and utilization of research findings. However, the rna-

@Seismicisolation
@Seismicisolation vii
viii PREFACE

jority of the research findings still exist only in the form of papers in research
journals, which are not generally available or familiar to the design engineer. There
is a need to digest and to bring together in a unified and coherent form the main
corpus of knowledge that has been accumulated and to disseminate it to the struc­
tural engineering profession. This task forms the main objective of this volume.
It is not possible to deal in a comprehensive manner in a single volume with all
aspects of tall building design and construction, and attention has been focused on
the building structure. Such important related topics as foundation design, con­
struction methods, fire resistance, planning, and economics have had to be omit­
ted. The intention has been to concentrate on the structllral aspects that are partic­
ularly affected by the quality of tallness; topics that are of equal relevance to low­
rise buildings have generally not been considered in any depth.
The major part of the book thus concentrates on the fundamental approaches to
the analysis of the behavior of different forms of tall building structures, including
frame, shear wall, tubular, core, and outrigger-braced systems. Both accurate com­
puter-based and approximate methods of analyses are included. The latter, al­
though being of value in their own right for the analysis of simplified regular
structures, serve also to highlight the most important actions and modes of behav­
ior of components and assemblies, and thus offer guidance to the engineer in de­
vising appropriate models for analytical purposes.
Introductory chapters discuss the forces to which the structure is subjected, the
design criteria that are of the greatest relevance and importance to tall buildings,
and the various structural forms that have developed over the years since the early
skyscrapers were first introduced at the tum of the century. A major chapter is
devoted to the modeling of real structures for both preliminary and final analyses.
Considerable attention is devoted to the assessment of the stability of the structure,
and the significance of creep and shrinkage in tall concrete buildings is discussed.
Finally, a chapter is devoted to the dynamic response of structures subjected to
wind and earthquake forces, including a discussion of the human response to tall
building motions.
In addition to the set of references appropriate to each chapter, a short bibli­
ography has also been presented. This has been designed to serve several purposes:
to note historically important papers, to recommend major works that themselves
contain large numbers of bibliographic references, and to refer to papers that offer
material or information additional to that contained in the different chapters. Space
has prevented the production of a comprehensive bibliography, since the literature
on the subject is now vast. Apologies are therefore due to the many authors whose
work has been omitted due to either the demand for brevity or the oversight of the
writers.
In view of the wide variations in practice in different countries, it was decided
not to concentrate on a single set of units in the numerical examples presented to
illustrate the theory. Thus both SI and US units will be found.
The book is aimed at two different groups. First, as a result of the continuing
activity in the design and construction of tall buildings throughout the world, it
will be of value to practicing structural engineers. Second, by treating the material

@Seismicisolation
@Seismicisolation
 PREFACE iX

in a logical, coherent, and unified form, it is hoped that it can form the basis of
an independent academic discipline, serving as a useful text for graduate student
courses, and as an introduction to the subject for senior undergraduates.
In writing the book, the authors are conscious of a debt to many sources, to
friends, colleagues, and co-workers in the field, and to the stimulating work of
those associated with the Council on Tall Buildings and Urban Habitat, the suc­
cessor to the International Committee for the Planning and Design of Tall Build­
ings, with whom they have been associated since its inception. A special privilege
of working in a university is the opportunity to interact with fresh young minds.
Consequently, above all, they acknowledge their indebtedness to the many re­
search students with whom they have worked over the years, who have done so
much to assist them in their progress. Many of their names figure in the References
and Bibliography, and many are now recognized authorities in this field. The au­
thors owe them much.
Although the subject material has altered considerably over the long period of
writing, the authors also wish to acknowledge the helpful discussions with Pro­
fessor Joseph Schwaighofer of the University of Toronto in the early stages of
planning this work.
Finally, the authors wish to express their gratitude to Ann Bless, Regina Gaiotti
and Marie Jose Nollet of McGill University, Andrea Green of Queens University.
and June Lawn and Tessa Bryden of Glasgow University, who have contributed
greatly to the production of this volume.

B. STAFFORD SMITH
A. COULL
Montreal, Quebec
Glasgow, Scotland
January 1991

@Seismicisolation
@Seismicisolation
@Seismicisolation
@Seismicisolation
 CeNTENTS

I. Tall Buildings 1

1.1 Why Tall Buildings? I

1.2 Factors Affecting Growth, Height, and Structural Form 2
1.3 The Tall Building Structure 4
I. 3.I The Design Process 5
1.4 Philosophy, Scope, and Content 6
1.5 Raisons D'Etre 7
Reference 8

2. Design Criteria 9

2.1 Design Philosophy 9

2.2 Loading 10
2.2.1 Sequential Loading 10
2.3 Strength and Stability II
2.4 Stiffness and Drift Limitations 11
2.5 Human Comfort Criteria 13
2.6 Creep, Shrinkage, and Temperature Effects 14
2.7 Fire 14
2.8 Foundation Settlement and Soil-Structure Interaction 15
Summary 16
References 17

3. Loading 18

3.1 Gravity Loading 18

3.1.1 Methods of Live Load Reduction 19
3.1.2 lmpact Gravity Loading 20
3.1.3 Construction Loads 20
3.2 Wind Loading 21
3.2.1 Simple Static Approach 22
3.2.2 Dynamic Methods 23
3.3 Earthquake Loading 25
3.3.1 Equivalent Lateral Force Procedure 26
3.3.2 Modal Analysis Procedure 29

@Seismicisolation
@Seismicisolation xi
Xii CONTENTS

3.4 Combinations of Loading 29

3.4.1. Working Stress Design 30

3.4.2 Limit States Design 30
3.4.3 Plastic Design 30
Summary 31
References 32

4. Structural Form 34

4.1 Structural Form 37

4.1.1 Braced-Frame Structures 37

4.1.2 Rigid-Frame Structures 38
4.1.3 lnfilled-Frame Structures 40
4.1.4 Flat-Plate and Flat-Slab Structures 41
4.1.5 Shear Wall Structures 41
4.1.6 Wall-Frame Structures 42
4.1.7 Framed-Tube Structures 44
4.1.8 Outrigger-Braced Structures 49
4.1 : 9 Suspended Structures 50
4.1.10 Core Structures 52
4.1.11 Space Structures 53
4.1.12 Hybrid Structures 54
4.2 Floor Systems-Reinforced Concrete 56
4.2.1 One-Way Slabs on Beams or Walls 57
4.2.2 One-Way Pan Joists and Beams 57
4.2.3 One-Way Slab on Beams and Girders 58
4.2.4 Two-Way Flat Plate 58
4.2.5 Two-Way Flat Slab 59
4.2.6 Waffle Flat Slabs 59
4.2.7 Two-Way Slab and Beam 59

4.3 Floor Systems-Steel Framing 60

4.3.1 One-Way Beam System 61
4.3.2 Two-Way Beam System 61
4.3.3 Three-Way Beam System 62
4.3.4 Composite Steel-Concrete Floor Systems 62
Summary 63

5. Modeling for Analysis 65

5.1 Approaches to Analysis 65

5.1.1 Preliminary Analyses 65
5.1.2 Intermediate and Final Analysis 66
5.1.3 Hybrid Approach to Preliminary and Final Analyses 67

@Seismicisolation
@Seismicisolation
 CONTENTS Xiii

5.2 Assumptions 67

5.2.1 Materials 68
5.2.2 Participating Components 68
5.2.3 Floor Slabs 68
5.2.4 Negligible Stiffnesses 68
5.2.5 Negligible Deformations 69
5.2.6 Cracking 69
5.3 High-Rise Behavior 69
5.4 Modeling for Approximate Analyses 70
5.4.1 Approximate Representation of Bents 71
5.4.2 Approximate Modeling of Slabs 73
5.4.3 Modeling for Continuum Analyses 77
5.5 Modeling for Accurate Analysis 78

5.5.1 Plane Frames 79

5.5.2 Plane Shear Wails 79
5.5.3 Three-Dimensional Frame and Wall Structures 83
5.5.4 P-Delta Effects 86
5.5.5 The Assembled Model 87
5.6 Reduction Techniques 88
5.6.1 Symmetry and Antisymmetry 88
5.6.2 Two-Dimensional Models of Nontwisting Structures 91
5.6.3 Two-Dimensional Models of Structures That
Translate and Twist 94
5.6.4 Lumping 99
5.6.5 Wide-Column Deep-Beam Analogies 103
Summary 104
References 105

6. Braced Frames 106

6.1 Types of Bracing 106

6.2 Behavior of Bracing 109
6.3 Behavior of Braced Bents Ill
6.4 Methods of Analysis 113
6.4.1 Member Force Analysis 113
6.4.2 Drift Analysis 115
6.4.3 Worked Example for Calculating Drift by
Approximate Methods 119
6.5 Use of Large-Scale Bracing 124
Summary 128
References 129

@Seismicisolation
@Seismicisolation
xlv CONTENTS

7. Rigid-Frame Structures 130

7.1 Rigid-Frame Behavior 131

7.2 Approximate Determination of Member Forces Caused by
Gravity Loading 133
7.2.1 Girder Forces-Code Recommended Values 133
7.2.2 Two-Cycle Moment Distribution 133
7.2.3 Column Forces 138
7.3 Approximate Analysis of Member Forces Caused by
Horizontal Loading 138
7.3.1 Allocation of Loading between Bents 138
7.3.2 Member Force Analysis by Portal Method 141
7.3.3 Approximate Analysis by Cantilever Method 146
7.3.4 Approximate Analysis of Rigid Frames with
Setbacks 150
7.4 Approximate Analysis for Drift 150
7.4.1 Components of Drift 152
7.4.2 Correction of Excessive Drift 156
7.4.3 Effective Shear Rigidity (GA) 157
7.5 Flat Plate Structure-Analogous Rigid Frame 158

7.5.1 Worked Example 159

7.6 Computer Analysis of Rigid Frames 161

7.7 Reduction of Rigid Frames for Analysis 161

7.7.1 Lumped Girder Frame 161

7.7.2 Single-Bay Substitute Frame 163

Summary 165
References 166

8. lnfilled-Frame Structures 168

8.1 Behavior of Infilled Frames 169

8.2 Forces in the Infill and Frame 172
8.2.1 Stresses in the Infill 172
8.2.2 Forces in the Frame 174
8.3 Development of the Design Procedure 174
8.3.1 Design of the Infill 175
8.3.2 Design of the Frame 177
8.3.3 Horizontal Deflection 178
8.4 Summary of the Design Method 178
8.4.1 Provisions 179
8.4.2 Design of the Infill 179
8.4.3 Design of the Frame 180
8.4.4 Deflections 180

@Seismicisolation
@Seismicisolation
 CONTENTS XV

8.5 Worked Example-Infilled Frame 180

Summary 182
References 183

9. Shear Wall Structures 184

9.1 Behavior of Shear Wall Structures 184

9.2 Analysis of Proportionate Wall Systems 186
9.2.1 Proportionate Nontwisting Structures 186
9.2.2 Proportionate Twisting Structures 187
9.3 Nonproportionate Structures 190
9.3.1 Nonproportionate Nontwisting Structures 190
9.3.2 Nonproportionate Twisting Structures 199
9.4 Behavior of Nonproportionate Structures 199
9.5 Effects of Discontinuities at the Base 202
9.6 Stress Analysis of Shear Walls 206
9.6.1 Membrane Finite Element Analysis 206
9.6.2 Analogous Frame Analysis 207
Summary 211
References 212

10. Coupled Shear Wall Structures 213

10.1 Behavior of Coupled Shear Wall Structures 213

10.2 Methods of Analysis 215
10.3 The Continuous Medium Method 216
10.3.1 Derivation of the Governing Differential Equations 216
10.3.2 General Solutions of Governing Equations 222
10.3.3 Solution for Standard Load Cases 223
10.3.4 Graphic Design Method 231
10.3.5 Coupled Shear Walls with Two Symmetrical Bands
of Openings 235
10.3.6 Worked Example of Coupled Shear Wall Structure 236
10.3.7 Coupled Shear Walls with Different Support
Conditions 243
10.4 Computer Analysis by Frame Analogy 246
10.4.1 Analysis of Analogous Frame 247
10.5 Computer Analysis Using Membrane Finite Elements 252
Summary 253
References 254

11. Wall-Frame Structures 255

11.1 Behavior df Symmetric Wall-Frames 257

11.2 Approximate Theory for Wall-Frames 260
11.2.1 Derivation of the Governing Differential Equation 260

@Seismicisolation
@Seismicisolation
XVI CONTENTS

11.2.2 Solution for Uniformly Distributed Loading 262

11.2.3 Forces in the Wall and Frame 264
11.2.4 Solutions for Alternative Loadings 266
11.2.5 Determination of Shear Rigidity ( GA) 266
11.3 Analysis by the Use of Graphs 268
11.4 Worked Example to Illustrate Approximate Analysis 271
11.5 Computer Analysis 277
11.6 Comments on the Design of Wall-Frame Structures 279
11.6.1 Optimum Structure 279
11.6.2 Curtailed or Interrupted Shear Walls 279
11.6.3 Increased Concentrated Interaction 280
Summary 281
References 282

12. Tubular Structures 283

12.1 Structural Behavior of Tubular Structures 283

12.1.1 Framed-Tube Structures 283
12.1.2 Bundled-Tube Structures 288
12.1.3 Braced-Tube Structures 289
12.2 General Three-Dimensional Structural Analysis 296
12.3 Simplified Analytical Models for Symmetrical Tubular
Structures 297
12.3.1 Reduction of Three-Dimensional Framed Tube to an
Equivalent Plane Frame 297
12.3.2 Bundled-Tube Structures 303
12.3.3 Diagonally Braced Framed-Tube Structures 305
Summary 306
References 307

13. Core Structures 308

13.1 Concept of Warping Behavior 310

13.2 Sectorial Properties of Thin-Walled Cores Subjected to
Torsion 315
13.2.1 Sectorial Coordinate w' 315
13.2.2 Shear Center 317
13.2.3 Principal Sectorial Coordinate ( w) Diagram 318
13.2.4 Sectorial Moment of Inertia Iw 320
13.2.5 Shear Torsion Constant J 321
13.2.6 Calculation of Sectorial Properties: Worked
Example 321
13.3 Theory for Restrained Warping of Uniform Cores Subjected
to Torsion 323
13.3.1 Governing Differential Equation 323

@Seismicisolation
@Seismicisolation
 CONTENTS XVIi

13.3.2 Solution for Unifonnly Distributed Torque 325

13.3.3 Warping Stresses 326
13.3.4 Elevator Cores with a Partially Closed Section 329
13.3.5 Forces in Connecting Beams 331
13.3.6 Solutions for Alternative Loadings 332
13.4 Analysis by the Use of Design Curves 332
13.5 ·worked Example to Analyze a Core Using Fonnulas and
Design Curves 333
13.6 Computer Analyses of Core Structures 341
13.6.1 Membrane Finite Element Model Analysis 341
13.6.2 Analogous Frame Analysis 344
13.6.3 Two-Column Analogy 345
13.6.4 Single Warping-Column Model 349
Summary 353
References 354

14. Outrigger-Braced Structures 355

14.1 Method of Analysis 356

14.1.1 Assumptions for Analysis 356
14.1.2 Compatibility Analysis of a Two-Outrigger
Structure 358
14.1.3 Analysis of Forces 362
14.1.4 Analysis of Horizontal Deflections 362
14.2 Generalized Solutions of Forces and Deflections 363
14.2.1 Restraining Moments 363
14.2.2 Horizontal Deflections 364
14.3 Optimum Locations of Outriggers 364
14.4 Perfonnance of Outrigger Structures 365
14.4.1 Optimum Locations of Outriggers 366
14.4.2 Effects of Outrigger Flexibility 368
14.4.3 "Efficiency" of Outrigger Structures 368
14.4.4 Alternative Loading Conditions 370
Summary 370
References 371

15. Generalized Theory 372

15.1 Coupled Wall Theory 373

15.2 Physical Interpretation of the Deflection Equation 377
15.3 Application to Other Types of Bent 378
15.3.1 Detennination of Rigidity Parameters 379
15.3.2 Calculation of Deflection 381
15.4 Application to Mixed-Bent Structures 381

@Seismicisolation
@Seismicisolation
XVIII CONTENTS

15.5 Accuracy of the Method 383

15.6 Numerical Example 384
Summary 386
References 387

16. Stability of High-Rise Buildings 388

16.1 Overall Buckling Analysis of Frames: Approximate Methods 389

16.1.1 Shear Mode 390
16.1.2 Flexural Mode 391
16.1.3 Combined Shear and Flexural Modes 391
16.2 Overall Buckling Analysis of Wall-Frames 392
16.2.1 Analytical Method 392
16.2.2 Example: Stability of Wall-Frame Structure 396
16.3 Second-Order Effects of Gravity Loading 398
16.3.1 The P-Delta Effect 398
16.3.2 Amplification Factor P-Delta Analysis 399
16.3.3 Iterative P-Delta Analysis 401
16.3.4 Iterative Gravity Load P-Delta Analysis 403
16.3.5 Direct P-Delta Analysis 405
16.4 Simultaneous First-Order and P-Delta Analysis 406
16.4.1 Development of the Second-Order Matrix 406
16.4.2 Negative Shear Area Column 408
16.4.3 Negative Flexural Stiffness Column 410
16.5 Translational-Torsional Instability 411
16.6 Out-of-Plumb Effects 414
16.7 Stiffness of Members in Stability Calculations 414
16.8 Effects of Foundation Rotation 415
Summary 416
References 417

17. Dynamic Analysis 419

17 .I Dynamic Response to Wind Loading 420

17.1.1 Sensitivity of Structures to Wind Forces 421
17.1.2 Dynamic Structural Response due to Wind Forces 422
17.1.3 Along-Wind Response 423
17 .1.4 Cross-Wind Response 429
17.1.5 Worked Example 430
17.2 Dynamic Response to Earthquake Motions 431
17.2.1 Response of Tall Buildings to Ground Accelerations 431
17.2.2 Response Spectrum Analysis 435

@Seismicisolation
@Seismicisolation
 CONTENTS XIX

17.2.3 Empirical Relationships for Fundamental Natural

Frequency 449
17.2.4 Structural Damping Ratios 451
17.3 Comfort Criteria: Human Response to Building Motions 452

17.3.1 Human Perception of Building Motion 452

17.3.2 Perception Thresholds 453
17.3.3 Use of Comfort Criteria in Design 457
Summary 458
References 459

18. Creep, Shrinkage, and Temperature Effects 461

18.1 Effects of Differential Movements 461

18.2 Designing for Differential Movement 462
18.3 Creep and Shrinkage Effects 464
18.3.1 Factors Affecting Creep and Shrinkage Movements
in Concrete 464
18.3.2 Determination of Vertical Shortening of Walls and
Columns 468
18.3.3 Influence of Reinforcement on Column Stresses,
Creep, and Shrinkage 471
18.3.4 Worked Example 472
18.3.5 Influence of Vertical Shortening on Structural
Actions in Horizontal Members 474
18.4 Temperature Effects 475
Summary 478
References 478

APPENDIX 1. Formulas and Design Curves for Coupled Shear

Walls 480

Al.l Formulas and Design Curves for Alternative

Load Cases 480
A I. I. I Formulas for Top Concentrated Load
and Triangularly Distributed Loading 480
A1.1.2 Design Curves 482
Al.2 Formulas for Coupled Shear Walls with Different
Flexible Support Conditions 487
Al.3 Stiffness of Floor Slabs Connecting Shear Walls 489
A I. 3.l Effective Width of Floor Slab 493
Al.3.2 Empirical Relationships for Effective
Slab Width 495
Al.3.3 Numerical Examples 497
References 500

@Seismicisolation
@Seismicisolation
XX CONTENTS

APPENDIX 2. Formulas and Graphs for Wall-Frame and Core

Structures 502

A2.1 Formulas and Graphs for Deflections and Forces 502

A2.1.1 Uniformly Distributed Horizontal
Loading 502
A2.1.2 Triangularly Distributed Horizontal
Loading 506
A2.1.3 Concentrated Horizontal Load at the
Top 506

Bibliography 512

Index 527

@Seismicisolation
@Seismicisolation
 CHAPTER 1

Tall Buildings

This book is concerned with tall building structures. Tallness, however, is a rel­
ative matter, and tall buildings cannot be defined in specific terms related just to
height or to the number of floors. The tallness of a building is a matter of a person's
or community's circumstance and their consequent perception; therefore, a mea­
surable definition of a tall building cannot be universally applied. From the struc­
tural engineer's point of view, however, a tall building may be defined as one that,
because of its height, is affected by lateral forces due to wind or earthquake actions
to an extent that they play an important role in the structural design. The influence
of these actions must therefore be considered from the very beginning of the design
process.

1.1 WHY TALL BUILDINGS?

Tall towers and buildings have fascinated mankind from the beginning of civili­
zation, their construction being initially for defense and subsequently for ecclesi­
astical purposes. The growth in modem tall building construction, however, which
began in the 1880s, has been largely for commercial and residential purposes.
Tall commercial buildings are primarily a response to the demand by business
activities to be as close to each other, and to the city center, as possible, thereby
putting intense pressure on the available land space. Also, because they form dis­
tinctive landmarks, tall commercial buildings are frequently developed in city cen­
ters as prestige symbols for corporate organizations. Further, the business and
tourist community, with its increasing mobility, has fuelled a need for more, fre­
quently high-rise, city center hotel accommodations.
The rapid growth of the urban population and the consequent pressure on lim­
ited space have considerably influenced city residential development. The high
cost of land, the desire to avoid a continuous urban sprawl, and the need to pre­
serve important agricultural production have all contributed to drive residential
buildings upward. In some cities, for example, Hong Kong and Rio de Janeiro,
local topographical restrictions make tall buildings the only feasible solution for
housing needs.

@Seismicisolation
@Seismicisolation 1
2 TALL BUILDINGS

1.2 FACTORS AFFECTING GROWTH, HEIGHT, AND

STRUCTURAL FORM

The feasibility and desirability of high-rise structures have always depended on

the available materials, the level of construction technology, and the state of de­
velopment of the services necessary for the use of the building. As a result, sig­
nificant advances have occurred from time to time with the advent of a new ma­
terial, construction facility, or form of service.
Multistory buildings were a feature of ancient Rome: four-story wooden tene­
ment buildings, of post and lintel construction, were common. Those built after
the great fire of Nero, however, used the new brick and concrete materials in the
form of arch and barrel vault structures. Through the following centuries, the two
basic construction materials were timber and masonry. The former lacked strength
for buildings of more than about five stories, and always presented a fire hazard.
The latter had high compressive strength and fire resistance, but its weight tended
to overload the lower supports. With the rapidly increasing number of masonry
high-rise buildings in North America toward the end of the nineteenth century, the
limits of this form of construction became apparent in 1891 in the 16-story Mon­
adnock Building in Chicago. With the space in its lower floors largely occupied
by walls of over 2 m thick, it was the last tall building in the city for which massive
load-bearing masonry walls were employed.
The socioeconomic problems that followed industrialization in the nineteenth
century, coupled with an increasing demand for space in the growing U.S. cities,
created a strong impetus to tall building construction. Yet the ensuing growth could
not have been sustained without two major technical innovations that occurred in
the middle of that century: the development of higher strength and structurally
more efficient materials, wrought iron and subsequently steel, and the introduction
of the elevator (cf. Fig. 1.1). Although the elevator had been developed some 20
years earlier, its potential in high-rise buildings was apparently not realized until
its incorporation in the Equitable Life Insurance Building in New York in 1870.
For the first time, this made the upper stories as attractive to rent as the lower
ones, and, consequently, made the taller building financially viable.
The new materials allowed the development of lightweight skeletal structures,
permitting buildings of greater height and with larger interior open spaces and
windows, although the early wrought-iron frame structures still employed load­
bearing masonry facade walls. The first high-rise building totally supported by a
metal frame was the 11-story Home Insurance Building in Chicago in 1883, fol­
lowed in 1889 by the first all-steel frame in the 9-story Rand-McNally Building.
Two years later, in the same city, diagonal bracings were introduced in the facade
frames of the 20-story Masonic Temple to form vertical trusses, the forerunner of
modem shear wall and braced frame construction. It was by then appreciated that
at that height wind forces were an important design consideration. Improved de­
sign methods and construction techniques allowed the maximum height of steel­
frame structures to increase steadily, reaching a height of 60 stories with the con­
struction of the Woolworth Building in New York in 1913. This golden age of

@Seismicisolation
@Seismicisolation
 1.2 FACTORS AFFECTING GROWTH, HEIGHT, AND STRUCTURAL FORM 3

100

"'
"
80
:J
..c
First wrought iron j First steel I
-<>
"'
rolled sections ! rolled sections !
�
Ql
60 I
"' I New York
-<>

" I I era

"' I I
Ql

...
40
Cast iron era I Chicago I
New York Schoo 1
I
0

"'
-<>

.....
0 I
20 Otis' elevator
0
z

1850 1870 1890 1910 1930

Fig. 1.1 Growth in height of the first great era of American skyscrapers.

American skyscraper construction culminated in 1931 in its crowning glory, the

Empire State Building, whose 102-story braced steel frame reached a height of
1250 ft (381 m).
Although reinforced concrete construction began around the tum of the century,
it does not appear to have been used for multistory buildings until after the end of
World War I. The inherent advantages of the composite material, which could be
readily formed to simultaneously satisfy both aesthetic and load-carrying require­
ments, were not then fully appreciated, and the early systems were purely imita­
tions of their steel counterparts. Progress in reinforced concrete was slow and
intermittent, and, at the time the steel-framed Empire State Building was com­
pleted, the tallest concrete building, the Exchange Building in Seattle, had attained
a height of only 23 stories.
The economic depression of the 1930s put an end to the great skyscraper era,
and it was not until some years after the end of World War II that the construction
of high-rise buildings recommenced, with radically new structural and architec­
tural solutions. Rather than bringing significant increases in height, however, these
modem developments comprised new structural systems, improved material qual­
ities and services, and better design and construction techniques. It was not until
1973 that the Empire State Building was eclipsed in height by the twin towers of
the 110-story, 1350 ft (412 m) high World Trade Center in New York, using
framed-tube construction, which was followed in 1974 by the 1450 ft (442 m)
high bundled-tube Sears Tower in Chicago.
Different structural systems have gradually evolved for residential and office
buildings, reflecting their differing functional requirements. In modem office
buildings, the need to satisfy the differing requirements of individual clients for

@Seismicisolation
@Seismicisolation
4 TALL BUILDINGS

floor space arrangements led to the provision of large column-free open areas to
allow flexibility in planning. Improved levels of services have frequently neces­
sitated the devotion of entire floors to mechanical plant, but the spaces lost can
often be utilized also to accommodate deep girders or trusses connecting the ex­
terior and interior structural systems. The earlier heavy internal partitions and ma­
sonry cladding, with their contributions to the reserve of stiffness and strength,
have largely given way to light demountable partitions and glass curtain walls,
forcing the basic structure alone to provide the required strength and stiffness
against both vertical and lateral loads.
Other architectural features of commercial buildings that have influenced struc­
tural form are the large entrances and open lobby areas at ground level, the mul­
tistory atriums, and the high-level restaurants and viewing galleries that may re­
quire more extensive elevator systems and associated sky lobbies.
A residential building's basic functional requirement is the provision of self­
contained individual dwelling units, separated by substantial partitions that provide
adequate fire and acoustic insulation. Because the partitions are repeated from story
to story, modem designs have utilized them in a structural capacity, leading to the
shear wall, cross wall, or infilled-frame forms of construction.
The trends to exposed structure and architectural cutouts, and the provision of
setbacks at the upper levels to meet daylight requirements, have also been features
of modem architecture. The requirement to provide adequately stiff and strong
structures, while accommodating these various features, led to radical develop­
ments in structural framing, and inspired the new generation of braced frames,
framed-tube and hull-core structures, wall-frame systems, and outrigger-braced
structures described in Chapter 4. The latest generation of "postmodern" build­
ings, with their even more varied and irregular external architectural treatment,
has led to hybrid double and sometimes triple combinations of the structural
monoforms used for modern buildings.
Speed of erection is a vital factor in obtaining a return on the investment in­
volved in such large-scale projects. Most tall buildings are constructed in con­
gested city sites, with difficult access; therefore careful planning and organization
of the construction sequence become essential. The story-to-story uniformity of
most multistory buildings encourages construction through repetitive operations
and prefabri�ation techniques. Progress in the ability to build tall has gone hand
in hand with the development of more efficient equipment and improved methods
of construction, such as slip- and flying-formwork, concrete pumping, and the use
of tower, climbing, and large mobile cranes.

1.3 THE TALL BUILDING STRUCTURE

·Ideally, in the early stages of planning a building, the entire design team, including
the architect, structural engineer, and services engineer, should collaborate to agree
on a form of structure to satisfy their respective requirements of function, safety
and serviceability, and servicing. A compromise between conflicting demands will

@Seismicisolation
@Seismicisolation
 1.3 THE TALL BUILDING STRUCTURE 5

be almost inevitable. In all but the very tallest structures, however, the structural
arrangement will be subservient to the architectural requirements of space arrange­
ment and aesthetics. Often, this will lead to a less-than-ideal structural solution
that will tax the ingenuity, and probably the patience, of the structural engineer.
The two primary types of vertical load-resisting elements of tall buildings are
columns and walls, the latter acting either independently as shear walls or in as­
semblies as shear wall cores. The building function will lead naturally to the pro­
vision of walls to divide and enclose space, and of cores to contain and convey
services such as elevators. Columns will be provided, in otherwise unsupported
regions, to transmit gravity loads and, in some types of structure, horizontal loads
also. Columns may also serve architecturally as, for example, facade mullions.
The inevitable primary function of the structural elements is to resist the gravity
loading from the weight of the building and its contents. Since the loading on
different floors tends to be similar, the weight of the floor system per unit floor
area is approximately constant, regardless of the building height. Because the grav­
ity load on the columns increases down the height of a building, the weight of
columns per unit area increases approximately linearly with the building heighL
The highly probable second function of the vertical structural elements is to
resist also the parasitic load caused by wind and possibly earthquakes, whose mag­
nitudes will be obtained from National Building Codes or wind tunnel studies. The
bending moments on the building caused by these lateral forces increase with at
least the square of the height, and their effects will become progressively more
important as the building height increases. On the basis of the factors above, the
relative quantities of material required in the floors, columns, and wind bracing of
a traditional steel frame and the penalty on these due to increasing height are ap­
proximately as illustrated in Fig. 4.1.
Because the worst possible effects of lateral forces occur rarely, if ever, in the
life of the building, it is imperative to minimize the penalty for height to achieve
an optimum design. The constant search for more efficient solutions led to the
innovative designs and new structural forms of recent years (cf. Chapter 4). In
developing a suitable system for resisting lateral forces, the engineer seeks to de­
vise stiff horizontal interconnections between the various vertical components to
form composite assemblies such as coupled walls and rigid frames, which, as dem­
onstrated in later chapters, create a total structural assembly having a lateral stiff­
ness many times greater than the sum of the lateral stiffnesses of the individual
vertical components.

1.3.1. The Design Process

·Once the functional layout of the structure has been decided, the design process
generally follows a well-defined iterative procedure. Preliminary calculations for
member sizes are usually based on gravity loading augmented by an arbitrary in­
crement to account for wind forces. The cross-sectional areas of the vertical mem­
bers will be based on the accumulated loadings from their associated tributary
areas, with reductions to account for the probability that not all floors will be

@Seismicisolation
@Seismicisolation
6 TALL BUILDINGS

subjected simultaneously to their maximum live loading. The initial sizes of beams
and slabs are normally based on moments and shears obtained from some simple
method of gravity load analysis, such as two-cycle moment distibution, or from
codified mid- and end-span values.
A check is then made on the maximum horizontal deflection, and the forces in
the major structural members, using some rapid approximate analysis technique.
If the deflection is excessive, or some of the members are inadequate, adjustments
are made to the member sizes or the structural arrangement. If certain members
attract excessive loads, the engineer may reduce their stiffness to redistribute the
load to less heavily stressed components. The procedure of preliminary analysis,
checking, and adjustment is repeated until a satisfactory solution is obtained.
Invariably, alterations to the initial layout of the building will be required as
the client's and architect's ideas of the building evolve. This will call for structural
modifications, or perhaps a radical rearrangement, which necessitates a complete
review of the structural design. The various preliminary stages may therefore have
to be repeated a number of times before a final solution is reached.
A rigorous final analysis, using a more refined analytical model, will then be
made to provide a final check on deflections and member strengths. This will usu­
ally include the second-order effects of gravity loads on the lateral deflections and
member forces (P-Delta effects). A dyn::mic analysis may also be required if, as
a result of wind loading, there is any likelihood of excessive deflections due to
oscillations or of comfort criteria being exceeded, or if earthquake loading has to
be considered. At some stage in the procedure the deleterious effects of differential
movements due to creep, shrinkage, or temperature differentials will also be
checked.
In the design process, a thorough knowledge of high-rise structural components
and their modes of behavior is a prerequisite to devising an appropriate load-re­
sisting system. Such a system must be efficient, economic, and should minimize
the structural penalty for height while maximizing the satisfaction of the basic
serviceability requirements. With the increasing availability of general-purpose
structural analysis programs, the formation of a concise and properly representa­
tive model has become an important part of tall building analysis; this also requires
a fundamental knowledge of structural behavior. Modeling for analysis is dis­
cussed in Chapter 5.

1.4 PHILOSOPHY, SCOPE, AND CONTENT

The iterative design process described above involves different levels of structural
analysis, ranging from relatively crude and approximate techniques for the prelim­
inary stages to sophisticated and accurate methods for the final check. The major
part of this book is devoted therefore to a discussion and comparison of the dif­
ferent practical methods of analysis developed for the range of structural forms
encountered in tall buildings. The emphasis throughout is on methods particular
to tall building structures, with less importance placed on methods for general

@Seismicisolation
@Seismicisolation
 1.5 RAISONS D'ETRE 7

structural analysis, which are treated comprehensively in other texts. It is thus

assumed that the reader is already familiar with the fundamentals of the stiffness
matrix method and the finite element method of analysis.
The methods of analysis presented are, almost without exception, static, and
assume linear elastic behavior of the structure. Although wind and earthquake
forces are transient in nature, it is reasonable and practical to represent them in the
majority of design situations by equivalent static force distributions, as described
in Chapter 3. Although recognizing that concrete and masonry behave in a nonlin­
ear manner, a linear elastic analysis is still the most important tool for deciding a
tall building's structural design. Techniques do exist for the prediction of inelastic
behavior, but they are not yet sufficiently well developed to be appropriate for
undertaking a detailed analysis of a highly indeterminate tall building structure.
The main emphasis of static linear analysis is applied to both components and
assemblies found in tall buildings, ranging from the primary rigid frames, braced
and infilled frames, and shear walls, to the more efficient composite systems that
include coupled shear walls, wall-frame and framed-tube structures, shear wall
cores, and outrigger-braced structures.
Methods suitable for both preliminary and final analyses are described and,
where appropriate, detailed worked examples are given to illustrate the steps in­
volved. Although computer-based matrix techniques form the most versatile and
accurate methods for practical structural analysis, attention is also devoted to the
more limited and approximate continuum techniques. These serve well to provide
an understanding of structural behavior and their generalized solutions indicate
more clearly and rapidly the influence of changes in structural parameters. Such
an understanding can be valuable in selecting a suitable model for computer anal­
ysis. The book concludes with a series of Appendices that include useful design
formulas and charts, and a selective Bibliography of significant references to the
subject matter of the various chapters.
It is impracticable to deal comprehensively in a single volume with all aspects
of tall building structures. Important associated topics, therefore, including foun­
dation systems, the detailed treatment of wind and earthquake forces and the as­
sociated dynamic structural analysis, and construction procedures, which form ma­
jor subjects in their own right, have had to be omitted. For a general discussion
on all aspects of tall buildings, architectural, social, and technical, the reader is
referred to the Reports and Proceedings of the Council on Tall Buildings and Urban
Habitat, particularly the five-volume series of definitive Monographs [I. 1].

1.5 RAISONS D'ETRE

The authors believe that a book devoted to the analysis and design of tall building
structures is merited on a number of counts. During the last few decades a large
body of knowledge on the subject has accrued from an intensive worldwide re­
search effort. The pace of this research has now abated, but the results are widely
dispersed and still generally available only in research journals. Many of the anal-

@Seismicisolation
@Seismicisolation
8 TALL BUILDINGS

ysis techniques that have been developed are virtually unique to tall buildings, and
they form the foundations of an academic discipline that has required the research
results to be digested, consolidated, and recorded in a coherent and unified form.
Meanwhile high-rise construction continues apace, and there is a continuous de­
mand for information from engineers involved in high-rise design, while structural
engineering graduate students are enrolled in courses and conducting further re­
search on tall building structures. This text is aimed to be of value to both the
design office and those in the classroom or laboratory.
The object of the book is therefore to offer a coherent and unified treatment of
the subject analysis and design of high-rise building structures, for practicing struc­
tural engineers concerned with the design of tall buildings, and for senior under­
graduate and postgraduate structural engineering students.

REFERENCE

1.1 Monograph 011 Planning and Design of Tall Buildings, Vols. CB, CL, PC, SB, and
SC, ASCE, 1980.

@Seismicisolation
@Seismicisolation
 CHAPTER 2

Design Criteria

Tall buildings are designed primarily to serve the needs of an intended occupancy,
whether residential, commercial, or, in some cases, a combination of the two. The
dominant design requirement is therefore the provision of an appropriate internal
layout for the building. At the same time, it is essential for the architect to satisfy
the client's expectations concerning the aesthetic qualities of the building's exte­
rior. The main design criteria are, therefore, architectural, and it is within these
that the engineer is usually constrained to fit his structure. Only in exceptionally
tall buildings will structural requirements become a predominant consideration.
The basic layout will be contained within a structural mesh that must be mini­
mally obtrusive to the functional requirements of the building. Simultaneously,
there must be an integration of the building structure with the various service sys­
tems-heating, ventilating, air-conditioning, water supply and waste disposal,
electrical supply, and vertical transportation-which are extensive and complex,
and constitute a major part of the cost of a tall building.
Once the functional layout has been established, the engineer must develop a
structural system that will satisfy established design criteria as efficiently and eco­
nomically as possible, while fitting into the architectural layout. The vital struc­
tural criteria are an adequate reserve of strength against failure, adequate lateral
stiffness, and an efficient performance during the service life of the building.
This chapter provides a brief description of the important criteria that must be
considered in the structural design of a tall building. Most of the principles of
structural design apply equally to low-rise as to high-rise buildings, and therefore,
for brevity, special attention is devoted to only those aspects that have particular
consequences for the designers of high-rise buildings.

2.1 DESIGN PHILOSOPHY

Chapter 1 described how radical changes in the structural form of tall buildings
occurred in the construction period that followed World War II. Over the same
period, a major shift occurred in design philosophy, and the Code formats have
progressed from the earlier working stress or ultimate strength deterministic bases
to modem more generally accepted probability-based approaches. The probabilis­

@Seismicisolation
tic approach for both structural properties and loading conditions has led to the

@Seismicisolation 9
10 DESIGN CRITERIA

limit states design philosophy, which is now almost universally accepted. The aim
of this approach is to ensure that all structures and their constituent components
are designed to resist with reasonable safety the worst loads and deformations that
are liable to occur during construction and service, and to have adequate durability
during their lifetime.
The entire structure, or any part of it, is considered as having "failed" when
it reaches any one of various ''limit states,'' when it no longer meets the prescribed
limiting design conditions. Two fundamental types of limit state must be consid­
ered: (l) the ultimate limit states corresponding to the loads to cause failure, in­
cluding instability: since events associated with collapse would be catastrophic,
endangering lives and causing serious financial losses, the probability of failure
must be very low; and (2) the serviceability limit states, which involve the criteria
governing the service life of the building, and which, because the consequences
of their failure would not be catastrophic, are permitted a much higher probability
of occurrence. These are concerned with the fitness of the building for normal use
rather than safety, and are of less critical importance.
A particular limit state may be reached as a result of an adverse combination of
random effects. Partial safety factors are employed for different conditions that
reflect the probability of certain occurrences or circumstances of the structure and
loading existing. The implicit objective of the design calculations is then to ensure
that the probability of any particular limit state being reached is maintained below
an acceptable value for the type of structure concerned.
The following sections consider the criteria that apply in particular to the design
of tall buildings.

2.2 LOADING

The structure must be designed to resist the gravitational and lateral forces, both
permanent and transient, that it will be called on to sustain during its construction
and subsequent service life. These forces will depend on the size and shape of the
building, as well as on its geographic location, and maximum probable values must
be established before the design can proceed.
The probable accuracy of estimating the dead and live loads, and the probability
of the simultaneous occurrence of different combinations of gravity loading, both
dead and live, with either wind or earthquake forces, is included in limit states
design through the use of prescribed factors.
The load systems that must be considered are described in Chapter 3.

2.2.1 Sequential Loading

For loads that are applied after completion of the building, such as live, wind, or
seismic loading, the analysis is independent of the construction sequence. For dead
loads, however, which are applied to the building frame as construction proceeds,
the effects of sequential loading should be considered to assess the worst conditions
to which any component may be subjected, and also to determine the true behavior
of the frame.
@Seismicisolation
@Seismicisolation
 2.4 STIFFNESS AND DRIFT LIMITATIONS 11

In multistory reinforced concrete construction, the usual practice is to shore the

freshly placed floor on several previously cast floors. The construction loads in the
supporting floors due to the weight of the wet concrete and formwork may appre­
ciably exceed the loads under service conditions. Such loads depend on the se­
quence and rate of erection.
If column axial deformations are calculated as though the dead loads are applied
to the completed structure, bending moments in the horizontal components will
result from any differential column shortening that is shown to result. Because of
the cumulative effects over the height of the building, the effects are greater in the
highest levels of the building. However, the effects of such differential movements
would be greatly overestimated because in reality, during the construction se­
quence, a particular horizontal member is constructed on columns in which the
initial axial deformations due to the dead weight of the structure up to that partic­
ular level have already taken place. The deformations of that particular floor will
then be caused by the loads that are applied subsequent to its construction. Such
sequential effects must be considered if an accurate assessment of the structural
actions due to dead loads is to be achieved.

2.3 STRENGTH AND STABILITY

For the ultimate limit state, the prime design requirement is that the building struc­
ture should have adequate strength to resist, and to remain stable under, the worst
probable load actions that may occur during the lifetime of the building, including
the period of construction.
This requires an analysis of the forces and stresses that will occur in the mem­
bers as a result of the most critical possible load combinations, including the aug­
mented moments that may arise from second-order additional deflections (P-Delta
effects) (cf. Chapter 16). An adequate reserve of strength, using prescribed load
factors, must be present. Particular attention must be paid to critical members,
whose failure could prove catastrophic in initiating a progressive collapse of part
of or the entire building. Any additional stresses caused by restrained differential
movements due to creep, shrinkage, or temperature must be included (cf. Chapter
18).
In addition, a check must be made on the most fundamental condition of equi­
librium, to establish that the applied lateral forces will not cause the entire building
to topple as a rigid body about one edge of the base. Taking moments about that
edge, the resisting moment of the dead weight of the building must be greater than
the overturning moment for stability by an acceptable factor of safety.

