APPROXIMATE

METHODS IN STRUCTURAL
SEISMIC DESIGN
Adrian S. Scarlat

TA
658 .44
S23
19SE
Approximate Methods in
Structural Seismic Design
RELATED BOOKS FROM E & FN SPON

Behaviour of Steel Structures in Seismic Areas

Edited by F.M. Mazzolani and V. Gioncu
Concrete under Severe Conditions: Environment and Loading
Edited by K. Sakai, N. Banthia and D.E. GjQrv
Earthquake Engineering
Y.X. Liu, S.c. Liu and W.M. Dong
Earthquake Resistant Concrete Structures
G.G. Penelis and A.J. Kappos
International Handbook of Earthquake Engineering Codes,
Programs and Examples
Edited by M.H. Paz
Natural Risk and Civil Protection
Edited by T. Horlick-Jones, R. Casale and A. Amendola
Nonlinear Dynamic Analysis of Structures
Edited by K.S. Virdi
Nonlinear Seismic Analysis and Design of Reinforced
Concrete Buildings
Edited by P. Fajfar and H. Krawinkler
Reliability and Optimization of Structural Systems
Edited by R. Rackwitz, G. Augusti and A. Borri
Soil Liquefaction
M. Jefferies and K. Been
Structural Design for Hazardous Loads
Edited by J.L. Clarke, F.K. Garas and G.S.T. Armer
Structures Subjected to Dynamic Loading: Stability and Strength
Edited by R. Narayanan and T.M. Roberts
Theory and Design of Seismic Resistant Steel Frames
F.M. Mazzolani and V. Piluso

For more information, contact the Promotion Department, E & FN Spon, 2-6
Boundary Row, London SE1 8HN, Tel: IntI + 44 171-8650066, Fax: IntI
+ 44171 5229623.
Approximate Methods in
Structural Seismic Design
ADRIAN S. SCA RLAT
Manager
Research and Technology
A. Scar/at. N. Salehi Engineers Ltd
Tel Aviv. Israel
and
Affiliate Professor
Tecbnion
Haifa,lsm el

E & FN SPON
An I"'pfln! 01 CMp/Noft a H8I

London' Glasgow ' Weinheim . New York · Tokyo' Melbourne' Madras

Published by E & FN Spon, an imprint of Chapman & Hall,
2-6 Boundary Row, London SEl 8HN, UK
Chapman & Hall, 2-6 Boundary Row, London SEI 8HN, UK

Blackie Academic & Professional, Wester Cleddens Road, Bishopbriggs,

Glasgow G64 2NZ, UK

Chapman & Hall GmbH, Pappelalle 3, 69469 Weinheim, Germany

Chapman & Hall USA, 115 Fifth Avenue, New York, NY 10003, USA

Chapman & Hall Japan, ITP-Japan, Kyowa Building, 3F, 2-2-1 Hirakawacho,
Chiyoda-ku, Tokyo 102, Japan

Chapman & Hall Australia, 102 Dodds Street, South Melbourne, Victoria 3205,
Australia

Chapman & Hall India, R. Seshadri, 32 Second Main Road, CIT East,
Madras 600 035, India

First edition 1996

( 1996 Adrian S. Scarlat
Typeset in 10/12 Times by Thompson Press (India) Ltd, Madras
Printed in Great Britain by T.J. Press (Padstow) Ltd, Padstow, Cornwall

ISBN 0 419 187502

Apart from any fair dealing for the purposes of research or private study, or
criticism or review, as permitted under the UK Copyright Designs and
Patents Act, 1988, this publication may not be reproduced, stored or
transmitted, in any form or by any means, without the prior permission in
writing of the publishers, or in the case of reprographic reproduction
only in accordance with the terms of the licences issued by the Copyright
Licensing Agency in the UK, or in accordance with the terms of licences
issued by the appropriate Reproduction Rights Organization outside
the UK. Enquiries concerning reproduction outside the terms stated here
should be sent to the publishers at the London address printed on this page.
The publisher makes no representation, express or implied, with regard
to the accuracy of the information contained in this book and cannot accept
any legal responsibility or liability for any errors or omissions that
may be made.