2.4 STIFFNESS A'ND DRIFT LIMITATIONS

The provision of adequate stiffness, particularly lateral stiffness, is a major con­

sideration in the design of a tall building for several important reasons. As far as

@Seismicisolation
the ultimate limit state is concerned, lateral deflections must be limited to prevent
@Seismicisolation
12 DESIGN CRITERIA

second-order P-Delta effects due to gravity loading being of such a magnitude as

to precipitate collapse. In terms of the serviceability limit states, deflections must
first be maintained at a sufficiently low level to allow the proper functioning of
nonstructural components such as elevators and doors; second, to avoid distress in
the structure, to prevent excessive cracking and consequent loss of stiffness, and
to avoid any redistribution of load to non-load-bearing partitions, infills, cladding,
or glazing; and third, the structure must be sufficiently stiff to prevent dynamic
motions becoming large enough to cause discomfort to occupants, prevent delicate
work being undertaken, or affect sensitive equipment. In fact, it is in the particular
need for concern for the provision of lateral stiffness that the design of a high-rise
building largely departs from that of a low-rise building.
One simple parameter that affords an estimate of the lateral stiffness of a build­
ing is the drift index, defined as the ratio of the maximum deflection at the top of
the building to the total height. In addition, the corresponding value for a single
story height, the interstory drift index, gives a measure of possible localized ex­
cessive deformation. The control of lateral deflections is of particular importance
for modern buildings in which the traditional reserves of stiffness due to heavy
internal partitions and outer cladding have largely disappeared. It must be stressed,
however, that even if the drift index is kept within traditionally accepted limits,
such as 5bo, it does not necessarily follow that the dynamic comfort criteria will
also be satisfactory. Problems may arise, for example, if there is coupling between
bending and torsional oscillations that leads to unacceptable complex motions or
accelerations. In addition to static deflection calculations, the question of the dy­
namic response, involving the lateral acceleration, amplitude, and period of oscil­
lation, may also have to be considered.
The establishment of a drift index limit is a major design decision, but, unfor­
tunately, there are no unambiguous or widely accepted values, or even, in some
of the National Codes concerned, any firm guidance. The designer is then faced
with having to decide on an appropriate value. The figure adopted will reflect the
building usage, the type of design criterion employed (for example, working or
ultimate load conditions), the form of construction, the materials employed, in­
cluding any substantial infills or claddings, the wind loads considered, and, in
particular, past experience of similar buildings that have performed satisfactorily.
Design drift index limits that have been used in different countries range from
0.001 to 0.005. To put this in perspective, a maximum horizontal top deflection
of between 0.1 and 0.5 m ( 6 to 20 in.) would be allowed in a 33-story, 100-m
(330-ft.) high building, or, alternatively, a relative deflection of 3 to 15 mm ( 0.12
to 0.6 in.) over a story height of 3 m ( 10 ft ). Generally, lower values should be
used for hotels or apartment buildings than for office buildings, since noise and
movement tend to be more disturbing in the former. Consideration may be given
to whether the stiffening effects of any internal partitions, infills, or claddings are
included in the deflection calculations.
The consideration of this limit state requires an accurate estimate of the lateral
deflections that occur, and involves an assessment of the stiffness of cracked mem­
bers, the effects of shrinkage and creep and any redistribution of forces that may

@Seismicisolation
@Seismicisolation
 2.5 HUMAN COMFORT CRITERIA 13

result, and of any rotational foundation movement. In the design process, the stiff­
ness of joints, particularly in precast or prefabricated structures, must be given
special attention to develop adequate lateral stiffness of the structure and to prevent
any possible progressive failure. The possibility of torsional deformations must
not be overlooked.
In practice, non-load-bearing infills, partitions, external wall panels, and win­
dow glazing should be designed with sufficient clearance or with flexible supports
to accommodate the calculated movements.
Sound engineering judgment is required when deciding on the drift index limit
to be imposed. However, for conventional structures, the preferred acceptable range
is 0.0015 to 0.003 (that is, approximately � to 3�0), and sufficient stiffness must
be provided to ensure that the top deflection does not exceed this value under
extreme load conditions. As the height of the building increases, drift index coef­
ficients should be decreased to the lower end of the range to keep the top story
deflection to a suitably low level. Succeeding chapters describe how deflections
may be computed.
The drift criteria apply essentially to quasistatic conditons. When extreme force
conditions are possible, or where problems involving vortex shedding or other
unusual phenomena may occur, a more sophisticated approach involving a dy­
namic analysis may be required.
If excessive, the drift of a structure can be reduced by changing the geometric
configuration to alter the mode of lateral load resistance, increasing the bending
stiffness of the horizontal members, adding additional stiffness by the inclusion of
stiffer wall or core members, achieving stiffer connections, and even by sloping
the exterior columns. In extreme circumstances, it may be necessary to add dam­
pers, which may be of the passive or active type.

2.5 HUMAN COMFORT CRITERIA

If a tall flexible structure is subjected to lateral or torsional deflections under the

action of fluctuating wind loads, the resulting oscillatory movements can induce a
wide range of responses in the building's occupants, ranging from mild discomfort
to acute nausea. Motions that have psychological or physiological effects on the
occupants may thus result in an otherwise acceptable structure becoming an un­
desirable or even unrentable building.
There are as yet no universally accepted international standards for comfort
criteria, although they are under consideration, and engineers must base their de­
sign criteria on an assessment of published data. It is generally agreed that accel­
eration is the predominant parameter in determining human response to vibration,
but other factors such as period, amplitude, body orientation, visual and acoustic
cues, and even past experience can be influential. Threshold curves are available
that give various limits for human behavior, ranging from motion perception
through work difficulty to ambulatory limits, in terms of acceleration and period.

@Seismicisolation
@Seismicisolation
14 DESIGN CRITERIA

A dynamic analysis is then required to allow the predicted response of the building
to be compared with the threshold limits.
The questions of human response to motion, comfort criteria, and their influ­
ence on structural design are considered in Chapter 17.

2.6 CREEP, SHRINKAGE, AND TEMPERATURE EFFECTS

In very tall concrete buildings, the cumulative vertical movements due to creep
and shrinkage may be sufficiently large to cause distress in nonstructural elements,
and to induce significant structural actions in the horizontal elements, especially
in the upper regions of the building. In assessing these long-term deformations,
the influence of a number of significant factors must be considered, particularly
the concrete properties, the loading history and age of the concrete at load appli­
cation, and the volume-surface ratio and amount of reinforcement in the members
concerned. The structural actions in the horizontal elements caused by the resulting
relative vertical deflections of their supports can then be estimated. The differential
movements due to creep and shrinkage must be considered structurally and accom­
modated as far as possible in the architectural details at the design stage. However,
by attempting to achieve a uniformity of stress in the vertical components, it is
possible to reduce as far as possible any relative vertical movement due to creep.
In the construction phase, in addition to creep and shrinkage, elastic shortening
will occur in the vertical elements of the lower levels due to the additional loads
imposed by the upper stories as they are completed. Any cumulative differential
movements will affect the stresses in the subsequent structure, especially in build­
ings that include both in situ and precast components.
In buildings with partially or fully exposed exterior columns, significant tem­
perature differences may occur between exterior and interior columns, and any
restraint to their relative deformations will induce stresses in the members con­
cerned. The analysis of such actions requires a knowledge of the differential tem­
peratures that are likely to occur between the building and its exterior and the
temperature gradient through the members. This will allow an evaluation of the
free thermal length changes that would occur if no restraint existed, and, hence,
using a standard elastic analysis, the resulting thermal stresses and deformations
may be determined.
Practical methods for analyzing the effects of creep, shrinkage, and temperature
are discussed in Chapter 18.

2.7 FIRE

The design considerations for fire prevention and protection, smoke control, fire­
fighting, and escape are beyond the scope of a book on building structures. How­
ever, since fire appears to be by far the most common extreme situation that will

@Seismicisolation
@Seismicisolation
 2.8 FOUNDATION SETTLEMENT AND SOIL-STRUCTURE INTERACTION 15

cause damage in structures, it must be a primary consideration in the design pro­

cess.
The characteristic feature of a fire, such as the temperature and duration, can
be estimated from a knowledge of the important parameters involved, particularly
the quantity and nature of combustible material present, the possibility and extent
of ventilation, and the geometric and thermal properties of the fire compartment
involved. Once the temperatures at the various surfaces have been determined,
from the gas temperature curve, it is possible to estimate the heat flow through the
insulation and structural members. A knowledge of the temperature gradient across
the member, and the degree of restraint afforded by the supports and surrounding
structure, enables the stresses in the member to be evaluated. The mechanical
properties of the structural materials, particularly the elastic modulus or stiffness
and strength, may deteriorate rapidly as the temperature rises, and the resistance
to loads is greatly reduced. For example the yield stress of mild steel at a temper­
ature of 700°C is only some 10-20% of its value at room temperature. Over the
same temperature range, the elastic modulus drops by around 40-50%. The critical
temperature at which large deflections or collapse occurs will thus depend on the
materials used, the nature of the structure, and the loading conditions.
The parameters that govern the approach are stochastic in nature, and the results
of any calculation can be given only in probabilistic terms. The aim should be to
achieve a homogeneous design in which the risks due to the different extreme
situations are comparable.
Designing against fire is, however, a specialist discipline, and the interested
reader is referred to the Monograph on Tall Buildings (Vol. CL) [2.1].

2.8 FOUNDATION SETTLEMENT AND SOIL-STRUCTURE

INTERACTION

The gravity and lateral forces on the building will be transmitted to the earth through
the foundation system, and, as the principles of foundation design are not affected
by the quality of tallness of the superstructure, conventional approaches will suf­
fice. The concern of the structural designer is then with the influence of any foun­
dation deformation on the building's structural behavior and on the soil-structure
interactive forces.
Because of its height, the loads transmitted by the columns in a tall building
can be very heavy. Where the underlying soil is rock or other strong stable
subgrade, foundations may be carried down to the stiff load-bearing layers by use
of piles, caissons, or deep basements. Problems are not generally encountered with
such conditions since large variations in column loadings and spacings can be
accommodated with negligible differential settlement. In areas in which soil con­
ditions are poor, loadings on foundation elements must be limited to prevent shear­
ing failures or excessive differential settlements. Relief may be obtained by ex­
cavating a weight of soil equal to a significant portion of the gross building weight.
Because of the high short-term transient moments and shears that arise from wind

@Seismicisolation
@Seismicisolation
16 DESIGN CRITERIA

loads, particular attention must be given to the design of the foundation system for
resisting moments and shears, especially if the precompression due to the dead
weight of the building is not sufficient to overcome the highest tensile stresses
caused by wind moments, leading to uplift on the foundation.
The major influences of foundation deformations are twofold. First, if the bases
of vertical elements yield, a stress redistribution will occur, and the extra loads
imposed on other elements may then further increase the deformation there. The
influence of the relative displacements on the forces in the horizontal elements
must then be assessed. Second, if an overall rotational settlement 8 of the entire
foundation occurs, the ensuing lateral deflections will be magnified by the height
H to give a top deflection of H8. As well as increasing the maximum drift, the
movement will have a destabilising effect on the structure as a whole, by increasing
any P-Delta effects that occur (cf. Chapter 16).
Soil-structure interaction involves both static and dynamic behavior. The for­
mer is generally treated by simplified models of subgrade behavior, and finite ele­
ment methods of analysis are usual. When considering dynamic effects, both in­
teractions between soil and structure, and any amplification caused by a coincidence
of the natural frequencies of building and foundation, must be included. Severe
permanent structural damage may be caused by earthquakes when large deforma­
tions occur due to the soil being compacted by the ground vibration, which under
certain conditions may result in the development of excess hydrostatic pressures
sufficient to produce liquefaction of the soil. These types of soil instability may be
prevented or reduced in intensity by appropriate soil investigation and foundation
design. On the other hand, the dynamic response of buildings to ground vibrations,
which is also affected by soil conditions, cannot be avoided and must be considered
in design.
A general discussion of all aspects of the design of foundations for tall buildings
is given in Reference 2.2.

SUMMARY

Probability-based limit states concepts form the basis of modem structural design
codes. This chapter summarizes the most important limit states involved in the
design of tall building structures. Ultimate limit states are concerned with the max­
imum load and carrying capacity of the structure, where the probability of failure
must be very low, whereas serviceability limit states are concerned with actions
that occur during the service life of the structure, and are permitted to have a much
higher probability of occurrence.
The most important ultimate limit state requirement is that the structure should
have adequate strength and remain stable under all probable load combinations
that may occur during the construction and subsequent life of the building. When
assessing stability, any second-order P-Delta effects in heavily loaded slender
members must be considered. Any stresses induced by relative movements caused
by creep, shrinkage, and temperature differentials must be included.

@Seismicisolation
@Seismicisolation
 REFERENCES 17

One major serviceability limit state criterion lies in the provision of adequate
stiffness, particularly lateral stiffness, to avoid excessive cracking in concrete and
to avoid any load transfer to non-load-bearing components, to avoid excessive
secondary P-Delta moments caused by lateral deflections, and to prevent any dy­
namic motions that would affect the comfort of the occupants. One measure of the
stiffness is the drift of the structure and this should be limited to the range of
0.0015 to 0.003 of the total height. Similar limits should be imposed on the ac­
ceptable interstory drift index.
The stresses and loss of stiffness that might result from a building fire must be
a major consideration, as this is not a remote possibility. However, designing
against fire is a specialist discipline that cannot be covered in any detail here.
Although the principles of foundation design are not affected by the height of a
building, the situation for tall buildings is different as a result of the high short­
term transient moments and shears that arise from wind loads. The high dead load
caused by the height of the building produces large compressive stresses on the
foundation, and excessive differential settlements must be avoided. Any lateral
deflections caused by rotational settlement will be magnified by the height of the
building, and the soil-structure interaction must be considered, particularly under
seismic actions.

REFERENCES

2.1 Tall Building Criteria and Loading. Vol. CL. Monograph on Planning and Design of
Tall Buildings, ASCE. 1980, pp. 251-390.
2.2 Tall Building Systems and Concepts. Vol. SC. Monograph on Planning and Design
of Tall Buildings, ASCE, 1980, pp. 259-340.

@Seismicisolation
@Seismicisolation
 CHAPTER 3

Loading

Loading on tall buildings differs from loading on low-rise buildings in its accu­
mulation into much larger structural forces, in the increased significance of wind
loading, and in the greater importance of dynamic effects. The collection of gravity
loading over a large number of stories in a tall building can produce column loads
of an order higher than those in low-rise buildings. Wind loading on a tall building
acts not only over a very large building surface, but also with greater intensity at
the greater heights and with a larger moment arm about the base than on a low­
rise building. Although wind loading on a low-rise building usually has an insig­
nificant influence on the design of the structure, wind on a high-rise building can
have a dominant influence on its structural arrangement and design. In an extreme
case of a very slender or flexible structure, the motion of the building in the wind
may have to be considered in assessing the loading applied by the wind.
In earthquake regions, any inertial loads from the shaking of the ground may
well exceed the loading due to wind and, therefore, be dominant in influencing
the building's structural form, design, and cost. As an inertial problem, the build­
ing's dynamic response plays a large part in influencing, and in estimating, the
effective loading on the structure.
With the exception of dead loading, the loads on a building cannot be assessed
accurately. While maximum gravity live loads can be anticipated approximately
from previous field observations, wind and earthquake loadings are random in
nature, more difficult to measure from past events, and even more difficult to pre­
dict with confidence. The application of probabilistic theory has helped to ration­
alize, if not in every case to simplify, the approaches to estimating wind and earth­
quake loading.
It is difficult to discuss approaches to the estimation of loading entirely in gen­
eralities because the variety of methods in the different Codes of Practice, although
rationally based, tend to be empirical in their presentation. Therefore, in some
parts of this chapter, methods from reasonably representative modem Codes are
given in detail to illustrate current philosophies and trends.

3.1 GRAVITY LOADING

Although the tributary areas, and therefore the gravity loading, supported by the
beams and slabs in a tall building do not differ from those in a low-rise building,

18 @Seismicisolation
@Seismicisolation
 3.1 GRAVITY LOADING 19

the accumulation in the former of many stories of loading by the columns and
walls can be very much greater.
As in a low-rise building, dead loading is calculated from the designed member
sizes and estimated material densities. This is prone to minor inaccuracies such as
differences between the real and the designed sizes, and between the actual and
the assumed densities.
Live loading is specified as the intensity of a uniformly distributed floor load,
according to the occupancy or use of the space. In certain situations such as in
parking areas, offices, and plant rooms, the floors should be considered for the
alternative worst possibility of specified concentrated loads.
The magnitudes of live loading specified in the Codes are estimates based on a
combination of experience and the results of typical field surveys. The differences
between the live load magnitudes in the Codes of different countries (some ex­
amples of which are shown in Table 3.I [3.I]) indicate a lack of unanimity and
consistency sufficient to raise questions about their accuracy. Load capacity ex­
periments have shown that even the Code values, which are usually accepted as
conservative, may in some circumstances underestimate the maximum possible
values.
Pattern distribution of gravity live loading over adjacent and alternate spans
should be considered in estimating the local maxima for member forces, while live
load reductions may be allowed to account for the improbability of total loading
being applied simultaneously over larger areas.

3.1.1 Methods of Live Load Reduction

The philosophy of live load reduction is that although, at some time in the life of
a structure, it is probable that a small area may be subjected to the full intensity

TABLE 3.1 Live Load Magnitudes

United Great U.S.S.R.

States Britain Japan (SN and
(ANSI (CP3-CH.V (AIJ [Link]-
A58.1-1972) PT.l: 1967) Standard) 62)

kPa psf kPa psf kPa psf kPa psf

Office buildings
Offices 2.4 50 2.5 52 2.9 61 2.0 41
Corridors 3.8 80 2.5 52 2.9 61
Lobbies 4.8 100 2.5 52 2.9 61 2.9 61

Residential
Apartments 1.9 40 1.5 31 1.8 37 1.5 31
Hotel 1.9 40 2.0 42 1.8 37 2.0 41
Corridors 3.8 80 1.8 37 2.9 61
Public rooms 4.8 100 2.0 42 3.5 74 2.0 41

From Ref. [3.1).

"Same values as for occupancy.

@Seismicisolation
@Seismicisolation
20 LOADING

of live load, it is improbable that the whole of a large area or a collection of areas,
and the members supporting them, will be subjected simultaneously to the full live
load. Consequently, it is reasonable to design the girders and columns supporting
a large tributary area for significantly less than the full live loading. The different
methods of live load reduction generally allow for the girders, columns, and walls
to be designed for a reduced proportion of the full live load with an increased
amount of supported area. An upper limit is usually placed on the reduction in
order to retain an adequate margin of safety.
The following three examples of methods of live load reduction serve to illus­
trate how the general philosophy may be applied [3.1].

I. Simple percentages may be specified for the reductions and for the limiting
amount. For example, the supporting members may be designed for 100%
of the live load on the roof, 85% of that on the top floor, and further reduc­
tions of 5% for each successive floor down to a minimum of 50% of the
live load.
2. A tributary area formula may be given, allowing a more refined definition
of the reduction, with the limit built into the formula. For example, the
supporting members may be designed for a live load equal to the basic live
load multiplied by a factor 0.3 + 10/ .JA, where A is the accumulated area
in square feet.
3. An even more sophisticated formula-type method may define the maximum
reduction in terms of the dead-to-live load ratio. For example, it may be
specified that the maximum percentage reduction shall not exceed [ I00 X
(D + L)]/4. 33L, in which D and L are the intensities of dead and live
loading, respectively. This particular limit is intended to ensure that if the
full live load should occur over the full tributary area, the element would
not be stressed to the yield point.

3.1.2 Impact Gravity Loading

Impact loading occurs as a gravity live load in the case of an elevator being ac­
celerated upward or brought to a rest on its way down. An increase of 100% of
the static elevator load has usually been used to give a satisfactory performance of
the supporting structure [3.I].

3.1.3 Construction Loads

Construction loads are often claimed to be the most severe loads that a building
has to withstand. Certainly, many more failures occur in buildings under construc­
tion than in those that are complete, but it is rare for special provision to be made
for construction loads in tall building design. If, however. in a building with an
unusual structure, a lack of consideration for construction loading could increase
the total cost of the project, an early liaison between the designer and contractor
on making some provision would obviously be desirable.

@Seismicisolation
@Seismicisolation
 3.2 WIND LOADING 21

Typically, the construction load that has to be supported is the weight of the
floor forms and a newly placed slab, which, in total, may equal twice the floor
dead load. This load is supported by props that transfer it to the three or four
previously constructed floors below. Now, with the possibility of as little as 3-day
cycle, or even 2-day cycle, story construction, and especially with concrete pump­
ing, which requires a more liquid mix, the problem is more severe; this is because
the newly released slab, rather than contributing to supporting the construction
loads, is still in need of support itself.
The climbing crane is another common construction load. This is usually sup­
ported by connecting it to a number of floors below with, possibly, additional
shoring in stories further below.

3.2 WIND LOADING

The lateral loading due to wind or earthquake is the major factor that causes the
design of high-rise buildings to differ from those of low- to medium-rise buildings.
For buildings of up to about 10 stories and of typical proportions, the design is
rarely affected by wind loads. Above this height, however, the increase in size of
the structural members, and the possible rearrangement of the structure to account
for wind loading, incurs a cost premium that increases progressively with height.
With innovations in architectural treatment, increases in the strengths of materials,
and advances in methods of analysis, tall building structures have become more
efficient and lighter and, consequently, more prone to deflect and even to sway
under wind loading. This served as a spur to research, which has produced signif­
icant advances in understanding the nature of wind loading and in developing
methods for its estimation. These developments have been mainly in experimental
and theoretical techniques for determining the increase in wind loading due to
gusting and the dynamic interaction of structures with gust forces.
The following review of some representative Code methods, which includes
ones that are relatively advanced in their consideration of gust loading, summarizes
the state of the art. The first method described is a static approach, in that it as­
sumes the building to be a fixed rigid body in the wind. Static methods are appro­
priate for tall buildings of unexceptional height, slenderness, or susceptibility to
vibration in the wind. The subsequently described dynamic methods are for ex­
ceptionally tall, slender, or vibration-prone buildings. These may be defined, for
example, as in the Uniform Building Code [3.2], as those of height greater than
400 ft ( 123 m ), or of a height greater than five times their width, or those with
structures that are sensitive to wind-excited oscillations. Alternatively, such ex­
ceptional buildings may be defined in a more rigorous way according to the natural
frequency and damping of the structure, as well as to its proportions and height
[3.3].
The methods are now explained with a level of detail intended to convey for
each its philosophy of approach. For more detailed information, sufficient to allow
the use of the methods, the reader is referred to the particular Codes of Practice.

@Seismicisolation
@Seismicisolation
22 LOADING

3.2.1 Simple Static Approach

Uniform Building Code (1988) Method [3.2]. The method is representative

of modem static methods of estimating wind loading in that it accounts for the
effects of gusting and for local extreme pressures over the faces of the building. It
also accounts for local differences in exposure between the open countryside and
a city center, as well as allowing for vital facilities such as hospitals, and fire and
police stations, whose safety must be ensured for use after an extreme windstorm.
The design wind pressure is obtained from the formula

( 3.1)

in which Ce is a coefficient to account for the combined effects of height, exposure,

and gusting, as defined in Table 3.2.
Cq is a coefficient that allows for locally higher pressures for wall and roof
elements as compared with average overall pressures used in the design of the
primary structure. For example, Cq has a value of 1.4 when using the projected
area method of calculating wind loading for structures over 40 ft in height, whereas
it has a local value of 2.0 at wall comers.
The pressure qs is a wind stagnation pressure for a minimum basic 50-year wind
speed at a height of 30 ft above ground, as given for different regions of the United
States in a wind speed contour map. Where local records indicate a greater than
basic value of the wind speed, this value should be used instead in determining q,.
The importance factor I is taken as 1.15 for postdisaster buildings and 1.00 for
all other buildings.

TABLE 3.2 Combined Height, Exposure, and Gust Factor

Coefficient (C.)

Height above Average

Level of Adjoining
Ground (ft) Exposure C" Exposure B"

0-20 1.2 0.7

20-40 1.3 0.8
40-60 1.5 1.0
60-100 1.6 1.1
100-150 1.8 1.3
150-200 1.9 1.4
200-300 2.1 1.6
300-400 2.2 1.8

Reproduced from the 1988 edition of the Unifonn Building Code. copyright© 198!!.
with the pennission of the publishers. the International Conference of Building Of­
ficials.

"Exposure C represents the most severe exposure with a flat and open terrain. Ex­
posure B has terrain with buildings. forest. or surface irregularities 20 ft or more in
height.

@Seismicisolation
@Seismicisolation
..

3.2 WIND LOADING 23

3.2.2 Dynamic Methods

If the building is exceptionally slender or tall, or if it is located in extremely severe

exposure conditions, the effective wind loading on the building may be increased
by dynamic interaction between the motion of the building and the gusting of the
wind. If it is possible to allow for it in the budget of the building, the best method
of assessing such dynamic effects is by wind tunnel tests in which the relevant
properties of the building and the surrounding countryside are modeled. For build­
ings that are not so extreme as to demand a wind tunnel test, but for which the
simple design procedure is inadequate, alternative dynamic methods of estimating
the wind loading by calculation have been developed. The wind tunnel experi­
mental method and one of the dynamic calculation methods will be reviewed
briefly.

Wind Tunnel Experimental Method. Wind tunnel tests to determine loading

may be quasisteady for determining the static pressure distribution or force on a
building. The pressure or force coefficients so developed are then used in calcu­
lating the full-scale loading through one of the described simple methods. This
approach is satisfactory for buildings whose motion is negligible and therefore has
little effect on the wind loading.
If the building slenderness or flexibility is such that its response to excitation
by the energy of the gusts may significantly influence the effective wind loading,
the wind tunnel test should be a fully dynamic one. In this case, the elastic struc­
tural properties and the mass distribution of the building as well as the relevant
characteristics of the wind should be modeled.
Building models for wind tunnel tests are constructed to scales which vary from
-dJo to Jboo, depending on the size of the building and the size of the wind tunnel,
with a scale of � being common. Tall buildings typically exhibit a combination
of shear and bending behavior that has a fundamental sway mode comprising a
flexurally shaped lower region and a relatively linear upper region. This can be
represented approximately in wind tunnel tests by a rigid model with a flexurally
sprung base. It is not necessary in such a model to represent the distribution of
mass in the building, but only its moment of inertia about the base.
More complex models are used when additional modes of oscillation are ex­
pected including, possibly, torsion. These models consist of lumped masses,
springs, and flexible rods, designed to simulate the stiffnesses and mass properties
of the prototype. Wind pressure measurements are made by flush surface pressure
taps on the faces of the models, and pressure transducers are used to obtain the
mean, root mean square (RMS), and peak pressures.
The wind characteristics that have to be generated in the wind tunnel are the
vertical profile of the horizontal velocity, the turbulence intensity, and the power
spectral density of the longitudinal component. Special ''boundary layer'' wind
tunnels have been designed to generate these characteristics. Some use long work­
ing sections in which the boundary layer develops naturally over a rough floor;

@Seismicisolation
@Seismicisolation
24 LOADING

other shorter ones include grids, fences, or spires at the test section entrance to­
gether with a rough floor, while some activate the boundary layer by jets or driven
flaps. The working sections of the tunnel are up to a maximum of about 6 ft2 and
they operate at atmospheric pressure [3.4].

Detailed Analytical Method. Wind tunnel testing is a highly specialized,

complex, and expensive procedure, and can be justified only for very high cost
projects. To bridge the gap between those buildings that require only a simple
approach to wind loading and those that clearly demand a wind tunnel dynamic
test, more detailed analytical methods have been developed that allow the dynamic
wind loading to be calculated [3. 5 , 3.6]. The method described here is based on
the pioneering work of Davenport and is now included in the National Building
Code of Canada, NBCC [3.7, 3.8].
The external pressure or suction p on the surface of the building is obtained
using the basic equation

(3.2)

in which the exposure factor Ce is based on a mean wind speed vertical profile,
which varies according to the roughness of the surrounding terrain. Three types
of exposure are considered: generally open terrain with minimal obstruction; semi­
obstructed terrrain such as suburban, urban, and wooded areas, and heavily
obstructed areas with heavy concentrations of tall buildings and at least 50%
of all the buildings exceeding four stories. A formula expressing the value of
Ce as a power of the height is given in the Code for each of the three exposure
conditions.
The gust effect factor Cg is the ratio of the expected peak loading effect to the
mean loading effect. It allows for the variable effectiveness of different sizes of
gusts and for the load magnification effect caused by gusts in resonance with the
vibrating structure. Cg is given in the Code by a series of formulas and graphs
that, although not difficult to use, are too complex to describe here. They can be
summarized briefly, however, as expressing the loading effect in terms of the in­
teraction .between the wind speed spectrum and the fundamental mode dynamic
response of the structure, which involves the natural frequency and damping of
the structure, using a transfer or admittance function.
Coefficient CP is the external pressure coefficient averaged over the area of the
surface considered. Its value is influenced by the shape of the building, the wind
direction, and the profile of the wind velocity, and is usually determined from the
wind tunnel experiments on small-scale models.
Details of the method are given in the National Building Code of Canada and
in its Supplement [3.7, 3.8]. A similar method by Simiu [3.6] is claimed to give
conservative wind loads, but of a significantly lower magnitude than those from
the NBCC method. Obviously scope exists for further verification and, possibly,
simplification of the dynamic load calculation methods.

@Seismicisolation
@Seismicisolation
 3.3 EARTHQUAKE LOADING 25

3.3 EARTHQUAKE LOADING

Earthquake loading consists of the inertial forces of the building mass that result
from the shaking of its foundation by a seismic disturbance. Earthquake resistant
design concentrates particularly on the translational inertia forces, whose effects
on a building are more significant than the vertical or rotational shaking compo­
nents.
Other severe earthquake forces may exist, such as those due to landsliding,
subsidence, active faulting below the foundation, or liquefaction of the local
subgrade as a result of vibration. These disturbances, however, which are local
effects, can be so massive as to defy any economic earthquake-resistant design,
and their possibility may suggest instead the selection of an alternative site.
Where earthquakes occur, their intensity is related inversely to their frequency
of occurrence; severe earthquakes are rare, moderate ones occur more often, and
minor ones are relatively frequent. Although it might be possible to design a build­
ing to resist the most severe earthquake without significant damage, the unlikely
need for such strength in the lifetime of the building would not justify the high
additional cost. Consequently, the general philosophy of earthquake-resistant de­
sign for buildings is based on the principles that they should

I. resist minor earthquakes without damage;

2. resist moderate earthquakes without structural damage but accepting the

probability of nonstructural damage;

3. resist average earthquakes with the probability of structural as well as non­

structural damage, but without collapse.

Some adjustments are made to the above principles to recognize that certain
buildings with a vital function to perform in the event of an earthquake should be
stronger.
The magnitude of earthquake loading is a result of the dynamic response of the
building to the shaking of the ground. To estimate the seismic loading two general
approaches are used, which take into account the properties of the structure and
the past record of earthquakes in the region.
The first approach, termed the equivalent lateral force procedure, uses a simple
estimate of the structure's fundamental period and the anticipated maximum ground
acceleration, or velocity, together with other relevant factors, to determine a max­
imum base shear. Horizontal loading equivalent to this shear is then distributed in
some prescribed manner throughout the height of the building to allow a static
analysis of the structure. The design forces used in this equivalent static analysis
are less than the actual forces imposed on the building by the corresponding earth­
quake. The justification for using lower design forces includes the potential for
greater strength of the structure provided by the working stress levels, the damping
provided by the building components, and the reduction in force due to the effec­
tive ductility of the structure as members yield beyond their elastic limits. The

@Seismicisolation
@Seismicisolation
26 LOADING

method is simple and rapid and is recommended for unexceptionally high buildings
with unexceptional structural arrangements. It is also useful for the preliminary
design of higher buildings and for those of a more unusual structural arrangement,
which may subsequently be analyzed for seismic loading by a more appropriate
method.
The second, more refined, procedure is a modal analysis in which the modal
frequencies of the structure are analyzed and then used in conjunction with earth­
quake design spectra to estimate the maximum modal responses. These are then
combined to find the maximum values of the responses. The procedure is more
complex and longer than the equivalent lateral force procedure, but it is more
accurate as well as being able to account approximately for the nonlinear behavior
of the structure.
The two procedures are now discussed in more detail.

3.3.1 Equivalent Lateral Force Procedure

In the United States there are various code methods with similarities in some re­
spects but having fundamental philosophical differences in the ways they express
the seismicity of a region and the effect of the type of structural system; for ex­
ample , theBOCA Basic Building Code [3.9], the National Building Code [3.10],
the Standard Building Code [3.11], and �he Uniform Building Code [3.2]. The
equivalent lateral force method in the Uniform Building Code (UBC), which is
used in the western United States and in many other locations will be discussed
here. It is based on the 1988 earthquake code of the Structural Engineers Associ­
ation of California [3.12].

Determination of th.e Minimum Base Shear Force. The UBC states that
the structure shall be designed to resist a minimum total lateral seismic load V,
which shall be assumed to act nonconcurrently in orthogonal directions parallel to
the main axes of the structure, where Vis calculated from the formula

ZJC
V=-W (3.3)
Rw

in which

l.25S
c = r2/3
(3.4)

The design base shear equation (3.3) provides the level of the seismic design
loading for a given structural system, assuming that the structure will undergo
inelastic deformation during a major earthquake. The coefficients in Eq. (3.3) take
into account the effects of the seismicity of the area, the dead load, the structural
type and its ability to dissipate energy without collapse, the response of the struc-

@Seismicisolation
@Seismicisolation
 3.3 EARTHQUAKE LOADING 27

ture, the interaction of the structure with the ground, and the importance of the
structure.
The zone coefficient Z corresponds numerically to the effective peak ground
acceleration (EPA) of a region, and is defined for the United States by a map that
is divided into regions representing five levels of ground motion [3.2]. As an EPA
value,.it is used to scale the spectral shape given by the coefficient C, Eq. (3.4),
so that the product of the coefficients Z and C represents an acceleration response
spectrum envelope having a 10% probability of being exceeded in 50 years.
The importance factor I is concerned with the numbers of people in the building
whose safety is directly at risk, and whether the building has an immediate post­
earthquake role in the safety and recovery of the community.
The coefficient C represents the response of the particular structure to the earth­
quake acceleration spectrum. The curve given by Eq. (3.4) is a simplified multi­
mode acceleration response spectrum normalized to an effective peak ground ac­
celeration of I basis. It is a function of the fundamental period of the structure T,
and a site coefficient S, which is included to adjust the shape of the appropriate
frequency response content of the site soil conditions. The UBC has categorized
the broad range of soil characteristics into four types, and a site coefficient has
been assigned to each of these depending on the soil type and depth. A maximum
limit on C = 2.75 for any structure and soil site condition is given to provide a
simple seismic load evaluation for design projects where it is not practical to eval­
uate the site soil conditions and the structure period. In addition, to assure that a
minimum base shear of 3% of the building weight is used in Seismic Zone 4, with
proportional values in the lower zones, a lower limit of C / Rw = 0.075 is pre­
scribed.
The structural system factor Rw is a measure of the ability of the structural
system to sustain cyclic inelastic deformations without collapse. It is in the de­
nominator of the design base shear equation (3.3) so that design loads decrease for
systems with large inelastic deformation capabilities. The magnitude of Rw de­
pends on the ductility of the type and material of the structure, the possibility of
failure of the vertical load system, the degree of redundancy of the system that
would allow some localized failures without overall failure, and the ability of the
secondary system, in the case of dual systems, to stabilize the building when the
primary system suffers significant damage.
The factor W is normally the total dead load of the building.
The value of V from Eq. (3.3) gives the magnitude of the total base shear that
must be distributed over the height of the structure for the equivalent static anal­
ysis.

Distribution of Total Base Shear. Having determined a value for the total
base shear it is necessary, in order to proceed with the analysis, to allocate the
base shear as effective hprizontal loads at the various floor levels. In deciding on
an appropriate distribution for the horizontal load the following factors are consid­
ered.:

@Seismicisolation
@Seismicisolation
28 LOADING

1. The effective load at a floor level is equal to the product of the mass assigned
to that floor and the horizontal acceleration at that level.
2. The maximum acceleration at any level of the structure in the fundamental
mode is proportional to its horizontal displacement in that mode.

3. The fundamental mode for a regular structure, consisting of shear walls and
frames, is approximately linear from the base.

A reasonable distribution of the total base shear V throughout the height would
be in accordance with a linear acceleration distribution, as given by

(3.5)

where w; and wx are those portions of W assigned to levels i and x, respectively;

that is, the weight at or adjacent to levels i and x, and assigned to those levels for
the purpose of the analysis.
For structures whose weight is distributed uniformly over their height, the hor­
izontal load distribution resulting from Eq. (3.5) forms a triangle, with a maximum
value at the top. Such a distribution has been found to be appropriate for buildings
of relatively stocky proportions where only the fundamental mode is significant.
In more slender, longer period buildings, however, higher modes become signif­
icant, causing a greater proportion of the total horizontal inertia forces to act near
the top; the intensity of this effect is related to the period of the building. Conse­
quently, this is reflected in the UBC [3.2], and in other Codes, by applying a part
of the total loading as a concentrated horizontal force F, at the top of the building.
The remainder of the total base shear is then distributed over the height of the
building as an inverted triangle.
Torsion in any story of the building is prescribed in the UBC [3.2], to be taken
as the product of the story shear and an eccentricity resulting from the addition of
a calculated eccentricity of the mass above, from the center of rigidity of the story,
and an accidental eccentricity of 5% of the plan dimension of the building perpen­
dicular to the direction of the force being considered. If torsional irregularities
exist, the accidental eccentricity is to be increased by an amplification factor re­
lating the maximum story drift at one end of the structure to the average of the
story drifts of the two ends of the structure.
Explanatory material and related technical information useful to the designer in
the application of the design procedure for this equivalent static approach is pro­
vided in the tentative commentary of the 1988 earthquake code of the Structural
Engineers Association of California [3. 12].
The Applied Technology Council (ATC) produced a report in 1978 with a sec­
ond printing in 1984, Tentative Provisions for the Development of Seismic Regu­
lations for Buildings, ATC 3-06 Amended [3.13], for the consideration of building
authorities across the United States. Its recommendations indicate the likely de-

@Seismicisolation
@Seismicisolation
 3.4 COMBINATIONS OF LOADING 29

velopments in the equivalent lateral force procedures of the major Codes. Details
in the approach of the ATC 3-06 have been reviewed by Berg [3.14). Many of the
ATC's provisions have been used by SEAOC, and consequently the International
Conference of Building Officials, as well as the National Research Council of
Canada, as key resource documents to develop their new code editions, Recom­
mended wteral Force Requirements and Tentative Commentary [3.12), Uniform
Building Code [3.2), and National Building Code of Canada [3. 7, 3.8), respec­
tively.

3.3.2 Modal Analysis Procedure

The equivalent static load type of analysis is suitable for the majority of high-rise
structures. If, however, either the lateral load resisting elements or the vertical
distribution of mass are significantly irregular over the height of the building, as
in buildings with large floor-to-floor variations of internal configuration, or with
setbacks, an analysis that takes greater consideration of the dynamic characteristics
of the building must be made. Usually, in such cases, a modal analysis would be
appropriate.
A detailed explanation of the theory and procedure of modal analysis is given
in Chapter 17 and in other texts [3.15, 3.16). Reviewing the method briefly, how­
ever, in a modal analysis a lumped mass model of the building with horizontal
degrees of freedom at each floor is analyzed to determine the modal shapes and
modal frequencies of vibration. The results are then used in conjunction with an
earthquake design response spectrum, and estimates of the modal damping, to
determine the probable maximum response of the structure from the combined
effect of its various modes of oscillation.
Buildings in which the mass at the floor levels is highly eccentric from the
corresponding centers of resistance will be subjected to torque, causing the pos­
sibility of significant torsional vibrations and of coupling between the lateral and
torsional mode. The modal method can also be applied to the analysis of such a
building, by adding to the structural model a third, rotational, degree of freedom
at each floor level.
The modal method is applicable, in the strictest sense, only to linear elastic
systems. Consequently, the results for a building structure's response are, at best,
only an approximate estimate, because of its typically being designed to suffer
significant inelastic deformations in only moderate earthquakes. More accurate
values of response may be obtained for some buildings by the modal analysis
method, using modified design response spectra for inelastic systems [3.16).

3.4 COMBINATIONS OF LOADING

Methods of accounting for load combinations and their effects on the design of
members vary according to the Code used and to the design philosophy. The com­
bination of dead and live loading with reductions in the live loading to allow for

@Seismicisolation
@Seismicisolation
30 LOADING

the improbability of fully loaded tributary areas, and considering patterned live
loading for the worst effects, have already been discussed.
The approaches to combinations of loading by two north American Codes, the
Uniform Building Code [3.2] and the National Building Code of Canada [3. 7],
will be referred to as representative of many of the major building Codes.

3.4.1 Working Stress Design

The UBC and NBCC both assume that wind and earthquake loading need not be
taken to act simultaneously. The UBC considers the improbability of extreme grav­
ity and wind, or earthquake, loadings acting simultaneously by allowing for the
combination a one-third increase in the permissible working stresses, which is
equivalent to a 25% reduction in the sum of the gravity and wind, or earthquake,
loading.
The NBCC approach to allowing for the improbability of the loads acting simul­
taneously is to apply a reduction factor to the combined loads rather than to allow
an increase in the permissible stresses, with greater reductions for the greater num­
ber of load types combined.

3.4.2 Limit States Design

In limit states design, the adequacy of the building and its members is checked
against factored loads in order to satisfy the various safety and serviceability limit
states.
The UBC requires that the strength must be able to resist the actions resulting
from the combination of the individually factored dead and live loads, where the
load factors take into account the variability of the load and load patterns.
If a wind load or earthquake load is to be included, a reduction factor is applied
to the combination of the individually factored loads to allow for the improbability
of the maximum values of the wind or earthquake, and other live loads occurring
simultaneously.
In the NBCC, three factors are required to account for combinations of loading
in limit states design: a load factor, which accounts for the variability of the loads
as before; a load combination factor, which is applied to loads other than dead
loads and accounts for the improbability of their extreme values acting simulta­
neously; and an importance factor, which allows a reduction where collapse is not
likely to have serious consequences.
In both the UBC and the NBCC the strength requirement is satisfied by ensuring
that the factored resistance of the members is not less than the corresponding ac­
tions caused by the factored loads.

3.4.3 Plastic Design.

In buildings in which plastic design is used for parts or the whole of the steel
framed structure, available methods of analysis are based on proportional systems

@Seismicisolation
@Seismicisolation
 SUMMARY 31

of loading, that is load combinations in which increasing loads maintain their rel­
ative magnitudes. Consequently, all the loads within a combination are given the
same load factor.

SUMMARY

Loading on high-rise buildings differs from loading on low-rise buildings mainly

in its accumulation over the height to cause very large gravity and lateral load
forces within the structure. In buildings that are exceptionally slender or flexible,
the building dynamics can also become important in influencing the effective load­
ing.
Gravity loading consists of dead loading, which can be predicted reasonably
accurately, and live loading, whose magnitudes are estimates based on experience
and field surveys, and which are predictable with much less accuracy. The prob­
ability of not all parts of a floor supported by a beam, and of not all floors sup­
ported by a column, being subjected to the full live loading simultaneously, is
provided for by reductions in the beam loading and in the column loading, re­
spectively, in accordance with various formulas. It is sometimes necessary to con­
sider also the effects of construction loads.
Wind loading becomes significant for buildings over about 10 stories high, and
progressively more so with increasing height. For buildings that are not very tall
or slender, the wind loading may be estimated by a static method. Modem static
methods of determining a design wind loading account for the region of the country
where the building is to be located, the exposure of the particular location, the
effects of gusting, and the importance of the building in a postwindstorm situation.
For exceptionally tall, slender, or flexible buildings, it is recommended that a
wind tunnel test on a model is made to estimate the wind loading. Boundary layer
wind tunnels, which simulate the variation of wind speed with height, and the
gusting are used for this purpose.
For buildings that do not quite fall into the category that demands a wind tunnel
test, or for those that are in that category but whose budget does not allow such a
test, dynamic methods of calculating the wind load have been developed.
Earthquake loading is a result of the dynamic response of the building to the
shaking of the ground. Estimates of the loading account for the properties of the
structure and the record of earthquakes in the region. For unexceptionally high
buildings with unexceptional structural arrangements an equivalent lateral force
method is recommended. In this, the loading is estimated on the basis of a simple
approximation for the structure's fundamental period, its dead load, the anticipated
ground acceleration or velocity, and other factors relating to the soil site condi­
tions, structure type and the importance of its use. The method gives the value of
the maximum horizontal base shear, which is then distributed as an equivalent
lateral load over the height of the building so that a static analysis can be per­
formed.

@Seismicisolation
@Seismicisolation
32 LOADING

If the building is exceptionally tall, or irregular in its structure or its mass dis­
tribution, a modal analysis procedure is recommended for estimating the earth­
quake loading. The modal shapes and frequencies of vibration are analyzed; these
are used in conjunction with an earthquake design response spectrum and estimates
of the modal damping to determine the probable maximum responses. The modal
method can also allow for the simultaneous torsional oscillation of the building.
Methods of combining types of loading vary according to the design method
and the Code of Practice concerned. Although dead load is considered to act in
full all the time, live loads do not necessarily do so. The probability of the full
gravity live loading acting with either the full wind, earthquake, or temperature
loading is low, and of all of them acting together is even lower. This is reflected
in the Codes by applying a greater reduction factor to those combinations incor­
porating more different types of loading. Wind and earthquakes are assumed never
to act simultaneously.

REFERENCES

3.1 Tall Building Criteria and Loading. Monograph on Planning and Design of Tall
Buildings, Vol. CL, ASCE, New York. 1980.
3.2 Uniform Building Code, /988, International Conference of Building Officials, Whit­
tier, California.

3.3 Cook, N. J. The Designer's Guide to Wind Loading of Building Structures, Parr I,
Building Research Establishment Report, Butterworths. London, 1985.
3.4 Simiu, E. and Scanlan, R. H. Wind Effects on Structures, 2nd ed., Wiley Intersci­
ence, New York, 1986.

3.5 Davenport, A. G. "Gust loading factors," J. Struct. Div., Proc. A.S.C.E. 93, June
1967, 12-34.
3.6 Simiu, E. "Equivalent static wind loads for tall building design," J. Struct. Div..
Proc. A.S.C.E. 102, April 1976, 719-737.
3.7 National Building Code of Canada, 1990, National Research Council of Canada,
Ottawa, Canada.
3.8 Supplement to the National Building Code of Canada, 1990, National Research
Council of Canada, Ottawa, Canada.

3.9 The BOCA Basic Building Code-1990, Building Officials and Code Administrators
International, Homewood, Illinois.

3.10 The National Building Code-1976, American Insurance Association, New York.
3.11 Standard Building Code, 1988 Edition, Southern Building Code Congress Interna­
tional, Birmingham, Alabama.

3.12 Recommended Lateral Force Requirements and Commentary, Seismology Commit­

tee, Structural Engineers Association of California, 1988.

3.13 Tentative Provisions for the Development of Seismic Regulations for Buildings, ATC
3-06 Amended, Applied Technology Council, National Bureau of Standards, Wash­
ington, D.C., 1984.

@Seismicisolation
@Seismicisolation
 REFERENCES 33

3.14 Berg, G. V. Seismic Design Codes and Procedures, Earthquake Engineering Re­
search Institute, Berkeley, California, 1982.

3.15 Clough, R. W. and Penzien, J. Dynamics of Structures, McGraw-Hill, New York,

1975.
3.16 Newmark, N. M. and Hall, W. J. Earthquake Spectra and Design, Earthquake En­
gineering Research Institute, Berkeley, California. 1982.

@Seismicisolation
@Seismicisolation
 CHAPTER 4

Structural Form

From the structural engineer's point of view, the determination of the structural
form of a high-rise building would ideally involve only the selection and arrange­
ment of the major structural elements to resist most efficiently the various com­
binations of gravity and horizontal loading. In reality, however, the choice of
structural form is usually strongly influenced by other than structural considera­
tions. The range of factors that has to be taken into account in deciding the struc­
tural form includes the internal planning, the material and method of construction,
the external architectural treatment, the planned location and routing of the service
systems, the nature and magnitude of the horizontal loading, and the height and
proportions of the building. The taller and more slender a building, the more im­
portant the structural factors become, and the more necessary it is to choose an
appropriate structural form.
In high-rise buildings designed for a similar purpose and of the same material
and height, the efficiency of the structures can be compared roughly by their weight
per unit floor area. In these terms, the weight of the floor framing is influenced
mainly by the floor span and is virtually independent of the building height, while
the weight of the columns, considering gravity load only, is approximately pro­
portional to the height (Fig. 4.1). Buildings of up to lO stories designed for gravity
loading can usually accommodate wind loading without any increase in member
sizes, because of the typically allowed increase in permissible stresses in Design
Codes for the combined loading. For buildings of more than 10 stories, however,
the additional material required for wind resistance increases nonlinearly with
height so that for buildings of 50 stories and more the selection of an appropriate
structural form may be critical for the economy, and indeed the viability, of the
building.
A major consideration affecting the structural form is the function of the build­
ing. Modem office buildings call for large open floor spaces that can be subdivided
with lightweight partitioning to suit the individual tenant's needs. Consequently,
the structure's main vertical components are generally arranged, as far as possible,
around the perimeter of the plan and, internally, in groups around the elevator,
stair, and service shafts (Fig. 4.2). The floors span the areas between the exterior
and interior components, leaving large column-free areas available for office plan­
ning. The services are distributed horizontally in each story above the partitioning
and are usually concealed in a ceiling space. The extra depth required by this space

34 @Seismicisolation
@Seismicisolation
 STRUCTURALFORM 35

64

.... 56
.....

40
"'
"' 32
....
"'

..... 24
0

....
..<:
16

I�..-.e::.;....-�
""

"' 8
3
Floor framing
0
0 10 20 30 40 50 60 70 80 90 1 00 110
Number of floors

Fig. 4.1 Weight of steel in tall buildings.

causes the typical story height in an office building to be ll ft-6 in. ( 3 .5 m) or

more.
In a residential building or hotel, accommodation is subdivided permanently
and usually repetitively from floor to floor. Therefore, continuously vertical col­
umns and walls can be distributed over the plan to form, or fit within, the parti­
tioning (Fig. 4.3). The services can then be run vertically, adjacent to the walls
and columns or in separate shafts, to emerge in each story either very close to
where required, or to be distributed horizontally from there to where required,
along the corridor ceiling spaces. With the exception of the corridors, therefore, a
ceiling space is not required, and the soffit of the slab can serve as the ceiling.
This allows the story heights in a typical residential building or hotel to be kept
down to approximately 8 ft-8 in. (2.7 m). A 40-story residential building is, there­
fore, generally of significantly less height than a 40-story office building.
In addition to satisfying the previously mentioned nonstructural requirements,
the principal objectives in choosing a building's structural form are to arrange to
support the gravity, dead and live, loading, and to resist at all levels the external

Central core

Fig. 4.2 Plan of office block (tube-type).

@Seismicisolation
@Seismicisolation
36 STRUCTURAL FORM

� •

"" - - •

x �
� �
.. - - •

Fig. 4.3 Plan of residential block. I_ •

horizontal load shear, moment, and torque with adequate strength and stiffness.
These requirements should be achieved, of course, as economically as possible.
With regard to horizontal loading, a high-rise building is essentially a vertical
cantilever. This may comprise one or more individually acting vertical cantilevers,
such as shear walls or cores, each bending about its own axis and acting in unison
only through the horizontal in-plane rigidity of the floor slabs. Alternatively, the
cantilever may comprise a number of columns or walls that are mobilized to act
compositely, to some degree, as the chords of a single massive cantilever, by
vertically shear-resistant connections such as bracing or beams. The lateral stiff­
ness and strength of both of these basic cantilever systems may be further enhanced
if the major vertical elements have different free deflection characteristics, in which
case they will interact horizontally through the connecting slabs and beams.
Within the constraints of the selected structural form, advantage may be taken
of locating the main vertical members on plan so that the dead load compressive
stresses suppress the lateral load tensile stresses, thereby avoiding the possibility
of net tension occurring in the vertical members and uplift on the foundations.
Particular emphasis is placed in some types of structural form on routing the grav­
ity load to the outer vertical members to achieve this purpose.
Steel framing has played a pioneering role in the history of tall buildings. It is
appropriate for all heights of structure and, because of its high strength-to-weight
ratio, it has always been the material of construction for the tallest buildings. It
allows the possibility of longer floor spans, and of partial prefabrication, leading
to reduced site work and more rapid erection. Its disadvantages, however, include
needing fire and rust protection, being expensive to clad, and requiring costly di­
agonal bracing or rigid-frame connections.
After the earlier use of steel through the first half of the century, in the form of
braced construction, it has evolved in its structural forms somewhat in parallel
with reinforced concrete to include rigid-frame, shear wall, wall-frame, tube and
braced-tube, and outrigger types of arrangements, as well as in forms more par­
ticular to steel such as the suspended structure and the highly efficient massive
space frame.

@Seismicisolation
@Seismicisolation
 4.1 STRUCTURAL FORM 37

Reinforced concrete tall buildings were introduced approximately two decades

after the first steel tall buildings. Understandably, the earlier concrete building
structures were influenced in form by the skeletal, column and girder arrangements
of their steel counterparts, but they differed in depending on the inherent rigid­
frame action of concrete construction to resist horizontal loading. Subsequently,
the flat plate and flat slab forms were introduced and these, with the moment­
resistant frame, continued as the main repertoire of reinforced concrete high-rise
structural form until the late 1940s.
A major step forward in reinforced concrete high-rise structural form came with
the introduction of shear walls for resisting horizontal loading. This was the first
in a series of significant developments in the structural forms of concrete high-rise
buildings, freeing them from the previous 20- to 25-story height limitations of the
rigid-frame and flat plate systems. The innovation and refinement of these new
forms, together with the development of higher strength concretes, has allowed
the height of concrete buildings to reach within striking distance of 100 stories.
Of the following structural forms, some are more appropriate to steel and others
to reinforced concrete; many are suitable for either material, while a few allow or
demand a combination of materials in the same structure. They are described in a
roughly historical sequence.
The structural form of tall buildings, as discussed so far, has concerned mainly
the arrangement of the primary vertical components and their interconnections.
This topic would not be complete, however, without including consideration of
floor systems, because some of them play an integral part with the vertical com­
ponents in resisting the lateral, as well as the gravity, loading. The last part of the
chapter is devoted, therefore, to a brief review of the floor systems used in tall
buildings. Many of these are commonly used also in low-rise buildings but are
included here for completeness.

4.1 STRUCTURAL FORM

4.1.1 Braced-Frame Structures

In braced frames the lateral resistance of the structure is provided by diagonal

members that, together with the girders, form the "web" of the vertical truss,
with the columns acting as the "chords" (Fig. 4.4). Because the horizontal shear
on the building is resisted by the horizontal components of the axial tensile or
compressive actions in the web members, bracing systems are highly efficient in
resisting lateral loads.
Bracing is generally regarded as an exclusively steel system because the diag­
onals are inevitably subjected to tension for one or the other directions of lateral
loading. Concrete bracing of the double diagonal form is sometimes used, how­
ever, with each diagonal designed as a compression member to carry the full ex­
ternal shear.
The efficiency of bracing, in being able to produce a laterally very stiff structure
for a minimum of additional material, makes it an economical structural form for

@Seismicisolation
@Seismicisolation
38 STRUCTURAL FORM

Chord members
......
Single diagonal
bracing
......

Double diagonal
-

Chevron
-

�
Story-height knee

Fig. 4.4 Braced frame-showing different types of bracing.

any height of building, up to the very tallest. An additional advantage of fully

triangulated bracing is that the girders usually participate only minimally in the
lateral bracing action; consequently, the floor framing design is independent of its
level in the structure and, therefore, can be repetitive up the height of the building
with obvious economy in design and fabrication. A major disadvantage of diagonal
bracing is that it obstructs the internal planning and the_ location of windows and
doors. For this reason, braced bents are usually incorporated internally along wall
and partition lines, and especially around elevator, stair, and service shafts. An­
other drawback is that the diagonal connections are expensive to fabricate and
erect.
The traditional use of bracing has been in story-height, bay-width modules (Fig.
4.4) that are fully concealed in the finished building. More recently, however,
external larger scale bracing, extending over many stories and bays (Fig. 4.5), has
been used to produce not only highly efficient structures, but aesthetically attrac­
tive buildings.
Bracing ·and its modes of behavior are described in more detail in Chapter 6.

4.1.2 Rigid-Frame Structures

Rigid-frame structures consist of columns and girders joined by moment-resistant

connections. The lateral stiffness of a rigid-frame bent depends on the bending
stiffness of the columns, girders, and connections in the plane of the bent (Fig.
4.6). The rigid frame's principal advantage is its open rectangular arrangement,
which allows freedom of planning and easy fitting of doors and windows. If used
as the only source of lateral resistance in a building, in its typical20 ft (6 m)-30
ft (9 m) bay size, rigid framing is economic only for buildings up to about 25
stories. Above 25 stories the relatively high lateral flexibility of the frame calls for
uneconomically large members in order to control the drift.

@Seismicisolation
@Seismicisolation
 4.1 STRUCTURAL FOAM 39

Fig. 4.5 Large-scale braced frame.

Rigid-frame construction is ideally suited for reinforced concrete buildings be­

cause of the inherent rigidity of reinforced concrete joints. The rigid-frame form
is also used for steel frame buildings, but moment-resistant connections in steel
tend to be costly. The sizes of the columns and girders at any level of a rigid frame
are directly influenced by the magnitude of the external shear at that level, and
they therefore increase toward the base. Consequently, the design of the floor
framing cannot be repetitive as it is in some braced frames. A further result is that
sometimes it is not possible in the lowest stories to accommodate the required
depth of girder within the normal ceiling space.

r: I"'� ,--. r.. ,--. �� t' "

��� \.. . J • .... J "'... ....

f , I I

Fig. 4.6 Rigid frame.

@Seismicisolation
@Seismicisolation
40 STRUCTURAL FORM

Gravity loading also is resisted by the rigid-frame action. Negative moments

are induced in the girders adjacent to the columns causing the mid-span positive
moments to be significantly less than in a simply supported span. In structures in
which gravity loads dictate the design, economies in member sizes that arise from
this effect tend to be offset by the higher cost of the rigid joints.
While rigid frames of a typical scale that serve alone to resist lateral loading
have an economic height limit of about 25 stories, smaller scale rigid frames in
the form of a perimeter tube, or typically scaled rigid frames in combination with
shear walls or braced bents, can be economic up to much greater heights. These
structural forms are described later in this chapter. The detailed behavior of rigid
frames is discussed in Chapter 7.

4.1.3 lnfilled-Frame Structures

In many countries infilled frames are the most usual form of construction for
tall buildings of up to 30 stories in height. Column and girder framing of reinforced
concrete, or sometimes steel, is infilled by panels of brickwork, blockwork, or
cast-in-place concrete.
When an infilled frame is subjected to lateral loading, the infill behaves effec­
tively as a strut along its compression diagonal to brace the frame (Fig. 4. 7).
Because the infills serve also as external walls or internal partitions, the system is
an economical way of stiffening and strengthening the structure.
The complex interactive behavior of the infill in the frame, and the rather ran­
dom quality of masonry, has made it difficult to predict with accuracy the stiffness
and strength of an infilled frame. Indeed, at the time of writing, no method of
analyzing infilled frames for their design has gained general acceptance. For these
reasons, and because of the fear of the unwitting removal of bracing infills at some
time in the life of the building, the use of the infills for bracing tall buildings has
mainly been supplementary to the rigid-frame action of concrete frames. An out­
line of a method for designing infilled frames is given in Chapter 8.

===- �
- lnfills
7
-
-
� -

......
-- �
� !==-
-- -
-
=-

Diagonal strut
� �� ac tion of infill
-
i=
- -

--
- --
--
- -=

, , , , / ,

Fig. 4.7 lnfilled frame.

@Seismicisolation
@Seismicisolation
 4.1 STRUCTURAL FORM 41

4.1.4 Flat-Plate and Flat-Slab Structures

The fiat-plate structure is the simplest and most logical of all structural forms in
that it consists of uniform slabs, of 5-8 in. ( 12-20 em) thickness, connected rig­
idly to supporting columns (Fig. 4.27). The system, which is essentially of rein­
forced concrete, is very economical in having a fiat soffit requiring the most un­
complicated formwork and, because the soffit can be used as the ceiling, in creating
a minimum possible floor depth.
Under lateral loading the behavior of a fiat-plate structure is similar to that of
a rigid frame, that is, its lateral resistance depends on the flexural stiffness of the
components and their connections, with the slabs corresponding to the girders of
the rigid frame. It is particularly appropriate for apartment and hotel construction
where ceiling spaces are not required and where the slab may serve directly as the
ceiling. The fiat-plate structure is economical for spans of up to about 25 ft ( 8 m),
above which drop panels can be added to create a flat-slab structure (Fig. 4.28)
for spans of up to 38 ft ( 12 m).
Buildings that depend entirely for their lateral resistance on flat-plate or flat­
slab action are economical up to about 25 stories. Previously, however, when Code
requirements for wind design were less stringent, many flat-plate buildings were
constructed in excess of 40 stories, and are still performing satisfactorily.

4.1.5 Shear Wall Structures

Concrete or masonry continuous vertical walls may serve both architecturally as

partitions and structurally to carry gravity and lateral loading. Their very high in­
plane stiffness and strength makes them ideally suited for bracing tall buildings.
In a shear wall structure, such walls are entirely responsible for the lateral load
resistance of the building. They act as vertical cantilevers in the form of separate
planar walls, and as nonplanar assemblies of connected walls around elevator,
stair, and service shafts (Fig. 4.8). Because they are much stiffer horizontally than
rigid frames, shear wall structures can be economical up to about 35 stories.
In contrast to rigid frames, the shear walls' solid form tends to restrict planning
where open internal spaces are required. They are well suited, however, to hotels
and residential buildings where the floor-by-floor repetitive planning allows the
walls to be vertically continuous and where they serve simultaneously as excellent
acoustic and fire insulators between rooms and apartments.
If, in low- to medium-rise buildings, shear walls are combined with frames, it
is reasonable to assume that the shear walls attract all the lateral loading so that
the frame may be designed for only gravity loading. It is especially important in
shear wall structures to try to plan the wall layout so that the lateral load tensile
stresses are suppressed by the gravity load stresses. This allows them to be de­
signed to have only the minimum reinforcement. Shear wall structures have been
shown to perform well in earthquakes, for which case ductility becomes an im­
portant consideration in their design. The behavior and methods of analysis of
shear wall structures are discussed in detail in Chapter 9.

@Seismicisolation
@Seismicisolation
42 STRUCTURALFORM

Fig. 4.8 Shear wall structure.

Coupled Wall Structures. A coupled wall structure is a particular, but very

common, form of shear wall structure with its own special problems of analysis
and design. It consists of two or more shear walls in the same plane, or almost the
same plane, connected at the floor levels by beams or stiff slabs (Fig. 4.9). The
effect of the shear-resistant connecting members is to cause the set of walls to
behave in their plane partly as a composite cantilever, bending about the common
centroidal axis of the walls. This results in a horizontal stiffness very much greater
than if the walls acted as a set of separate uncoupled cantilevers.
Coupled walls occur often in residential construction where lateral-load resist­
ant cross walls, which separate the apartments, consist of in-plane coupled pairs,
or trios, of shear walls between which there are corridor or window openings.
Coupled shear walls are considered in detail in Chapter 10.
Although shear walls are obviously more appropriate for concrete construction,
they have occasionally been constructed of heavy steel plate, in the style of mas­
sive vertical plate or box girders, as parts of steel frame structures. These have
been designed for locations of extremely heavy shear, .such as at the base of ele­
vator shafts.

4.1.6 Wall-Frame Structures

When shear walls are combined with rigid frames (Fig. 4.10) the walls, which

@Seismicisolation
@Seismicisolation
tend to deflect in a flexural configuration, and the frames, which tend to deflect in
 4.1 STRUCTURAL FORM 43

Coupled shear walls

I

'·

Fig. 4.9 Coupled shear wall structure.

Rigid frames

@Seismicisolation
Fig. 4.10

@Seismicisolation Wall-frame structure.

44 STRUCTURAL FORM

a shear mode, are constrained to adopt a common deflected shape by the horizontal
rigidity of the girders and slabs. As a consequence, the walls and frames interact
horizontally, especially at the top, to produce a stiffer and stronger structure. The
interacting wall-frame combination is appropriate for buildings in the 40- to
60 -story range, well beyond that of rigid frames or shear walls alone.
An additional, less well known feature of the wall-frame structure is that, in a
carefully "tuned" structure, the shear in the frame can be made approximately
uniform over the height, allowing the floor framing to be repetitive.
Although the wall-frame structure is usually perceived as a concrete structural
form, with shear walls and concrete frames, a steel counterpart using braced frames
and steel rigid frames offers similar benefits of horizontal interaction. The braced
frames behave with an overall flexural tendency to interact with the shear mode of
the rigid frames.
Detailed descriptions of the behavior and methods of analysis for wall-frame
structures are given in Chapter II.

4.1. 7 Framed-Tube Structures

The lateral resistance of framed-tube structures is provided by very stiff moment­

resisting frames that form a "tube" around the perimeter of the building. The
frames consist of closely spaced columns, 6-12 ft (2-4 m) between centers, joined
by deep spandrel girders (Fig. 4.11). Although the tube carries all the lateral load­
ing, the gravity loading is shared between the tube and interior columns or walls.
When lateral loading acts, the perimeter frames aligned in the direction of loading
act as the "webs" of the massive tube cantilever, and those normal to the direction
of the loading act as the "flanges. "
The close spacing of the columns throughout the height of the structure is usu­
ally unacceptable at the entrance level. The columns are therefore merged, or ter­
minated on a transfer beam, a few stories above the base so that only a few, larger,
more widely spaced columns continue to the base. The tube form was developed
originally for buildings of rectangular plan, and probably its most efficient use is
in that shape. It is appropriate, however, for other plan shapes, and has occasion­
ally been used in circular and triangular configurations.
The tube is suitable for both steel and reinforced concrete construction and has
been used for buildings ranging from 40 to more than 100 stories. The highly
repetitive pattern of the frames lends itself to prefabrication in steel, and to the use
of rapidly movable gang forms in concrete, which make for rapid construction.
The framed tube has been one of the most significant modem developments in
high-rise structural form. It offers a relatively efficient, easily constructed struc­
ture, appropriate for use up to the greatest of heights. Aesthetically, the tube's
externally evident form is regarded with mixed enthusiasm; some praise the logic
of the clearly expressed structure while others criticize the grid-like facade as small­
windowed and uninterestingly repetitious.
The tube structure's structural efficiency, although high, still leaves scope for
improvement because the "flange" frames tend to suffer from "shear lag"; this

@Seismicisolation
@Seismicisolation
 4.1 STRUCTURAL FORM 45

to carry gravity loads

Framed-tube to carry gravity

and entire l ateral loading

Fig. 4.11 Framed-tube.

results in the mid-face "flange" columns being less stressed than the comer col­
umns and, therefore, not contributing as fully as they could to the flange action.

Tube-in-Tube or Hull-Core Structures. This variation of the framed tube

consists of an outer framed tube, the "hull," together with an internal elevator
and service core (Fig. 4.12). The hull and core act jointly in resisting both gravity
and lateral loading. In a steel structure the core may consist of braced frames,
whereas in a concrete structure it would consist of an assembly of shear walls.
To some extent, the outer framed tube and the inner core interact horizontally
as the shear and flexural components of a wall-frame structure, with the benefit of
increased lateral stiffness. However, the structural tube usually adopts a highly
dominant role because of its much greater structural depth.

Bundled-Tube Structures. This structural form is notable in its having been

used for the Sears Tower in Chicago-the world's tallest building. The Sears Tower
consists of four parallel rigid steel frames in each orthogonal direction, intercon­
nected to form nine "bundled" tubes (Fig. 4.13a). As in the single-tube structure,
the frames in the direction of lateral loading serve as "webs" of the vertical can­
tilever, with the normal frames acting as "flanges."

@Seismicisolation
@Seismicisolation
46 STRUCTURAL FORM

Core (or inner tube)

(or outer tube)

Fig. 4.12 Tube-in-tube.

The introduction of the internal webs greatly reduces the shear lag in the flanges;
consequently their columns are more evenly stressed than in the single-tube struc­
ture, and their contribution to the lateral stiffness is greater. This allows columns
of the frames to be spaced further apart and to be less obtrusive. In the Sears
Tower, advantage was taken of the bundled form to discontinue some of the tubes,
and so reduce the plan of the building at stages up the height (Fig. 4.13b, c, and
d).

Braced- Tube Structures. Another way of improving the efficiency of the

framed tube, thereby increasing its potential for use to even greater heights as well
as allowing greater spacing between the columns, is to add diagonal bracing to the
faces of the tube. This arrangement was first used in a steel structure in 1969, in
Chicago's John Hancock Building (Fig. 4.14), and in a reinforced concrete struc­
ture in 1985, in New York's 780 Third Avenue Building (Fig. 4.15). In the steel
tube the bracing traverses the faces of the rigid frames, whereas in the concrete
structure the bracing is formed by a diagonal pattern of concrete window-size
panels, poured integrally with the frame.
Because the diagonals of a braced tube are connected to the columns at each
intersection, they virtually eliminate the effects of shear lag in both the flange and

@Seismicisolation
@Seismicisolation
 Jwo tubes omitted

5 s
S
\0��; II ( 't (I )l \._) 11 If I
X I II
0
'
�� ���s
0

I
I

(a) (b)

Two additional Three additional

tubes omitted tubes omitted
I
Stories 67-90 t 91-110
S ,,,
If r r 1

@Seismicisolation
@Seismicisolation
(c) (d)

Fig. 4.13 (a-d) Bundled-tube.

�
....
 Window openings
�r-r-r--, omitted to create
diagonal bracing
t:- _system
� t:-�

-
v"'
� �
I' v

v
� v
"'
�
" v I

� '""
/
v \
\ I

,... v I
� I

\
"" v;;> \
\ --
I

-
Dv
-----

"'
�v " I
:"'-. \
7
['I'- v
� /.t-t:::=
t-

-
\....:::: _ ___

Fig. 4.14 Steel-braced tube. Fig. 4.15 Concrete-braced tube.

48 @Seismicisolation
@Seismicisolation
 4.1 STRUCTURAL FORM 49

web frames. As a result, the structure behaves under lateral loading more like a
braced frame, with greatly diminished bending in the members of the frames. Con­
sequently, the spacing of the columns can be larger and the depth of the spandrels
less, thereby allowing larger size windows than in the conventional tube structure.
In the braced-tube structure the bracing contributes also to the improved per­
formance of the tube in carrying gravity loading: differences between gravity load
stresses in the columns are evened out by the braces transferring axial loading from
the more highly to the less highly stressed columns.

4.1.8 Outrigger-Braced Structures

This efficient structural form consists of a central core, comprising either braced
frames or shear walls, with horizontal cantilever "outrigger" trusses or girders
connecting the core to the outer columns (Fig. 4.16a). When the structure is loaded
horizontally, vertical plane rotations of the core are restrained by the outriggers
through tension in the windward columns and compression in the leeward columns
(Fig. 4.16b). The effective structural depth of the building is greatly increased,
thus augmenting the lateral stiffness of the building and reducing the lateral de­
flections and moments in the core. In effect, the outriggers join the columns to the
core to make the structure behave as a partly composite cantilever.
Perimeter columns, other than those connected directly to the ends of the out­
riggers, can also be made to participate in the outrigger action by joining all the
perimeter columns with a horizontal truss or girder around the face of the building

Outrigger
trusses

I
1'\./ V"-..I'.. V

r-

Braced
core r-

\ "-../ V"-i"-..V

f--

,, .,

(a) (b)
Fig. 4.16 (a) Outrigger-braced structure; (b) outrigger-braced structure under load.

@Seismicisolation
@Seismicisolation
 ;

50 STRUCTURALFORM

at the outrigger level. The large, often two-story, depths of the outrigger and pe­
rimeter trusses make it desirable to locate them within the plant levels in the build­
ing.
The degree to which the perimeter columns of an outrigger structure behave
compositely with the core depends on the number of levels of outriggers and their
stiffnesses. Multilevel outrigger structures show a considerable increase in their
effective moment of resistance over single outrigger structures. This increase di­
minishes, however, with each additional level of outriggers, so that four or five
levels appears to be the economic limit. Outrigger-braced structures have been
used for buildings from 40 to 70 stories high, but the system should be effective
and efficient for much greater heights.

4.1.9 Suspended Structures

The suspended structure consists of a central core, or cores, with horizontal can­
tilevers at roof level, to which vertical hangers of steel cable, rod, or plate are
attached. The floor slabs are suspended from the hangers (Fig. 4.17a).
The advantages of this structural form are primarily architectural in that, except
for the presence of the central core, the ground story can be entirely free of major
vertical members, thereby allowing an open concourse; also, the hangers, because
they are in tension and consequently can be of high strength steel, have a minimum
sized section and are therefore less obtrusive. The potential of this latter benefit
tends to be offset, however, by the need to proof the hangers against fire and rust,
thereby significantly increasing their bulk. The suspended structure has some con­
struction advantages in allowing the core, cantilevers, and hangers to be con­
structed while the slabs are being poured on top of each other at ground level; the
slabs are then lifted in sets and fixed in position (Fig. 4.17b).
The structural disadvantages of the suspended structure are that it is inefficient
in first transmitting the gravity loads upward to the roof-level cantilevers before
returning them through the core to the ground, and that the structural width of the
building at the base is limited to the relatively narrow depth of the core, which
restricts the system to buildings of lesser height. A further problem is caused by
the vertical extension of the slender hangers that, over the range from zero to full
live loading, can result in significant changes in the levels of the edges of the slabs.
This effect increases at each level down the length of the hanger and, consequently,
is worst at the lowest hung floor. The problem can be limited by restricting the
maximum number of floors supported by a single length of hanger to about 10,
and by having multilevel cantilever systems (Fig. 4.18). Similarly to outrigger
structures, and for the same reasons, the cantilevers are normally incorporated
within the plant levels.
Variations from the single-core hanging structure include two- and four-core
structures, in which vertical hangers are suspended fr�m massive girders that span
between the cores, or in which hangers are draped, catenary fashion, between the
cores. The benefits of such multicore hanging structures include large open floor
spaces at all levels, and the possiblity of a column-free ground story.

@Seismicisolation
@Seismicisolation
 ';:!

Located and
fixed separately

Braced
core

Hoisted in sets

Temporary enclosure

r
Slabs cast on top
I of each other
I 1
/1;;"";

(a) (b)

@Seismicisolation
Fig. 4.17 (a) Suspended structure; (b) sequence of construction-suspended structure.

@Seismicisolation
U1
...
52 STRUCTURAL FORM

Cantilever supporting upper

/I'- """" � region of floors
A
/'-.
A
-

1--v
V"-
V'-
Cantilever supporting lower
/I'- """" !"..
region of floors
""
A
A
"
-

-
'
/'-.
V'-

� !.'
·� V'-
;j ... V"-
. .�
Fig. 4.18 Two-tiered suspended structure.

4.1.1 0 Core Structures

In these structures a single core serves to carry the entire gravity and horizontal
loading (Fig. 4.19). In some, the slabs are supported at each level by cantilevers

� Cantilever support for

each floor

-·
- Cantilever supporting
....___ = a number of floors
-

- - ::::;;;
-
,....-- Core
" ,. "' -,-y-

@Seismicisolation
@Seismicisolation Fig. 4.19 Core structure.
 4.1 STRUCTURAL FORM 53

from the core. In others, the slabs are supported between the core and perimeter
columns, which terminate either on maj or cantilevers at intervals down the height,
or on a single massive cantilever a few stories above the ground.
Similarly to the suspended building, the merits of the system are mainly archi­
tectural, in providing a column-free perimeter at the ground level and at other
levels j ust below the cantilevers. The structural penalties are considerable, how­
ever, in having only the small effective structural depth of the core and, therefore,
being inefficient in resisting lateral loading, as well as in supporting the floor load­
ing by cantilevers-a highly inefficient structural component.

4.1.11 Space Structures

The primary load-resisting system of a space structure consists essentially of a

three-dimensional triangulated frame-as distinct from an assembly of planar
bents-whose members serve dually in resisting both gravity and horizontal load­
ing. The result is a highly efficient, relatively lightweight structure with a potential
for achieving the greatest heights. The 76-story Hong Kong Bank of China Build­
ing (Fig. 4.20) is a classic example of this structural form.

@Seismicisolation
@Seismicisolation Fig. 4.20 Space structure.
54 STRUCTURAL FORM

Inner, braced core

Outer, space frame

Fig. 4.21 Space structure.

Although simple in their overall concept, space structures are usually geomet­
rically complex, which calls for considerable structural ingenuity in transferring
both the gravity loading and the lateral loading from the floors to the main struc­
ture. One solution is to have an inner braced core, which serves to collect the
lateral loading, and the inner region gravity loading, from the slabs over a number
of multistory regions. At the bottom of each region, the lateral and gravity loads
are transferred out to the main joints of the space frame (Fig. 4.21 ).
Although the multidirectional inclined members of the space frame are struc­
turally awkward and costly to connect, as well as making the fenestration difficult,
the structural form is visually interesting and aesthetically very pleasing in its ap­
parent simplicity.

4.1.12 Hybrid Structures

Many of the previously described structural arrangements are particularly suitable

for prismatically shaped, tower or block, so-called "modem" buildings, which
can be completely structured by a single identifiable system, for example, a tube
or a wall-frame.
Partly as a reaction to an increasingly monotonous urban environment consist­
ing of regularly shaped and repetitive "modem" buildings, and partly because the
analysis and design of much more complex structures have become feasible, ar-

@Seismicisolation
@Seismicisolation
 4.2 FLOOR SYSTEMS-REINFORCED CONCRETE 55

chitects have responded with a new generation of "postmodern" buildings that

are emphatically nonregular in shape, with large-scale cut-outs, flutings, facets,
and crowns that defy classification in their intricacy and variety.
Buildings of a nonprismatic shape are less amenable to a single form of structure
and, therefore, the engineer has to improvise in developing a satisfactory structural
solution. In such situations combinations of two or even more of the basic struc­
tural forms have often been used in the same building, either by direct combination
as, for example, in a superimposed tube and outrigger system (Fig. 4.22), or by
adopting different forms in different parts of the structure as, for example, in a
tube system on three faces of the building and a space frame on a faceted fourth
face (Fig. 4.23).
During the earlier period of the rapid development of structural form, that is
from the 1950s until the mid-1970s, the single form high-rise structure had the
advantage that it was usually possible to make an approximate but acceptable struc­
tural analysis either by hand or by the use of a small computer. Now, with the
ready availability of powerful computers and highly efficient structural analysis
programs, an engineer possessing a sound knowledge of structural form and be­
havior should be able to devise and analyze a structure to suit a building of almost
any conceivable irregularity.

Core and outrigger

system

Framed-tube

Fig. 4.22 Hybrid structure.

@Seismicisolation
@Seismicisolation
56 STRUCTURAL FORM

4.2 FLOOR SYSTEMS-REINFORCED CONCRETE

An appropriate floor system is an important factor in the overall economy of

the building. Some of the factors that influence the choice of floor system are
architectural. For example, in residential buildings, where smaller permanent di­
visions of the floor space are required, shorter floor spans are possible; whereas,
in modem office buildings, that require more open, temporarily subdivisible floor
spaces, longer span systems are necessary. Other factors affecting the choice of
floor system are related to its intended structural performance, such as whether it
is to participate in the lateral load-resisting system, and to its construction, for
example, whether there is urgency in the speed of erection.
Reinforced concrete floor systems are grouped into two categories: one-way, in
which the slab spans in one direction between supporting beams or walls, and two­
way, in which the slab spans in orthogonal directions. In both systems, advantage
is taken of continuity over interior supports by providing negative moment rein­
forcement in the slab.

Framed-tube on three
faces of building

Space frame on
fourth face

Fig. 4.23 Hybrid structure.

@Seismicisolation
@Seismicisolation
 4.2 FLOOR SYSTEMS-REINFORCED CONCRETE 57

Fig. 4.24 One-way slab.

4.2.1 One-Way Slabs on Beams or Walls

A solid slab of up to 8 in. (0.2 m) thick, spanning continuously over walls or

beams up to 24 ft (7.4 m) apart (Fig. 4.24), provides a floor system requiring
simple formwork, possibly flying formwork, with simple reinforcement. The sys­
tem is heavy and inefficient in its use of both concrete and reinforcement. It is
appropriate for use in cross-wall and cross-frame residential high-rise construction
and, when constructed in a number of uninterrupted continuous spans, lends itself
to prestressing.

4.2.2 One-Way Pan Joists and Beams

A thin, mesh-reinforced slab sits on closely spaced cast-in-place joists spanning

between major beams which transfer the load to the columns (Fig. 4.25). The slab
may be as thin as 2.5 in. ( 6 em) while the joists are from 6 in. ( 15 em) to 20 in.
(51 em) in depth and spaced from 20 to 30 in. ( 76 em) centers. The compositely
acting slab and joists form in effect a set of closely spaced T-beams, capable of
large, up to 40ft ( 12.3 m), spans. The joists are formed between reusable pans
that are positioned to set the regular width of the joist, as well as any special
widths.

4.2.3 One-Way Slab on Beams and Girders

A one-way slab spans between beams at a relatively close spacing while the beams
are supported by girders that transfer the load to the columns (Fig. 4.26). The

Fig. 4.25 One-way pan joists.

@Seismicisolation
@Seismicisolation
58 STRUCTURAL FORM

Fig. 4.26 One-way slab on beams and girders.

short spanning slab may be thin, from 3 to 6 in. (7.6-15 em) thick, while the
system is capable of providing long spans of up to 46 ft ( 14 m). The principal
merits of the system are its long span capability and its compatibility with a two­
way lateral load resisting rigid-frame structure.

4.2.4 Two-Way Flat Plate

A uniformly thick, two-way reinforced slab is supported directly by columns or

individual short walls (Fig. 4.27). It can span up to 26 ft ( 8 m) in the ordinary
reinforced form and up to 36 ft (II m) when posttensioned. Because of its sim­
plicity, it is the most economical floor system in terms of formwork and reinforce­
ment. Its uniform thickness allows considerable freedom in the location of the
supporting columns and walls and, with the possibility of using the clear soffit as
a ceiling, it results in minimum story height.

4.2.5 Two-Way Flat Slab

The flat slab differs from the flat plate in having capitals and/or drop panels at the
tops of the columns (Fig. 4.28). The capitals increase the shear capacity, while

Fig. 4.27 Two-way flat plate.

@Seismicisolation
@Seismicisolation
 4.2 FLOOR SYSTEMS-REINFORCED CONCRETE 59

Fig. 4.28 Two-way flat slab.

the drop panels increase both the shear and negative moment capacities at the
supports, where the maximum values occur. The flat slab is therefore more appro­
priate than the flat plate for heavier loading and longer spans and, in similar situ­
ations, would require less concrete and reinforcement. It is most suitably used in
square, or near-to-square, arrangements.

4.2.6 Waffle Flat Slabs

A slab is supported by a square grid of closely spaced joists with filler panels over
the columns (Fig. 4.29). The slab and joists are poured integrally over square,
domed fonns that are omitted around the columns to create the filler panels. The
fonns, which are of sizes up to 30 in. ( 76 m) square and up to 20 in. (50 em)
deep, provide a geometrically interesting soffit, which is often left without further
finish as the ceiling.

4.2. 7 Two-Way Slab and Beam

The slab spans two ways between orthogonal sets of beams that transfer the load
to the columns or walls (Fig. 4.30). The two-way system allows a thinner slab and
is economical in concrete and reinforcement. It is also compatible with a lateral
load-resisting rigid-frame structure. The maximum length-to-width ratio for a slab
to be effective in two directions is approximately 2.

Fig. 4.29 Waffle flat slab.

@Seismicisolation
@Seismicisolation
60 STRUCTURAL FORM

Fig. 4.30 Two-way slab and beam.