A catalogue record for this book is available from the British Library

Library of Congress Catalog Card Number: 95-74654

@l Printed on permanent acid-free text paper, manufactured in accordani.:e 'l'.ith ANSI NISO
Z.WAX-J991 and ANSI NISO ZJ9. 4X-19S4 (Permanence of Paper).
Dedicated to
My wife Evemina
My sons Alex, Dan and Yuval
Contents

page
Preface xiii

1 Multi-storey building frames 1

1.1 Introduction 1
1.2 Multi-storey, one-bay frames 2
1.2.1 General approach 2
1.2.2 The zero moment point (ZMP) procedure 2
1.2.3 Continuum approach 9
1.2.4 Pin-supported frames 14
1.2.5 Approximate assessment of deflections 17
1.3 The substitute (equivalent) frame method 23
1.3.1 General approach 23
1.3.2 Remarks dealing with the application of the
substitute frame method 25
1.3.3 Effect of soil deformability 33
1.4 First screening of existing moment-resisting frame structures 34
1.4.1 Purpose of first screening methods 34
1.4.2 Reinforced concrete frames 34
1.4.3 Steel frames 36
Bibliography 37

2 Structural walls 38
2.1 Introduction 38
2.2 Structural walls without openings 39
2.2.1 Effect of shear forces 39
2.2.2 Effect of soil deformability 42
2.3 Structural walls with openings (coupled walls) 47
2.3.1 Slab coupling: effective width of lintels 48
2.3.2 Approximate analysis by frames with finite joints 48
2.3.3 Approximate analysis of uniform coupled structural
walls by the continuum approach 50
2.3.4 Assessment of vertical shear forces in lintels 55
2.3.5 Assessment of maximum horizontal deflections 59
2.3.6 Distribution of vertical reactions 60
2.3.7 Effect of soil deformability 60
2.4 First screening of existing structures laterally supported
by RC structural walls 65
2.4.1 Introduction 65
viii Contents

2.4.2 Proposed relationships 66

2.5 Masonry structural walls 68
2.5.1 Introduction 68
2.5.2 Plain masonry (in-plane bending) 70
2.5.3 Infilled frames (in-plane bending) 70
2.5.4 Reinforced masonry (in-plane bending) 75
2.6 Out-of-plane bending of masonry walls 77
2.7 First screening of existing buildings laterally
supported by masonry walls 78
2.7.1 Plain masonry 78
2.7.2 Infilled frames 80
2.7.3 Reinforced masonry 80
2.8 Connection forces between facade panels in one-storey
industrial buildings 81
Bibliography 84

3 Dual systems (structural walls and frames) 86

3.1 Introduction 86
3.2 Analysis of dual systems: classical approach 86
3.2.1 Analysis of dual systems on planar schemes 86
3.2.2 Approximate methods for the distribution of
lateral forces 88
3.3 Effect of soil deformability and cracking of RC elements
on the distribution of lateral forces 94
Bibliography 98
4 Space structures 100
4.1 Torsional forces: Lin's theory 100
4.2 Approximate torsional analysis of dual structures based
on Lin's theory 110
4.3 A critical review of Lin's theory 116
4.3.1 Effect of vertical deformability of slabs 116
4.3.2 Effect of soil deformability on the distribution
of torsional forces 118
4.4 First screening of existing dual structures subject to
torsion 121
4.5 Vertical grids composed of structural walls and slabs 123
4.6 One-storey industrial buildings: bracing conditions 127
4.7 Framed tube structures: approximate assessment of stresses 129
4.7.1 Introduction 129
4.7.2 Approximate analysis by reducing the space
structure to a planar structure 130
4.7.3 Approximate analysis based on equivalent flanges 132
4.8 Axi-symmetric system of bars: evaluation of maximum
stresses and deflections 135
 Contents ix