4.3 FLOOR SYSTEMS-STEEL FRAMING

The steel-framed floor system is characterized by a reinforced concrete slab sup­

ported on a steel framework consisting variously of joists, beams, and girders that
transfer the gravity loading to the columns. The slab component is usually one­
way with either a cast-in-place solid reinforced concrete slab from 4 in. ( lO em)
to 7 in. ( 18 em) thick, or a concrete on metal deck slab with a variety of possible
section shapes and a minimum slab thickness from 2.5 in. ( 6 em) (Fig. 4.31 ), or
a slab of precast units laid on steel beams and covered by a thin concrete topping
(Fig. 4.32).
A major consideration in the weight and cost of a steel frame building is the
weight of the slab. A floor arrangement with shorter spanning, thinner slabs is
desirable. Longer span, closer spaced beams supporting a short-spanning slab is a
typical arrangement meeting these requirements. The following types of steel floor
framing are categorized according to the spanning arrangement of the supporting
steel framework.

4.3.1 One-Way Beam System

A rectangular grid of columns supports sets of parallel longer span beams at a

relatively close spacing, with the slab spanning the shorter spans transversely to
the beams (Figs. 4.33). In cross-frame structures, the beams at partition lines may
be deepened to participate in lateral load resisting rigid frames or braced bents.

Fig. 4.31 Concrete on metal deck .

@Seismicisolation
@Seismicisolation
 4.3 FLOOR SYSTEMS-STEEL FRAMING 61

Fig. 4.32 Precast units with topping.

Fig. 4.33 One-way beam system in steel.

4.3.2 Two-Way Beam System

In buildings in which columns are required to be farther apart in both directions,

a two-way frame system of girders and beams is often used, with the slab spanning
between the beams (Fig. 4.34). To minimize the total structural depth of the floor
frame, the heavily loaded girders are aligned with the shorter span and the rela­
tively lightly loaded secondary beams with the longer span.

Fig. 4.34 Two-way beam system in steel.

@Seismicisolation
@Seismicisolation
62 STRUCTURAL FORM

Fig. 4.35 Three-way beam system in steel.

4.3.3 Three-Way Beam System

In buildings in which the columns have to be very widely spaced to allow large
internal column-free areas, a three-way beam system may be necessary (Fig. 4.35).
A deep lattice girder may form the primary component with beams or open web
joists forming the secondary and tertiary systems. In each case the system is ar­
ranged to provide relatively short spans for the supported concrete slab.

4.3.4 Composite Steel-Concrete Floor Systems

The use of steel members to support a concrete floor slab offers the possibility of
composite construction in which the steel members are joined to the slab by shear
connectors so that the slab serves as a compression flange.
In one simple and constructionally convenient slab system, steel decking, which
is often used to act merely as rapidly erected permanent formwork for a bar-rein­
forced slab, serves also as the reinforcement for the concrete slab in a composite
role, using thicker wall sections with indentations or protrusions for shear connec­
tors (Fig. 4.36).

Fig. 4.36 Steel decking composite slab.

@Seismicisolation
@Seismicisolation
 SUMMARY 63

Fig. 4.37 Composite frame system.

Fig. 4.38 Composite frame and steel decking.

Slabs may also be designed to act compositely with the supporting beams by
the more usual forms of stud, angle, or channel shear connectors, so that the slab
alone spans the short distance between the beams while the compositely acting
slab and beam provide the supporting system (Fig. 4.37). The further combination
of a concrete slab on metal decking with shear connectors welded through to the
supporting beam or truss is an efficient floor system (Fig. 4.38).

SUMMARY

The structural form of a high-rise building is influenced strongly by its function,

while having to satisfy the requirements of strength and serviceability under all
probable conditions of gravity and lateral loading. Other influential factors include
the building's material of construction, its accommodation of services and, of
course, its overall economy. The taller a building, the more important it is eco­
nomically to select an appropriate structural form.
The basic structural forms of the first half of the twentieth century were the
braced frame, which is unrestricted in height but limited to steel structures, and

@Seismicisolation
@Seismicisolation
64 STRUCTURAL FORM

the rigid frame or the flat plate, which are economical to only about 25 stories in
height and appropriate particularly to concrete structures. These forms have now
been augmented by a variety of other forms that allow structures of greater effi­
ciency and height to be achieved in both steel and concrete. Advances have oc­
curred mainly in the use of shear walls, framed tubes, large-scale braced systems,
and space frames, and in better recognizing and accounting for the various types
of vertical and horizontal interaction between the major vertical components.
The single structural forms used in the vertically prismatic "modem" high-rise
buildings of the 1950s. 1960s. and early 1970s have given way to some extent to
hybrid, or mixed, forms in the less regularly shaped "postmodem" buildings of
the later 1970s and 1980s. In these mixed forms, combinations of two or more of
the single forms are used to fit the "postmodem" buildings' irregular shapes or
cut-outs.
Floors slabs are invariably of reinforced concrete. The most appropriate type
of floor framing system may depend on the material of construction of the building,
whether the building is for office use-requiring larger spans, or residential use­
allowing shorter spans, and whether the floor system is expected to participate in
the lateral load resistance of the building.
Reinforced concrete systems include one- or two-way spanning slabs on a sys­
tem of beams or beams and girders. Alternatively, .two-way spanning slabs or
waffle slabs with or without drop panels, and supported directly by columns. allow
the possibility of lesser floor depths and a nonuniform column grid.
Steel-framed floor systems consist of a slab, which may be of solid one-way
reinforced concrete, or of concrete on metal decking, or of precast concrete units,
supported by a one-, two- or three-way steel beam system.
Composite steel-concrete floor systems consist of a steel frame supporting either
a solid reinforced-concrete slab joined to the frame by shear connectors, or a con­
crete on steel decking slab with or without shear connectors joining it to the frame.

@Seismicisolation
@Seismicisolation
 CHAPTER 5

Modeling for Analysis

A building's response to loading is governed by the components that are stressed

as the building deflects. Ideally, for ease and accuracy of the structural analysis,
the participating components would include only the main structural elements: the
slabs, beams, girders, columns, walls, and cores. In reality, however, other, non­
structural, elements are stressed and contribute to the building's behavior; these
include, for example, the staircases, partitions, and cladding. To simplify the prob­
lem it is usual, in modeling a building for analysis, to include only the main struc­
tural members and to assume that the effects of the nonstructural components are
small and conservative.
To identify the main structural elements, it is necessary to recognize the dom­
inant modes of action of the proposed building structure and to assess the extent
of the various members' contribution to them. Then, by neglecting consideration
of the nonstructural components, and the less essential structural components, the
problem of analyzing a tall building structure can be reduced to a more viable size.
For extremely large or complex building structures, it may be essential to re­
duce even further the size of the analysis problem by representing some of the
structure's assemblies by simpler analogous components. This chapter reviews the
more usual approaches to analysis, the most commonly made assumptions, and
the principles and techniques employed in forming a model for structural analysis.

5.1 APPROACHES TO ANALYSIS

The modeling of a tall building structure for analysis is dependent to some extent
on the approach to analysis, which is in tum related to the type and size of structure
and the stage of design for which the analysis is made. The usual approach is to
conduct approximate rapid analyses in the preliminary stages of design, and more
detailed and accurate analyses for the final design stages. A hybrid approach is
also possible in which a simplified model of the total structure is analyzed first,
after which the results are used to allow part by part detailed analyses of the struc­
ture.

5.1.1 Preliminary Analyses

The purpose of preliminary analyses, that is analyses for the early stages of design,

@Seismicisolation
may be to compare the performance of alternative proposals for the structure, or

@Seismicisolation 65
66 MODELING FOR ANALYSIS

to determine the deflections and major member forces in a chosen structure so as

to allow it to be properly proportioned. The formation of the model and the pro­
cedure for a preliminary analysis should be rapid and should produce results that
are dependable approximations. The model and its analysis should therefore rep­
resent fairly well, if not absolutely accurately, the principal modes of action and
interaction of the major structural elements.
The simplifications adopted in making a preliminary analysis are often in the
formation of the structural model. Sometimes the approximation is large, as, for
example, when numerous hinges are inserted at assumed points of contraflexure
in the beams and columns of a rigid frame to convert it from a highly statically
indeterminate into a statically determinate system, thus allowing a simple solution
using the equilibrium equations. Or the approximation may be to assume a simple
cantilever to represent a complex bent, or that a bent is uniform throughout its
height and that its beams are "smeared" to allow a continuum solution. These are
just a few of the gross approximations that may be made in a structural model to
allow a relatively simple preliminary analysis to be achieved.
Alternative and sometimes additional simplifications adopted for the prelimi­
nary analysis concern the loading. For example, it is common to make an as­
sumption for the distribution, between the individual bents, of the total external
loading on the building, and then to analyze each bent in tum for its assumed
loading. In structures with different types of bents, this is a highly uncertain ap­
proach. Or, if a continuum analysis is to be made, it will be assumed that the load
is applied in some continuous distribution over the height of the structure rather
than as it really occurs, at discrete cladding connection levels.
Even with the gross approximations made in simplifying the structure and the
loading, it is generally expected that a preliminary analysis should give results for
deflections and main member forces that are dependably within about 15% of the
values from an accurate analysis.

5.1.2 Intermediate and Final Analysis

The requirement of intermediate and final analyses is that they should give, as
accurately as possible, results for deflections and member forces. The model
should, therefore, be as detailed as the analysis program and computer capacity
will allow for its analysis. All the major modes of action and interaction, and as
many as possible of the lesser modes, should be incorporated. Except where a
structure is symmetrical in plan and loading, the effects of the structure's twisting
should be included.
The most complete approach to satisfying the above requirements would be a
three-dimensional stiffness matrix analysis of a fully detailed finite element model
of the structure. The columns, beams, and bracing members would be represented
by beam elements, while shear wall and core components would be represented
by assemblies of membrane elements.
Certain reductions in the size or complexity of the model might be acceptable
while allowing it to still qualify in accuracy as a final analysis; for example, if the

@Seismicisolation
@Seismicisolation
 5.2 ASSUMPTIONS 67

structure and loading are symmetrical, a three-dimensional analysis of a half-struc­

ture model, or even a two-dimensional analysis of a fully interactive two-dimen­
sional model, would be acceptable. Or, if repetitive regions up the height of a
structure can be simplified by a lumping technique, this also would be acceptable.
In contrast to the reductions above, however, certain final analyses may require
separate, more detailed analyses of particular parts, using the forces or applied
displacements from the main analysis, for example, in deep beams at transition
levels of the structure, or around irregularities or holes in shear walls.

5.1.3 Hybrid Approach to Preliminary and Final Analyses

If the three-dimensional analysis of a fully detailed model of a structure presents

too formidable a task of bookkeeping or computation, an alternative might be to
use a hybrid, two-stage approach that would serve dually for the preliminary and
final analyses.
When a structure consists of bents that are representable by simple equivalent
cantilevers, a three-dimensional model of the structure can be formed by an as­
sembly of the cantilevers that can be analyzed to find the approximate deflections
and bent loadings. Detailed, two-dimensional models of the individual bents sub­
jected to the determined loadings are then analyzed individually to find the member
forces and to allow the member sizes to be adjusted. This first cycle of the overall
three-dimensional and individual two-dimensional analyses would be considered
preliminary.
In the second cycle, the three-dimensional model, with the cantilevers modified
to represent the adjusted bents, would then be analyzed to obtain the corrected
bent loadings. These bent loadings would then be used to reanalyze the two-di­
mensional bents to obtain the final member forces and structure deflections.
Using this two-stage procedure, a single large three-dimensional analysis of a
detailed model can be avoided, and replaced by a number of simpler analyses.
All approaches to analysis call for a sound understanding of high-rise structural
behavior and a knowledge of modeling techniques. The hybrid approach in partic­
ular requires special care in forming the three-dimensional cantilever model to
obtain reliable results. An understanding of high-rise behavior and modeling is
valuable not only for analysis, but also for deciding on and developing the struc­
tural forms of proposed tall buildings.

5.2 ASSUMPTIONS

An attempt to analyze a high-rise building and account accurately for all aspects
of behavior of all the components and materials, even if their sizes and properties
were known, would be virtually impossible. Simplifying assumptions are neces­
sary to reduce the problem to a viable size.
Although a wide variety of assumptions is available, some more valid than
others, the ones adopted in forming a particular model will depend on the arrange-

@Seismicisolation
@Seismicisolation
68 MODELING FOR ANALYSIS

ment of the structure, its anticipated mode of behavior, and the type of analysis.
The most common assumptions are as follows.

5.2.1 Materials

The material of the structure and the structural components are linearly elastic.
This assumption allows the superposition of actions and deflections and, hence,
the use of linear methods of analysis. The development of linear methods and their
solution by computer have made it possible to analyze large complex statically
indeterminate structures.
Although nonlinear methods of analysis have been and are still being devel­
oped, their use at present for high-rise buildings is more for research than for the
design office.

5.2.2 Participating Components

Only the primary structural components participate in the overall behavior. The
effects of secondary structural components and nonstructural components are as­
sumed to be negligible and conservative. Although this assumption is generally
valid, exceptions occur. For example, the effects of heavy cladding may be not
negligible and may significantly stiffen a structure; similarly, masonry infills may
significantly change the behavior and increase the forces unconservatively in a
surrounding frame.

5.2.3 Floor Slabs

Floor slabs are assumed to be rigid in plane. This assumption causes the hori­
zontal plane displacements of all vertical elements at a floor level to be definable
in terms of the horizontal plane rigid-body rotation and translations of the .floor
slab. Thus the number of unknown displacements to be determined in the analysis
is greatly reduced.
Although valid for practical purposes in most building structures, this assump­
tion may not be applicable in certain cases in which the slab plan is very long and
narrow, or it has a necked region, or it consists of precast units without a topping.

5.2.4 Negligible Stiffnesses

Component stiffnesses of relatively small magnitude are assumed negligible.

These often include, for example, the transverse bending stiffness of slabs, the
minor-axis stiffness of shear walls, and the torsional stiffness of columns, beams,
and walls. The use of this assumption should be dependent on the role of the
component in the structure's behavior. For example, the contribution of a slab's
bending resistance to the lateral load resistance of a column-and-beam rigid-frame
structure is negligible, whereas its contribution to the lateral resistance of a flat
plate structure is vital and must not be neglected.

@Seismicisolation
@Seismicisolation
 5.3 HIGH-RISE BEHAVIOR 69

5.2.5 Negligible Deformations

Deformations that are relatively small, and of Little influence, are ne­
glected. These include the shear and axial deformations of beams, the previously
discussed in-plane bending and shear deformations of floor slabs, and, in low- to
medium-rise structures, the axial deformations of columns.

5.2.6 Cracking

The effects of cracking in reinforced concrete members due to flexural tensile

stresses are assumed representable by a reduced moment of inertia. The gross
inertias of beams are usually reduced to 50% of their uncracked values, while the
gross inertias of columns are reduced to 80%.

5.3 HIGH-RISE BEHAVIOR

A reasonably accurate assessment of a proposed high-rise structure's behavior is

necessary to form a properly representative model for analysis. A high-rise struc­
ture is essentially a vertical cantilever that is subjected to axial loading by gravity
and to transverse loading by wind or earthquake.
Gravity live loading acts on the slabs, which transfer it horizontally to the ver­
tical walls and columns through which it passes to the foundation. The magnitude
of axial loading in the vertical components is estimated from the slab tributary
areas, and its calculation is not usually considered to be a difficult problem. Hor­
izontal loading exerts at each level of a building a shear, a moment, and some­
times, a torque, which have maximum values at the base of the structure that
increase rapidly with the building's height. The response of a structure to horizon­
tal loading, in having to carry the external shear, moment, and torque, is more
complex than its first-order response to gravity loading. The recognition of the
structure's behavior under horizontal loading and the formation of the correspond­
ing model are usually the dominant problems of analysis. The principal criterion
of a satisfactory model is that under horizontal loading it should deflect similarly
to the prototype structure.
The resistance of the structure to the external moment is provided by flexure of
the vertical components, and by their axial action acting as the chords of a vertical
truss. The allocation of the external moment between the flexural and axial actions
of the vertical components depends on the vertical shearing stiffness of the "web"
system connecting the vertical components, that is, the girders, slabs, vertical dia­
phragms, and bracing. The stiffer the shear connection, the larger the proportion
of the external moment that is carried by axial forces in the vertical members, and
the stiffer and more efficiently the structure behaves.
The described flexural and axial actions of the vertical components and the shear
action of the connecting members are interrelated , and their relative contributions
define the fundamental characteristics of the structure. It is necessary in forming
a model to assess the nature and degree of the vertical shear stiffness between the

@Seismicisolation
@Seismicisolation
70 MODELING FOR ANALYSIS

vertical components so that the resulting flexural and axially generated resisting
moments will be apportioned properly.
The horizontal shear at any level in a high-rise structure is resisted by shear in
the vertical members and by the horizontal component of the axial force in any
diagonal bracing at that level. If the model has been properly formed with respect
to its moment resistance, the external shear will automatically be properly appor­
tioned between the components.
Torsion on a building is resisted mainly by shear in the vertical components,
by the horizontal components of axial force in any diagonal bracing members, and
by the shear and warping torque resistance of elevator, stair, and service shafts. If
the individual bents, and vertical components with assigned torque constants, are
correctly simulated and located in the model, and their horizontal shear connec­
tions are correctly modeled, their contribution to the torsional resistance of the
structure will be correctly represented also.
A structure's resistance to bending and torsion can be significantly influenced
also by the vertical shearing action between connected orthogonal bents or walls.
It is important therefore that this is properly included in the model by ensuring the
vertical connections between orthogonal components.
The preceding discussion of a high-rise structure's behavior has emphasized the
importance of the role of the vertical shear interaction between the main vertical
components in developing the structure's lateral load resistance. An additional
mode of interaction between the vertical components, a horizontal force interac­
tion, can also play a significant role in stiffening the structure, and this also should
be recognized when forming the model. Horizontal force interaction occurs when
a horizontally deflected system of vertical components with dissimilar lateral de­
flection characteristics, for example, a wall and a frame, is connected horizontally.
In constraining the different vertical components to deflect similarly, the connect­
ing links or slabs are subjected to horizontal interactive forces that redistribute the
horizontal loading between the vertical components. For this reason, in a tall wall­
frame structure the wall tends to restrain the frame near the base while the frame
restrains the wall near the top. Similarly, horizontal force interaction occurs when
a structure consisting of dissimilar vertical components twists. In constraining the
different vertical components to displace about a center of rotation and to twist
identically at each level, the connecting slabs are subjected to horizontal forces
that redistribute the torque between the vertical components and increase the torque
resistance of the structure.
Having assessed a proposed structure's dominant modes of behavior, the for­
mation of an appropriate model requires next a knowledge of the available mod­
eling elements and their methods of connection.

5.4 MODELING FOR APPROXIMATE ANALYSES

Approximate analyses are often made at the preliminary design stage to estimate
quickly a proposed structure's stiffness and hence its feasibility. They are also used

@Seismicisolation
@Seismicisolation
 5.4 MODELING FOR APPROXIMATE ANALYSES 71

to estimate the allocation of external loading between the bents to allow for more
detailed individual bent analyses.
The requirements of simplicity and rapidity for a preliminary analysis usually
call for large approximations in forming the model. An approximate analysis may
be a numerical analysis of a very simplified, discrete member model or, for certain
types of structure, the analysis may consist of a closed solution to the characteristic
differential equation of an equivalent continuum structure. Some approximations
used in these two types of model are now described. The accuracy of an approx­
imate solution depends on how closely the approximations made in forming the
model represent the real structure.

5.4.1 Approximate Representation of Bents

Bents consisting of shear walls or of moment-resisting frames can be modeled

approximately provided that the flexural and shear characteristics of the original
assembly are reproduced in the model.
An axially concentric tall shear wall (Fig. [Link]), consisting of relatively uni­
form regions, can be modeled by a column located on the centroidal axis of the
wall (Fig. [Link]). The column segments are assigned to have the inertias and shear
areas of the corresponding regions of the wall. If the centroidal axis of the wall is
not concentric, as in Fig. 5.2a, the analogous columns on the respective wall axes
should be connected by horizontal rigid arms (Fig. 5.2b). When using a column
to model a wall, the wall stresses are evaluated by applying the resulting column
moment and shear to the appropriate sectional properties of the wall.
A multibay rigid frame (Fig. 5.3a) can be modeled very closely with regard to
its lateral behavior by a single-bay rigid frame (Fig. 5.3b). The criteria for equiv­
alence are that the racking shear rigidity ( GA) as defined by the column and beam

Fig. 5.1 (a) Axially concentric shear

(a) (b) wall; (b) equivalent column.

@Seismicisolation
@Seismicisolation
72 MODELING FOR ANALYSIS

Rigid I3. AS3

arms I

(a) (b)

Fig. 5.2 (a) Axially eccentric shear wall; (b) equivalent column.

flexural inertias, the sum of the column inertias, I; and the overall flexural inertia,
lx, as defined by the column sectional areas, are at each level the same in the
equivalent single-bay frame as they are in the multibay frame. These properties
and their equivalence are discussed in Chapter 7. Rigid-frame and braced-frame
bents (Figs. 5. 3a and 5 .4a) whether single or multibay, can be represented in a
very approximate way by single-column models (Fig. 5.4b). In these, the shear
area of the analogous column is assigned to provide the same shear rigidity GA as
the racking shear rigidity ( GA) of the bent. Formulas for evaluating the racking
shear rigidities ( GA) for braced frames are given in Chapter 6. The flexural inertia
of the equivalent column is assigned to have the same value as the inertia of the

Story i l:I
1: I i, I , ( GA) -1-----t---+---1
gi ;

,
(a) (b)
Fig. 5.3 (a) Multibay rigid frame; (b) equivalent single-bay frame.

@Seismicisolation
@Seismicisolation
 5.4 MODELING FOR APPROXIMATE ANALYSES 73

Story i
I , (GA)
gi i

(a) (b)

Fig. 5.4 (a) Multibay braced frame; (b) equivalent column.

column areas about their common centroid in the braced or the rigid frame. In this
approximation, the single curvature flexure of the columns in the braced and rigid
frames, which usually has only a minor influence on the frames' overall behavior,
is neglected in the column model.
If a shear wall has beams connecting to it in-plane, causing it to interact verti­
cally, as well as horizontally, with another shear wall or with other parts of the
structure (Fig. 5.5a) the wall can be represented by an analogous "wide column."
This is a column placed at the wall's centroidal axis and assigned to have the wall's
inertia and axial area, and having rigid arms that join the column to the connecting
beams at each framing level (Fig. 5.5b). In this way the rotations and vertical
displacements at the edges of the wall are transferred to the connecting beams.

Shear wa11 Shear wall 2 Rigid frame Column Column Rigid frame
A
!1, Sl
A
\ I A A
,.
A A
1
I A A
52' 2
I
I
• l z· sz· z Sl' z·
•
� �
1--
1--- Rigid
arms T T �
1---
1---
1---
1---
1--
1--
1--

(a) (b)

Fig. 5.5 (a) Shear walls and frame joined by beams; (b) equivalent wide-column model.

@Seismicisolation
@Seismicisolation
74 MODELING FOR ANALYSIS

Nonplanar assemblies of shear walls that form elevator cores (Fig. 5.6a and b)
in structures that translate but do not twist under lateral loading, can be simulated
by a single column located at the shear center of the section and assigned to have
the principal second moments of area of the core section (Fig. 5.7a). If the struc­
ture twists as well as translates, and the core has an effectively closed, box-like
section, as in Fig. 5.6b, the single column should be additionally assigned the
torsion constant J of the core (Fig. 5.7b).
If the structure twists and translates, and the core walls form an I, U, as in Fig.
5.6a, or more complex open-section shape, warping torsional effects may be im­
portant, in which case it is possible to use a two-column model (Fig. 5.7c) to give
an approximate representation of all the bending and torsional properties. Details
of such a model are given in Chapter 13.

5.4.2 Approximate Modeling of Slabs

In-Plane Effects. In structures that do not depend on the transverse bending

resistance of slabs as part of their lateral load resisting system, the slabs are taken
to serve only as rigid diaphragms that distribute the horizontal loading to the vertical

t
Shear center axis --1
z

_..,v

(a) (b)

Fig. 5.6 (a) Open section nonplanar shear wall assembly; (b) closed section shear wall
assembly.

@Seismicisolation
@Seismicisolation
 5.4 MODELING FOR APPROXIMATE ANALYSES 75

f/
z Shear center
axis

...... _ I _,/Y

(a)

Fig. 5. 7 (a) Equivalent flexural column; (b) equivalent flexural-torsional column; (c)
equivalent two-column flexural-torsional-warping model.

elements and that hold the building plan in shape as the structure translates and
twists. The slab then serves to constrain the horizontal displacements of the vertical
components at each floor to be related to the horizontal two displacements and
rotation of the slab. In a three-dimensional analysis of a structure (Fig. 5.8a) the
in-plane rigidity of the slab can be represented at each floor by a horizontal frame
of rigid beams joining the vertical elements (Fig. 5.8b) or, if the computer program
includes a "rigid-floor" option for simulating a rigid in-plane slab, its usc is simpler
and more accurate.

Transverse Bending Effects. Flat plate structures, and structures with shear
walls coupled by slabs, employ the transverse bending stiffness of slabs as part of
the lateral load-resisting system, similar to the girders of a rigid frame, as well as
using the in-plane rigidity of the slabs to hold the plan shape of the building. In
modeling the structure, the bending action of a slab between in-line columns or
walls can be represented by a connecting beam of equivalent flexural stiffness (Fig.
5.9). This model will result in the correct horizontal deflections, and forces in the
vertical members, but it gives only the concentrated moments and shears applied
to the slabs. The inertia of equivalent connecting beams to represent the slab bend­
ing action is discussed in Chapter 7 and Appendix I.

@Seismicisolation
@Seismicisolation
 z
-1
z
Shear center t Shear center axis
aXlS �
y
___.,-v -- /
--

----X -._X

Two columns
in combination
represent
Ix' Iy' Jz'
w

,
�
..__ . /

� ----
(b) (c)

Fig. 5.7 continued.

Rigid
frames

t
(a)

Horizontal-plane
rigid frame
representing floor slab

Columns Columns representing

representing shear walls
rigid frames

(b)

Fig. 5.8 (a) Plan asymmetric structure; (b) representation of slab diaphragm action.

76
@Seismicisolation
@Seismicisolation
 5.4 MODELING FOR APPROXIMATE ANALYSES 77

Fig. 5.9 Equivalent beam representation of slab

bending action.

5.4.3 Modeling for Continuum Analyses

So far, all the considered approximations have been for discrete member models,
that is incorporating individual vertical and horizontal members, for solution by a
stiffness matrix analysis. For certain structures with relatively uniform properties
over the height, alternative continuum analogy models may be formed that can be
analyzed by a closed solution of the characteristic differential equation. In a con­
tinuum model, the horizontal slabs and beams connecting the vertical elements are
assumed to be smeared as a continuous connecting medium-a continuum-having
equivalent distributed stiffness properties. Although continuum methods are lim­
ited in their facility to represent variations of a structure over its height, they can
give very rapid approximate solutions and are valuable in providing a general un­
derstanding of a structure's behavior. Two examples of the types of structure that
can be solved using continuum techniques are a coupled wall and a wall-frame
structure (Figs. 5.10a and [Link]). In the coupled wall, the connecting beams are
represented by a continuum with equivalent bending and shear properties (Fig.
5.10b). In a wall-frame structure, the connecting links between the wall and the
frame are represented by a horizontally incompressible medium, while the beams
in the frame are smeared into the general shear property of the equivalent shear
column (Fig. [Link]).

Continuum with
equivalent flexural
Connecting and shear stiffness
Shear beams
walls .-----,r+,,.....,

(a)
rn (b)

Fig. 5.10 (a) Coupled shear walls; (b) equivalent continuum model.

@Seismicisolation
@Seismicisolation
78 MODELING FOR ANALYSIS

Shear wa11 c onnee 1ng

1 inks ( Rigl.d
frame
Equivalent
shear
column

Continous
-linking
media

.I.�.J:..l.
Fig. 5.11
(a) (b)

(a) Wall-frame structure; (b) equivalent continuum model.

5.5 MODELING FOR ACCURATE ANALYSIS

It is necessary for the intermediate and final stages of design to obtain a reasonably
accurate estimate of the structure deflections and member forces. With the wide
availability of structural analysis programs and powerful computers it is now pos­
sible to solve very large and complex structural models. Some of the more gross
approximations used for a preliminary analysis, such as representing braced frames
and rigid frames by single columns, are too approximate for a detailed analysis,
and they do not yield the detailed forces necessary for sizing and reinforcing the
individual members. The structural model for an accurate analysis should represent
in a more detailed way all the major active components of the prototype structure.
The principal ones are the columns, walls, and cores, and their connecting slabs
and beams.
The major structural analysis programs typically offer a variety of finite ele­
ments for structural modeling. As an absolute minimum for accurately representing
high-rise structures, a -
th ree dimensiona l program with beam elements and quad­
rilateral membrane elements (Fig. 5.12a and b) will suffice. Beam elements are
used to represent beams and columns and, by making their inertias negligibly small
or by releasing their end rotations, they can also be used to represent truss mem­
bers. Membrane elements, which are used for shear walls and wall assemblies,
should preferably include an incompatible mode option to better allow for the char­
acteristic in-plane bending of shear walls.
If truss elements (Fig. 5.12c), quadrilateral plate elements, (Fig. 5.12d), and
combined membrane-plate elements are also available, they can be used to advan­
tage in representing, respectively, truss members, slabs in bending, and shear walls
subjected to out of plane bending.
Some typical high-rise structural components and assemblies, and their repre­
sentation by finite elements, will now be discussed.

@Seismicisolation
@Seismicisolation
 5.5 MODELING FOR ACCURATE ANALYSIS 79

(b)

(a)

(c) (d)

Fig. 5.12 (a)'Beam element; (b) quadrilateral membrane element; (c) truss element; (d)
quadrilateral plate bending element.

5.5.1 Plane Frames

A plane rigid frame, which is probably the simplest assembly to be modeled, has
both its column and beam members represented by beam elements (Fig. 5.13).
Shear defonnations of the members are nonnally neglected except for beams with
a span-to-depth ratio of less than about 5. The results of the analysis include the
vertical and horizontal displacements, and the vertical plane rotations of the nodes,

Beam
/ ? elements

,II "t

..! ,1, ; ' J '

' Fig. 5.13 Rigid frame using beam elements.

@Seismicisolation
@Seismicisolation
80 MODELING FOR ANALYSIS

Truss
elements

Beam elements
(outer ends with
rotational releases)

Fig. 5.14 Braced frame using truss and beam elements.

together with the members' axial force, shear force, and bending moments. In a
braced frame (Fig. 5.14) the braces are represented by truss elements or small­
inertia beam elements, the columns by beam elements, and the beams by beam
elements with their end rotations released. The results for the truss elements give
axial forces only.

5.5.2 Plane Shear Walls

Similar to the modeling of walls for an approximate analysis, a tall slender shear
wall that is not connected by beams to other parts of the structure (Figs. 5.l a and
5.2a) can be modeled for an accurate analysis by a stack of beam elements (Figs.
5.1 b and 5.2b) located on the centroidal axis of the wall, and assigned to have the
principal inertia and corresponding shear areas of the wall. Shear walls connected
by beams to other parts of the structure (Fig. 5.5a) can be similarly represented
by vertical stacks of beam elements located on the centroidal axes of the walls with
rigid horizontal beam elements attached at the framing levels to represent t h e effect
of the walls' width (Fig. 5 .5b). In the case of a beam-connected wall, axial forces
will be induced in the wall, so it is necessary to assign to the analogous column
an axial area as well as an inertia and a shear area.
Walls that are not slender, or that have openings, cannot be well represented
by simple equivalent columns and are better represented by an assembly of plane­
stress membrane elements (Fig. 5.15a). Because the segments of a shear wall and
the membrane elements that are used to model it are subjected to in-plane bending,
incompatible mode elements that are formed to include this deformation invariably
give more accurate results, as well as allowing the use of rectangular elements of
much greater height-to-width proportions with acceptabl-y accurate results. The
results for a plane-stress element typically include the horizontal and vertical dis-

@Seismicisolation
@Seismicisolation
 5.5 MODELING FOR ACCURATE ANALYSIS 81

Shell or
membrane
plana r ,...
elemen ts

Beam elements

Links

Rigid beams

[>(
, , , , , -, ,

(a) (b)

Fig. 5.15 (a) Shear wall: membrane element model; (b) shear wall: analogous frame model.

placements of the nodes, and the vertical and horizontal direct stresses and shear
stresses at either the corners or the mid-sides of the element.
If the available structural analysis computer program does not include plane­
stress elements, a shear wall can be modeled alternatively using an analogous
frame, such as in Fig. 5.15b, which can be assembled entirely from beam ele­
ments. The stresses resulting from such a model are usually within 1 or 2% of
those from a membrane element model analysis. Details of an analogous frame are
given in Chapter 9.
Nonrectangular walls can be modeled using quadrilateral elements, and, if more
detailed stresses are required in a particular region of the wall, a finer mesh can
be used in that area, with quadrilateral elements being used to make the transition
(Fig. 5.16). For greater accuracy, quadrilateral elements should be proportioned
to be as close as possible to equal-sided parallelograms.
When modeled by membrane elements, shear walls with in-plane connecting
beams require special consideration. Membrane elements do not have a degree of
freedom to represent an in-plane rotation of their corners; therefore, a beam ele­
ment connected to a node of a membrane element is effectively connected only by
a hinge. A remedy for this deficiency is to add a fictitious, ftexurally rigid, aux­
iliary beam to the edge wall element, in one of the ways shown in Fig. 5.17. The
adjacent ends of the auxiliary beam and the external beam are both constrained to
rotate with the wall-edge node. Consequently, the rotation of the wall, as defined
by the relative transverse displacements of the ends of the auxiliary beam, and a
moment, are transferred to the external beam.

@Seismicisolation
@Seismicisolation
82 MODELING FOR ANALYSIS

Finer
mesh

Fig. 5.16 Nonrectangular shear wall with transition, represented by quadrilateral ele­
ments.

......--Au xiliary beam

_£_ ceonnecting beam

I1 Au xiliary beam
_.--
I
GCo
I
nnecting beam

Fig. 5.17 Connection of beams to mem­

brane element shear wall.

5.5.3 Three-Dimensional Frame and Wall Structures

The high-rise rigid frame structure has moment-resisting joints, and its columns
and beams are modeled by three�dimensional beam elements (Fig. 5.18). These
elements deform axially, in shear and bending in two transverse directions. and in
twist. Generally, therefore, they have to be assigned an axial area. two shear areas.
two flexural inertias, and a torsion constant. Often, however, shear deformations
of the columns and beams, and axial deformations of the beams, are assumed

@Seismicisolation
@Seismicisolation
 5.5 MODELING FOR ACCURATE ANALYSIS 83

Three dimensional
beam elements

Fig. 5.18 Three-dimensional rigid-frame model using beam elements.

negligible. These are usually allowed for by omitting the assignment of a shear
area and by assigning either a fictitiously large axial area, or constraints between
the axial displacements of the member ends. In addition, the torsional stiffness of
practically proportioned beams and columns is usually negligible, which is allowed
for by omitting the assignment of a torsion constant. The usual results of signifi­
cance are, therefore, the translations and rotations of the nodes, the shear forces,
bending moments and axial force in the columns, and the shear forces and mo­
ments in the beams.
Three-dimensional shear wall assemblies often form the most important major
lateral load-resisting components in a high-rise building. They occur variously in
multibranch open sectional shapes (Fig. 5 .19a), in effectively closed sections (Fig.
5.19b), and in beam-connected sections (Fig. 5.19c). Whether of closed or open­
section form, the principal actions of the individual walls in an assembly are in­
plane shear and flexure, and the principal interaction between the walls of an as­
sembly is vertical shear along the joints. Consequently, plane stress membrane
elements are highly suitable for modeling three-dimensional shear wall compo­
nents (Fig. 5.20a and b). Story-height wall-width elements give an acceptably
accurate representation for most purposes.
Plane stress elements alone are not adequate for modeling three-dimensional
wall systems because they lack the transverse stiffness necessary at orthogonal wall
connections to allow a stiffness matrix analysis of the problem. Nor, when used
alone, can plane stress membrane elements provide the out-of-plane rigidity re­
quired to maintain the sectional shape of the core, as it is held in reality by the in-

@Seismicisolation
@Seismicisolation
 CIO
�

(a) (b) (c)

Fig. 5.19 (a) Open section shear wall assembly; (b) partially closed section shear wall assembly;
(c) nonplanar walls connected by beams.

@Seismicisolation
@Seismicisolation
 Horizontal rigid frame Horizontally and vertically
of auxiliary beams rigid auxiliary beams

torsi on
column plane stress
elements

plane stress Real connecting beams

elements assigned to be
horizontally rigid

(a) (b)

Fig. 5.20 (a) Membrane element and auxiliary beam model; (b) model for beam-connected shear
walls.

@Seismicisolation
@Seismicisolation
Q)
Ul
86 MODELING FOR ANALYSIS

plane rigidity of the floor slabs. The remedy for these deficiencies is to add at each
nodal level a horizontal frame of fictitious, rigid auxiliary beams (Fig. 5.20a). If
any of the walls are connected in-plane to each other, or to other parts of the
structure, by beams, the auxiliary beams adjacent to the wall edges can be made
vertically rigid also, to cause the transfer of moment (Fig. 5.20b) as described in
Section 5.5.2.
Another action, which would automatically be accounted for if shell elements
were used for the model, but not in the case of plane stress elements, is the tor­
sional stiffness corresponding to twisting of the walls. Although this is usually
relatively insignificant, in open-section wall assemblies it can be important and
should be incorporated. It is introduced by adding to the model a fictitious column
located on any one of the vertical sets of nodes (Fig. 5.20a) and assigning it a
torsion constant with a value equal to the sum of the individual walls' torsion
constants, as discussed in Chapter 13. The axial area and inertia of the column are
assigned to be zero.
An alternative way of representing beams connecting shear walls in the same
plane is to represent them by story-height membrane elements with a vertical
shearing stiffness equal to the vertical-displacement stiffness of the represented
beam, as shown in Fig. 5.21. In such a model, auxiliary beams are still required
to form a horizontally rigid frame around each level of the wall assembly, but the
beams adjacent to the openings do not have to be vertically rigid.

Membrane elements
representing connecting
beams

Membrane elements
representing walls

Fig. 5.21 Beam-connected wall assemblies: membrane elements representing beams.

5.5.4 P-Delta Effects

Second order P-Delta effects of gravity loading can be included in a single first­
order computer analysis of the structure by adding to the first-order model a ficti­
tious column with a negative stiffness.
The translational P-Delta effects in a nontwisting structure can be incorporated
in the two-dimensional model by adding a shear column, connected to the model
by rigid links at the framing levels (Fig. 5.22a). The column is assigned a negative

@Seismicisolation
@Seismicisolation
 5.5 MODELING FOR ACCURATE ANALYSIS 87

Actual
str ucture
model
(
Axially rigid

�- links
Actual
structure
Axially rigid
links
model

h
�

_ _ __ __,._- ---.._
-<:l! =fn
I
--¢1
I -<:�:
Negative
-�
:v
Negative
shear area I
..-� / inertia
column ��
...._;.--- co 1 umn
-01
• -<FT-:ii
�-=r-s.
-01
I
--q
• �
�=r�
- <II
I

-gr"T3-
-01
I
-d
I -<fi-7-=:a
i
' I '

(a) (b)
Fig. 5.22 (a) P-Delta negative shear column model; (b) P-Delta negative inertia column
model.

shear area to simulate the lateral softening of the structure due to gravity loading.
The column is assigned to be rigid in flexure. Alternatively, the same result can
be achieved by using a flexural column with its rotation restrained at the framing
levels (Fig. 5.22b) and its inertia assigned a negative value. The column is spec­
ified to be rigid in shear. The resulting deflections and member forces in the model
then include the P-Delta effects of gravity loading. Details of the technique are
given Chapter 16.
When making a full three-dimensional analysis of an asymmetric structure, the
P-Delta effects of twisting, as well as of translating parallel to the building's major
axes, can also be represented in the model by a fictitious negative stiffness column.
The column is located in each story at the centroid of the resultant gravity loading
acting through the story, and is assigned to have either negative shear areas, or
negative inertias, as described before, corresponding to the directions of the build­
ing's two major axes. The column is additionally assigned a negative torsion con­
stant to allow for the twisting P-Delta effects. This technique also is discussed in
detail in Chapter 16.

5.5.5 The Assembled Model

By combining the previously described techniques, a complete three-dimensional

model can be formed for any high-rise structure consisting of a combination of
frames, walls, and cores with beam and slab connections.
If the bending resistance of the slabs contributes to the lateral load resistance
of the structure, it is usual to model the slabs by beams of equivalent flexural
stiffness connecting the vertical components. Although an even more accurate
model could be formed by representing each slab as an assembly of plate elements,
such a detailed representation would vastly increase the size of the problem.

@Seismicisolation
@Seismicisolation
88 MODELING FOR ANALYSIS

In the complete detailed model, therefore, beam elements are used to represent
beams and columns, and story-height plane-stress membrane elements are used to
represent shear walls and cores. At all floor levels an auxiliary beam is added to
the top of each membrane element. The auxiliary beams, and the real beams, are
assigned extremely high axial areas and horizontal bending inertias in order to
simulate the rigid diaphragm effect of the slab. Auxiliary beams are also used at
each floor level to interconnect frames, walls and cores, as well as any isolated
columns. Where a real beam connects in plane with a wall, the auxiliary beam on
the connected wall element is assigned to be rigid in the vertical, as well as the
horizontal, plane so as to transfer moment between the wall and the external beam.
For each open section shear wall assembly, a vertical column assigned to have the
walls' torsion constant is added to the assembly.
The requirement for providing auxiliary beams, joining the columns, walls, and
cores to form a rigid horizontal diaphragm at each floor, and to connect shear walls
to beams in their planes, has been avoided in at least one tall building structure
analysis program [5.1].

5.6 REDUCTION TECHNIQUES

When the detailed model of a high-rise structure is so large and complex that its
analysis presents a formidable task of bookkeeping and computation, it may be
preferable to try to simplify the model, provided the accuracy of the results is not
seriously compromised. The following techniques are among those used to sim­
plify the model. Some of the techniques do not diminish at all the accuracy of the
analysis, while others, although losing a little in accuracy, are still good enough
for a final design analysis. The reductions are therefore applicable to both detailed
and to simplified models for anlaysis.

5.6.1 Symmetry and Antisymmetry

A structure that is symmetric in plan about the axis of horizontal loading (Fig.
5.23a) can be analyzed as a half-structure, to one side of the line of symmetry,
subjected to half the loads (Fig. 5.23b). The ends of the members cut by the line
of symmetry must be constrained to represent the omitted half of the structure.
That is, they must be constrained against rotation and horizontal displacement in
the plane perpendicular to the direction of loading, and against rotation about a
vertical axis, while simultaneously being free to displace vertically and to translate
in the direction of the loading. The results for the deflections and forces for the
analyzed half-structure will apply symmetrically to the corresponding nodes and
members in the omitted half-structure.
A structure that is symmetric in plan about a horizontal axis perpendicular to
the axis of horizontal loading (Fig. 5.24a) behaves antisymmetrically about the
axis of symmetry. In this case only half of the structure, to one side of the axis of
symmetry, and subjected to loads of half value, needs to be analyzed (Fig. 5.24b).

@Seismicisolation
@Seismicisolation
 Axis of symmetry
Axis of symmetry for
structure and loading
v
j/ /'/
/

Restrained
against x­
translation and
y-rotation

�: X

(a) (b)

Fig. 5.23 (a) Plan symmetric structure with symme1ric loading; (b) half-structure model.

@Seismicisolation
@Seismicisolation
00
co
 U)
Q

Axis of
Axis of anti-symmetrical
, behavior
--� --

--

Restrained against
vertical displacement

(a) (b)

Fig. 5.24 (a) Antisymmetrically behaving structure; (b) half-structure model.

@Seismicisolation
@Seismicisolation
 5.6 REDUCTION TECHNIQUES 91

The ends of the cut members are constrained on the line of symmetry to represent
their connection to the omitted antisymmetrically behaving other half of the struc­
ture. That is, they are constrained against vertical displacement, but are free to
rotate in the vertical plane parallel to the direction of loading. The values of the
results for the analyzed half-structure apply antisymmetrically to the omitted half­
structure.
Thus, if a structure is doubly symmetric on plan and subjected to horizontal
loading along one of its axes of symmetry, it can be analyzed by considering just
one-quarter of the structure, with appropriate constraints applied to the ends of
members cut on the lines of symmetry, to represent the symmetrical and antisym­
metrical aspects of behavior.

5.6.2 Two-Dimensional Models of Nontwisting Structures

The assumption that the floor slabs are rigid in plane, which permits the horizontal
displacements of all vertical elements at a floor level to be defined in terms of the
slab's horizontal translation and rotation, allows the possibility of representing a
three-dimensional structure by a two-dimensional model. An explanation of this
can be developed by first considering techniques for the planar representation of
nontwisting structures, and then extending them to twisting structures.

Symmetrical Structure Consisting of Parallel Bents. A structure that is

symmetric on plan and symmetrically loaded does not twist. Adding to this the
assumption of the slab's in-plane rigidity means that the horizontal displacements
of all the vertical components at a floor level are identical. Now considering the
symmetrical structure in Fig. 5.25a, and allowing for symmetry by analyzing only
one-half of the structure, the identity of displacements at the floor levels can be
established in a planar model by assembling the bents in the same plane, in any
order and at an arbitrary spacing, as in Fig. 5.25b, and providing a horizontal
constraint between the bents at each level. The constraint can be formed in two
alternative ways. If the analysis program has a dependent node option, sets of
nodes, one in each bent, at the same level, can be assigned to have the same
horizontal displacements. If a dependent node option is not available, pairs of
nodes at the same level in adjacent bents may be joined by axially rigid pin-ended
links, as in Fig. 5.25b. The half-structure model is then subjected at the floor levels
to loads of half the value of the total load per level.
As far as the validity of the assumptions allow, the resulting moments, shears,
and vertical axial forces in the model will correctly represent those in the structure.
The shear in the slabs between bents must be found by considering the differences
between the shears in successive stories of each bent, and the relative plan location
of the bents. The axial forces in the beams and links of the model are not mean­
ingful because both the application of the loading and the horizontal connections
in the planar model do not properly represent their on-plan locations in the real
structure.

@Seismicisolation
@Seismicisolation
92 MODELING FOR ANALYSIS

Axis of symmetry

L----/
B D F I
•
�
I
•

I I
II
II
II
II
II
• I
II
II
II
•
I
II
" II
II II II "

I I
,, II II II
II
•
II
•
I' II
•
II
•
A c I
t Load
resultant

(a)

A
�

Half- �
loading
--

_,.,

Half-structure model

(b)

Fig. 5.25 (a) Symmetric structure with parallel bents; (b) equivalent two-dimensional
model.

Symmetrical Structure with Connected Orthogonal Bents. Structures

that consist of an orthogonal system of connected bents, which are symmetrically
located about the axis of horizontal loading, as in Fig. 5.26a, can be modeled for
analysis by an extension of the planar modeling technique described above.
Considering half the structure, and assuming that, perpendicular to their planes,
the bents have negligible stiffness, the structure's shear resistance in the direction
of loading is provided by bents AB and CD, as they displace horizontally in their
planes parallel to the direction of loading (Fig. 5.26a). Bents AE and BF, perpen­
dicular to the loading, do not displace horizontally in their planes, but interact
vertically with bents AB and CD along their vertical lines of connection A, B, C,
and D. This vertical interaction causes the perpendicular bents to act as "flanges"
to the parallel bent "webs," as part of the structure's overall flexural action.
In the equivalent half-structure planar model, half the parallel bents and the
perpendicular half-bents are assembled in plane, with the parallel bents in one

@Seismicisolation
@Seismicisolation
 !�Axis of
r-symmetry
'
B D F
===-: -=·=-== "::::=-=== •=:=

20 :: 15' :: ' :: II
� lO' I �II
II I I
20 :: 15' :: :: ::
t �
•===•= :�=== ===��=·::::.:1:
1 ::=11
A
1.22' 122' lc22' f
(a)

Axially rigid
( link
A B j � v n � u u
-+ +> r
... "'
Half- I 0 " ...
-- - .<= 0 t
"' ·�
load I +>
.... - _.
a;;:; V'l
t
u E
� - .....
> ..... 10
+> .....
.. ... l
"' .c
_.. - 0::4-
...
........ -� t
.. ·� �
-- +> +>
t-- t
-. �� t
� 3Q_ 4 45 46 47 48 49 50 -- 51 52 53 54... 30 '::1.: 55 56 57 .58.� __lQ
.....
2Q.:_ 29 ,/"'30 31 32-:/-'\� =� - 5�� :in-- '\7 "0. 38 39·;
e- 3 ·�\ 20'-1" 40 41 42 43:; 20
lQ_:__ 14 15 16 17 18 � 20 21 22 23 24 -
10'-1" 25 26 27 28 _j_Q
..... -�
.,
1 2 3 4 5 6 7 8 9 10 11 12 13

@Seismicisolation
@Seismicisolation
_j � � � � � _j j _J
59' 69' 84'
�0' �20' �40' *�44' _jj 0'-1" 22'-1" 44'-1" 66'-1" 40' 62' 84' 106'
Bents parallel to wind Bents perpendicular to wind
(b)
co
w
Fig. 5.26 (a) Symmetric structure with orthogonal interacting bents; (b ) equivalent two�dimen­
sional model.
94 MODELING FOR ANALYSIS

group and the perpendicular half-bents in another (Fig. 5.26b). A column at the
intersection of orthogonal bents appears twice in the planar model, once in a par­
allel bent and once in a perpendicular bent. In each bent the column is assigned
an inertia appropriate to its bending in the plane of that bent. So that the axial area
of an intersection line column is not represented twice, it is arbitrarily assigned
entirely to the column in the parallel bent with a zero area assigned to that in the
perpendicular frame. The nodes in the model are numbered so that those on the
vertical lines of intersection, which are represented twice, are assigned two differ­
ent numbers, as in Fig. 5.26b.
The identical horizontal displacements of the parallel bents are established in
the model as before, either by using the dependent node facility or by including
fictitious axially rigid links, as between Band C in Fig. 5.26b.
The compatibility of vertical displacements between the parallel and perpendic­
ular bents may also be achieved in alternative ways. If a dependent node option is
available, vertical compatibility can be established by constraining the connection
nodes that are duplicated in the parallel and horizontal bents to have the same
vertical displacements. The zero horizontal in-plane displacement of the perpen­
dicular bents is arranged by constraining horizontally at least one vertical line of
nodes in each of those bents.
If a dependent node option is not available, there are two alternative ways of
using fictitious members to establish the connection in the planar model. The first
is to dimension the model horizontally so that the vertical intersection lines on
each perpendicular frame, as for example lines A and lines C in Fig. 5.26b, are
located immediately adjacent, say as close as 11200 of the adjacent span, to the
duplicate intersection lines of the connected "parallel" frames. Each pair of du­
plicated connection nodes is then joined by a very stiff horizontal beam with a
horizontal and rotational release at one end, as, for example, nodes 32 and 38 in
Fig. 5.26b. The alternative is to dimension the model horizontally so that the
vertical connection lines on the perpendicular frames are, in effect, coincident with
those on the parallel frames, as, for example, lines B and lines D in Fig. 5.26b,
and to dimension them vertically so that the connection nodes on the perpendicular
frame are displaced upward slightly, say I II 00 story height, from the correspond­
ing nodes on the parallel frame. Each pair of duplicated connection nodes, as, for
example, nodes 46 and 55 in Fig. 5.26b, is then joined by a vertical axially rigid
link. In either of these ways, the fictitious links establish vertical compatibility,
while avoiding horizontal interaction and vertical plane rotational interaction be­
tween the orthogonal bents.
The technique can be used for structures whose bents consist of walls, or frames,
or combinations of both.

5.6.3 Two-Dimensional Models of Structures That Translate and

Twist

The common assumption for analysis, that the floor slabs are rigid in their planes,
implies that for an arbitrary origin and a pair of axes parallel to the orthogonally

@Seismicisolation
@Seismicisolation
oriented bents of a laterally loaded structure (Fig. 5.27) the resulting displaced
 5.6 REDUCTION TECHNIQUES 95

y
CD Displaced in x direction to final position

Final Displaced in direction

position y

Initial position

Fig. 5.27 Displacements of bending and twisting structure.

location of any floor slab can be defined in terms of the rotation of the slab about
the origin, and two displacements parallel to the axes. Further, for the horizontal
equilibrium of any slab, the external X- and Y-direction forces on the slab and their
combined moment about the vertical axis through the origin must be in equilibrium
with, respectively, the X- and Y-direction resultants of the reactions from the bents
and their resultant moment about the origin.
Assuming that the structure consists of a plan-asymmetric system of orthogonal
bents that are stiff in their planes but have zero transverse and torsional stiffnesses
(Fig. 5.28a), a two-dimensional model can be formed to satisfy the above condi­
tions of displacement and equilibrium, as follows.
First, select an arbitrary origin 0 (Fig. 5.28a) that is located to the left of and
below the lower left-hand comer of the structural plan. Bents AB and CD are
parallel to, and at distances x1 and x2 from the Y axis, while the orthogonal bents
AC and BD are parallel to, and at y1 and y2 from the X axis.
Next, form the two-dimensional model by assembling all the bents in the same
plane with the X-direction bents in one group and theY-direction bents in the other,

,tr,l,
y

�
II
II 0'

..,1
• "
X

L:::
'

---
y
1 ..__A_ ___=_=_-
_-_-_-=-_=_ ..C::...
..; _..
._ _P_l_ X

o xz --!..._ I•
�-..
(a)

@Seismicisolation
Fig. 5.28

@Seismicisolation
(a) Plan of nonsymmetric structure.
 U)
Q)
145
h 155
hI I
x Y2
z

p
D � I
nx�.K
149 158B

20
p1 p2
.c 12 13
31 ll 14 5 16 17 114
1�*14� 1�1
�*�
--- -------.,.-
'---__J
�----�r-
Y-direction bents X-direction bents
* indicates arbitrary space between bents
=

@Seismicisolation
@Seismicisolation
h1 1st story height

(b)
Fig. 5.28 (b) Equivalent planar model.
 5.6 REDUCTION TECHNIQUES 97

"-=----:.--,
II
11 II
Level
•
i
II
II II

l-:.-::.-=:.:.1
p
lX

t piy

Fig. 5.28 (c) Equivalent loading on model.

as shown in Fig. 5.28b. To make the viewed faces of the bents consistent with the
location of the origin as specified above, the bents are displayed in the model (Fig.
5.28b) as viewed in Fig. 5.28a looking negatively along the X and Y axes, re­
spectively (i.e., A to the left of B in bent AB, and C to the left of A in bent CA).
This is to ensure that a horizontal plane rigid body rotation of a slab about the
origin 0, in Fig. 5.28a, corresponds to all the bents moving in the same direction
in the planar model (Fig. 5.28b). For example, a counterclockwise rotation of the
slab about 0 in Fig. 5.28a corresponds to a rightward displacement of the bents in
the model of Fig. 5.28b.
Because the bents are shown separately from each other, the columns on lines
of intersection of orthogonal bents will appear twice in the model, as, for example,
column A appearing in each of bents AB and CA. The flexural inertias and axial
areas of the duplicated columns are assigned in the way described previously for
intersecting bent structures.
Establish on the model, for the left-hand edge of each bent and on the same
vertical line as the edge, a set of "governing" nodes, one node for each floor
level. Each governing node is located above its associated floor level by a height
equal to the distance on-plan of the bent from the X or Y axis to which it is parallel.
For example, governing node 141, for the top floor of bent AB, is on the vertical
line A at a height x1 above the top floor, while governing node 101, for the third­
to-top floor of bent DB, is on line D at a height y2 above the third-to-top floor
(Fig. 5.28b).
In Fig. 5.28b, the governing nodes are shown, for clarity, offset to the left from
the left-hand edges of their associated bents, but they are assigned horizontal co­
ordinates to locate them in the model on the same vertical line.
Now connect each governing node to its corresponding floor-level node by an
effectively rigid vertical arm, with a rotational release at the floor-level node. All
the nodes of the structure are then numbered in sequence, starting from left to right
across the base of the model, and then level by level upward, including the gov­
erning nodes, as shown in Fig. 5.28b.
Consider now, for example, the top levels of bents AB and CD in the model.

@Seismicisolation
@Seismicisolation
98 MODELING FOR ANALYSIS

By constraining governing nodes 141 and 145 to displace horizontally together,

using the dependent node option, a horizontal translation without rotation of gov­
erning node 141, and hence of 145, will cause through their connecting arms a
translation in the Y direction of the whole top floor. Similarly, in bents CA and
DB, constraining governing nodes 150 and 155 to translate identically means that
a horizontal translation without rotation of governing node 150, and hence of 155,
will cause a translation in the X direction of the top floor. Further, by constraining
nodes 141, 145, 150, and 155 to rotate together means that a rotation without
translation of node 141, and hence of 145, 150, and 155, will cause, through their
connecting arms, the top floors of the bents in the planar model to translate with
the same relative displacements as they would in rotating in plan as a rigid body
about the origin 0 in Fig. 5.27. Specifying similar translational and rotational con­
straints between the governing nodes for each of the other floor levels will cause
the horizontal displacements of the planar model to properly represent in-plane
displacements of each level of each bent, due to the translations and rotations of
the structure.
Considering again the top levels of bents AB and CD in the planar model, their
horizontal reactions will be transmitted to governing nodes 141 and 145 as hori­
zontal forces and vertical plane moments, having the same magnitudes, respec­
tively, as the Y-direction reactions, and the horizontal plane moments of those
reactions about the origin, of the top levels of bents AB and CD in the plan view
(Fig. 5.28a). Similarly, the horizontal reactions of bents CA and DB in the planar
model will be transmitted to nodes 150 and 155 as horizontal forces and vertical
plane moments having the same magnitudes as the X-direction reactions, and the
horizontal plane moments of those reactions about the origin, as the top levels of
bents CA and DB in the plan view (Fig. 5.28a).
Because nodes 141 and 145 are constrained to displace identically in the Y
direction, the sum of the Y-direction reactions Pny can be assumed to act at node
141. Similarly, the sum of the X-direction reactions P,r:c can be assumed to act at
node 150. Since nodes 141, 145, 150, and 155 are constrained to rotate together,
the sum of the moments from all the X- and Y-direction bents can be assumed to
occur at node 141. By this reasoning, the resulting horizontal reactions at govern­
ing nodes 141 and 150, and the resulting vertical plane moment of the reactions
at node 141, are the same as the resultant Y- and X-direction reactions, and their
resultant horizontal plane moment about the origin, in Fig. 5.28a.
When orthogonal bents in a structure intersect, as they do in Fig. 5.28a, the
vertical interaction along their vertical lines of intersection is an important factor
in the structure's behavior and must be represented in the model. For this, addi­
tional constraints have to be applied in the planar model to establish vertical com­
patibility between the orthogonal bents. In the model (Fig. 5.28b) these constraints
are on the vertical lines of intersection A, B, C, and D. The constraints are applied
using the dependent node option by assigning each pair of duplicated connection
nodes, for example, 16 and 28 at A, 18 and 32 at B, 20 and 25 at C, and 23 and
30 at D, to have the same vertical displacements.
Having formed the planar model, it remains only to transform the loading on

@Seismicisolation
@Seismicisolation
 5.6 REDUCTION TECHNIQUES 99

the structure into equivalent loads for application to the model. Referring to Fig.
5.28a, the load in the Y direction at level i, P;, .• acts at a distance x1,; from the
origin 0. This may be transformed into a forceP;" acting at the origin and a torque
P;rXp; (Fig. 5.28c). Similarly, the loadP;, in the X direction, at a distance y1,; from
the origin, can be transformed into a forceP;, at the origin and a torque -P;,y1,;.
These equivalent actions may be applied to the model (Fig. 5.28b) asP;, to one
of the governing nodes for level i of the Y-direction bents, P;, to one of the gov­
erning nodes for level i of the X-direction bents, and a torque P;,x1,; - P;,Y,; to
one of the governing nodes for level i of all the bents. For example, at the top,
nth level, P,.,. is applied to node 141, P,r is applied to node 150, and a counter­
clockwise torque M, equal toP,,.x,, - P,ry,, is applied to node 141. A similar
transformation of the loads at each other level to equivalent loads and a torque
about the origin, and their application to the corresponding governing nodes in the
planar model, will make the model ready for analysis. Note that loads that act in
the A-to-B and A-to-C directions of the bents on the plan of the structure are
applied in the A-to-B and A-to-C directions to the governing nodes of the planar
model. A horizontal plane counterclockwise torque on the plan of the structure is
applied as a vertical plane counterclockwise torque on the planar model.
A two-dimensional stiffness method analysis of the planar model subjected to
the transformed loads will yield results for deflections and member forces identical
to those from a full three-dimensional analysis of the structure, provided that in
the latter analysis the assumptions of the slabs' in-plane rigidity and the bents'
zero transverse and torsional stiffness are also adopted. If a structure includes a
core consisting of an assembly of shear walls, this can also be included in the
planar model by treating each individual wall of the core as a bent, representing it
by a stack of plane-stress finite elements, and assigning to it a set of governing
nodes, rigid arms, and constraints, as though it were just another bent.
A final necessary comment concerns the flexural stiffnesses to be assigned to
the connecting arms to cause them to behave as rigid. It is recommended that each
should be assigned an inertia such that, if the arm were considered as a vertical
cantilever fixed at its governing node, its lateral stiffness at the lower end would
be of an order 1000 times greater than the estimated lateral stiffness of the bent at
the level where the ann connects. If the ann stiffnesses were assigned to be not
stiff enough, they would bend and not enforce proper translations on the bent,
whereas, if they were excessively stiff they could cause numerical instability in
the analysis.
An explanation of this modeling technique is given in Ref. [5.2].

5.6.4 Lumping

"Lumping" means the combination of several of a structure's similar, and simi­

larly behaving, components or assemblies of components into an equivalent single
component or assembly in order to reduce the size of the model for analysis. The
resulting forces in the equivalent component or assembly are subsequently distrib­
uted to give the forces in the original units.

@Seismicisolation
@Seismicisolation
100 MODELING FOR ANALYSIS

IAxis of
Three frames: Two walls 'symmetry
Members I '
f
A
f
I
w I
• •
II ,,
II II
II
• � I '
f
A
f
" II
II II
,, II
• •

t Load
resultant

(a)
One equivalent One equivalent frame

I
wa11 2I members 3If' 3Af
w
,
I�

(b)

Fig. 5.29 (a) Symmetric structure with repetitive bents; (b) equivalent lumped model.

Lateral Lumping. Consider as an example the symmetrical and symmetrically

loaded (and therefore nontwisting) structure in Fig. 5.29a, which consists of two
identical shear walls and three identical rigid frames. The walls can be lumped
laterally into a single wall, with twice the inertia of an individual wall, and the
frames lumped into a single frame with member properties three times those of an
individual frame. The lumped wall and frame can then be assembled as a planar
model (Fig. 5.29b) and analyzed relatively simply. The resulting forces in the wall
and frame of the lumped structure are divided by two and three, respectively, to
give the forces in the individual walls and frames. This simple lateral lumping
technique may be applied only to structures that do not twist, because the forces
in the bents of a nontwisting structure are independent of their lateral location.

Vertical Lumping. More usual examples of lumping occur in tall multistory

coupled-wall or rigid-frame structures in which the story heights and beam sizes
are repetitive, as in Figs. 5.30a and 5.3la. The detailed models can be simplified
by vertically combining groups of three or five beams into single beams, at the
middle beam location, and assigning to them the lumped properties of inertia and

@Seismicisolation
@Seismicisolation
 Walls represented
by membrane
Lumped Beams represented
element
U'S;;UII by membrane elements
Load: H H
- -- ....
H per
floor - � 11 s represented
-- � by membrane
elements
--
3H
�
} 3I
3H_ b
-

--
3H
�
}
� 1!!.
�
H
�
} - �

(a) (b) (c)

Fig. 5.30 (a) Coupled walls with repetitive beams; (b) equivalent lumped beam model; (c) equiv­
alent membrane element reduced model.

@Seismicisolation
@Seismicisolation
....
0
....
1 02 MODELING FOR ANALYSIS

--

l
}
}
(a) (b)

Fig. 5.31 (a) Rigid frame with repetitive beams; (b) equivalent lumped beam model.

shear area (Figs. 5.30b and 5.3lb). It is advisable to leave the bottom one or two
beams, and the top one or two beams, of the structure in their original locations
to better represent the localized effects at the base and the top.
In the case of lumping beams that connect shear walls, as in Fig. 5.30a, the
sectional properties of the membrane elements, or the analogous wide columns,
representing the walls would be the same in the lumped model (Fig. 5.30b) as in
the nonlumped model, because of the predominantly single-curvature behavior of
the walls. In a rigid frame, however (Fig. 5.3la), the predominantly story-height
double-curvature bending of the columns would require their inertias to be in­
creased in the lumped model with its increased story heights, to make the lateral
racking stiffnesses of the two models identical. The axial areas of the columns in
the two models would, however, be the same. The lateral loads are also lumped
and applied at the lumped beam levels. Details of this technique are given in Chap­
ter 7.
When coupled walls are being represented by membrane finite elements, a vari­
ation of the lumping technique is to represent sets of n successive connecting
beams, as well as the shear walls by n-story-height membrane elements (Fig.
5.30c). The wall elements are assigned the same sectional dimensions as the walls,
while the elements representing the beams are assigned a thickness to represent
the distributed vertical flexural and shear stiffnesses of the connecting beams. De­
tails of this technique are also given in Chapter 7.
The results for the member forces of a lumped model analysis must be inter­
preted to obtain the forces in the members of the original structure. The resulting
moment and shear in the original middle beam of a lumped set of n beams are one­
nth of the resulting values for the lumped beam. The forces in the other beams of
the original structure must be estimated by interpolation between the values ob­
tained for the middle beams above and below. The distribution of horizontal shear
between the vertical members at any level in the original structure will be in the
same ratio as between the corresponding members in the lumped model structure,

@Seismicisolation
@Seismicisolation
 5.6 REDUCTION TECHNIQUES 103

while the sum of shears will be equal to the external shear at that level. The mo­
ment at any level in a shear wall of the original structure will be given by the
moment at that level of the wall in the lumped structure, while moments in the
columns of a rigid frame will be given approximately by the product of the column
shear, determined as above, and the original half-story height.

5.6.5 Wide-Column Deep-Beam Analogies

It has been explained earlier how horizontally loaded shear walls connected by
beams, as in Fig. 5.32a, can be modeled by equivalent wide columns that consist
of a column on the centroidal axis of the wall, with rigid arms at the beam levels
to represent the effects of the walls' width (Fig. 5.32b). Some frame analysis pro­
grams include a rigid-end member option that includes the wide-column effects
and therefore allows the beam to be considered as a single member between the
column axes. If the available program does not have such an option, the rigid-end
beam may be simplified in the model to a full-span uniform beam with an increased
inertia to allow for the wide-column effects (Fig. 5.32c). An expression for the
increase in effective inertia, which is given in Chapter 10, is dependent on the
assumption of the wall cross sections rotating in-plane identically at the same level.
This is generally valid for coupled shear walls and for rigid frames with a pattern
of regularly spaced equally sized columns such as occur in framed-tube structures.
In rigid-frame systems with deep beams (Fig. 5.33a), the stiffening effect of the
beam depth on the columns can be represented by rigid vertical arms (Fig. 5.33b).
This also can be accommodated in an analysis by a rigid-end member option.
If the analysis program does not have a rigid-end member facility, however,
the rigid-end column can be replaced in the model by a uniform full-height column
between the beam axes (Fig. 5.33c) with modified stiffness properties to allow for
the deep beam effect. The inertia of the full-height column will be increased to
allow for the rigid-end effect by a factor that depends on whether the vertical

Equivalent uniform beams

\ Ib

Ic I
c

(a) (b) (c)

Fig. 5.32 (a) Coupled shear walls; (b) equivalent wide-column model; (c) equivalent uni­
form beam model.

@Seismicisolation
@Seismicisolation
104 MODELING FOR ANALYSIS

Rigid Equivalent
uniform co 1 umns

I'
c

(a) (b) (c)

Fig. 5.33 (a) Columns joined by deep beams; (b) equivalent deep beam model; (c) equiv­
alent unifonn column model.

Equivalent
uniform
columns

(a) (b) (c)

Fig. 5.34 (a) Wide-column, deep beam frame; (b) wide-column, deep beam model; (c)
equivalent unifonn member model.

members are deflecting primarily in single curvature, as would shear walls, or in

story-height double curvature, as would slender columns.
A frame combining wide columns and deep beams, such as a reinforced con­
crete frame tube (Fig. 5.34a), can be represented either by an analogous wide­
column deep-beam frame (Fig. 5.34b) or more simply by a frame of equivalent
full length beams and columns with appropriately increased stiffnesses (Fig. 5.34c).

SUMMARY

In modeling a structure for analysis it is usual to represent only the main structural
members and to assume that the effects of nonstructural members are small and
conservative. Additional assumptions are made with regard to the linear behavior

@Seismicisolation
@Seismicisolation
 REFERENCES 1 05

of the material, the in-plane rigidity of the floor slabs, and the neglect of certain
member stiffnesses and deformations, in order to further simplify the model for
analysis.
The extent to which a model will be simplified is related to the stage of analysis:
a simple model will be used for an approximate preliminary analysis, and a rela­
tively detailed one for a more accurate final analysis. In approximate modeling,
whole bents, which may be rigid frames, braced frames, shear walls, or cores,
may be reduced to equivalent single-column members, for a computer stiffness
matrix analysis. Or sets of connecting beams or links between major vertical com­
ponents may be represented by an equivalent continuous medium to allow a closed
solution of the governing differential equation.
In more accurate modeling, the columns and beams of frames will be repre­
sented individually by beam finite elements, while shear walls and cores will be
represented by assemblies of membrane finite elements. In cases where the trans­
verse bending of slabs is important, they will be represented by equivalent beams.
For an accurate solution, a computer analysis using a general structural analysis
program is usually accepted as the best method.
Certain reductions of a detailed model are possible while still producing an
acceptably accurate solution. These reductions include halving the model to allow
for symmetrical or antisymmetrical behavior, or representing the structures by a
planar model and conducting a two-dimensional analysis, or lumping similar frames
together in a nontwisting structure, or lumping vertically adjacent beams in a frame
or connected wall structure.
The ability to model high-rise structures successfully for analysis requires an
understanding of their behavior under load, while a good grasp of the techniques
of modeling serves in return as an aid in generally assessing a tall building's be­
havior, as well as assisting in the selection and development of structural forms
for tall buildings.

REFERENCES

5.1 ETABS, Three Dimensional Analysis of Building Systems. Computers and Structures
Inc., Berkeley, California, 1989.

5.2 Stafford Smith, B. and Cruvellier, M. "Planar Modelling Techniques for Asymmetric
Building Structures." Proc. lnst. Civil Engineers Part 2, 89, March 1990, 1-14.

@Seismicisolation
@Seismicisolation
 CHAPTER 6

Braced Frames

Bracing is a highly efficient and economical method of resisting horizontal forces

in a frame structure. A braced bent consists of the usual columns and girders,
whose primary purpose is to support the gravity loading, and diagonal bracing
members that are connected so that the total set of members forms a vertical can­
tilever truss to resist the horizontal loading. The braces and girders act as the web
members of the truss, while the columns act as the chords. Bracing is efficient
because the diagonals work in axial stress and therefore call for minimum member
sizes in providing stiffness and strength against horizontal shear.
Historically, bracing has been used to stabilize laterally the majority of the
world's tallest building structures, from the earliest examples at the end of the
nineteenth century to the present time. The Statue of Liberty, constructed in New
York in 1883, was one of the first major braced structures. In the following three
decades large numbers of braced steel-frame tall buildings were erected in Chicago
and New York. The 57-story, 792-ft-high, braced steel Woolworth Tower, com­
pleted in 1913, established a height record, which it held until the 77-story, 1046-
ft-high Chrysler Building and the 102-story, 1250-ft-high Empire State Building
(Fig. 6.1) were completed in 1930 and 1931, respectively.
One- or two-story-height bracing, as used generally in the earlier high-rise steel
structures, is an effective and still widely used arrangement. Recently, however,
a much larger scale form of bracing, traversing many stories and bays, has also
been used to considerable structural and architectural advantage in medium- and
high-rise buildings, thereby extending significantly the repertoire of bracing con­
cepts.

6.1 TYPES OF BRACING

Diagonal bracing is inherently obstructive to the architectural plan and can pose
problems in the organization of internal space and traffic as well as in locating
window and door openings. For this reason it is usually concentrated in vertical
panels or bents that are located to cause a minimum of obstruction while satisfying
the structural requirements of resisting the shear and torque on the building. In
many locations the type of bracing has to be selected primarily on the basis of
allowing the necessary openings through the bay, often at the expense of efficiency
in resisting the lateral forces.

106 @Seismicisolation
@Seismicisolation
 6.1 TYPES OF BRACING 1 07

Street level

Fig. 6.1 Empire State Building: typical braced bent.

In low- or moderate-rise buildings that are not particularly slender, it is usually

possible for the engineer to arrange the bracing in the structure without the archi­
tect having to consider it in planning the building. In a slender, moderate-rise
building or a truly high-rise building, the location of the lateral load-resisting bents
is more important and, indeed, might be all important to the viability of the struc­
ture. In such cases the architect and the structural engineer should liaise in the
early stages of design.
The most efficient, but also the most obstructive, types of bracing are those that
form a fully triangulated vertical truss. These include the single-diagonal, double­
diagonal, and K -braced types (Fig. 6. 2a-e). The full-diagonal types of braced bent
are usually located where passage is not required, such as beside and between
elevator, service, and stair shafts, which are unlikely to be relocated in the lifetime
of the building.
Other types of braced bent that allow window and door openings, but whose

@Seismicisolation
@Seismicisolation
 D

.D

108 @Seismicisolation
@Seismicisolation
 6.2 BEHAVIOR OF BRACING 109

arrangements cause bending in the girder, are shown in Fig. 6.2f-l. Some other
types, which introduce bending in both the columns and the girders, are shown in
Fig. 6.2m, n, and p. Generally, the types of braced bent that respond to lateral
loading by bending of the girders, or of the girders and columns, are laterally less
stiff and, therefore, less efficient, weight for weight, than the fully triangulated
trusses, which respond with axial member forces only.

6.2 BEHAVIOR OF BRACING

Because lateral loading on a building is reversible, braces will be subjected in tum

to both tension and compression; consequently, they are usually designed for the
more stringent case of compression. For this reason, bracing systems with shorter
braces, for example the K-types, may be preferred to the full-diagonal types. As
an exception to designing braces for compression, the braces in the double-diag­
onal system are sometimes assumed to buckle in compression, and each diagonal
is designed to carry in tension the full shear in the panel.
A significant advantage of the fully triangulated bracing types (Fig. 6.2a-e) is
that the girder moments and shears are independent of the lateral loading on the
structure. Consequently, the floor framing, which, in this case, is designed for
gravity loading only, can be repetitive throughout the height of the structure with
obvious economy in the design and construction.
In bracing systems in which the diagonals connect to the girder at a significant
distance from the girder ends, for example, those in Fig. 6.2c, d, e, and h, the
girder can be designed more economically as continuous over the connection, thus
helping to offset the cost of the bracing. A further advantage of this type of bracing
system is that the braces, in having one or both ends connected to the beam, which
is relatively flexible vertically, do not attract a significant load as the columns
shorten under gravity loading.
Eccentric bracing systems (i.e., systems in which the braces are not concentric
with the main joints) may be used to design a ductile structure for an earthquake­
resistant steel-framed building. The bracing acts in its usual elastic manner when
controlli�g drift against wind or minor earthquakes. In the event of an overload
during a major earthquake, the short link in the beam between the brace connection
and the column in Fig. 6.2f, g, k, and I, and the link in the beam between brace
connections in Fig. 6.2h, serves as a "fuse" by defonning plastically in shear to
give a ductile response of the structure. Such braced systems combine high elastic
stiffness and a large inelastic energy dissipation capacity that can be sustained over
many cycles.
The roles of the "web" members in resisting shear on a bent can be understood
by following the path of the horizontal shear down the bent from story to story.
Referring to Fig. 6.3 and considering four typical types of bracing subjected to
the total external shear, that is, neglecting the lesser effects of the horizontal forces
applied locally at the floor levels, the vertical transmission of horizontal shear can
be traced. In Fig. 6.3a the diagonal in each story is in compression, causing the
beams to be in axial tension; therefore, the shortening of the diagonals and exten-

@Seismicisolation
@Seismicisolation
11 0 BRACED FRAMES

(a)

(c) (d)

Fig. 6.3 Path of horizontal shear through web members. (a) Single-diagonal bracing; (b)
double-diagonal bracing; (c) K-bracing; (d) story-height knee bracing.

sion of the beams give rise to the shear deformation of the bent. In Fig. 6.3b, the
forces in the braces connecting to each beam end are in equilibrium horizontally,
with the beam carrying an insignificant axial load. In Fig. 6.3c half of each beam
is in compression and the other half in tension, whereas in Fig. 6.3d the end parts
of the beam are in compression and tension with the whole beam subjected to
double curvature bending. With a reverse in the direction of the horizontal load on
the structure the actions and deformations in each member of the bracing will also
be reversed.
The roles, if any, of the web members in picking up compressive force as the
structure shortens vertically under gravity loading can be traced similarly. As the
columns in Fig. 6.4a and b shorten, the diagonals are subjected to compression,
which can be developed because of the tying action of the beams. In Fig. 6.4c the
ends of the beams where diagonals are not connected are not stiffly restrained by
the columns' bending rigidity; therefore the beams cannot provide the horizontal
restraint that the diagonals need to develop a force. Consequently, the diagonals
will not attract significant gravity load forces. Similarly, in Fig. 6.4d the vertical
restraint from the flexural stiffness of the beam is not large; therefore, as in the
previous case, the diagonals experience only negligible gravity load forces. If the
type of bracing system allows the diagonals to attract compressive loading due to
gravity loading on the structure, the diagonals should be either designed to carry
the compressive forces or, to avoid backlash in the lateral load behavior of the
structure due to the braces having buckled, they must be detailed short and pre­
stressed in tension during erection.

@Seismicisolation
@Seismicisolation
 6.3 BEHAVIOR OF BRACED BENTS 111

(a) (b)

(c) (d)

Fig. 6.4 Path of gravity loading down bent. (a) Single-diagonal. single-direction bracing;
(b) double-diagonal bracing; (c) single-diagonal. alternate-direction bracing; (d) K-bracing.

6.3 BEHAVIOR OF BRACED BENTS

A braced bent behaves under horizontal loading as a vertical cantilever truss. The
columns act as the chords in carrying the external load moment, with tension in
the windward column and compression in the leeward column. The diagonals and
girders serve as the web members in carrying the horizontal shear, with the diag­
onals in axial tension or compression depending on their direction of inclination.
The girders act axially and, in some cases, in bending also.
The effect of the chords' axial deformations on the lateral deflection of the frame
is to tend to cause a "flexural" configuration of the structure, that is, with con­
cavity downwind and a maximum slope at the top (Fig. 6.5a). The effect of the
web member deformations, however, is to tend to cause a "shear" configuration
of the structure (i.e., with concavity upwind, a maximum slope at the base, and a
zero slope at the top; Fig. 6.5b). The resulting deflected shape (Fig. 6.5c) is a
combination of the effects of the flexural and shear curves with a resultant config­
uration depending on their relative magnitudes, as determined mainly by the type
of bracing.
In bents that are braced in a single bay, horizontal loading causes a maximum
tension at the base of the windward column of the braced bay. The more slender
the bay, the larger the tensile force. Depending on the tributary area of slab sup­
ported by the column, the tension will be partly or wholly suppressed by the dead
load of the structure. For height-to-width ratios of braced bays greater than about
10, however, the probability arises of uplift forces that are too large to handle. In

@Seismicisolation
@Seismicisolation
 c
.Q
tJ
<!.)
a::
<!.)
"0
� "0
<!.)
c:
:0
E
0
u

�
c:
.Q
tJ
<!.)
a::
<!.)
"0
...
"'
<!.)
� ..c:
"'

e
c:
.Q
tJ
<!.)
a::
<!.)
"0

""§
:::>
><
<!.)
u:
�
Ill
�
oil
�

ttttttttttttftt

112 @Seismicisolation
@Seismicisolation
 6.4 METHODS OF ANALYSIS 113

�
/
/
7
/
z
7
�
/
,
/
, , , I f ,, ,,,,,,,,,, ,

Fig. 6.6 Bracing in different bays of a bent.

multibay bents this problem can be avoided by placing successive story bracing in
different bays of the bent, as in Fig. 6.6. In this arrangement the column axial
forces caused by horizontal loading will be significantly smaller.
In providing for architectural requirements it is sometime necessary to use dif­
ferent types of bracing in different bays of the same bent, or in bays of different
parallel bents. This does not present a particular problem, except that care should
be taken to ensure that the laterai stiffnesses of the individual braced bays are
comparable. Combinations of full-diagonal or K-type braced panels, both of which
are usually very stiff in shear, with knuckle-type braced panels, which are usually
much less stiff, may prove unsatisfactory, because the stiff panels will attract an
unacceptably large proportion of the lateral load. In determining the individual
panel stiffness, the total height behavior of the braced panel should be considered.
This means that the lateral flexural flexibility due to axial deformations of the
columns, as well as the lateral shear flexibility due to deformations of the braces
and girders, should be taken into account.
In some situations, because of setbacks or transition levels, it is not possible to
locate the braces in a single vertical plane throughout the entire height of the struc­
ture. In these cases the shear can be transferred from the braced bents above the
setback or transition to those below by the horizontal-plane rigidity of the floor
slab or by horizontal bracing in the plane of the floor.

6.4 METHODS OF ANAL VSIS

6.4.1 Member Force Analysis

In the majority of modem design offices all but the simplest of braced high-rise
structures are now analyzed by computer using a frame analysis program. To re­
mind the reader of other possibilities, however, simple hand methods of analysis

@Seismicisolation
@Seismicisolation
114 BRACED FRAMES

that may be used for statically determinate, or certain low-redundancy, braced

structures will be reviewed.
An analysis of the forces in a statically determinate triangulated braced bent
can be made using the method of sections. For example, in the single-diagonal
braced panel of Fig. 6. 7, subjected to an external shear Q; in story i and external
moments M; and M; 1 at floor levels i and i - I, respectively, and assuming the
_

frame to be pin-jointed so that the members carry only axial forces, the force in
the brace can be found by considering the horizontal equilibrium of the free body
above section XX, thus,

F8c cos () = Q; ( 6.1)

hence,

Q;
Fs c = ( 6.2)
cos()
-­

The force F80 in the column BD is found by considering moment equilibrium

of the upper free body about C, thus

( 6.3)

hence

M;-J
Fso = -- ( 6.4)
L

while the force FAc in column AC is obtained similarly from the moment equilib­
rium of the upper free body about B, to give

M;

� A��--���8
Floor level i

I.
Fig. 6.7 Single diagonal braced panel.

@Seismicisolation
@Seismicisolation
 6.4 METHODS OF ANALYSIS 115

Floor level i
A B
Q
�
Story i
M(i-1) "
c D Floor level (i-1)

Fig. 6.8 Story-height knee-braced panel.

(6.5)

This procedure can be repeated for the members in each story of the bent.
The member forces in more complex types of braced bents can also be obtained
by taking horizontal sections. For example, in the story-height knee-braced bent
of Fig. 6.8, it could be assumed that the shear in story i is shared equally between
the braces. Then, from horizontal equilibrium of the upper free body,

- __lL__
FEC - 2
COS()
(6.6)

and, from moment equilibrium of the upper free body about D.

M_
; 1
(FAc +FEe sin O)L = (6.7)

from which

; 1
FAC = M
L - - FEC Sin
. () (6.8)

As the frame responds to horizontal shear, the girder in this type of bracing
system is subjected to bending throughout its length and to axial forces Q;/2 in
the lengths AE and FB. The bending is caused by the vertical components of the
forces in the braces, while the axial forces are caused by the horizontal compo­
nents.

6.4.2 Drift Analysis

In considering the deflected shape of a braced frame it is important to appreciate

the relative influence of the flexural and shear mode contributions, due to the col-

@Seismicisolation
@Seismicisolation
116 BRACED FRAMES

umn axial deformations and to the diagonal and girder deformations, respectively.
In typically proportioned low-rise braced structures, the shear mode displacements
are the most significant and, incidentally, will largely determine the lateral stiffness
of the structure. In medium- to high-rise structures, however, the higher axial
forces and deformations in the columns, and the accumulation of their effects over
a greater height, cause the flexural component of displacement to become domi­
nant. In a panel with single diagonal bracing and a height-to-width ratio of 8, the
total drift may be typically 60-70% attributable to the flexural component, with
the remainder due to the shear component. In knee-braced bents, in which lateral
loading subjects the girders-and in some arrangements the girders and columns­
to bending, as well as the braces to axial deformation, the proportion of the total
drift attributable to the shear component would be significantly greater.
The story drift, that is, the increment of lateral deflection in a story height,
which is often the limiting drift criterion and which in a braced bent is a maximum
at or close to the top of the structure, is more strongly influenced by the flexural
component of deflection. This is because the inclination of the structure caused by
the flexural component accumulates up the structure, while the story shear com­
ponent diminishes toward the top. Consequently, in a single-diagonal braced frame,
such as the one previously cited, the flexural component may contribute as much
as 95% of the top-story drift.
One virtue of a hand analysis for drift is that it easily allows the drift contri­
butions of the individual frame members to be seen, thereby providing guidance
as to which members should be increased in size to most effectively reduce an
excessive total drift or story drift.