4.8.1 Introduction 135

4.8.2 Fixed end moments 136
4.8.3 Maximum axial forces 137
4.8.4 Maximum horizontal deflections 139
Bibliography 145

5 Pile foundations and retaining walls 146

5.1 The 'equivalent pile length' concept 146
5.2 Distribution of lateral forces in a pile group 152
5.3 Retaining walls: earth pressure during earthquakes 156
Bibliography 159

6 Earthquake design: basic concepts, approximate methods 160

6.1 Evaluation of seismic forces: regular structures 160
6.1.1 General data, performance criteria, seismic
coefficients 160
6.1.2 Seismic intensity factor 161
6.1.3 Soil or site factor 164
6.1.4 Rigidty and mass factors; fundamental period 164
6.1.5 Ductility factor; reduction factor 175
6.1.6 Importance factor 180
6.1. 7 Vertical seismic forces 180
6.2 Evaluation of seismic forces: special structures 181
6.2.1 General approach 181
6.2.2 Response spectra 181
6.2.3 Modal analysis 185
6.2.4 Approximate methods in modal analysis 188
6.3 Structural irregularities 190
6.3.1 Plan irregularities 190
6.3.2 Vertical irregularities 193
6.3.3 Irregularities and structural design 197
6.4 Bounds on the seismic coefficient 199
6.4.1 General approach 199
6.4.2 The upper bound 200
6.4.3 The lower bound 200
6.5 Problems of deformability 202
6.5.1 General approach 202
6.5.2 Storey drift limitations 203
6.5.3 P L1 effect 203
6.5.4 Pounding of adjacent buildings 205
6.5.5 Effect of resonance 206
6.6 Energy approach: application to soft storeys 207
6.7 Non-structural elements and non-building structures 209
6.7.1 Non-structural elements 209
x Contents

6.7.2 Non-building structures 210

Bibliography 210

7 Evaluation of existing buildings for seismic hazard 213

7.1 Introduction 213
7.2 Level I: oustide inspection of existing buildings 214
7.3 Level II: first screening of existing buildings 216
7.3.1 General data 216
7.3.2 Buildings exempted from first screening 216
7.3.3 Data needed for computing the seismic index 216
7.3.4 Classification of structures 217
7.3.5 Computation of the seismic index Is 220
7.3.6 Classification of buildings 224
7.4 Level III: accurate analysis of exisitng buildings 224
7.4.1 Existing documentation 224
7.4.2 Present condition of the structure 225
7.4.3 Precast structural elements 226
7.4.4 Effect of pounding (hammering) 226
7.4.5 Structural analysis 226
7.4.6 Non-structural elements 228
Bibliography 228

Postscript 230
1 Choice of structural solution 230
2 Choice of structural design method 230

Appendix A Structural dynamics: main formulae 232

A.l Single-degree-of-freedom (SDOF) systems 232
A.l.1 Undamped free vibrations 232
A.1.2 Damped free vibrations 233
A.1.3 Undamped forced vibrations 234
A.l.4 Damped forced vibrations 235
A.2 Multi-degree-of-freedom (MDOF) systems 237
A.2.1 Undamped free vibrations: flexibility
formulation 237
A.2.2 Undamped free vibrations: stiffness
formulation 239
A.2.3 Orthogonality of modes hand.i (h #.il 240
A.2.4 Undamped free vibrations: uniform
cantilever 241
A.2.5 Undamped forced vibrations 241
A.2.6 Damped forced vibrations 243
A.2.7 Arbitrary excitation 244
 Contents xi

A.3 Seismic forces 244

A.3.1 Single-degree-of-freedom structures 244
A.3.2 M ulti-degree-of-freedom structures 245
Bibliography 247

Appendix B Techniques for finite element computations 248

B.l Approximate analysis of structural walls by finite elements 248
8.2 Use of symmetry properties 249

Appendix C Methods of quantifying the structural rigidity 256

Cl Quantifying the effect of soil deformability 256
Cl.l Spread and mat foundations 256
C1.2 Foundations on piles 259
C2 Quantifying the effect of cracking of RC elements 261
C2.1 Moment-resisting frames 261
C2.2 Structural RC walls without openings 261
C2.3 Structural RC coupled walls 262
C3 Final remarks 262
Bibliography 263