Virtual Work Drift Analysis. In this method a force analysis of the structure
subjected to the design horizontal loading is first made in order to determine the
axial force P
1 in each member j, as well as the bending moment
along those members subjected to bending (Fig. 6.9a). A second force analysis is
M,1 at sections X

then made with the structure subjected to only a unit imaginary or "dummy"
horizontal load at the level N whose drift is required (Fig. 6.9b) to give the axial
force PJN• and moment m,JN at section X in the bending members. The resulting
horizontal deflection at N is then given by

LlN 2:; p-J.N (PL) + 2:; [Lj m ( )

Mx d.x
Jo XJ
.N
= ( 6.9)
EA 1 El 1

in which L1, A1, and 11 are the length, sectional area, and moment of inertia for
each member j, and E is the elastic modulus. The first summation in (6.9) refers
to all members subjected to axial loading, while the second refers to only those
members subjected to bending.
If the drift is required at another level, n, of the structure, another dummy unit
load analysis will have to be made, but with the unit load applied only at level n.
The resulting values Pin and mxJn will be substituted in Eq. (6.9) to give the drift.

@Seismicisolation
@Seismicisolation
 6.4 METHODS OF ANALYSIS 117

(a) (b)
Fig. 6.9 (a) Member forces due to design horizontal loading; (b) member forces due to
unit dummy loading.

The virtual work method is exact and can easily be systematized by tabulation.
An adequate assessment of the deflected configuration, the total drift, and the story
drifts can be obtained by plotting the deflection diagram from the deflections at
just three or four equally spaced points up the height of the structure, requiring
one design load force analysis plus three or four "dummy" unit load analyses.

Combined Moment-Area and Shear Formula Approximate Drift Analysis.

An approximate calculation of the drift can be made by using the moment-area
method to obtain the flexural component (i.e., the component resulting from col­
umn axial deformations) and by applying a shear deflection formula to calculate
the shear component. The method is appropriate for braced bents in which the
flexural mode stiffness is entirely attributable to the axial areas of the columns;
these include the majority of bracing types. It has the advantage that a detailed
member force analysis of the frame is not necessary; only the external moment
and the total shear force at each level are required.

Flexural Component. The procedure for obtaining the flexural component of

drift is to first calculate for the structure (Fig. 6.10a) the external moment diagram
(Fig. 6.10b). Then, to compute for the different vertical regions of the bent, the
second moments of area I of the column sectional areas about their common cen­
troid. For example, the value for the lower region of the braced bent in Fig. 6.1 Oa
is

@Seismicisolation
@Seismicisolation
 ....
....
Cll

Level N
-

Region 3
-
column
_...,. areas A3

� I \ I \ I ZN
_...,.
� -----

-
I \
- Region 2
column
- areas A2

-
�''"''"'d of � d''''''
- EI
- A area of _!1__ d1agram
Region 1 N
EI
---- column
areas A1
-
\"'""

@Seismicisolation
@Seismicisolation
(a) (b) (c)

Fig. 6.10 (a) Braced frame: approximate deflection analysis; (b) extemal load moment diagram;
(c) M / El diagram.
 6.4 METHODS OF ANALYSIS 119

(6.10)

The moment diagram and the values of I are used to construct an M I El diagram,
as in Fig. 6.10c.
The story drift in story i, O;r, due to the flexure of the structure, is then obtained
from

(6.11)

in which h; is the height of story i, and O;r is the inclination of story i, which is
equal to the area under the M I El curve between the base of the structure and the
mid-height of story i.
The total drift at floor n, due to flexure, is then given by the sum of the story
drifts from the first to the nth stories.

II

.6.11f = L; O;r ( 6.12)

I

Shear Component. The shear component of the story drift in story i, o;,, is a
function of the external shear and the properties of the braces and girder in that
story. The shear component of the total drift at floor level n, .6.11" is equal to the
sum of the story shear components of drift from the first to the nth stories, that is

II

.6,11s = L; 0 is (6.13)
I

Formulas for the shear component of the story drift, o;,, are given for various types
of braced bent in Table 6.1.
Having obtained the flexural and shear components of drift, the total drift at
level n is given by

(6.14)

6.4.3 Worked Example for Calculating Drift by Approximate

Methods

A IS-story single-diagonally braced frame (Fig. 6.11) consists of three 5-story

regions. It is required to determine the drift at floors 5, 10, and 15 (i.e., where
floor n is at the top of story n) for a uniform wind load of I 0 kips per story. Assume
2
the elastic modulus E = 4.2 x I 06 kip I ft .
The flexural and shear components of drift will be determined separately, as
follows.

@Seismicisolation
@Seismicisolation
120 BRACED FRAMES

TABLE 6.1 Braced Bents: Shear Deflection per Story

TYP E OF SHE A R DEFLECTION

DIME NSIONS PER STORY
BRACING

� [
65 =Q L+_L j
I
SINGLE
h
DIAGONAL E L2 A A
d 9
L

~
DOUBLE
DIAGONAL
s
6 =_Q_ L [
2E L 2A
d
J

K- BRACE

� s
[
6 =Q 2d' +-L-
E L2 A
d
4A
g
j

~
STORY HEIGHT o s= Q [ m
d 3 +-+ h2 ( L -2m)'
KNEE-BRACE E 2m 2A
--

d
2A
g
121 L
g
J

~ [
OFFSET s m2
6 =Q d' �
+h2
DIAGONAL
E(L-2m)2A
d
+
A
g
3I L
g
J
Q is the story shear
A is the sectional area of a diagonal
d
A and I are, respectively, the sectional area and inertia of the upper girder
g g
E is the elastic modulus

Flexural Component. Using Table 6.2 to record the steps of the computation:

1. Compute the moment of inertia of the column sectional areas about their
common centroid for each of the three height regions and record the values in
column 3.
In the frame under consideration the column areas are equal, therefore their
common centroid is mid-way between the columns

( L )2 A2cL2-
1=2XAc 2- =-
As an example, for the lowest region, stories 1-5, where Ac = 35 in.2

=AcL2 = XX2202 =
I
2
35
144
48.6 ft4

@Seismicisolation
@Seismicisolation
 6.4 METHODS OF ANALYSIS 121

5 kips
......
� Leve1 15

-t
A area of co1umn�10in2
10--
� � :
<1J
A area of diagona1�5in
2

10
......
2
'- A area of girder�30in
� g
V>

10
�

� A 20 in2
0 c
=

10
�

10
_.

10
�

Fig. 6.11 Example frame.

2. Compute the value of the external moment M at each mid-story level and
enter the values in column 4. For example, in story 12

M = I 0 (5 + 15 + 25) + 5 X 35 = 625 kip ft

3. Determine for each story the value of hM/EI, retaining E as a symbol, and
enter the result in column 5. These are the changes in inclination in each story i
due to flexure, OO;r·
For example, in story 5,

hM/EI = 005r = 10 X 5525/48.6£ = 1136.8/£

4. Determine for each story i the accumulation of 00;1, from story I up to and
including story i, 8;1 and record it in column 6.
For example, the accumulation of oB;r up to story 5 is

s M
l::h- (2165.6 + 1877.6 + 1610.1 + 1363.2 + 1136.8)/£ 8153.3/£
I
= =

EI

Such accumulated values give the inclination of each story i due to flexure, B;r·

@Seismicisolation
@Seismicisolation
122 BRACED FRAMES

TABLE 6.2 Evaluation of Flexural Components of Drift

Story Frame External Story Story

Drift
Height Inertia Moment Inclination Drift
h; I; M; oB; 8;r O;r Eo;r a"r
Story (ft) (ft4) (k . ft) ( rads/E) (rads/E) (ft/E) (ft/E) (ft)

15 10 13.9 25 18.0 14942.8 149428 159735 0.380

14 10 13.9 125 89.9 14924.8 149248
13 10 13.9 325 233.8 14834.9 148349
12 10 13.9 625 449.6 14601.1 146011
II 10 13.9 1025 734.4 14151.5 141515

10 10 27.8 1525 548.6 13414.1 134141 862791 0.205

9 10 27.8 2125 764.4 12865.5 128655
8 10 27.8 2825 1016.1 12101.1 121011
7 10 27.8 3625 1304.0 11085.5 no855
6 10 27.8 4525 1627.7 9781.0 97810

5 10 48.6 5525 1136.8 8153.3 81533 270319 0.064

4 10 48.6 6625 1363.2 7016.5 70165
3 10 48.6 7825 1610.1 5653.3 56533
2 10 48.6 9125 1877.6 4043.2 40432
10 48.6 10525 2165.6 2165.6 21656

5. Record the product of h; and B;r in column 7. h;B;r is the drift in story i, O;r.
due to flexure.
For example, the drift in story 5 due to flexure is

o5r = lO X 8153.3/E = 81533/Eft

6. At each level where the value of the lateral drift is required, evaluate the
accumulation of the story drifts, O;r. from story -1 up to the considered nth floor,
to give the drift -1nr due to flexure. Enter these in column 8.
For example, at floor 5:

.15r = (21656 + 40432 + 56533 + 70165 + 81533)/E = 270319/4.2

X 106 = 0.064 ft

Shear Components. Using Table 6.3 to record the steps of the computation:

1. Compute the value of the external shear Q; acting 'in each story i and enter
in column 2.
2. Compute for each story i the story drift due to shear, O;" by substituting the
value of the story shear and member properties into the appropriate formula from
Table 6.1. Record the resulting values of o;, in column 3.

@Seismicisolation
@Seismicisolation
 6.4 METHODS OF ANALYSIS 123

TABLE 6.3 Evaluation of Shear Components of Drift

Shear Story Drift Drift

Q 0;, .:lns
Story (kips) ( ft) ( ft)

15 5 0.0011 0.126
14 15 0.0032
13 25 0.0054
12 35 0.0075
II 45 0.0097

10 55 0.0065 0.099
9 65 0.0077
8 75 0.0089
7 85 0.0101
6 95 0.0113

5 105 0.0091 0.054

4 115 0.0100
3 125 0.0109
2 135 0.0117
145 0.0125

For example, the shear deflection formula for the single-diagonally braced ex­
ample frame is

and using this to compute the drift in story 8 due to shear

0
Ss
=
75

4.2 X 10
6
(22.363
2
X 144 +

20 X 10
20 X 144 )
30
=
0.0089 ft

3. Sum the story drifts due to shear up to and including stories 5, 10, and 15
to obtain the total shear drift at floor levels 5, 10, and 15, and record the values
in column 4.
For example, the drift due to shear at floor 5

+
0.0125 + 0.0117 + 0.0109 + 0.0100
= =

.6ss 0.0091 0.054 ft

Total Drift. The total drift at any floor level is the sum of the flexural and shear
drifts at that level; for example, the total drift at the top of the 15-story frame in
question is

@Seismicisolation
@Seismicisolation
124 BRACED FRAMES

d15 = dl5f + dl5s = 0.380 + 0.126 = 0.506 ft

A computer stiffness matrix analysis of the same structure gave the result d15
= 0.477 ft; hence, in this case the approximate hand method was +6.1% in error.
This error was probably due to the assumption implicit in the method of calculating
the flexural component of drift, that the axial forces in the two columns of any
particular story of a single-bay frame are equal in value. In the single-diagonally
braced frame considered, the column axial forces in each story do not have exactly
the same value. One is always smaller than assumed in the calculation, because of
the vertical component of the force in the bracing member; hence the deflection
calculated by the approximate method is larger.
Figure 6.12a shows the relative contributions of the columns', diagonals', and
girders' deformations to the drift of the example structure. It is evident that al­
though the diagonals have the largest influence on the drift in the lowest region,
the column axial deformations tend to dominate the drift further up the structure,
thus causing an overall flexural mode of behavior of the structure.
Figure 6.12b shows the relative contributions of the columns, diagonals, and
girders to the story drifts. In the upper part of the structure, the axial deformations
of the columns dominate the story drifts even more than they do the total drift.

6.5 USE OF LARGE-SCALE BRACING

Traditionally, and indeed currently, the typical arrangement of bracing in tall

building structures is in story-height, bay-width modules. In this form it is usually
possible to conceal the bracing within the walls or facade of a building to leave
little evidence of its being a braced structure.
Over the last two decades the high efficiency of bracing in resisting lateral load­
ing has been further exploited by using it on a larger modular scale, both within
the building and externally across the faces. In the latter form the massive diago­
nals have sometimes been emphasized as an architectural feature of the facade.
A simple and elegant example of the use of massive K-braced trusses in resist­
ing wind loading is in the 35-story Mercantile Tower in St. Louis, Missouri (Fig.
6.13). Four vertical trusses, each consisting of three-story height K-braced panels,
are aligned diagonally in plan across the cut-off comers of the building. Each pair
of vertical trusses at the ends of the building is joined by a rigid frame. The trusses
are also connected to a single-bay rigid frame on each of the wide faces to form a
stiff vertical U-section assembly at each end of the building. These provide resis­
tance to wind in both the transverse and longitudinal directions of the building.
The 27-story Alcan Building in San Francisco (Fig. 6.14) uses six-story height
panels of double-diagonal bracing between the main full-height columns on each
of the building's four faces. At each mid-panel crossover point the braces connect
to intermediate columns that rise from the first floor, transition girder level. In this
arrangement the braces serve several roles:

@Seismicisolation
@Seismicisolation
 15 �----,---�--� 15n-----,---,

I ,
i
r:
I /, r
/ L,
i J
�
10 � I 10
/ r· �
i. i
.I I , r.
i. J.
., ., r
: ,
>
>
I .,
., J
i.
, r ·
.._ .._
0 I
·I :;
0
0 0 I _j
i.
I 1 r
LL. due to girder deformations I
I ·J
5 5 r .. J
!:·; " diagonal
I
r
.
" column r .J
f.v total deflection
\ �
� r··j
..
I
�J :·
\r--J� .
1:
O L---�----�--L---�
0 1 2 3 4 5 6 0.1 0.2 0.3 0.4 0.5
Deflection (in) Deflection per story (in)

@Seismicisolation
@Seismicisolation
(a) (b)

Fig. 6.12 (a) Components of total drift. (b) components of story drift.
....
N
U1
 � �

��
r.l
IJ
IJ

�\ IJ

��
IJ
!) '

��. � :��
�
�
�
Fig. 6.13 Mercantile Tower, St. Louis, Missouri.

Fig. 6.14 Alcan Building, San Francisco.

126 @Seismicisolation
@Seismicisolation
 6.5 USE OF LARGE-SCALE BRACING 127

1. to carry the lateral shear on the building;

2. to mobilize the intermediate columns axially so that they participate with

the main columns in resisting the lateral load moment;
3. to shift gravity loading from the intermediate columns to the main columns
and thus reduce the load on the transfer girder.

The 100-story John Hancock Building in Chicago is a braced tube (Fig. 4.14).
In this hybrid form of structure the four rigid frame faces of the building are stiff­
ened by overall diagonal bracing. The rigid frames form a vertical tube-type can­
tilever in which the frames parallel to the wind act as the webs of the cantilever,
while the frames normal to the wind act as the flanges. The role of the bracing is
again multi-purpose in:

1. resisting the horizontal shear;

2. reducing the shear lag in the flange column axial forces and hence making
the whole cross-section of the building structure stiffer against horizontal
load bending;
3. helping to equalize the gravity load stresses in the columns.

An important consequence of the reduced shear lag in the flange frames of the
braced-tube structure is that the demand on the rigid-frame action is reduced so
much that the columns can be spaced further apart, and the spandrel beams can be
shallower than in unbraced tube structures, thereby allowing larger window open­
ings.
The 914-ft-tall Citicorp Building in New York City has a frame structure (Fig.
6.15), which, although completely concealed by cladding, depends heavily on di­
agonal members. The square plan tower is supported by a full-height central core
and four nine-story braced legs that are located under the middle of the tower faces.
Each braced leg supports a two-story transfer truss from the top of which a "ma­
jor" mast column extends in line with the leg to the top of the tower. "Minor"
columns are located at the comers and quarter points of the tower faces. The col­
umn system is K-braced by eight-story-high major diagonals that form chevron­
like eight-story tiers supported by the mast columns. Gravity loads are shared be­
tween the core and the outer frames. In the frames the load is transferred from the
minor columns to the mast column by the diagonals at eight-story intervals. Wind
shear is collected by the core over eight-story-height regions and transferred to the
braced outer frames at the base of each tier. At the base of the tower the entire
shear is transferred back to the core and hence to the ground. Wind moment is
carried mainly by the mast columns and legs in the faces normal to the wind, and
partly by the core. The unique structure of the Citicorp Building was developed to
satisfy a requirement for the building to overhang an existing church on the site.
Since the diagonals carry a significant part of the gravity loading, the structure
may be classified as either a space truss or a braced frame.

@Seismicisolation
@Seismicisolation
128 BRACED FRAMES

Fig. 6.15 Citicorp Building, New York City.

SUMMARY

A braced frame is an efficient structural form for resisting horizontal loading. It

acts as a vertical truss, with the columns as chords and the braces and girders as
web members. The most efficient type of bracing, using full diagonals, is also the
most obstructive to door and window openings. Other arrangements are available
that are more amenable to allowing openings but that, weight for weight, are less
stiff horizontally. The bracing arrangement is usually dictated by the requirements
for openings.
An advantage of some types of braced framing is that horizontal loading does
not contribute significantly to the girder forces; consequently, the girders can be
uniform over the height of the structure with economy in design and construction.

@Seismicisolation
@Seismicisolation
 REFERENCES 129

Some forms, in which the braces connect part way along the girder, allow the
girder to be designed for gravity loading as continuous over the brace connections,
again with resulting economy.
Braced bents deflect with a combination of flexural and shear components: the
flexural component results from the column axial deformations, and the shear com­
ponent from the brace and girder deformations. Low-rise structures deflect in a
predominantly shear mode while high-rise braced bents deflect in a predominantly
flexural mode.
Braced-frame member forces may usually be analyzed by the method of joints
or by the method of sections. To allow a statically determinate analysis, it is usu­
ally assumed either that the shear is shared equally between the tension and
compression braces, or that the compression brace has buckled and the tension
brace carries all the shear.
Deflections may be analyzed by hand, either exactly, using the virtual work
method, or approximately, using a combination of the moment area method and a
shear deflection formula. An advantage of the virtual work method is that it indi­
cates which members contribute most significantly to the deflection, therefore pro­
viding guidance as to which members should be adjusted to control the deflection.
Although bracing has been used typically in story-height bay-width modules, a
recent development for very tall buildings has been to incorporate it in larger scale,
multistory multibay arrangements. The effect of these has been to cause a more
integral behavior of the column-girder system in resisting both gravity and hori­
zontal loading, creating highly efficient structural forms for very tall buildings. In
some notable examples, the large scale bracing has been exposed on the buildings'
faces to give a characteristic architectural effect.

REFERENCES

6.1 Rathbun, J .C. "Wind Forces on a Tall Building." Proc. ASCE Paper 2056, Septem­
ber 1938, 1-41.

6.2 Morris, Clyde T. "Practical Design of Wind Bracing." Proc. Am. Jnst. Steel Const.,
October 1927.

6.3 Wind Bracing in Steel Buildings. Second Progress Report of Sub-Committee No. 31,
Committee on Steel of the Structural Division, Proc. A.S.C.E., February 1932, pp.
214-230.
6.4 Hart, F., Henn, W., and Sontag, H. Multi-Story Buildings in Steel. Crosby, Lock­
wood, Staples, 1978.

@Seismicisolation
@Seismicisolation
 CHAPTER 7

Rigid-Frame Structures

A rigid-frame high-rise structure typically comprises parallel or orthogonally ar­

ranged bents consisting of columns and girders with moment resistant joints. Re­
sistance to horizontal loading is provided by the bending resistance of the columns,
girders, and joints. The continuity of the frame also contributes to resisting gravity
loading, by reducing the moments in the girders.
The advantages of a rigid frame are the simplicity and convenience of its rect­
angular form. Its unobstructed arrangement, clear of bracing members and struc­
tural walls, allows freedom internally for the layout and externally for the fenes­
tration. Rigid frames are considered economical for buildings of up to about 25
stories, above which their drift resistance is costly to control. If, however, a rigid
frame is combined with shear walls or cores, the resulting structure is very much
stiffer so that its height potential may extend up to 50 stories or more. A flat plate
structure is very similar to a rigid frame, but with slabs replacing the girders. As
with a rigid frame, horizontal and vertical loadings are resisted in a flat plate struc­
ture by the flexural continuity between the vertical and horizontal components.
As highly redundant structures, rigid frames are designed initially on the basis
of approximate analyses, after which more rigorous analyses and checks can be
made. The procedure may typically include the following stages:

l. Estimation of gravity load forces in girders and columns by approximate

method.
2. Preliminary estimate of member sizes based on gravity load forces with ar­
bitrary increase in sizes to allow for horizontal loading.
3. Approximate allocation of horizontal loading to bents and preliminary anal­
ysis of member forces in bents.
4. Check on drift and adjustment of member sizes if necessary.

5. Check on strength of members for worst combination of gravity and hori­

zontal loading, and adjustment of member sizes if necessary.
6. Computer analysis of total structure for more accurate check on member
strengths and drift, with further adjustment of sizes where required. This
stage may include the second-order P-Delta effects of gravity loading on the
member forces and drift.
7. Detailed design of members and connections.

130 @Seismicisolation
@Seismicisolation
 7.1 RIGID FRAME BEHAVIOR 131

This chapter considers methods of analysis for the deflections and forces for
both gravity and horizontal loading. The methods are included in roughly the order
of the design procedure, with approximate methods initially and computer tech­
niques later. Stability analyses of rigid frames are discussed in Chapter 16.

7.1 RIGID FRAME BEHAVIOR

The horizontal stiffness of a rigid frame is governed mainly by the bending resis­
tance of the girders, the columns, and their connections, and, in a tall frame, by
the axial rigidity of the columns. The accumulated horizontal shear above any story
of a rigid frame is resisted by shear in the columns of that story (Fig. 7.I). The
shear causes the story-height columns to bend in double curvature with points of
contraflexure at approximately mid-story-height levels. The moments applied to a
joint from the columns above and below are resisted by the attached girders, which
also bend in double curvature, with points of contraflexure at approximately mid­
span. These deformations of the columns and girders allow racking of the frame
and horizontal deflection in each story. The overall deflected shape of a rigid frame
structure due to racking has a shear configuration with concavity upwind, a max­
imum inclination near the base, and a minimum inclination at the top, as shown
in Fig. 7.1.
The overall moment of the external horizontal load is resisted in each story level
by the couple resulting from the axial tensile and compressive forces in the col­
umns on opposite sides of the structure (Fig. 7 .2). The extension and shortening
of the columns cause overall bending and associated horizontal displacements of
the structure. Because of the cumulative rotation up the height, the story drift due
to overall bending increases with height, while that due to racking tends to de­
crease. Consequently the contribution to story drift from overall bending may, in.
the uppermost stories, exceed that from racking. The contribution of overall bend­
ing to the total drift, however, will usually not exceed 10% of that of racking,

Points of
contraflexure

.... _
Typical column
moment diagram

Fig. 7.1 Forces and deformations caused by external shear.

@Seismicisolation
@Seismicisolation
132 RIGID-FRAME STRUCTURES

r
I --,
-

c:
0
"'
- VI c:
c:
"' c:

...
->0 "'
)( ->0
w
- 0
�
Vl

-
�compression

Tension � �
I I I I , I , ,

Fig. 7.2 Forces and deformations caused by external moment.

except in very tall, slender, rigid frames. Therefore the overall deflected shape of
a high-rise rigid frame usually has a shear configuration.
The response of a rigid frame to gravity loading differs from a simply connected
frame in the continuous behavior of the girders. Negative moments are induced
adjacent to the columns, and positive moments of usually lesser magnitude occur
in the mid-span regions. The continuity also causes the maximum girder moments
to be sensitive to the pattern of live loading. This must be considered when esti­
mating the worst moment conditions. For example, the gravity load maximum
hogging moment adjacent to an edge column occurs when live load acts only on
the edge span and alternate other spans, as for A in Fig. 7.3a. The maximum
hogging moments adjacent to an interior column are caused, however, when live
load acts only on the spans adjacent to the column, as forB in Fig. 7.3b. The
maximum mid-span sagging moment occurs when live load acts on the span under
consideration, and alternate other spans, as for spans AB and CD in Fig. 7..3a.
The dependence of a rigid frame on the moment capacity of the columns for
resisting horizontal loading usually causes the columns of a rigid frame to be larger
than those of the corresponding fully braced simply connected frame. On the other
hand, while girders in braced frames are designed for their mid-span sagging mo-

A B c D

�sssssss1
(a)

A B C D
fSSS\S S\,SS\\\\\\1
(b)

Fig. 7.3 (a) Live load p attern for maximum positive moment in AB and CD, and maxi­
mum negative moment at A; (b) live load pattern for maximum negative moment at B.

@Seismicisolation
@Seismicisolation
 7.2 DETERMINATION OF MEMBER FORCES CAUSED BY GRAVITY LOADING 133

ment, girders in rigid frames are designed for the end-of-span resultant hogging
moments, which may be of lesser value. Consequently, girders in a rigid frame
may be smaller than in the corresponding braced frame. Such reductions in size
allow economy through the lower cost of the girders and possible reductions in
story heights. These benefits may be offset, however, by the higher cost of the
more complex rigid connections.

7.2 APPROXIMATE DETERMINATION OF MEMBER FORCES

CAUSED BY GRAVITY LOADING

A rigid frame is a highly redundant structure; consequently, an accurate analysis

can be made only after the member sizes are assigned. Initially, therefore, member
sizes are decided on the basis of approximate forces estimated either by conserv­
ative formulas or by simplified methods of analysis that are independent of member
properties. Two approaches for estimating girder forces due to gravity loading are
given here.

7 .2.1 Girder Forces-Code Recommended Values

In rigid frames with two or more spans in which the longer of any two adjacent
spans does not exceed the shorter by more than 20%, and where the uniformly
distributed design live load does not exceed three times the dead load, the girder
moment and shears may be estimated from Table 7. I. This summarizes the rec­
ommendations given in the Uniform Building Code [7. I]. In other cases a conven­
tional moment distribution or two-cycle moment distribution analysis should be
made for a line of girders at a floor level.

7.2.2 Two-Cycle Moment Distribution [7 .2]

This is a concise form of moment distribution for estimating girder moments in a
continuous multibay span. It is more accurate than the formulas in Table 7. I,
especially for cases of unequal spans and unequal loading in different spans.
The following is assumed for the analysis:

I. A counterclockwise restraining moment on the end of a girder is positive

and a clockwise moment is negative.
2. The ends of the columns at the floors above and below the considered girder
are fixed.
3. In the absence of known member sizes, distribution factors at each joint are
taken equal to I/ n, where n is the number of members framing into the joint
in the plane of the frame.

Two-Cycle Moment Distribution-Worked Example. The method is dem­

onstrated by a worked example. In Fig. 7 .4, a four-span girder AE from a rigid-

@Seismicisolation
@Seismicisolation
134 RIGID-FRAME STRUCTURES

TABLE 7.1 Gravity Load Forces in Girders

Location on Girder Value of Moment"

Sagging End spans: discontinuous end

moment unrestrained

End spans: discontinuous end integral

with support

Interior spans

16

Hogging At exterior face of first interior support:

moment for two spans

At exterior face of first interior support:

for more than two spans

At other faces of interior supports

At face of all supports where, at each end

of each span E column stiffnesses/
beam stiffness > 8
At interior face of exterior support for
member built integrally with spandrel
beam or girder
At interior face of exterior support for
member built integrally with column

Shear In end members at face of first interior wL

115-
2
.
support

At face of all other supports wL

2

•w is load per unit length of distributed load, L is the clear span for sagging moment or shear, and the
average of adjacent clear spans for hogging moment.

30 kN/
20 kN/m

A 4.0m B 2.33 2.33 0

I. 8.0m
• 1 •
7
.o .I . 6.7

Fig. 7.4 Gravity loading on girders of rigid frame.

@Seismicisolation
@Seismicisolation
 7.2 DETERMINATION OF MEMBER FORCES CAUSED BY GRAVITY LOADING 135

Uniformly distributed
Concentrated
i loadf:w
ll�l loading w/unit length

A �-------�-�----�
� �8 A� � l � ! i� � � � i ,B

:I I' I I
a b

I: L/
2 I
' L/2 L /2 L/2

MA = Wab'fL 2 MA =wl' /12

MB = -Wa2bfL 2 MB = -wl' 112

Me =Wa' 12L (for a< L/2) Me =wL' 124

Fig. 7.5 Fonnulas for fixed-end moments.

frame bent is shown with its loading. The fixed-end moments in each span are
calculated for dead loading and total loading using the formulas given in Fig. 7 .5.
The moments are summarized in Table 7. 2.
The purpose of the moment distribution is to estimate for each support the max­
imum girder moments that can occur as a result of dead loading and pattern live
loading. A different load combination must be considered for the maximum mo­
ment at each support, and a distribution made for each combination.
The five distributions are presented separately in Table 7. 3, and in a com­
bined form in Table 7.4. Distributions a in Table 7.3 are for the exterior sup­
ports A and E. For the maximum hogging moment at A, total loading is ap­
plied to span AB with dead loading only on BC. The fixed-end moments are
written in rows I and 2. In this distribution only the resulting moment at A is
of interest. For the first cycle, joint B is balanced with a correcting moment of
- ( -867 + 315) I 4 -
U I 4 assigned to M8A where U is the unbalanced mo-
=

TABLE 7.2 Fixed-End Moments for Two-Cycle Moment Distribution Worked

Example

Dead Load Dead + Live Load

Span Loading Moment (kNm) Moment ( kNm)

AB Concentrated 200 600

Unifonn distribution 107 267
Total 307 867

BC Concentrated 233 544

Unifonn distribution 82 184
Total 315 728

CD Concentrated 195 585

Unifonn distribution 101 228
Total 296 813

DE Concentrated 0 0
Unifonn distribution 75 187
Total 75 187

@Seismicisolation
@Seismicisolation
 136 RIGID-FRAME STRUCTURES

TABLE 7.3 Two-Cycle Moment Distribution

a. Maximum Moments at A and E

A B D E

Distribution MaA Mac Moe MoE MED

Factors 1/4 1/4 1/4 1/4 1/3
I. D.L. FEMU 315 -296
2. T.L. FEM" 867 � -867 187 -----
- 187
3. Carryover 69 14
4. Addition 936 -173
5. Distribution -312 58
6. Maximum 624 -115
moments

b. Maximum Moment at B

A B c

Distribution MaA Mac Mea Mco

Factors 1/4 1/4 1/4 1/4
I. D.L. FEM 296
2. T.L. FEM 867 � -867 728 �-728
3. Carryover -145 54
4. Addition -1012 782
5. Distribution 58 58
6. Maximum -954 840
moments

c. Maximum Moment at C

B c D

Distribution MaA Mac Mea Mco Moe MoE

Factors 1/4 1/4 1/4 1/4 1/4 1/4
I. D.L. FEM -307 75
2. T.L. FEM 728 ----...-728 813 �-813
3. Carryover -53 92
4. Addition -781 905
5. Distribution -31 -31
6. Maximum -812 874
moments

ment. This is not recorded, but half of it, (-U/4) /2. is carried over to MAs·
This is recorded in row 3 and then added to the fixed-end moment and the result
recorded in row 4.
The second cycle involves the release and balance of joint A. The unbalanced

@Seismicisolation
@Seismicisolation
moment of 936 is balanced by adding -U/3 = -936/3 = -312 to MAs
 7.2 DETERMINATION OF MEMBER FORCES CAUSED BY GRAVITY LOADING 137

TABLE 7.3 (Continued)

d. Maximum Moment at D

c D E

Distribution Mea Mco Moe MoE Mw

Factors 1/4 1/4 I /4 I /4 1/3
I. D.L. FEM -315
2. T.L. FEM 813 -813 187 ....... -187
-......
3. CarryQver -62 31 .....-
4. Addition -875 218
5. Distribution 164 164
6. Maximum -711 382
moments

"D.L. . dead loading; T.L . . total loading; FEM, fixed-end moments.

(row 5), implicitly adding the same moment to the two column ends at A. This
completes the second cycle of the distribution. The resulting maximum moment
at A is then given by the addition of rows 4 and 5, 936 - 312 = 624. The distri­
bution for the maximum moment at E follows a similar procedure.
Distribution b in Table 7.3 is for the maximum moment at B. The most severe
loading pattern for this is with total loading on spans AB and BC and dead load
only on CD. The operations are similar to those in Distribution a, except that the
first cycle involves balancing the two adjacent joints A and C while recording only
their carryover moments to B. In the second cycle, B is balanced by adding
-( -1012 + 782)/4 = 58 to each side of B. The addition of rows 4 and 5 then
gives the maximum hogging moments at B. Distributions c and d, for the moments
at joints C and D, follow patterns similar to Distribution b.
The complete set of operations can be combined as in Table 7.4 by initially
recording at each joint the fixed-end moments for both dead and total loading.
Then the joint, or joints, adjacent to the one under consideration are balanced for

TABLE 7.4 Combined Two-Cycle Moment Distribution

A B c D E

Distribution MAB MBA Mac Me a Mco Moe MoE Mw

Factors 1/3 1/4 1/4 1 /4 1/4 1/4 1/4 1/3
I. D. L . FEMa, 307 -307><315 -315><296 -296.>< 75 -75
2. T.L. FEM0 867 ><'-867 728><-728 813'>-<.-813 187 �87
3. Carryover 69 -145 54 -53 92 -62 31 14
4. Addition 936 -1012 782 -781 905 -875 218 -173
5. Distribution -312 58 58 -31 -31 164 164 58
6. Maximum 624 -954 840 -812 874 -711 382 -115
moments

@Seismicisolation
@Seismicisolation
"For abbreviations, see the footnote to Table 7-3.
138 RIGID-FRAME STRUCTURES

the appropriate combination of loading, and carryover moments assigned to the

considered joint and recorded. The joint is then balanced to complete the distri­
bution for that support.

Maximum Mid-Span Moments. The most severe loading condition for a max­
imum mid-span sagging moment is when the considered span and alternate other
spans carry total loading. A concise method of obtaining these values may be
included in the combined two-cycle distribution, as shown in Table 7.5. Adopting
the convention that sagging moments at mid-span are positive, a mid-span total
loading moment is calculated for the fixed-end condition of each span and entered
in the mid-span column of row 2. These mid-span moments must now be corrected
to allow for rotation of the joints. This is achieved by multiplying the carryover
moment, row 3, at the left-hand end of the span by (I + 0.5 D.F. )/2, and the
carryover moment at the right-hand end by -(I + 0.5 D. F. )/2, where D. F. is
the appropriate distribution factor, and recording the results in the middle column.
For example, the carryover to the mid-span of AB from A = [ (I + 0.5/3)/2]
x 69 = 40 and from B = - [ (I + 0.5/4)/2] x ( -145) = 82. These correction
moments are then added to the fixed-end mid-span moment to give the maximum
mid-span sagging moment, that is, 733 + 40 + 82 = 855.

7.2.3 Column Forces

The gravity load axial force in a column is estimated from the accumulated trib­
utary dead and live floor loading above that level, with reductions in live loading
as permitted by the local Code of Practice. The gravity load maximum column
moment is estimated by taking the maximum difference of the end moments in the
connected girders and allocating it equally between the column ends just above
and below the joint. To this should be added any unbalanced moment due to ec­
centricity of the girder connections from the centroid of the column, also allocated
equally between the column ends above and below the joint.

7.3 APPROXIMATE ANALYSIS OF MEMBER FORCES CAUSED

BY HORIZONTAL LOADING

7.3.1 Allocation of Loading between Bents

A first step in the approximate analysis of a rigid frame is to estimate the allocation
of the external horizontal force to each bent. For this it is usual to assume that the
floor slabs are rigid in plane and, therefore, constrain the horizontal displacements
of all the vertical bents at a floor level to be related by the horizontal translations
and rotation of the floor slab.

Symmetric Plan Structures Subjected to Symmetric Loading. A sym­

metric structure subjected to symmetric loading (Fig. 7.6a) translates but does not
twist. From the assumption of slab rigidity, the bents translate identically. The

@Seismicisolation
@Seismicisolation
 TABLE 7.5 Two-Cycle Moment Distribution with Maximum Mid-Span Moments

A B c D E
-- -

MAB MBA Msc Mcs Mco Moe M DE Mw

Distribution Factors 1 3
/ 1 4
/ 1 4
/ 1 4
/ 1 4
/ I 4
/ 1 4
/ 1 3
/
I. D.L. FEM" 307 -307 315 -315 296 -296 75 -75
2. T.L. FEM" 867 733 -867 728 364 -728 813 699 -813 187 94 -187
3. Canyover 69- 40 -145 54 30 -53 92-52 -62 31-17
4. Addition 30 905 35
� 14
936 82 �.:..1011 782 �-781 �-875 218 -8 -173
5. Distribution -312 58 58 -31 -31 164 164 62

6. Maximum 624 855 -954 840 424 -812 874 786 -711 382 103 -115
moments

"For abbreviations. see the footnote to Table 7-3 .

@Seismicisolation
@Seismicisolation
....
w
CD
140 RIGID-FRAME STRUCTURES

q-
l�-r+o
ti=-i-l-1I
!_l_LJ_l_l:
to
(a)

Arbitrary
I It�
--.-.,.0
or1g1n x,
x2
x3 etc c3 :1
cz· cs
I
I c, c6
x

(b)

Fig. 7.6 (a) Symmetric-plan rigid frame; (b) asymmetric-plan rigid frame.

total external shear at a level will be distributed between the bents in proportion
to their shear rigidities (GA) at that level. An explanation of the shear rigidity
parameter (GA) is given in a later section but, for now, it may be obtained for
level i in a bent simply by using

( GA) =

h
I
(I I)12£

G
+-
c i
(7.1)

in which h; is the height of story i, G = f. (Jg / L) for all the girders of span L
across floor i of the bent, and C = f. (Jj h;) for all the columns in story i of the
bent. E is the modulus of elasticity, and /,. and /11 are the moments of inertia of the
columns and girders, respectively.

@Seismicisolation
@Seismicisolation
 7.3 ANALYSIS OF MEMBER FORCES CAUSED BY HORIZONTAL LOADING 141

Asymmetric Plan Structures. The effect of lateral loading on a structure hav­

ing an asymmetric plan is to cause a horizontal plane torque in addition to trans­
verse shear. Therefore, the structure will twist as well as translate.
Referring to the asymmetric structure in Fig. 7.6b, and defining the location of
the center of shear rigidity of the set of parallel bents in story i, relative to an
arbitrary origin 0, as given by

· = [2::(GA)xj]
XI (7.2)
[Link]( GA)
;

An estimate of the shear Qji carried by bent j at level i is given by

Q; (GA )j; Q;e; [ (GA )c)
Qji = 2:: (GA) ; + {
2:: [ (GA)c2
(7 .3)

in which for level i,

Q; is the total shear, ( GA )ji is the shear rigidity of bent j in
story i, e; is the eccentricity of Q; from the center of shear rigidity in story i, cj is
the distance of bent j from the center of shear rigidity, and the two summations
refer to the full set of bents parallel to the direction of loading. The signs of c and
e are the same when they are on the same side of the center of rigidity.

7.3.2 Member Force Analysis by Portal Method

The portal method [7 .3] allows an approximate hand analysis for rigid frames
without having to specify member sizes and, therefore, it is very useful for a pre­
liminary analysis. The method is most appropriate to rigid frames that deflect pre­
dominantly by racking. It is suitable, therefore, for structures of moderate slen­
derness and height, and is commonly recommended as useful for structures of up
to 25 stories in height with a height-to-width ratio not greater than 4: I [7 .4]. Its
name is derived from the analogy between a set of single-bay portal frames and a
single story of a multibay rigid frame (Fig. 7.7a and b). When each of the separate
portals carries a share of the horizontal shear, tension occurs in the windward
columns and compression in the leeward columns. If these are superposed to sim­
ulate the multibay frame, the axial forces of the interior columns are eliminated,
leaving axial forces only in the extreme windward and leeward columns.
The reduction of the highly redundant multistory frame to allow a simple anal­
ysis is achieved by making the following assumptions:

I. Horizontal loading on the frame causes double curvature bending of all the
columns and girders, with points of contraftexure at the mid-height of col­
umns and mid-span of girders (Fig. 7 .1).
2. The horizontal shear at mid-story levels is shared between the columns in
proportion to the width of aisle each column supports.

@Seismicisolation
@Seismicisolation
142 RIGID-FRAME STRUCTURES

(a)

Zero ax ia 1 force
in internal columns
(b)
Fig. 7.7 (a) Separate portals analogy for portal method; (b) separate portals superposed.

The method may be used to analyze the whole frame, or just a portion of the
frame at a selected level. The analysis of the whole frame considers in tum the
equilibrium of separate frame modules, each module consisting of a joint with its
column and beam segments extending to the nearest points of contraftexure. The
sequence of analyzing the modules is from left to right, starting at the top and
working down to the base.
The procedure for a whole frame analysis is as follows:

1. Draw a line diagram of the frame and indicate on it the horizontal shear at
each mid-story level (Fig. 7. 8).

2. In each story allocate the shear to the columns in proportion to the aisle
widths they support, indicating the values on the diagram.
3. Starting with the top-left module (Fig. 7.9a), compute the maximum mo­
ment just below the joint from the product of the column shear and the half­
story height.
4. Find the girder-end moment just to the right of the joint from the equilibrium
of the column and girder moments at the joint. The moment at the other end
of the girder is of the same magnitude but corresponds to the opposite cur­
vature.
5. Evaluate the girder shear by dividing the girder end-moment by half the
span.
6. Consider next the equilibrium of the second joint (Fig. 7.9b), repeating steps
3 to 5 to find the maximum moment in the second column, and the moment
and shear in the second girder from the left.

This is repeated for each successive module working across to the right, and is
then continued in the level below, starting again from the left. The values of shear

@Seismicisolation
@Seismicisolation
 7.3 ANALYSIS OF MEMBER FORCES CAUSED BY HORIZONTAL LOADING 143

Wind Externa 1
load shear A B 0
Floor
kN kN level

� 20

18.4
---

� 19

55.2
--

� 18

92.0
--....

E
�
423.2 II
- E

..;
36.8 "" 8
-
"'
.,
460.0 L.

..,
0
.._..
"'
a
N

680.8
--....

[Link]-

6.5m 7.5m 6.0m

:I
·I I·
20.0m

Fig. 7.8 Example: Portal method of analysis.

and moment are recorded and a bending moment diagram drawn on the diagram
of the structure as the analysis progresses (Fig. 7 .8). The bending moments are
recorded on the girders above the left-hand end and below the right-hand end, and
similarly on the columns as viewed from the right. The shears are written perpen­
dicular to the columns and beams at the mid-heights and mid-spans, respectively.
The bending moment diagram is drawn here on the tension side of the member.
If member forces are required only at a particular level in the structure, the
horizontal row of modules at that level, consisting of the girders and half-columns
above and below, can be analyzed separately by the above procedure without hav­
ing to start the analysis at the top (Fig. 7.9c and d).
The consideration of vertical equilibrium of a joint module should give the

@Seismicisolation
@Seismicisolation
increment of axial load picked up by a column at that level. However, the assumed
144 RIGID-FRAME STRUCTURES

1-
"'
.,
.s;;;
"'

A 1-
., B
"'0

Girder
1-

(moment
<.!)

-5.23 ��

1
N

�s
Column
moment 6.44
3. 25m 3. 25m 3. 75
-----

I ·+· 1
m
•I • "

(a) (b)

68.8 148.1
-:JI""' ----

;t - 8

74.8 161.0
--- ._.....-

3. 25m • 3.25m •I 3. 75m

I
(c) (d)
Fig. 7.9 Equilibrium of modules: portal method.

distribution of shear between the columns results in a zero increment for all except
the two exterior columns. The axial force in the exterior columns in any story is
equal, therefore, to the moment of the external loading about the mid-height level
of that story, divided by the distance between the columns. The portal method
tends to overestimate the axial force in the exterior columns and is incorrect in
estimating zero axial force for the interior columns. However, when these forces
are added to the gravity load axial forces, the effect of the discrepancies on the
resultant axial force is generally negligible.
The simplicity of the portal method and the advantage that it allows a direct
analysis of member forces at intermediate levels make it the most useful of the
approximate methods for rigid-frame analysis. If, however, the frame is taller and
more slender, so that overall bending of the structu.-e by axial deformations of the
columns becomes significant, it may be more appropriate to analyze it by the can­
tilever method.

Portal Method- Worked Example. It is required to determine the member

7.8. The story height is typically 3.5 m, to
forces in the 20-story frame of Fig.

@Seismicisolation
@Seismicisolation
 7.3 ANALYSIS OF MEMBER FORCES CAUSED BY HORIZONTAL LOADING 145

give a total height of 70 m. The bents are spaced at 7.0 m. The intensity of the
wind loading is I. 5 kN/ m2 throughout the height.

Wind load per floor: At typical levels 1.5 x 7.0 x 3.5 = 36.8 kN
At the roof level 1.5 x 7.0 x 1.75 = 18.4 kN
Shear in the top story = 18.4 kN

Distributing this shear between the top-story columns in proportion to the widths
of aisle supported:

For column A: 18.4 x 3.25/20 = 2.99 kN

For column B: 18.4(3.25 + 3.75)/20 = 6.44 kN

The shear in columns C and D and in the columns of the stories below are allocated
similarly. The values are recorded on Fig. 7.8.
Starting with the top-left module A20 (Fig. 7.9a) and considering its free-body
equilibrium:
Moment at top of column = column shear X half-story height
= 2.99 X 1.75 = 5.23 kNm

From moment equilibrium of the joint, the moment at left end of the first girder

= -5.23 kNm
Shear in girder = girder-end moment/half girder length
= 5.23/3.25 = 1.61 kN.
Because of the mid-length point of contraftexure, the moment at the right end
of the girder has the same value as at the left end. Similarly, the column moments
at the top and bottom of a story are equal. The sign convention for numerical
values of the bending moment is that an anticlockwise moment applied by a joint
to the end of a member is taken as positive.
The values of the moments and shears are recorded on Fig. 7.8. Continuing
with the next module to the right, B20, in Fig. 7.9b:

Moment at top of column = column shear x half-story height

= 6.44 X 1.75 = 11.27 kNm

From moment equilibrium of joint, moment at end of second girder

= -(11.27 - 5.23) = -6.04 kNm

Then shear in second girder = girder moment/half-girder length
= 6.04/3.75 = 1.61 kN

@Seismicisolation
@Seismicisolation
146 RIGID-FRAME STRUCTURES

The above procedure is repeated for successive modules to the right, and then
continued on the floor below, starting again from left.
For the direct analysis of forces at an intermediate level, consider floor level 8
(Fig. 7.8).
Starting with the left module AS (Fig. 7.9c):

Moment in column above joint = 68.8 X 1.75 120.4 kNm

Moment in column below joint = 74.8 X 1.75 130.9 kNm

From moment equilibrium of joint, moment at end of first girder

= -(120.4 + 130.9) = -251.3 kNm

Then shear in first girder= 251.3/3.25 = 77.3 kN

Continuing with the next-right module B8 (Fig. 7.9d):

Moment in column above joint= 148.1 x l. 75= 259.2 kNm

Moment in column below joint= 161.0 x 1.75 = 281.8 kNm

From moment equilibrium of the joint, the moment at the end of the second
girder

= -(259.2 + 281.8 - 251.3) = -289.7 kNm

Then the shear in the second girder= 289.7/3.75 = 77.3 kN

The above procedure is repeated for successive modules to the right, as in Fig.
7.8.

7 .3.3 Approximate Analysis by Cantilever Method [7 .5)

The cantilever method is based on the concept that a tall rigid frame subjected to
horizontal loading deflects as a flexural cantilever (Fig. 7 .2). The validity of this
concept increases for taller, more slender frames, and for frames with higher girder
stiffness. The method is recommended [7.4] as suitable for the analysis of struc­
tures of up to 35 stories high with height-to-width ratios of up to 5 : 1.
It is similar to the portal method in considering the equilibrium of joint modules
in sequence. It differs, however, in starting by assuming values for the axial forces,
rather than the shears, in the columns. It is less versatile than the portal method in
not allowing a direct analysis of intermediate stories.
The assumptions for the cantilever method are as follows:

I. Horizontal loading on the frame causes double curvature bending of all the
columns and girders with points of contraflexure at the mid-heights of col­

@Seismicisolation
umns and mid-spans of girders.
@Seismicisolation
 7.3 ANALYSIS OF MEMBER FORCES CAUSED BY HORIZONTAL LOADING 147

2. The axial stress in a column is proportional to its distance from the centroid
of the column areas.

The procedure for the analysis is as follows:

I. Draw a line diagmm of the frame and record on it the external moment M
at each mid-story level (Fig. 7. I 0).
2. Find the centroid of the column areas and compute the second moment of
the column areas about the centroid using

(7.4)

where cj is the distance of column j from the centroid. In a case where the
column areas Aj are not known, they are to be taken as unity. Calculate the
column axial forces Fj in each story using

(7 .5)

Record these on the diagram of the structure.

Wind Resu 1ting

A B c 0
load external
kN moment
184 kN·m

r;;,

6.5m .1 7.5m 1 6.0m

I I
10. 13m I 9.87 m
l�:entroid of
20.0m column areas

Axial force in column indicated in adjacent box:

Otherwise member forces indicated as for Portal Method

@Seismicisolation
@Seismicisolation
Fig. 7.10 Example: cantilever method of analysis.
148 RIGID-FRAME STRUCTURES

A B

-4.65

·I
3. 25m 3. 25m 3. 75m
I· ·I
(a) (b)
Fig. 7.11 Equilibrium of modules: cantilever method.

3. Starting with the top-left module (Fig. [Link]) find the vertical shear in the
girder from the vertical equilibrium of the module.
4. Compute the girder-end moments from the product of the girder shear and
its half-span.
5. Compute the moment in the column just below the joint from the equilibrium
of the girder and column moments at the joint.
6. Evaluate the column shear by dividing the column-top moment by half the
story height.
7. Considering the next-right module (Fig. 7 .IIb) find the shear and moment
in the second girder and column by repeating steps 3 to 6.

This is repeated for each module in tum, moving to the right across the top
level, and then continuing from left to right in the level below. The values of shear
and moment are recorded on the diagram of the structure (Fig. 7.10).
The convention for indicating forces in the members is the same as in Fig. 7.8,
with the column axial forces written in boxes.

Cantilever Method-Worked Example. Analysis of the same 20-story,

70-m-high frame considered in the portal analysis. Referring to Fig. 7.10, external
moments due to wind are

At mid-height of story 20 = 18.4 x 1.75 = 32.2 kNm

At mid-height of story 19 = 18.4 x 5.25 + 36.8 x 1.75

161.0 kNm

Continue to calculate the external moment for each story down to the base and
record the values on Fig. 7.10.
Assuming a unit sectional area for each column:
Location of centroid of areas= 1 X (6.5 + 14.0 + 20 ) / 4 = 10.13 m from left

@Seismicisolation
@Seismicisolation
 7.3 ANALYSIS OF MEMBER FORCES CAUSED BY HORIZONTAL LOADING 149

The second moment of area

= I X (10.132 + 3.632 + 3.872 + 9.872)

4
= 228.2 m

Column axial forces:

Top story first column = MeA/I= 32.2 x 10.13 x I /228.2

= 1.43 kN tension

second column= MeA/!= 32.2 X 3.63 X 1/228.2

= 0.51 kN tension

Continue to find the axial forces in all the columns down to the base. The values
are recorded (in boxes) on Fig. 7.10.
Starting with the top-left module, A20 (Fig. 7.1la):
From vertical equilibrium of module, shear in first girder

= 1.43 kN

Moment at left end of girder= shear x half length of girder

= - (1.43 X 3.25) = -4.65 kN m

From moment equilibrium of joint, moment at top of column

= 4.65 kNm

Shear in column = moment at top/half story height

= 4.65/1.75 = 2.66 kN

The moments at opposite ends of the girders and columns are of the same value.
The moments and shears and a bending moment sketch are recorded, as for the
portal method.
Considering the next-right module, B20, Fig. 7.11b.
From vertical equilibrium of module, shear in second girder

= 1.43 + 0.51 = 1.94 kN

Moment in left end of second girder= shear x half length of girder

= -( 1.94 X 3.75) = - 7 . 2 8 kNm

From moment equilibrium of joint, moment at top of column

@Seismicisolation
@Seismicisolation
150 RIGID-FRAME STRUCTURES

= - ( -4.65 - 7.28 ) = I 1.93 kNm

Shear in column = moment at column top/half story height

= 11.93 / 1.75 = 6.82 kN

This procedure is repeated for successive modules to the right, then on the level
below, working again from left to right.

7.3.4 Approximate Analysis of Rigid Frames with Setbacks

A rigid frame bent with setbacks, as shown in Fig. 7.12a, can be analyzed ap­
proximately by applying the cantilever method to the upper and lower parts as
though they were two separate frames (Fig. 7.12b).
An analysis is made first of the upper part down to and including those parts of
the setback girder that form the upper frame. A moment distribution is then carried
out for the setback girder supported on the lower columns and subjected to the
calculated vertical forces from the columns of the upper structure. Because the
setback girder is so much stiffer in bending than the columns, it may be assumed
for this part of the analysis that the girder rests on simple supports. The moment
distribution yields the girder moments and shears and, hence, the vertical forces
that the girder applies to the supporting columns; these forces are assumed to carry
all the way to the foundation.
The lower structure, including the setback girder, may then be analyzed by the
cantilever method applying, in addition to its story increments of wind load, a
concentrated horizontal load at the setback level equal to the total horizontal force
above that level. The column axial forces calculated from the moment of the ex­
ternal horizontal loading are added to those determined from the setback beam
distribution to start the cantilever analysis for the lower part.
The total moments and shears in the setback girder due to wind forces are the
superposed results of the three analyses: the cantilever analysis of the upper part,
the cantilever analysis of the lower part, and the moment distribution of the girder.
If the complete setback structure has a low height-to-width ratio, it would be
more appropriate to use the portal method of analysis. As described above, the
two parts of the structure would be analyzed separately with the total shear from
the upper structure applied as a concentrated load at the setback level for the anal­
ysis of the lower structure. The forces in the setback girder would be obtained by
superposing the girder results from the two portal analyses, and the results of a
moment distribution using vertical forces from the upper columns as for the can­
tilever method.

7.4 APPROXIMATE ANALYSIS FOR DRIFT

When the initial sizes of the frame members have been selected, an approximate
check on the horizontal drift of the structure can be made. The drift in a nonslender

@Seismicisolation
@Seismicisolation
rigid frame is mainly caused by racking (Fig. 7.1). The racking may be considered
 7.4 APPROXIMATE ANALYSIS FOR DRIFT 151

�
�

�
( Transfer girder
-
Setback
�

-.

_,..

_,..

, , . , ., ,

(a)

--

--

-
Vertical forces from
Transfer girder upper structure

I:Horiz. forces from

upper structure --...__ Vertical forces from
� • transfer beam
·�
-

, , , ,

(b)

Fig. 7.12 (a) Rigid frame with setback; (b) setback structure separated for analysis.

as comprising two components: the first is due to rotation of the joints, as allowed
by the double bending of the girders (Fig. 7.13a and b), while the second is caused
by double bending of the columns (Fig. 7.13c). If a rigid frame is slender, a
contribution to drift caused by the overall bending of the frame, resulting from

@Seismicisolation
@Seismicisolation
axial deformations of the columns, may be significant (Fig. 7.2). If the frame has
152 RIGID-FRAME STRUCTURES

{a)

i- 1

(b)

"·[f.._',_
_ .._/_.<_ _: 7__/ __
___.__ /__, : ::: :-I
(c)
Fig. 7.13 (a) Joint rotation due to girder flexure; (b) story drift due to girder flexure; (c)
story drift due to column flexure.

a height:width ratio less than 4: l, the contribution of overall bending to the total
drift at the top of the structure is usually less than 10% of that due to racking.
The following method of calculation for drift allows the separate determination
of the components attributable to beam bending, column bending, and overall can­
tilever action.

7.4.1 Components of Drift [7.6]

It is assumed for the drift analysis that points of contraflexure occur in the frame
at the mid-story level of the columns and at the mid-span of the girders. This is a
reasonable assumption for high-rise rigid frames for all stories except near the top
and bottom.

Story Drift due to Girder Flexure. Consider a story-height segment of a frame

at floor level i consisting of a line of girders and half-story-height columns above

@Seismicisolation
@Seismicisolation
 7.4 APPROXIMATE ANALYSIS FOR DRIFT 153

and below each joint (Fig. 7.13a). To isolate the effect of girder bending, assume
the columns are flexurally rigid.
The average rotation of the joints can be expressed approximately as

total moment carried by the joints

(7.6)
B;g == total rotational stiffness of the joints

The total moment = 2 Q;h; +1 h;+ + Q; --

2
1
(7.7)

and the total rotational stiffness

= 12£
� (t} (7 .8)

From Eqs. (7.6) to (7.8)

(7 .9)

A similar expression may be obtained for the average joint rotation in the floor ( i
- I ) below, but with subscripts ( i + I ) replaced by i, and i by (i - I) .

Referring to Fig. 7.13b, the drift in story i due to the joint rotations is

(7.10)

that is

0·
lfi
=!!.!
2
[Qi-lhi-(1 /)Q;h; Q;h; Qi+(/lh)i+l]
24£ L: ...!.
+
+
+

24£ L: ...!.
(7 .II)

L ;-1 L ;

Assuming that the girders in floors i - I and i are the same, the story heights are
the same, and the average of
Q; Q; 2 Q; + 1 and _
1 is equal to

Qh(/) O·
lfi
=
L:
I I
(7.12)

i 12£
t
@Seismicisolation
@Seismicisolation
154 RIGID-FRAME STRUCTURES

Story Drift due to Column Flexure. Referring to Fig. 7.13c, in which the
drift due to bending of the columns is isolated by assuming the girders are rigid,
the drift of the structure in story i is

( 7 .13)

from which

( 7.14 )

Story Drift due to Overall Bending. Although the component of total drift
due to overall bending may be small relative to that caused by racking, the bending
inclination increases cumulatively throughout the height. Consequently, in the up­
per stories, where the story shear drift tends to be less than in the lower region,
the bending drift may become a significant part of the story drift. An estimate of
the bending drift can be made by assuming the structure behaves as a flexural
cantilever with a moment of inertia equal to the second moment of the column
I; E (Ac2); (Fig. 7.14a and b). If the
areas about their common centroid, that is =

moment diagram (Fig. 7.14c) is used to construct an MIEI diagram [in which I
= E (Ac2)] (Fig. 7.14d). the area of the diagram A� between the base and the
mid-height of story i gives the average slope of story i due to bending action, that
is

(7.15)

Then the bending component of drift in story i is given by

( 7.16 )

Story Drift and Total Drift. The resulting drift in a single story i is the sum of
the components,

( 7.17 )
or

( 7.18 )

Denoting E Ux I L ); by G;, and E Uc I h); by C;, this may be rewritten

�.
u =
Q;h� (.!. .!.)
+ +
.
h;Ao ( 7.19 )
I 12£ G c i

@Seismicisolation
@Seismicisolation
 __,..

I =l:(Ac2);
;
� Flcor i
Column Mid-height of
F' area story i
A A A Stc ry ;
�
l 2 3

"---A----'
;
Area A
__. 0
(shaded)

- Fl
� oor

Stc ry
- 0'----- 0 '-------......
Value of moment M
.. /
c3
,
r�: � I .I
column areas

(a) (b) (c) (d)

Fig. 7.14 (a) Frame structure; (b) distribution of inertia /; (c) distribution of external moment

@Seismicisolation
@Seismicisolation
M; (d) M/El diagram .

.....
Ul
Ul
156 RIGID-FRAME STRUCTURES

The assumption of a mid-story-height point of contraflexure is not valid for the

first story of a rigid frame because of the fully fixed or hinged conditions at the
base of the columns. Therefore, special expressions should be used for the first
story drift attributable to column and girder bending [7.7]. If the columns have
rigid base connections, the first story drift may be estimated by

(7.20)

If the columns have pinned base connections

(7.21)

The total drift at the nth floor of a building may then be found from

(7 .22)

A check on the story drift should be made for the top story and for intermediate
stories where member size reductions occur. If, on the basis of the initially sized
frame, the calculated drifts are well within the allowable values, these spot checks
will probably be adequate.

7.4.2 Correction of Excessive Drift

The typical proportioning of member sizes in tall rigid frames is such that girder
flexure is the major cause of drift, with column flexure a close second. Therefore,
increasing the girder stiffness is usually the most effective and economical way of
correcting excessive drift. If the girder in any single bay is substantially smaller
than the others at that level, it should be increased first.
An estimate of the modified girder sizes required at level i to correct the drift
in that story can be obtained by neglecting the contribution due to overall bending
and rewriting Eq. (7.18) in the form

(7.23)

in which O; is assigned the value of the allowable story drift. If the frame is un­
usually proportioned so that column flexure contributes a major part of the drift,

@Seismicisolation
@Seismicisolation
 7.4 APPROXIMATE ANALYSIS FOR DRIFT 157

Eq. (7 .23) may be rewritten to allow an estimate of the required column sizes by
interchanging E (lg I L ) ; and E (/,.I h);.
A relatively simple check on whether girders or columns should be adjusted
first has been proposed as follows [7. 8]. Compute for each joint across the floor
levels above and below the story whose drift is critical, the value of a parameter
1/; where

(7.24)

in which El1J L refers to the girders connecting into the joint.

If a scan of the resulting values of 1/; indicates that

l. 1/; >> 0.5, adjust the girder sizes;

2. 1/; << 0.5, adjust the column sizes;
3. 1/; == 0.5, adjust both column and girder sizes.

This test should preferably be accompanied by an inspection of the drift com­

ponents of Eq. (7 .18) to ascertain whether the allowable story drift is exceeded by
any one component alone, as might occur in a grossly undersized initial design. If
it is exceeded by any one component, whether as a result of undersized columns
or of undersized beams, that component must be remedied first.

7 .4.3 Effective Shear Rigidity ( GA)

This parameter expresses the racking stiffness of a frame on a story-height average

basis. It is a useful parameter when considering the allocation of loading between
rigid frame bents, and the horizontal interaction of frames with walls. The com­
posite symbol ( GA) is used because it corresponds with the shear rigidity of an
analogous shear cantilever of sectional area A and modulus of rigidity G. A story­
height segment of such a cantilever may be compared (Fig. 7.l5a) with a corre­
sponding portion of a rigid frame (Fig. 7.l5b).

Surface area A
Q
, >-

-r
,
l --r
,
Fig. 7.15 (a) Story-height segment of analogous shear wall; (b) single story of rigid frame.
@Seismicisolation
@Seismicisolation
158 RIGID-FRAME STRUCTURES

When the cantilever segment is subj ected to a shear Q, its deflection is given
by

Qh
0 = (7.25)
GA

from which the shear rigidity is given by

Qh Q
(GA) =- =- (7.26)
0 cf>

where cf> is the angle of inclination. That is, ( GA) is the shear force necessary to
cause unit inclination of the shear structure.
For the corresponding portion of frame, using Eq. (7.19) and neglecting drift
caused by overall bending

(GA) =
Qh
b =

Qh2
Qh
+(.!. .!.) (7.27)
12£ G C
then

( )
12£
h -
(GA) = (7.28)
I I
+­
G C

If the value of ( GA) ; at level i of a frame is known, the horizontal displacement

in story i is given by

0·I = (GA
Qh
) i
(7.29)

7.5 FLAT PLATE STRUCTURE-ANALOGOUS RIGID FRAME

Flat plate structures, in which the columns are cast integrally with the floor slabs,
behave under horizontal loading similarly to rigid frames. The lateral deflections
of the structure a i-e a result of simple double curvature bending of the columns,
and a more complex three-dimensional form of double bending in the slab. If the
columns are on a regular orthogonal grid (Fig. 7.16), the response of the structure
can be studied by considering each bay-width replaced by an equivalent rigid frame
bent. The slab is replaced for the analysis by an equivalent beam with the same
double bending stiffness. The hand methods of estimating drift, outlined in Sec­
tions 7.4.1 to 7.4.3, or a computer analysis, can then be applied.
The flexural stiffness of the equivalent beam depends mainly on the width-to­
length spacing of the columns and on the dimension of the column in the direction

@Seismicisolation
@Seismicisolation
 7.5 FLAT PLATE STRUCTURE-ANALOGOUS RIGID FRAME 159

Fig. 7.16 Flat plate structure.

of drift. In Fig. 7.17, these parameters are used to present the effective width of
the equivalent beam [7.9], that is, the width of the uniform-section beam having
the same double curvature flexural stiffness as the slab, with the same depth, span,
and modulus of elasticity as the slab. This equivalent beam may be used only in
the lateral loading analysis of flat plate structures. It is not appropriate for gravity
or combined loading analyses.
Figure 7.17 shows the equivalent beam stiffness to be very sensitive to the width
of the column in the direction of drift. This is because of the "wide-column" effect
that is demonstrated even more markedly by coupled shear walls (cf. Chapter 10).
When the slab width-to-span ratio b I a exceeds 1.5, the effective width becomes
virtually constant because the slab boundary regions parallel to the direction of
drift deform negligibly and therefore contribute little to the stiffness. The apparent
reduction in effective width shown by Fig. 7.17 as b I a increases is caused by
plotting the effective width as a fraction of the transverse span. The curves in Fig.
7.17 were obtained for square section columns; however, they are equally appli­
cable to rectangular section columns since additional analyses [7.9] have shown
that variations in the column transverse dimension from one-half to two times the
longitudinal dimension cause less than a 2% change in effective width.

7.5.1 Worked Example

A flat plate multistory structure consists of a regular rectangular grid of columns

spaced at 8.0 m by 6.0 m ctrs. The columns are 0.6 m square and the slab is
0.2 m thick. For horizontal loading acting parallel to the 8 m dimension, determine
the moment of inertia of an equivalent beam to replace the slab.
Referring to Fig. 7. 17

a = 8.0 m, b = 6.0 m, u = 0.6 m

u 0.6 b 6.0
-= - = 0.075; - = - = 0.75
a 8.0 a 8.0

@Seismicisolation
@Seismicisolation
160 RIGID-FRAME STRUCTURES

a
r ·I
- -ffi- - -y-
__ L ___ �;l--
a�

�§��E
Lo

--- .
- - --------
, -

YAwlmYAIJ�·
L -t-
- -
. ----- -

0.25

0 �----����--�----��----_.
0 0.05 0.10 0 .15 0.20
Value of u/a

Fig. 7.17 Effective width of equivalent beam.

Referring to the graph, the above values give

b'
- = 0 . 61
b

Effective slab width b' = 0.61 x 6.0 = 3.66 m. Therefore, moment of inertia of
equivalent beam

3.66 X 0.23 4
I= = 0.0024 m
12

This value would normally be reduced in the analysis by 50% to allow for the
reduction in stiffness due to cracking as the slab bends.

@Seismicisolation
@Seismicisolation
 7.7 REDUCTION OF RIGID FRAMES FOR ANA t YSIS 161

7.6 COMPUTER ANALYSIS OF RIGID FRAMES

Although the previously described hand methods of detennining deflections and

forces in rigid frames have served engineers well for the design of rigid frames,
and are still useful for preliminary analysis and checking, they have now been
superseded for most practical purposes by computer analysis. A computer analysis
is more accurate, and better able to analyze complex structures. A wide variety of
commercial structural analysis programs, invariably based on the stiffness matrix
method, are available.
Fonning the model of the rigid frame for computer analysis has been described
in Chapter 5. Briefly, it consists of an assembly of beam-type elements to represent
both the beams and columns of the frame. The columns are assigned their principal
inertias and sectional areas. The beams are assigned their horizontal axis inertia
while their sectional areas are assigned to be effectively rigid. Torsional stiffnesses
of the columns and beams are usually small and, therefore, neglected. Shear de­
fonnations of columns and beams are also usually neglected unless the member
has a length-to-depth ratio of less than about 5, in which case a shear area is
assigned.
If the frame is of reinforced concrete, reduced inertias are assigned to the mem­
bers to allow for cracking: 50% of their gross inertia to the beams and 80% of
their gross inertia to the columns.
Some analysis programs include the option of considering the slabs to be rigid
in-plane, and some the option of including P-Delta effects. If a rigid slab option
is not available, the effect can be simulated by interconnecting all vertical elements
by a horizontal frame at each floor, adding fictitious beams where necessary, and
assigning the beams to be effectively rigid axially and in flexure in the horizontal
plane. Slabs are usually assumed to have a negligible transverse rigidity unless a
flat plate or flat slab action is intended, in which case the slab is represented as a
connecting grid of equivalent stiffness beams.

7.7 REDUCTION OF RIGID FRAMES FOR ANALYSIS

The reduction of a rigid frame to a simpler equivalent frame is a useful way of

simplifying its analysis when it is not essential to obtain the exact member forces.
The two techniques described below can be used separately, or in combination, to
give large-scale reductions in the size of the computational problem.

7.7.1 Lumped Girder Frame

A repetitive floor system offers scope for the lumping of girders in successive floors
to fonn a model with fewer stories. The Jumped girder frame allows an accurate
estimate of the drift and a good estimate of the member forces. The girders are
usually lumped in threes or, if the frame is very tall, in fives. In the example of
Fig. 7 .18, three sets of three girders are lumped into single girders that converts

@Seismicisolation
@Seismicisolation
162 RIGID-FRAME STRUCTURES

J
girders

I,
�

1c 1
ce
9
1
ca

I
gS

)
I
g4 ---
-
(I
Ig g
3
Ig 1c
2 3
I 1c
c
I
gl 2
Ic 1c
l
., ,, ,
(a) (b)

Fig. 7.18 (a) Prototype rigid frame; (b) equivalent lumped girder frame.

the 13-story frame into a 7-story equivalent frame. The first floor and roof girders
must not be included in the lumping because the frame behavior near the top and
the base differ significantly from that in middle regions. In Fig. 7.l8b the second
floor and next-to-roof girders are also left as in the original to give an even closer
representation of the boundary conditions.
The requirement of a substitute frame is that, for horizontal loading, joint trans­
lations should be the same as those of the original structure. For translations caused
by girder flexure, Eq. (7.12) shows this requirement to be satisfied by assigning
the inertia of each equivalent girder to be equal to the sum of the lumped n-girder
inertias in the original frame, that is,

II

( 7 .30)

To determine the properties of the columns in a lumped girder equivalent frame,

reference is made to Eq. (7.14). Equating the component of drift caused by double
curvature column bending in a story of height nh in the equivalent frame to the
corresponding drift over n stories in the original frame

2 2
Q( n h) Qh 11 I
-=:=::--'--=-2:=--- ( 7.31 )
12£ 2:1"./nh 12£ 2:(1,./h) I

from which the equivalent column inertias are

@Seismicisolation
@Seismicisolation
 7.7 REDUCTION OF RIGID FRAMES FOR ANALYSIS 163

lc,, = " (I) (7.32)

L: � i

For example, in the structure shown in Fig. 7 .!Sa, the vertical stack of three
equal-height columns lcs• /,-9, and /('10, would be replaced in the equivalent lumped
girder model (Fig. 7.18b) by a three-story-height equivalent column having an
inertia

33
Ice = ----­ (7.33)
I I I
- + - + ­

/,.g /,'9 1,. 10

If within each region to be lumped the inertia of a column in successive single

stories is constant, the inertia of the equivalent column of height 3h would be nine
times that of the original single-story-height columns, while the inertia of the
equivalent column in an intermediate region (Fig. 7.18b), whose length is 2h,
would be four times that of the original single-story-height columns.
The columns' sectional areas, which control the cantilever component of de­
flection, must have the same second moment about their common centroid in the
lumped and original structures. Consequently, the areas of the equivalent columns
remain the same as those in the original frame. The horizontal loading on the
equivalent frame is applied as equivalent concentrated loads at the lumped girder
levels, taking the half new"story-height regions above and below the lumped gir­
ders as tributary areas.
When the lumped girder frame has been analyzed, the results must be trans­
formed back to the original frame. The moments in the original girders at the
lumped girder levels should be taken as I/ n of the resulting moment in the cor­
responding lumped girders. The moments in the original girders between the
lumped girder levels should then be estimated by vertical interpolation. Girder
shears are estimated by dividing the end moments by the half-span lengths.
For original column shears in a particular story, the actual external shear at the
mid-height of the story should be distributed between the columns in the same
ratio as that between the resulting shears at that level in the equivalent frame. The
moments at the top and bottom of a column should be taken as the product of the
column shear and the original half-story height.

7. 7.2 Single-Bay Substitute Frame (7 .1 O]

The reduction of a multibay rigid frame to a single-bay equivalent frame provides

a model that closely simulates the response of the structure to horizontal loading.
It is useful, therefore, in estimating deflections for stability analyses and for dy­
namic analyses of frames whose member forces are not required. It can also be
used in a two-stage member force analysis of a large multibent, multibay frame,

@Seismicisolation
@Seismicisolation
164 RIGID-FRAME STRUCTURES

for which a first-stage overall analysis is made of the structure represented by an

assembly of equivalent single-bay bents. The resulting shears in the bents are used
to obtain the individual bent loadings, which are then used in second-stage anal­
yses of the individual multibay bents to obtain their member forces (cf. Chapter
5, Section 5.1).
Figure 7.19a and b shows a multibay rigid frame and its single-bay analogy.
Member sizes are assigned to the single-bay frame to cause it to deflect horizon­
tally in the same manner as the prototype. First assume an arbitrary width /, for
the single-bay frame. Then, equating the component of drift in story i of the pro­
totype, caused by double bending of the girders, to that in the single-bay frame
structure, and using Eq. (7.12)

{7 .34)

Therefore, the girder in floor i of the single-bay frame is assigned an inertia

! I I (Jge)i
Ql <l2 <l3 F loor i

A 1 Ac 1c A 1 A 1�Story i (A ce l; 1 ce)i (Ace)· (I c

cl cl z 2 c3 c3 c4 .£i...t
c4

1 2 3 4
, , ,,,
c ,c
z
c, c
i_ 4

� Centroid of
column areas

(a) (b)
Fig. 7.19 (a) Multibay rigid frame; (b) equivalent single-bay frame.

@Seismicisolation
@Seismicisolation
 SUMMARY 165

Considering similarly the component of drift due to double bending of the col­
umns, and applying Eq. (7.14)

(7 .36)
12£ 2:: ((.I h ) I
.

therefore, the moment of inertia of the equivalent single-bay column is

(7.37)

Finally, equating the components of drift resulting from cantilever action in the
prototype and the single-bay frames, and using Eq. (7.16)

h; l Area ( E
l::M
Ace
2
)l ;
0
= h; l Area ( � ;z)l;
EA n· 2
0
(7.38)

then

(7.39)

therefore, the equivalent columns must be assigned sectional areas

(7 .40)

Although the single-bay frame results for horizontal deflections will be fairly
accurate, the resulting member forces for the single-bay frame are not transform­
able back to the multibay frame.
The lumped girder and single-bay frame techniques can also be used in com­
bination to reduce an extremely large frame structure to one that is much more
amenable to a first-stage, displacement and bent shear, analysis.

SUMMARY

The flexural continuity between the members of a rigid frame enables the structure
to resist horizontal loading as well as to assist in carrying gravity loading. The
probable worst combined effects of gravity and horizontal loading have to be es­
timated for the design of the frame.
Gravity loading causes regions of sagging moment near the mid-span of the
girders and of hogging moment beside the columns. Pattern live loading must be

@Seismicisolation
@Seismicisolation
166 RIGID-FRAME STRUCTURES

used to estimate the worst effects of gravity loading. The girder maximum mo­
ments may be evaluated approximately from formulas or more accurately from
conventional or shortened forms of moment distribution.
Horizontal loading causes racking of the frame due to double bending of the
columns and girders, resulting in an overall shear mode of deformation of the
structure. The portal and cantilever methods of analysis provide an estimate of the
horizontal loading member forces that, when combined with the gravity loading
member forces, allow a preliminary design of the frame members. The portal and
cantilever methods may be used also for the analysis of rigid frames with setbacks.
The lateral displacement of rigid frames subjected to horizontal loading is due
to three modes of member deformation: girder flexure, column flexure, and axial
deformation of the columns. The horizontal displacements in each story attribut­
able to these three components can be calculated separately and summed to give
the total story drift. The sum of the story drifts from the base upward gives the
horizontal displacement at any level. If the total drift, or the drift within any story,
exceeds the allowable values, an inspection of the components of drift will indicate
which members should be increased in size to most effectively control the drift.
A flat-plate structure responds to loading in a manner similar to a rigid frame
but with the transversely varying vertical flexure of the floor slab replacing the
single-plane vertical flexure of the rigid frame girder. A horizontal deflection anal­
ysis of a regular flat-plate structure can be made by considering the slabs replaced
by equivalent girders, and treating it as a rigid frame.
When a rigid frame includes many repetitive stories it may be reduced for a
horizontal loading analysis by lumping the girders in three, or five, successive
floors to give an equivalent simpler structure. The properties of the girders and
columns must be transformed initially in formulating the equivalent structure, and
the resulting forces subsequently transformed back to give the forces in the mem­
bers of the original structure. A multibay rigid frame may be reduced to an equiv­
alent single-bay frame for a horizontal loading analysis. This model is useful for
representing the horizontal response of the bent and for determining its horizontal
deflections. The two reduction methods may be used, either separately or in com­
bination, to simplify extremely large rigid frame structures for analysis.

REFERENCES

7.1 Uniform Building Code 1988. International Conference of Building Officials, Whit·
tier, California 90601.
7.2 Continuity in Concrete Building Frames. Portland Cement Association, Skokie, II·
linois 60076.
7.3 Smith, A. and Wilson, C. A. "Wind Stresses in the Steel Frames of Office Build­
ings. " J. Western Soc. Engineers April 19 15, 365-390.
7.4 Wind Bracing in Steel Buildings. Final Report of Sub-Committee No. 31 on Steel of
the Structural Division, Trans. ASCE 105, 1940, 1713-1739.
7.5 Wilson, A. C. "Wind Bracing with Knee-Braces or Gusset Plates." Engineer. Rec.
September 1908, 227-274.

@Seismicisolation
@Seismicisolation
 REFERENCES 167

7.6 Goldberg, J. E. "Approximate Methods in Stress and Stability Analysis of Tall

Building Frames." Proc. JABS£, ASCE Regional Conference on Tall Buildings,
Bangkok, January 1974, 177-194.
7.7 Goldberg, J. E. "Analysis of Two-Column Symmetrical Bents and Vierendeel
Trusses Having Parallel and Equal Chords." J. Am. Cone. Inst. 19(3), November
1947, 225-234.
7.8 Cheong-Siat-Moy, F. "Stiffness Design of Unbraced Steel Frames." AISC Engi­
neer. Journal, 1st Quarter, 1976, pp. 8-10.

Wong, Y. C. and Coull, A. "Effective Slab Stiffness in Flat Plate Structures." Proc.
Jnstn. Civ. Engineers Part 2, 69, September 1980, 721-735.

7.10 Goldberg, J. E. Structural Design of Tall Steel Buildings, Council on Tall Buildings
and Urban Habitat, Monograph, Vol. SB, 1979, p. 53.

@Seismicisolation
@Seismicisolation
 CHAPTER 8

lnfilled-Frame Structures

The infilled frame consists of a steel or reinforced concrete column-and-girder

frame with infills of brickwork or concrete blockwork (Fig. 8.1). In addition to
functioning as partitions, exterior walls, and walls around stair, elevator, and ser­
vice shafts, the infills may also serve structurally to brace the frame against hori­
zontal loading. In nonearthquake regions where the wind forces are not severe, the
masonry infilled concrete frame is one of the most. common structural forms for
high-rise construction. The frame is designed for gravity loading only and, in the
absence of an accepted design method, the infills are presumed to contribute suf­
ficiently to the lateral strength of the structure for it to withstand the horizontal
loading. The simplicity of construction, and the highly developed expertise in
building that type of structure have made the infilled frame one of the most rapid
and economical structural forms for tall buildings.
In countries with stringently applied Codes of Practice the absence of a well­
recognized method of design for infilled frames has severely restricted their use
for bracing. It has been more usual in such countries, when designing an infilled­
frame structure, to arrange for the frame to carry the total vertical and horizontal
loading and to include the infills on the assumption that, with precautions taken to
avoid load being transferred to them, the infills do not participate as part of the
primary structure. It is evident from the frequently observed diagonal cracking of
such infill walls that the approach is not always valid. The walls do sometimes
attract significant bracing loads and, in so doing, modify the structure's mode of
behavior and the forces in the frame. In such cases it would have been better to
design the walls for the lateral loads. and the frame to allow for its modified mode
of behavior.
In this chapter a design method is presented to allow the use of infilled frames
as bracing. It is based on theoretical and experimental studies of interactive wall­
frame behavior. Rather than being a method for the direct design of the frame
members and the wall, it is intended for use more as a method of checking and
adjusting an infilled frame that has already been designed to satisfy other criteria.
The frame is sized initially to be adequate for gravity loading, while the thickness
ofthe infill wall is probably decided on the basis of the acoustic, fire, and climatic
requirements.
To brace a structure, the arrangement of infill walls within the three-dimen­
sional frame must satisfy the same requirements as for the layout of bracing in a

168 @Seismicisolation
@Seismicisolation
 8.1 BEHAVIOR OF INFILLED FRAMES 169

Steel or reinforced
concrete frame

Brickwork or concrete
blockwork infills

Fig. 8.1 Structural frame infilled with masonry.

steel structure. Within any story the infills must be statically capable of resisting
horizontal shear in two orthogonal directions, as well as resisting a horizontal
torque. To achieve this there must be at least three infills that may not be all
parallel or all concurrent. They must, of course, also be able to satisfy the strength
and stiffness requirements.
Certain reservations arise in the use of infilled frames for bracing a structure.
For example, it is possible that as part of a renovation project, partition walls are
removed with the result that the structure becomes inadequately braced. Precau­
tions against this, either by including a generously excessive number of bracing
walls, or by somehow permanently identifying the vital bracing walls, should be
considered as part of the design. A reservation against their use where earthquake
resistance is a factor is that the walls might be shaken out of their frames trans­
versely and, consequently, be of little use as bracing in their own planes. On the
basis of substantial field evidence this fear is well justified. Their use in earthquake
regions, therefore, should be with the additional provision that the walls are rein­
forced and anchored into the surrounding frame with sufficient strength to with­
stand their own transverse inertial forces.

8.1 BEHAVIOR OF INFILLED FRAMES

The use of a masonry infill to brace a frame combines some of the desirable struc­
tural characteristics of each, while overcoming some of their deficiencies. The high
in-plane rigidity of the masonry wall significantly stiffens the otherwise relatively
flexible frame, while the ductile frame contains the brittle masonry, after cracking,
up to loads and displacements much larger than it could achieve without the frame.
'fhe result is, therefore, a relatively stiff and tough bracing system.
The wall braces the frame partly by its in-plane shear resistance and partly by
its behavior as a diagonal bracing strut in the frame. Figure 8.2a illustrates these

@Seismicisolation
@Seismicisolation
 ....

Shear deformation
of infills

Leeward
columns

Windward
Frame bearing � columns in �Equivalent
on infill tension diagonal
strut

(a) (b)

@Seismicisolation
@Seismicisolation
Fig. 8.2 (a) Interactive behavior of frame and infills; (b) analogous braced frame.
 8.1 BEHAVIOR OF INFILLED FRAMES 171

modes of behavior. When the frame is subjected to horizontal loading, it defonns

with double-curvature bending of the columns and girders. The translation of the
upper part of the column in each story and the shortening of the leading diagonal
of the frame cause the column to lean against the wall as well as to compress the
wall along its diagonal. It is roughly analogous to a diagonally braced frame (Fig.
8.2b).
Three potential modes of failure of the wall arise as a result of its interaction
with the frame, and these are illustrated in Fig. 8.3a. The first is a shear failure
stepping down through the joints of the masonry, and precipitated by the horizontal
shear stresses in the bed joints. The second is a diagonal cracking of the wall
through the masonry along a line, or lines, parallel to the leading diagonal, and

---=-:--=:':: . �
�

Leng th of -· -· . ------,

bearing

��r
Diagonal cracking

h s hear cracking

�FF/ / r / //

Com p res s iv e, crus h·n

1 g
failure
(a)

(b)
Fig. 8.3 (a) Modes of infill failure; (b) modes of frame failure.