Appendix D Glossary of earthquake engineering terms 265

Appendix E Design codes and standards 270

Index 274
Preface

Three distinct stages may be defined in the structural design of buildings:

1. selection of the structural solution and assessment of the dimensions of
the main elements;
2. detailed computation of stresses and displacements;
3. checking of the main results.
Note that computerized analysis suits only stage 2, whereas approximate
methods have to be used for stages 1 and 3; this points out the exceptional
importance of approximate methods in structural design.
This book deals with the most difficult problem arising in the design of
multi-storey buildings and similar structures: the effect of horizontal forces,
mainly seismic loads and similar loads such as wind pressure. In fact, the
general stability of a structure capable of resisting horizontal forces is also
ensured for vertical loads; possible accidents are generally only local. Unfortu-
nately, the most difficult task is to define clearly the concept of approximate
methods. Their reverse, the accurate methods, also elude definition: in order to
apply them, several simplifying assumptions are required. Hence we are ob-
liged to accept a rather vague definition: an approximate method in structural
analysis is a method that permits the assessment of stresses or displacements in
a much shorter time than the commonly used design methods.
We shall refer in the following to two different types of approximate
methods: (1) methods designed to determine the stresses and displacements of
a given structure by using substitute structures; (2) methods based on 'global
parameters' (seismic coefficients, total area of structural walls, etc.). The advent
of computerized structural analysis and, in particular, of finite element pro-
grams, has opened up vast possibilities for a fundamental reappraisal of exist-
ing approximate methods, including a more accurate definition of their scope
and limits. We have taken advantage of these possibilities, have examined
several 'classical' approximate methods, and have proposed a number of new
techniques intended to complement the existing ones.
This book is the result of many years spent in structural design and teach-
ing. One of its main sources is a course in structural dynamics taught by
Professor P. Mazilu at the Institute of Civil Engineering of Bucharest, Romania,
which the author was privileged to attend.
The target audience of the book is first of all design engineers, but it should
be of use for non-specialist engineers too; it is assumed that it will serve also as
a teaching aid for undergraduate students as well as for advanced high-rise
buildings courses. It is hoped, by citing Mozart (route proportion ?Jardee), that
xiv Preface

'here and there are sections that only connoisseurs will enjoy, but these sec-
tions have been written so that even a layman will have to enjoy them, albeit
without knowing it.'
Adrian S. Scarlat

ACKNOWLEDGEMENT

The author is especially grateful to Professor A. Rutenberg for his careful

screening of the manuscript and for suggesting valuable improvements.
1 Multi-storey building frames

1.1 Introduction

In this chapter we propose to assess the bending moments and the deflections
of a multi-storey building frame acted upon by lateral forces (Figure 1.1a), by
using approximate methods.
As will be shown in the next section (1.2), the approximate analysis of this
frame may be performed on a 'substitute' (equivalent) one-bay, symmetrical
frame (Figure 1.1 b). Therefore we shall deal mainly with this latter frame, as a
first step in analysing the actual multi-storey, multi-bay frame.
In our analysis, we assume several hypotheses aimed at simplifying the
computation. These are as follows.

• It is assumed that all the horizontal loads are concentrated at floor levels.
• The effect of the shear forces (V) on the deformations is neglected. This
hypothesis is acceptable as long as we deal with the usual systems of bars.
The analysis of structural walls, where this effect is important, will be
dealt with separately (Chapter 2).
• The effect of axial forces (N) on the deformations is neglected. This hypo-
thesis, too is acceptable as long as the total length of the multi-bay
frame (L) is not small with respect to its total height (H). In most practical
cases this assumption is justified. We point out, in order to be consistent,
that we also have to neglect the effect of axial forces in the analysis of the
substitute frame. This means that we have to assume cross-sectional areas

1
~l
n
I
~ I
i I
hi H
I
3 I
I
2 I Me"}

~
1=G
"
,,. 1" ",.
.",..~~--- L ---~J
(a) (b) (c)

Figure 1.1
2 Multi-storey building frames

A ---> CfJ (independently of the ratio L/ H). Consequently, the asymmetrical

load of Figure 1.1 b may be replaced by the anti-symmetrical load shown
in Figure l.lc.
In the following, we shall deal mainly with columns fixed in the foundations.
The case of pin-supported frames will be considered separately (section 1.2.4).
The effect of soil deformability is dealt with separately, too (section 1.3.3).

1.2 Multi-storey, one-bay frames

1.2.1 GENERAL APPROACH

The approximate analysis of multi-storey, one-bay, symmetrical frames subjec-

ted to horizontal loads is performed by one of the following procedures:
1. the zero moment point procedure;
2. the continuum procedure (replacement of the beams by a continuous
medium). This latter technique is only suitable for multi-storey frames
with a large degree of uniformity.
In the case of very irregular substitute frames (from the aspect of both ge-
ometry and rigidity), it is advisable to perform the analysis by computer (the
'approximation' will stem from the conversion of the substitute frame to the
actual one). We have to assume in this case, too, that the effect of axial forces is
negligible (by assuming A ---> CfJ).

1.2.2 THE ZERO MOMENT POINT (ZMP) PROCEDURE

Let us consider the one-bay symmetrical frame shown in Figure 1.2. It is

assumed that we know the position of the point where the moment diagram M
intersects the column's axis (zero moment point, ZMP), i.e. the height ho. The
problem becomes statically determinate:
V ho V(h o - h)
M bOtt =-2-; M top = 2 (1.1)

where V represents the sum of horizontal forces above the given column.
The position of ZMP may be defined by the ratio
ho
8-------~- (1.2a)
. - h - 1 - (M top/ M bott)
(M bott and M top have the same sign if they tension the same fibre).
It therefore follows:
Vf,h
M bott =-2-
 Multi-storey, one-bay frames 3

M~
I I I I

~li
I , 1_.
r1 :-'i
\1 I I
~
I I
hI I
~ .....,
I' '.'
~ ~ ---==-
II ::
JL .........

Figure 1.2

(a) (b)

Figure 1.3

-V(I-8)h
M top = 2 ( 1.2 b)

For a frame of given geometry and for a given storey i, [, depends mainly on
the ratio

(1.3)

where kb = I b / I; kc = I c/ h; and I b(l J are moments of inertia of beam i (column

i). We n~te that in'mos't practical cases the ratio v lies between 0.1 and 5.
The type of loading has some effect on the position of ZMP, but we may
safely neglect this effect and assume that for any laterally distributed load
ZMP is invariant.
Figure 1.3 presents diagrams for two extreme situations. Figure 1.3a shows
extremely stiff beams (v -> oc), where ZMP lies at the mid-height of each storey
(E = 0'5). Figure 1.3b illustrates extremely flexible beams (v -> 0); the diagram of
4 Multi-storey building frames

bending moments M is of the cantilever type. In the case of stiff beams

ZMP lies within the given storey. In the case of flexible beams ZMP may be
positioned above the given storey.
We shall now determine the approximate position of ZMP by referring to
two cases: uniform and non-uniform frames.

( a) Uniform frames
Uniform frames have equal heights, Ie = constant and Ib = constant at each storey.
Analyses were performed for n = 6,10 and 15 storeys subjected to uniform
and inverted triangular loads (Figure 1.4); for each type of frame and loading, eight
ratios v = kb/ke were considered: 0·01,0·1,0·5,1,2,5,10,1000 (a total of 48 cases).
Ratios t: were computed at three levels: 8 1 = t:G (at the ground floor), 8 2 (one
floor above) and 8 m (at the mid-height of the structure). The ratio 8 at the top
floor is not significant, as the corresponding moments are usually very small.
Average curves for 8 versus v are shown in Figure 1.5 for F.l = F.G' 1:2 and I:m :
• F.G is close to 0.5-0.6 for v> 2 and greater than 1 for v < 0.2 .
• t:m remains close to 0.5 for v > 0.5.