@Seismicisolation
@Seismicisolation
172 INFILLED-FRAME STRUCTURES

caused by tensile stresses perpendicular to the leading diagonal. The "perpendic­

ular'' tensile stresses are caused by the divergence of the compressive stress tra­
jectories on opposite sides of the leading diagonal as they approach the middle
region of the infill. The diagonal cracking is initiated at and spreads from the
middle of the infill, where the tensile stresses are a maximum, tending to stop near
the compression comers, where the tension is suppressed. In the third mode of
failure, a comer of the infill at one of the ends of the diagonal strut may be crushed
against the frame due to the high· compressive stresses in the comer.
The nature of the forces in the frame can be understood by referring to the
analogous braced frame (Fig. 8.2b). The windward column is in tension and the
leeward column is in compression. Since the infill bears on the frame not as a
concentrated force exactly at the comers, but over short lengths of the beam and
column adjacent to each compression comer, the frame members are subjected
also to tl1}nsverse shear and a small amount of bending. Consequently, the frame
members or their connections are liable to fail by axial force or shear, and espe­
cially by tension at the base of the windward column (Fig. 8.3b).

8.2 FORCES IN THE INFILL AND FRAME

A concept of the behavior of infilled frames has been developed from a combi­
nation of results of tests [8.1-8.8], very approximate analyses [8.9], and more
sophisticated finite element analyses [8.10]. An understanding of infilled-frame
behavior is far from complete and further research needs to be done, especially
with full-scale tests. Consequently, opinions about the approach to the design of
infilled frames differ, especially as to whether it should be elastically or plastically
based. The method presented here draws from a combination of test observations
and the results of analyses. It may be classified as an elastic approach except for
the criterion used to predict the infill crushing, for which a plastic type of failure
of the masonry infill is assumed.

8.2.1 Stresses in the lnfill

Relating to Shear Failure. Shear failure of the infill is related to the

combination of shear and normal stresses induced at points in the infill when the
frame bears on it as the structure is subjected to the external lateral shear. An
extensive series of plane-stress membrane finite-element analyses [8.11] has shown
that the critical values of this combination of stresses occur at the center of the
infill and that they can be expressed empirically by

l.43Q
Shear stress TX\'
. = ( 8.1)
Lt
--

(0.8 h/L- 0.2)Q

Vertical compressive stress (8.2)
Lt
where Q is the horizontal shear load applied by the frame to the infill of length L,

@Seismicisolation
height h, and thickness t.

@Seismicisolation
 8.2 FORCES IN THE IN FILL AND FRAME 173

Relating to Diagonal Tensile Failure. Similarly, diagonal cracking of the

infill is related to the maximum value of diagonal tensile stress in the infill. This
also occurs at the center of the infill and, based on the results of the analyses, may
be expressed empirically as

0.58Q
Diagonal tensile stress a --­
(8.3)
d- Lt

These stresses are governed mainly by the proportions of the infill. They are
little influenced by the stiffness properties of the frame because they occur at the
center of the infill, away from the region of contact with the frame.

Relating to Compressive Failure of the Corners. Tests on model infilled

frames have shown that the length of bearing of each story-height column against
its adjacent infill is governed by the flexural stiffness of the column relative to the
inplane bearing stiffness of the infill. The stiffer the column, the longer the length
of bearing and the lower the compressive stresses at the interface. Tests to failure
have borne out the deduction that the stiffer the column, the higher the strength of
the infill against compressive failure. They have also shown that crushing failure
of the infill occurs over a length approximately equal to the length of bearing of
the column against the infill (Fig. 8.3a).
As a crude approximation, an analogy may be drawn with the theory for a beam
on an elastic foundation [8 .12], from which it has been proposed that [8. 9] the
length of column bearing a may be estimated by

7r
a=- (8.4)
2A
where

A=
.rE.7
�4EJh- (8.5)

in which Em is the elastic modulus of the masonry and El the flexural rigidity of
the column.
The parameter A expresses the bearing stiffness of the infill relative to the flex­
ural rigidity of the column: the stiffer the column, the smaller the value of A and
the longer the length of bearing.
If it is assumed that when the comer of the infill crushes, the masonry bearing
against the column within the length a is at the masonry ultimate compressive
stress/:,, then the corresponding ultimate horizontal shear Q;. on the infill is given
by

Q; = j:,ar (8.6)

or
@Seismicisolation
@Seismicisolation
174 INFILLED-FRAME STRUCTURES

7r
Q;. = J:nt · 2 ·
( 8.7)

Considering now the allowable horizontal shear Q

.
. on the infill, and assuming
a value forE/Em of 30 in the case of a steel frame and 3 in the case of a reinforced
concrete frame, the allowable horizontal shear on a steel framed infill correspond­
ing to a compressive failure is given by

Q.. = 5.2fm � ( 8.8)

and for a reinforced concrete framed infill

Q.
. = 2.9fm � ( 8.9)

in which fm is the masonry allowable compressive stress.

These semiempirical formulas indicate the significant parameters that influence
the horizontal shear strength of an infill when it is governed by a compressive
failure of one of its comers. The masonry compressive strength and the wall thick­
ness have the most direct influence on the infill strength, while the column inertia
and infill height exert control in proportion to their fourth roots. The infill strengths
indicated by Eqn. (8.8) and (8.9) are very approximate. Experimental evidence
has shown them to overestimate the real values; therefore, they will be modified
before being used in the design procedure.

8.2.2 Forces in the Frame

Experiments on horizontally loaded model infilled frames, and finite-element stress

analyses, have shown that the axial forces in the beams and columns of an infilled
frame can be estimated reasonably well by a simple analysis of the analogous
braced frame (Fig. 8.2b), assuming hinges at all joints. A conservative estimate
of the shear in the columns is given by the horizontal component of the force in
the diagonal strut and, similarly, an estimate of the shear in the beams is given by
the vertical component of that force. The analyses have indicated that the bending
moments in the columns and beams caused by the perpendicular thrust from the
infill are small relative to the moments that would occur in a similarly loaded rigid
frame without infills. A conservative nominal moment of Q - h / 20 is suggested
as a maximum value.

8.3 DEVELOPMENT OF THE DESIGN PROCEDURE

The main factors to be provided for in the design method are as follows:

I. In the weakest of its three modes of failure (i.e., shear, diagonal tensile,
and compressive) the infill must be capable of withstanding the stresses in­
duced by the frame bearing on it under the action of the external shear.
@Seismicisolation
@Seismicisolation
 8.3 DEVELOPMENT OF THE DESIGN PROCEDURE 175

2. The frame must be able to transmit to the infill the external shear imposed
on it, as well as be strong enough to withstand the reactions it receives from
the infill.

The following discussion concerns the development of a design procedure that

attempts to-satisfy the above criteria. It is assumed, conservatively, on the basis
that the lateral stiffness of the infill is much greater than that of the frame, that the
infill carries the total applied shear.

8.3.1 Design of the lnfill

Shear Failure. Shear failure, which occurs along the masonry bed joints, is
assumed to be initiated at the point in the infill where the ratio of horizontal shear
stress to available shear strength is a maximum. As noted before, theoretical
analyses have indicated, and tests have verified, that this occurs at the center of
the infill.
The shear strength of masonry has commonly been represented in Codes of
Practice [8.13, 8.14, 8.15] by a static friction type of equation

!s = ih_,. + p.a_,. (8.10)

together with a limiting maximum value.

The bond shear strength A, is similar in action to an adhesive shear strength
while [Link] is a frictional component of resistance, in which p. is a coefficient of
internal friction and cJ,. is the vertical compressive stress across the horizontal joint.
The allowable value off, depends on factored values of ih_, and p., to allow for the
type of masonry and a factor of safety.
Equating the shear stress at the wall center [Eq. (8.1)] to the allowable masonry
shear stress [Eq. (8.10)], and substituting Eqn. (8.2) for the vertical stress a,., gives

l.43Q, =Jibs + p.Q, ( 0.8h

- 0.2 ) ( 8.11)
Lt Lt L

Then, at any level of the structure, the allowable horizontal shear force based on
the shear failure criterion is

ih_,Lt
Q
s
= ---

1.43 -
,.,.
(
-"'""- ---

0.8h
- o.2
--,-
) ( 8.12)

Considering also the maximum allowable masonry shear stress

1.43 Qs
'f hmax (8.13)
Lt

from which
@Seismicisolation
@Seismicisolation
176 INFILLED-FRAME STRUCTURES

Q, -:f 0. 7Ltfsmax (8.14)

in whichhmax is the specified maximum allowable shear stress.

Diagonal Tensile Failure. The diagonal tensile strength of masonry is some­

what uncertain in value. Tests [8.11] have shown, however, that it can be esti­
mated conservatively as approximately equal to one-tenth of the mortar compres­
sive strength. Codes of Practice give an allowable flexural tensile stress in masonry
equal to approximately one-fortieth of the compressive strength of the weakest
allowable mortar. Assuming a typical factor of safety of 4 for brickwork, it is
reasonable to take the allowable diagonal tensile stress in masonry as equal to its
allowable flexural tensile stress, that is

(8.15)

Then, equating the maximum diagonal tensile stress [Eq. (8.3)] to the permissible
diagonal tensile stress [Eq. (8.15)]

0. 58 Q
j, (8.16)
Lt
= I

from which the allowable horizontal shear Q", based on the diagonal tensile failure
criterion, is given by

Q" = I. 7Ltj, (8. 17)

Comparing the allowable horizontal shear based on the maximum allowable

shear stress criterion [Eq. (8.14)] with that based on diagonal tensile failure [Eq.
(8.17)] by substituting values for unreinforced brick masonry from the Uniform
Building Code [8.13], the latter always results in a higher allowable horizontal
shear and, therefore, is less critical. Consequently, Eq. (8.17) may be dropped as
a design consideration.
Comparing also the allowable horizontal force based on the friction-formula
allowable shear stress [Eq. (8.12)], with that based on the maximum allowable
shear stress [Eq. (8.14)], and substituting in the equations values of ih_,, IJ., and
!smax• implicit in the Uniform Building Code, the friction-formula expression al­
ways gives lower values of allowable horizontal force for practically proportioned
frames and is therefore the more critical condition. Consequently, the maximum
allowable shear stress condition [Eq. (8.14)] can also be dropped. Therefore Eq.
(8.12) for the shear failure remains as one of the design criteria for a satisfactory
infill.

Compressive Failure. Equation (8. 7) demonstrates how the relative stiffness

of the column and infill influence the magnitude of the shear load required to cause

@Seismicisolation
@Seismicisolation
 8.3 DEVELOPMENT OF THE DESIGN PROCEDURE 177

compressive failure of the inti!!. This influence occurs because of the effect that
the column stiffness has on its length of bearing against the infill.
Tests to failure of model masonry infilled frames [8.16] have shown that com­
pressive failure shear loads may be represented more accurately by

(8.18)

in which 8 is the angle of the infill diagonal to the horizontal (Fig. 8.2b).
Substituting for>-. from Eq. (8.5) yields

( 4£/ )0.22
Q;. 1.12 J:nht cos2 0 (8.19)
Emth3
=

Then, assuming forE/Em a value of 30 in the case of a steel frame, and 3 for a
reinforced concrete frame, and using the allowable compressive stress fm, Eq.
(8.19) gives the allowable shear on a masonry infilled steel frame approximately
as

3.2/m cos 8 4vlhr

r;;-j
Qc =
2 (8.20)

and that on a reinforced concrete infilled frame approximately as

(8.21)

Equations (8.20) and (8.21), which are more conservative than the theoretically
deduced Eqs. (8.8) and (8.9), will be used in the design procedure.

8.3.2 Design of the Frame

Because an infilled frame behaves under horizontal loading in approximately the
same way as the analogous diagonally braced truss, the members can be designed
directly on the basis of the dead, live, and wind loading.

Columns. The design axial forces in the columns will be the worst combinations
of the forces from gravity and wind loading acting on the analogous braced frame.
In addition, on the basis of the results of the stress analyses, columns should be
assigned to have a bending moment with a conservative value of Q · h /20. The
shear force in the ends of a column should be assumed equal to the horizontal
component of the infill force at that level, that is Q.

Beams. The axial force in a beam may also be obtained from the analysis of the
analogous frame. Theoretically, this will be a tensile or compressive force equal
to the external shear at that level; however, this will be a conservative value be­
cause in reality the force will be shared with the floor slab.
If a beam has an infill above and below, it will be restrained against bending in

@Seismicisolation
@Seismicisolation
178 INFILLED-FRAME STRUCTURES

the vertical plane. If, however, a beam does not have an infill above, or it does
not have one below, the vertical thrust from the infill will cause a bending moment
in the beam. As for the columns, this bending moment may be taken conservatively
to be equal to Q - hl20.
The shear force in the ends of a beam may be taken to be equal to the vertical
component of the infill diagonal force. that is Q · hI L.
In designing for beam moments and shears caused by an infill above, when
there is no infill below, these moments and shears must be added to those from
gravity loading on the beam.

Connections. These should be designed to carry the axial and shear forces in
the connected members. Since moment resistance of the joints has been found to
make only a small difference in the overall behavior of the structure, it is not
necessary for the beam-to-column connections to be designed for moment.

8.3.3 Horizontal Deflection

In contrast to the shear configuration of a laterally loaded rigid frame without

infills, an infilled frame deflects in a flexural shape. This difference in deflected
shape occurs because the infill greatly reduces the shear mode deformations. At­
tempts have been made to estimate the diagonal stiffness of the infill for the pur­
pose of including it in a deflection analysis of the analogous truss. It has been
proposed by some to take a stiffness based on an equivalent width of strut equal
to a fraction of the diagonal length of the infill, while others have modified this
concept to assess the equivalent width as a function of the column stiffness. Un­
fortunately, the correlation between predicted and observed experimental deflec­
tions has been poor, probably because of unpredictable factors such as differences
in the tightness of fit of the infill in the frame, or the possible adherence of the
infill to the frame at low load levels.
In braced and infilled bents, the more slender the structure, the relatively greater
the influence of the column axial stiffness on the horizontal top deflection com­
pared with that of the diagonal bracing stiffness. In view of this, together with the
uncertainties about the diagonal stiffness of the infill, it is proposed here that the
deflection should be calculated as for the analogous diagonally braced frame, tak­
ing the area of the equivalent diagonal struts as the product of one-tenth of the
infills' diagonal length and their thickness. Assuming an elastic modulus of 7 x
103 N I mm2, ( 1 x 106 lb I in. 2) for the equivalent diagonal strut, an analysis of
the analogous braced frame will yield a conservative, that is excessive, estimate
of the deflection.

8.4 SUMMARY OF THE DESIGN METHOD

On the basis of the previous discussion, procedures for checking the strengths of
the frame and infills of an infill-braced structure, as well as the drift, can be for­
mulated.
@Seismicisolation
@Seismicisolation
 8.4 SUMMARY OF THE DESIGN METHOD 179

In the case of a tall infill-braced building structure, the initial design of the
frame would probably be on the basis of the gravity loading, and the design of the
infills on the basis of their acoustic and fire requirements. The number and direc­
tion of infills in each story must be arranged so they at least equal, and preferably
exceed significantly, the minimum requirements for static stability of the structure.
The loads carried by the individual bents should then be assessed so that the most
heavily loaded bents can be checked for the strength of their infills and frames.
The recommended design procedure for an individual bent would be as follows.

8.4.1 Provisions

1. The axis of the frame member sections should lie within the middle third of
the thickness of the infill to ensure the effective interaction of the frame and
infill.
2. The height-to-length ratio of the wall should be within the range of 0.3 to
3.
3. Care should be taken during construction to ensure a tight fit of the infill in
the frame.
4. The slenderness ratio of the wall should conform with the relevant Masonry
Code, assuming an effective height equal to the height of the infill.
5. Openings should not be allowed in the infill except at the edges, within the
middle third of the length of the sides. The maximum dimension of such
openings must not exceed one-tenth of the height or length of the infill,
whichever is the lesser value.

8.4.2 Design of the lnfill

Two modes of infill failure may cause collapse of the structure. The first is a shear
failure, stepping down diagonally through the bed joints of the masonry, and the
second is by spalling and crushing of the masonry in the comers of the infill. The
lesser of the two strengths should be taken as the critical value.

Shear Failure. The shear strength of the structure based on the shear failure of
the infill should be estimated from:

(8.12)
Q, = 1.43- �L(0.8(h/L)- 0.2)

in which As and I! are the allowable values of the bond shear stress and the coef­
ficient of internal friction, respectively, as given in the relevant Code formula for
the allowable shear stress in masonry [see Eq. (8.10)].

Compressive Failure. If the infill is bounded by a steel frame, the shear

strength of the structure relating to a compressive failure of the infill should be
estimated from
@Seismicisolation
@Seismicisolation
180 INFILLED-FRAME STRUCTURES

2 ( 8.22 )
Q,. = 3fm cos () '!.fihi3

and, if it is bounded by a reinforced concrete frame, from

( 8.23 )

8.4.3 Design of the Frame

Axial Forces. The gravity load forces in the columns should be calculated from
the tributary areas, applying reduction factors to the live load forces as appropriate.
Axial forces in the columns and beams resulting from the horizontal loading should
be estimated by a simple static analysis of the analogous braced frame, considering
each infill as a diagonal strut.

Bending Moments and Shear Forces

Columns. In addition to the axial forces determined as above, columns should

be able to withstand a design bending moment of Q · h 1 20 and a shear of Q.

Beams. The beams and their connections should be designed to carry an upward
shear force of Q · h I L, less the shear force due to dead load, and a downward
shear force of Q · hI L, plus the shear force due to dead and live load.
Where an upper beam of an infilled panel is not restrained by an infill above,
it should be designed to carry a negative (i.e., "hogging") moment of Q · hl20,
less the moment due to dead load. Where a lower beam of an infilled panel is not
restrained by an infill below, it should be designed to carry a mid-span positive
moment of Q · h l 20, in addition to the moment caused by vertical dead and live
load.

8.4.4 Deflections

A conservative estimate of the horizontal deflection of an infilled frame would be

given by the calculated deflection of the equivalent pin-jointed braced frame, as­
suming each infill to be replaced by a diagonal strut with a cross-sectional area
equal to the product of one-tenth of its diagonal length and its thickness, and taking
3 2 6 2
an elastic modulus equal to 7 x 10 N l mm (1 x 10 lb l in. ). Methods for
calculating the lateral deflection of a braced frame are given in Section 6.4.2.

8.5 WORKED EXAMPLE-INFILLED FRAME

A reinforced concrete, rigid-frame structure consists of a system of parallel three­

bay bents, as shown in Fig. 8.4, at 20 ft centers. The outer bays of each bent are
2
infilled by 8-in.-thick walls of 10,000 lb / in. clay brickwork.
It is required to assess the adequacy of the walls' strength to serve as bracing
for a horizontal wind pressure of 30 lb / ft2.

@Seismicisolation
@Seismicisolation
 8.5 WORKED EXAMPLE-INFILLED FRAME 181

Columns
18 [Link].

--

--
12 stories
Wind pressure�
24
" @ 11ft 4in
2
--
30 lb/ft =136 ft

-
"
24
--

Bents at 20ft spacing

lnfills. 8 in thick. 10.000 lb/in 2 clay masonry

Fig. 8.4 Example infilled frame.

Properties of the Frame

18 X 183
Inertia of columns I= in.4 = 8748 in.4
12

Reduced by 20% for cracking I= 6998 in.4

For slope() of infill diagonal cos() = 0.906

Properties of Masonry Infill. Taking, for example, values of 10,000 lb/in.2

clay masonry properties given implicitly in the UBC [8.13]

Allowable compressive stress fm = 3300 1b/in. 2

Allowable coefficient of friction J.L = 0.2

11
Allowable bond shear strength = 0. 3 (!:,) 2
A,
= 0.3( 10,000)112

= 30 1b/in.2

Wind Shear at Base of Structure

Q1 = 20 X 12 X 11.33 X 30 = 81576lb

= 81.6kip

@Seismicisolation
@Seismicisolation
182 INFILLED-FRAME STRUCTURES

Structure Shear Strength-/nfi/1 Shear Failure. Using Eqn. (8.12): strength

for two infills

s
2 X 30 X 240 X 8
Q= �----�--��----�
--------

1.43- 0.2[0.8 X (112/240)- 0.2 ]

= 82,561 lb= 82.6 kip

Structure Shear Strength-/nfi/1 Compressive Failure. Using Eqn. (8.23):

strength for two infills

Qc= 2 X 2 X 3300 X 0.9062 X �6998 X 112 X 83

= 1,533,542 = 1533.5 kip

Conclusion. The infill is just adequate to carry the external shear on the basis
of the shear failure criterion (strength= 82.6 kip compared with load of 81.6),
and more than adequate on the basis of the compressive failure criterion (strength
= 1533.5 kip). In addition to these calculations for the strength of the infill, the
members of the frame should be checked to see that they are adequate to carry the
forces described in Section 8.4.3. This is not included here, however, because the
procedure would be the same as for the members of a low-rise structure, as well
as being particular to the local Code-recommended method.

SUMMARY

An infilled frame consists of a steel or reinforced concrete frame of columns and

beams containing panels of brickwork or concrete blockwork. When an infilled
frame is subjected to horizontal loading, the infills behave as diagonal struts to
brace the structure and restrain its lateral deflection.
A method of design is developed that considers three possible modes of failure
of the infill: shear along the bedding planes of the masonry, diagonal cracking
through the masonry, and crushing of a comer of the infill against a column. The
estimated strengths of the three modes are based on a combination of experimental
evidence and the results of theoretical stress analyses.
The forces in the frame are estimated by a simple static analysis of the analo­
gous braced frame, considering the infills to be diagonal bracing struts. Failure of
the columns at the base of the structure, by tension on the windward side and by
shear on the leeward side, are of particular concern.
It is proposed that a conservative estimate of the lateral deflection of an infilled
frame would be made by calculating the deflection of the analogous braced frame,
assuming the equivalent diagonal strut to have an effective width equal to one­
tenth of the diagonal length of the infill.

@Seismicisolation
@Seismicisolation
 REFERENCES 183

REFERENCES

8.1 Polyakov,S. V. Masonry in Jnfilled Framed Buildings, 1956,G. L. Cairns,(trans.).

B.R.S. (U.K.) Publication. 1963.
8.2 Thomas, F. G. "The Strength of Brickwork." Struct. Engineer (U.K.), February
1953,35-46.
8.3 Benjamin, J. R. and Williams, H. A. "The Behavior of One-Story Brick Shear
Walls." Proc. A.S.C.E. 84., ST4, 1958,1723-1-1723-30.

8.4 Esteva,L. "Behavior under Alternating Loads of Masonry Diaphragms Framed by

Reinforced Concrete Members." Int. Symp. on the Effects of Repeated Loading of
Materials and Structures, Vol. V,RILEM,Mexico City,September 15-17,1966.
8.5 Stafford Smith,B. "Model Tests Results of Vertical and Horizontal Loading of In­
filled Frames." J. A.C.J. August 1968,618-624.

8.6 Meli,R. and Salgardo,G. Componamiento de muros de mamposteria sujetos a cargo

lateral (in Spanish). National University of Mexico,September 1969.
8.7 Mainstone,R. J. and Weeks,G. A. "Influence of a Bounding Frame on the Racking
Stiffness and Strengths of Brick Walls." Proc. 2nd. Int. Brick Masonry Conference,
Stoke-on-Trent,April 1970,pp. 165-171.

8.8 Fiorato,A. E., Sozen,M. A., and Gamble,W. L. Investigation of the Interaction
of Reinforced Concrete Frames with Masonry Filler Walls. Technical Report to the
Department of Defense,University of Illinois,November 1970.

8.9 Stafford Smith,B. "The Composite Behaviour of Infilled Frames." Proc. Symp. on
Tall Buildings, University of Southampton. 1966. pp. 481-492.
8.10 Riddington,J. R. and Stafford Smith, B. "Analysis of lnfilled Frames Subject to
Racking with Design Recommendations." The Struct. Engineer (U.K.),June 1977,
263-268.

8.11 Riddington,J. R. "Composite Behaviour of Walls Interacting with Flexural Mem­

bers." Ph.D. Thesis,University of Southampton,1974.

8.12 Hetenyi,M. Beams on Elastic Foundations, Vol. XVI. University of Michigan Stud­
ies,Scientific Series,1946.
8.13 Uniform Building Code. Int e rna tiona l Conference of Building Officials. Whittier.
California. 1976.

8.14 Code of Practice for Use of Masonry, BS 5628,Part I Structural Use of Unreinforced
Masonry,British Standards Institution, 1978.
8.15 Masonry Design for Buildings. National Standard of Canada,CAN3-S304-M84,Ca­
nadian Standards Association,1984.

8.16 Mainstone, R. J. "Supplementary Note on the Stiffness and Strengths of Infilled

Frames." Current Paper CP 13/74, Building Research Establishment, U.K . • Feb­
ruary 1974.

@Seismicisolation
@Seismicisolation
 CHAPTER 9

Shear Wall Structures

A shear wall structure is considered to be one whose resistance to horizontal load­

ing is provided entirely by shear walls. The walls may be part of a service core or
a stairwell, or they may serve as partitions between accommodations (Fig. 9.1 ).
They are usually continuous down to the base to which they are rigidly attached
to form vertical cantilevers. Their high inplane stiffness and strength makes them
well suited for bracing buildings of up to about 35 stories, while simultaneously
carrying gravity loading. It is usual to locate the walls on plan so that they attract
an amount of gravity dead loading sufficient to suppress the maximum tensile
bending stresses in the wall caused by lateral loading. In this situation, only min­
imum wall reinforcement is required. The term "shear wall" is in some ways a
misnomer because the walls deform predominantly in Hexure. Shear walls may be
planar, but are often of L-, T -, 1-, or U-shaped section to better suit the planning
and to increase their Hexural stiffness.
This chapter is concerned with the behavior of single walls and "linked-wall"
systems, that is, walls that are connected by Hoor slabs or beams with negligible
bending resistance, so that only horizontal interactive forces are transmitted. Walls
connected by bending members, termed "coupled walls," are considered sepa­
rately in Chapter 10.

9.1 BEHAVIOR OF SHEAR WALL STRUCTURES

A tall shear wall building typically comprises an assembly of shear walls whose
lengths and thicknesses may change, or which may be discontinued, at stages up
the height. The effects of such variations can be a complex redistribution of the
moments and shears between the walls, with associated horizontal interactive forces
in the connecting girders and slabs. As an aid to understanding the behavior of
shear wall structures, it is useful to categorize them as proportionate or nonpro­
portionate systems.
A proportionate system is one in which the ratios of the Hexural rigidities of
the walls remain constant throughout their height, as in Fig. 9.2a. For example, a
set of walls whose lengths do not change throughout their height, but whose chang­
ing wall thicknesses are the same at any level, is proportionate. Proportionate
systems of walls do not incur any redistribution of shears or moments at the change

184 @Seismicisolation
@Seismicisolation
 9.1 BEHAVIOR OF SHEAR WALL STRUCTURES 185

Fig. 9.1 Shear wall structure.

Wall 2
Connecting
links

/
Region A

�. �. � not equa 1
12A 128 12c

(a) (b)

Fig. 9.2 (a) Proportionate shear walls; (b) nonproportionate shear walls.

@Seismicisolation
@Seismicisolation
186 SHEAR WALL STRUCTURES

levels. The statical determinacy of proportionate systems allows their analysis to

be made by considerations of equilibrium, with the external moment and shear on
nontwisting structures distributed between the walls simply in proportion to their
flexural rigidities.
A nonproportionate system is one in which the ratios of the walls' flexural
rigidities are not constant up the height (Fig. 9.2b). At levels where the rigidities
change, redistributions of the wall shears and moments occur, with corresponding
horizontal interactions in the connecting members and the possibility of very high
local shears in the walls. Nonproportionate structures are statically indeterminate
and therefore much more difficult to visualize in behavior, and to analyze.
In this chapter, hand methods of analysis are described for proportionate non­
twisting and twisting structures. A hand method of analysis is presented for non­
proportionate, nontwisting structures also. However, it is generally more expedient
for nonproportionate, nontwisting structures, and essential for nonproportionate
twisting structures, to be analyzed by computer.

9.2 ANALYSIS OF PROPORTIONATE WALL SYSTEMS

The problem of analyzing a proportionate wall system is relatively uncomplicated

because of its statical determinacy. It will be considered in two subcategories of
structure-those that do not twist and those that twist.

9.2.1 Proportionate Nontwisting Structures

A structure that is symmetrical on plan about the axis of loading, as in Fig. 9.3,
will not twist. At any level i, the total external shear Q;, and the total external
moment M;, will be distributed between the walls in the ratio of their flexural
rigidities. The resulting shear and moment in a wallj at a level i can be expressed
as

(EI
)ii
Qp Q; (9.1)
= 2:: (El).
I

*
Fig. 9.3 Symmetric shear wall structure.

@Seismicisolation
@Seismicisolation
 9.2 ANALYSIS OF PROPORTIONATE WALL SYSTEMS 187

and

(9.2)

where (El )Ji is the flexural rigidity of wall j at level i and E(El ) ; represents the
summation of the flexural rigidities of all the walls at level i.
In such a proportionate nontwisting structure, there is no redistribution of shear
or moment at the change levels, and no redistributive interactive forces between
the walls.

9.2.2 Proportionate Twisting Structures

A structure that is not symmetric on plan about the axis of loading will generally
twist as well as translate. In a proportionate shear wall structure that twists under
the action of horizontal loading(Fig. 9.4) the resulting horizontal displacement of
any floor is a combination of a translation and a rotation of the floor about a center
of twist, which, in a proportionate structure, is located at the "centroid" of the
flexural rigidities of the walls. Referring to the asymmetric cross-wall structure in
Fig. 9.5, and assuming that the stiffness of a planar wall transverse to its plane is
negligible, the X-location of the center of twist from an arbitrary origin is

2: (Eix)
-;=;----.!'
X =
(9.3)
2:; (El). I

in which(£/); and (Eix); are, respectively, the sum of the flexural rigidities and
the sum of the first moments of the flexural rigidities about the origin, for all the
walls parallel to the Y axis at level i.

Translation
� � Centroid of
r1g1d1t1
e5

+'- �
wa

- --- �- --
11

-- L-- Lr - - .lJ
t
Fig. 9.4 Displacements of asymmetric structure.

@Seismicisolation
@Seismicisolation
188 SHEAR WALL STRUCTURES

y
Center of twist
I

v
1
!, 3 I
I

�------_.. x

Fig. 9.5 Asymmetric structure with walls parallel to loading.

In a proportionate structure, the center of twist and the shear center axis of the
structure coincide. Consequently, the effect of horizontal loading on the structure
Q;. and a resultant horizontal torque, which
is to produce at level i a resultant shear
is equal to the product of the resultant shearQ; and its eccentricity e from the shear
center, that is Q;e. The resultant shear in any wall j at level i is a combination of
its share of the external shear and the shear due to resisting its share of the external
torque at that level, which may be expressed as

(EI)Jl (Elc)}I
QJi =
Q; .
e
£ -) + Q; L: ( /
=L::--(-1 £ c2 ) (9.4)
I I

in which C;; is the distance of wall j from the shear center.

Noting that the moment in a wall can be obtained by integrating the shear (M
= f� Qdz), integrating Eq. 9.4 leads to an expression for the moment in wallj at
level i,

(9.5)

The first terms on the right-hand sides of Eqs. 9.4 and 9.5 are the shear and
moment, respectively, associated with bending translation of the structure, while
the second terms are associated with bending of the walls as the structure twists.
In Eqs. 9.4 and 9.5, ci; is taken as positive when on the same side of the center
of twist as the eccentricity e. Consequently. walls on the same side of the center

@Seismicisolation
@Seismicisolation
 9.2 ANALYSIS OF PROPORTIONATE WALL SYSTEMS 189

of twist as the resultant loading will have their shears and moments increased by
the twisting behavior. while those on the opposite side will have their shears and
moments reduced.
If a proportionate structure also includes walls perpendicular to the direction of
external loading, that is, aligned in the X direction, as in Fig. 9.6, the Y location
of the center of twist can be defined by

2.; (Ely )
y = -=------'' (9.6)
2: (El) I

in which the flexural rigidities refer to only the "perpendicular" walls.

As the structure twists under the action of horizontal loading, the total set of
orthogonally oriented walls will rotate about the axis of twist.
The effect of the "perpendicular" walls will be to stiffen the structure in tor­
sion, to reduce the twist, and, in doing so, to influence the contributions to the
"parallel" walls' shears and moments that result from the structure's twisting.
The denominator of the second terms in Eqs. (9.4) and (9.5}, for the shears and
moments in the "parallel" walls, must then be modified to 'f.(E/c2) + 'f.(Eid2),
in which E/c2 is the second moment of the "parallel" walls· flexural rigidities
2
about the center of twist while Eld refers correspondingly to the ''perpendicular''
walls.
Shears and moments will result in the "perpendicular" walls only from twisting
of the structure. The shear at level i in a "perpendicular" wall r will be

(Eid ) ; ,

(9.7)

and the moment will be

'Perpendicular' walls

walls

M
�
I�
�p-
===£i
t
Fig. 9.6 Asymmetric structure including "perpendicular" walls.

@Seismicisolation
@Seismicisolation
190 SHEAR WALL STRUCTURES

(Eid ) ,.;

(9.8)

Shear walls that are not aligned with the structure axes can be incorporated into
such an analysis by resolving their rigidities into components along the axes at the
shear centers of the walls, and treating the components of rigidity as those of walls
aligned parallel or perpendicular to the direction of loading.

9.3 NONPROPORTIONATE STRUCTURES

Nonproportionate structures consist of walls whose flexural rigidity ratios are not
constant throughout the height, and that consequently have different load-deflec­
tion characteristics. When the system of walls is subjected to horizontal loading,
so that the structure deflects and possibly twists, the rigidity of the floor slabs
constrains the dissimilar walls to deflect with similar configurations thereby in­
ducing horizontal interactive forces between them. The horizontal interactions play
a significant role in redistributing the horizontal shears and moments between the
walls.

9.3.1 Nonproportionate Nontwisting Structures

A nonproportionate, plan symmetric, nontwisting structure, such as shown in Fig.

9.7a, could be analyzed using a plane frame analysis program by assembling half
of the walls in a single plane, representing them by column elements, and con­
necting them at floor levels by axially rigid links (Fig. 9.7b) and then subjecting
them to half of the loading.
A hand method exists, however, that is accurate and can also be easily pro­
grammed for a small computer [9.1]. It is an iterative relaxation method, some­
what similar in its derivation to the well-known moment distribution method. The
iterations are reduced, however, by a series expression to make it a relatively
concise two-step operation. Rather than presenting the derivation, which is lengthy,
the principles of the method are explained, and the resulting procedure illustrated
with a worked example.
Referring again to Fig. 9.7a. in which a plan symmetric set of shear walls
changes flexural rigidity ratios at levels A and B, the external moment at each floor
level is allocated initially between the individual walls in proportion to their flex­
ural rigidities. At each of the change levels A and B two allocations are made, one
above the change and one below. Considering any single wall, because of the
nonproportional system the moments that have been allocated just above and below
level A are out of balance. Equilibrium of the wall is then restored at A by applying
a correcting moment that is shared above and below A in proportion to the respec­
tive wall rigidities. This is repeated at each change level in each wall.

@Seismicisolation
@Seismicisolation
 9.3 NONPROPORTIONATE STRUCTURES 191

Half symm. structure

Wa 11 2 3
--

--

.....
Load
1
2
-

..... A
-

__.

__.

__. B
�

....

" . ,.
(a)
(!:>)
Fig. 9.7 (a) Nonproportionate, plan-symmetric structure; (b) half-structure model for
computer analysis.

Now returning to level A, the correcting moments applied to the set of walls
just above A will not sum to zero. This means that at that level the wall moments
are not in equilibrium with the total external moment. The same situation exists
just below A with a residual moment of the same amount. A correcting moment is
then applied to the set of walls just above A, and similarly just below A, and
distributed between the walls in proportion to their rigidities. In any one wall at
level A, because of nonproportionality, the distributed moments just above and
below A are again out of balance, and so the cycle of balancing begins over.
The final moment in a wall just above level A consists of the sum of the initially
allocated "primary" moment, and a secondary moment that comprises all the al­
locations of correcting moments from the vertical and horizontal distributions. The
iteration converges to a solution in which the moments in each wall just above and
below each change level balance, while the sum of moments in the walls at each
change level balance with the external moment. Fortunately, the steps of the it­
eration can be written as a mathematical series that can be represented by a simple
expression.
In each wall the moments at levels other than the change points receive car­
ryover moments from the change level correction moments; however, these di­
minish so rapidly with each story further from a change level that it is necessary
to calculate them only for two stories above and two below each change level.
This reveals the interesting information that, in nonproportionate structures at lev-

@Seismicisolation
@Seismicisolation
192 SHEAR WALL STRUCTURES

els more than two stories away from change levels, the external moment is shared
between the walls almost exactly in proportion to their flexural rigidities, as in
proportionate structures.
A procedure with a worked example for illustration is now presented. The pro­
cedure as given provides for the determination of wall moments at all levels of the
structure; such a complete analysis, however, would be lengthy and tedious.
Therefore, a shortened form of the analysis, which would be adequate for most
design purposes, is used in the accompanying worked example. In this, the mo­
ments are found in the walls at the change levels, at one story above and one below
the change levels, and at the base.

Procedure and Worked Example. Consider the structure shown in Fig. 9.8a
and b, which consists of 20 3.5-m stories with a total height of 70 m. The five
shear walls include two symmetrical pairs (Types I and 2) and a central core

f cp
f �
Wall
type

I
I
15m
m
14om

{a)

Uniform wind
:o '

D
:
E

pressure 1. 5 kN/m2 Choog• ''"'' 0

....

= 30 kN/m height A@45,5 m II

.... U'>

( ha 1f structure) M

· ' '
--

" "
�

� 0
"'
"'

...
0
....
"' "'
"'
0
... N
'0

�D,
....
"'

...,

(b)

Fig. 9.8 (a) Example structure-plan; (b) example structure-end view.

@Seismicisolation
@Seismicisolation
 9.3 NONPROPORTIONATE STRUCTURES 193

TABLE9.1 Wall Dimensions and Inertias

Wall I Wall2 Wall 3

Total Half
Inertia Inertia Inertia Inertia
Dimensions /I Dimensions /, Dimensions /.1 1, / 2
4
(m) (m ) ( m) ( .;4) (m)
4
(m )
4
(m )

Top region, 8 X 0.2 8.533 5 X 0.2 2.083 Outside 26.046 13.023

45.5-70 m 6 X 6, walls
0.2 m thick
Middle region, 8 X 0.3 12.800 5 X 0.3 3.125 Outside 26.046 13.023
21-45.5 m 6 X 6, walls
0.2 m thick
Bottom region, 8 X 0.45 19.200 7 X 0.5 14.292 Outside 47.070 23.535
0-21 m 6 x 6, walls
0.4 m thick

(Type 3). Two change levels, A and B, divide the structure into three regions. The
wall dimensions and inenias are given in Table 9.1. Making use of plan symmetry,
one-half of the structure, comprising one Type I wall, one Type 2 wall, and a half
of the core, will be analyzed. Each stage of the worked example consists of a
procedural step in algebraic terms, together with a corresponding numerical step
for the example structure.

Procedure. For a nontwisting system of nonproponionate shear walls numbered

l, 2, 3, ... , j, . . . , n, and with change levels A, B, . . . , x.

I. Determine for each wall j at each change level x, the parameters

a.
and (9.9)

where I�j and 1;j are, respectively, the inenias of wall j just above and
just below change point x. For example, for wall I at change level A

8.533
------ = 0.361
8.533 + 2.083 + 13.023
and

12.800
k:l = = 0.442
12.800 + 3.125 + 13.023

The other values, for change level B and for other walls, are obtained
similarly.

@Seismicisolation
@Seismicisolation
194 SHEAR WALL STRUCTURES

b. (9.10)

For example, for wall I at change level A

tHai = 0.442 - 0.361 = 0.081

The other values, for change level B and for other walls, are obtained
similarly.

c. - I�; /�;
I ., h - ..:.•-.,...
Pxj and Pxj- __
(9.11)
fl. + b
11.j + I':j
=

X) JX)

For example, for wall I at change level A

I
-8.533
Pal -0.400
8.533 + 12.800
h
12.800
Pal = 0.600
8.533 + 12.800

The other values for change level B and for other walls are obtained
similarly. It should be noted that as a check, at a change level, f. k1.j =

Ek�j I for the set of walls, and I P�jI + I P�jI

= I for each wall. =

The values of k, t.k, and p, obtained from steps Ia, l b, and l c are
entered in Table 9.2.
2. Determine also for each change level

II

(XX =
_.l: P�j t.k.j (9.12)
;=I

For example, considering change level A, and using values from Table 9.2.

TABLE 9.2 Parameters for Analysis

Change
Level Wall kl.; k�; Ilk,; P'.; P�; f3'.; {31:;

A I 0.361 0.442 0.081 -0.400 0.600 -0.036 0.045

2 0.088 0.108 0.020 -0.400 0.600 -0.009 0.011
3 0.551 0.450 -0.101 -0.500 0.500 0.045 -0.056
01, = 0.0101

B I 0.442 0.336 -0.106 -0.400 0.600 0.030 -0.076

2 0.108 0.251 0.143 -0.179 0.821 -0.030 0.113
3 0.450 0.413 -0.037 -0.356 0.644 0 -0.037
(Jih = 0.0296

@Seismicisolation
@Seismicisolation
 9.3 NONPROPORTIONATE STRUCTURES 195

(XCI = ( -0.400) ( 0.081) + ( -0.400)(0.020) + (-0.500)(-0.10 I )

= 0.0101

The other value for change level B is obtained and entered in Table 9.2.
3. Using the parameters evaluated in Steps 1 and 2, determine for each wall j
at change level x:

I
I
f3xJ. = --
( PxJI t:.kXJ
. - (Xx k ·9.)
I

I (XX
_

and

b I_
f3 =
_
(/'X) t:.kX}. - aX e)
X) (9.13)
X) I - (XX

For example, for wall I at change level A, and using the values from Table
9.2

I
(-0.400 X 0.081 - 0.0101 X 0.361) = -0.036
[Link]
_

I
(0.600 X 0.081- 0.0101 X 0.442) = 0.045
[Link]
_

The other values, for change level B and for other walls, are obtained sim­
ilarly and the results entered in Table 9.2.
4. Calculate the total external moment M; on the structure at each level i, des­
ignating as Ma, Mh, ... , M" the external moments at change levels, A, B,
. . . ,X.
For example, at levels A + I, A, and A- I

2
Ma+ 1 = 30(70- 49) /2 = 6615 kNm
2
M" = 30(70 - 45.5) /2 = 9004 kNm
2
Ma-l= 30(70- 42) /2 = 11760 kNm

The other values, for change level B, plus and minus one story, and for
the base, are obtained similarly. The results for this step and for all subse­
quent steps are entered in Table 9.3.
5. Determine the primary moments in each wall j
a. just above and below each change level X, using respectively

and (9.14)

@Seismicisolation
@Seismicisolation
 ....

CD
Q)

TABLE 9.3 Bending Moments in Shear Walls (in kNm)