~l
~ H=nh

J
Ie

Figure 1.4

10

1.5
1."
1.3 ........
1.2 ............
1.1
.~

0.9' t-.... '"""" f""'-.....

-
0.8
"- I.r-I
0.7
............
~
-~ ~
0.6 Em
--- --
0.5
0-0.1
--
.-'----
0.2 0.3 0.4 0.5 2 3 4 5 10 V

Figure 1.5
 Multi-storey, one-bay frames 5

We note that even for reinforced concrete structures the column reinforcement
is usually uniform along the height of the storey. In cases where em f= 0.5 (either
em> 0.5 or t;m < 0.5), the maximum moment is greater than the moment we
have computed on the assumption that em = 0.5.
Therefore, in design it is advisable to consider the bending moments around
the mid-height of the frame, increased by 10-20% with respect to the moments
based on em = 0.5.
(;2 lies between eG and em; for v = 0.1-0.3, e2 varies between 0·5 and 0·9.
For ratios v < 0.1, the spread of results is too wide to permit a reasonable
average value to be accepted. In such cases, it is more convenient to relate the
maximum moment acting on the columns (M max above the foundations) direc-
tly to the maximum moment acting on a cantilever (Meant)' due to loads F /2;
see Figure 1.6.
From the frames we have analysed we obtain:

v = kb/ ke = 0·001 0,8-0,9 }

0·01 0-4-0'6 (1.4)
0·1 0,15-0,3

We note that even for very flexible beams (v = 0'01) the effect of the beams
remains significant (it ensures a decrease of the maximum moments acting on
the columns by ~ 50% with respect to the cantilever moments).

( b) Non-uniform frames
The position of ZMP for several cases of lO-storey frames has been examined.
The results are as follows .

• The moments of inertia of the columns vary along the height of the frame
(Figure 1.7), the beams having constant moments of inertia (various ratios
v between 0.1 and 10 have been considered). The computations show that,

Figure 1.6
6 Multi-storey building frames

Figure 1.7

if we denote
(1.5)

then we may use the curves shown in Figure 1.5 .

• The height of the ground floor (hd is greater than the height of the
remaining floors (h): see Figure 1.8. The moment of inertia is assumed
constant (I J, so that keG = I j hG < ke = I j h. Ground-floor heights
hG = 1.5 h-2 h have been taken into account.
Ratios v between 0·1 and 10 have been checked.
The results show that we may use the curves of Figure 1.5 on condition that
for ce;, we refer to

(1.6)

for £2' we refer to

(1.7)

More accurate results may be obtained in this latter case by using the continu-
um approach (see section 1.2.3).
In the case of variable heights and moments of inertia, the results obtained
by the ZMP procedure are only reliable for the ground floor.
 Multi-storey, one-bay frames 7

Figure 1.8

o Numerical example 1.1

Consider the uniform one-bay, 10-storey frame shown in Figure 1.9, loaded by horizon-
tal identical concentrated loads F = 1 kN. Ie = 1, i.e. ke = 0·333. Assume Ib = 1'333, i.e.
kb= 1'333/4=0'333; v=kb/ke= I.
The accurate moments are shown in Figure 1.9b (within the brackets). To compute
the same moments approximately, we use the curves given in Figure 1.5: v = I,
LG = 0'66, f.2 = 0'53, Lm = 0·50. The corresponding moments are shown in Figure 1.9b
(outside the brackets).
For instance, at the ground floor:
I 10 10
M bOil = -
2
IF'
1
f. G · hG = - x 0·66 x 3·0 = 9.9 kN m
2
(accurate: 9·5 kN m).

I 10
M top =--IF-(l-t:')'h =- (10)
-2 xO'34x3'0=-5'lkNm
2 1 G G

(accurate: - 5·5 kN m).

By assuming a different ratio, v = kb/ke = 0'1, the accurate moments are as shown in
Figure 1.9c. From the curves in Figure 1.5, EG = 1'30,1: 2 = 0'88; Em is uncertain. The
approximate moments are shown outside the brackets.
At the ground floor:
10
M=-x J-30x3·0=19·5kNm (accurate: 18'9kNm)
2
10
M= -2 x (1-1'3) x 3·0=4·5kNm (accurate: 4'5kNm)
8 Multi-storey building frames

Icr (-1)
(0.5)
1.5( -1.8)
1.5(1.2)
-2.25( -2.5)
2.25(2)
o -3(-3.2) (-5.3)
q 3(2.8) (0.7)
~ -3.75(-4) (-5.9)
o 3.75(3.5) (1.6)
M - 4.5( - 4.8) (-6.3)
.; 4.5(4.2) (2.7)
x .,.... - 5.25( - 5.5) (-6.3)
o.,.... (4.2)
5.25(5)
(-5.4)

1
-6(-6.2)
6(5.8) (6.6)
- 6.34( - 6.7) -1.6( -2.7
7.16(6.8) 11.9(10.8)
-5.1(-5.5)
9.9(9.5) 4.5(3.9o.£----::L
19.5(18.9)
v=1 v=0.1
f- J. £,,=0.66; E,=0.53; EG=1.30; E,=0.88
1= 4.0 m
Em=0.50
(a) (b) (c)

I
(-4)
(-30)
14(12.2) (-40)
-20( -22.1)
20(18.4) (5)
o -25.5(-27.1
(11 )
(-40)

~
25.5 (23.9)
-30(-31.4 (-40)
M 30(28.6) (19)
x -34(-34.9 (-41)
o
.,.... 34(32.6) (26)
-37(-37.6 (-38)
37(35.9) (36)
-39(-39.4 (-30)

1
39(38.6) (48)
-38(-39.3) 10(-11)
39(41.7) .----t_ _ 71(70)
-28(-29.4 25(28)
55(53.1) 107(111)

v =1 v =0.1
£,,=0.66; E,=0.53; EG=1.30; £,=0.88
1= 4.0m
Em=0.50
(d) (e) (f)

Figure 1.9
 Multi-storey, one-bay frames 9

If we refer to the cantilever moment:

10
Meant = 2 X 3·0 X (1 + 2 + '" + 10) = 82·5

yielding:

Similar cases are shown in Figure 1.9d-f, but for an inverted triangular load: the same
ratios v as considered in the first case (uniformly distributed loads) have been assumed.
Referring to the cantilever moment:

3'0) x(1·1+2·2+,,·+IO·IO)=577-5
M eant = ( 2

yielding:

Mmax III
--=--=0'19
Meant 577·5
o

o Numerical example 1.2

The frame shown in Figure 1.10 has variable heights, as well as variable moments of
inertia. The approximate moments for the ground floor only are computed as follows.

From Figure 1·5: t:G = 0·63

10
M bolt = 2 x 0'63 x 4'5 = 14·2 kN m (accurate: 14·1)

10
M top = - -2 x 0·37 x 4·5 = - 8·3 kN m (accurate: -8-4) 0

1.2.3 CONTINUUM APPROACH

This procedure is based on replacing the beams of the one-bay, multi-storey

building frame by a continuous medium. It yields satisfactory results provided
the frame is uniform, or nearly uniform: i.e. identical moments of inertia of
columns, Ie; identical moments of inertia of beams, I b (except the top slab,
where I ~ = I b/2); and not very different heights of columns (hj hj = 2/3 ... 3/2).
Let us consider a completely uniform frame subjected to a laterally distrib-
uted load of intensity Px (Figure 1.11a). The beams are replaced by a continu-
ous medium formed by an infinite number of very thin horizontal laminae at
10 Multi-storey building frames

1kN Ie
-+

U
1=3.0
Accurate Approximate

Figure 1.10

distances dx, having the moment of inertia (Figure 1.11 b):

I~ = Ibdx (1.8)
h

By using the flexibility method, we may formulate the differential equation of

compatibility of displacements and subsequently determine the couples C act-
ing on the columns at each floor (Figure Ulc).
For a uniformly distributed load:

p=-
"LF (1.9)
H
we obtain (Csonka, 1962a):

p) . smh
C i = ( rJ.2 2 . [O·cos h(rJ.x;) - smh(rJ..\";)]
. (rJ.h) . ph:'<j
+ -2-

(i=1,2, ... ,n-l) (1.10)

 Multi-storey, one-bay frames 11

)Cn=Cto p

I;'
1 =i dX
~+
P~-+
1
T-T
Xt

Xb

5' .1
ho /2

J.-j
(a) (b) (c) (d)

Figure 1.11

where:

We note that the coordinates Xi originate at the top.

Each column is now acted upon by the couples C i and the given loads
p/2kNm- 1 , i.e. ph/2kN at each joint, except the top joint, where ph/4 and
COP are acting (Figure U1c). The problem is statically determinate, and we
may compute the bending moments.:

_Mp,
Mi-T+M c, ( 1.11)

2
where M = PXi • (1.11)
p, 2'

In the case of a different height at ground floor (hG =1= h), the corresponding
couple to be considered is (Figure l.l1d)

C bott = (2~2)- {O[sin h (ctX B) - sinh (ctX T )] - (cosh(ctx B ) - cos h(ctx T )}

p(x~ - xi)
+ 4 ( 1.12)

The results in this case are not very accurate.

In fact, we may use the results obtained for the uniformly distributed loads
p (namely, the positions of the ZM Ps) for an approximate analysis of the frame
subjected to any lateral distributed load by assuming that the ZMPs are the
same.
12 Multi-storey building frames

As such, we solve the problem in two stages.

1. We determine the positions of the ZMP at each storey as for a uniformly
distributed load.
2. We compute for the given load the bending moments acting on the
columns corresponding to the ZMPs obtained in stage l.
An alternative procedure has been proposed, where the differential equation of
compatibility of displacements is replaced by a difference equation. This pro-
cedure entails an excessive volume of computations and requires a very high
degree of precision. It is therefore seldom used in design.

o Numerical example 1.3

Let us refer to the frame shown in Figure 1.12 (a nearly uniform frame): n = 10;
1= 3·0m; h = 3·0m; hG = 4·5m; H = 31·5m; Iblle = 1·5.
1·0
p = 3.0 = 0·333kNm- 1 (or: F;= I kN; Flop = 0·5kN; F boll = 1·25kN).

rJ.= 6x 1·5 )=1;

(Ix 3x 2

p) .sinh (rJ.)
C;= ( rJ.2 phx;
2: .[O·cosh(rJ.x;l-sinh(rJ.x;)]+-2-=···
= 0·709· [cosh (rJ.x;) - sinh (rJ.x;)] + 0·5 x;

i=2 x;=24m C;= 11·99kNm

3 21 10-49
4 18 8·99
5 15 7·50
6 12 6·00
7 9 4·50
8 6 3·00
9 3 1·53

ClOP =
( 2:1(2 . 0 smh 2
p ) [ . (rJ.h) - cosh(rJ.hI2) + I + 16
p·hJ 2

0.333) 0·333 x Y
= ( 2XJ2 .(1 x 2·1293-2·3524+ \)+ 16 =0.32kNm

CbOIl = (2: 2)- {e· [sinh(rJ.x B ) - sinh(rJ.x T )] - (cosh(rJ.x B ) - cosh(rJ.x T )]

p(x~ -xi)
+ 4
 Multi-storey, one-bay frames 13

p
0.75

3
6.75

12

co 18.75

27
p = 0.333 kN m- 1
36.75

48

60.75

1.25 k
I----.L". 82.69

,j.?om.,l.. Ie =1 Mp/2
Ib =0.5

-0.32 (-a.50) £ -6.7 (-9.4)

0.43 (0.25) 0.573 9.0 (6.4)
-1.10 (-1.30) -14.8 (-16.5)
1.13 (0.90) 0.507 15.2 (13.5)
-1.85 (-2.00) -21.1 (-22.7)
1.90 (1.70) 0.507 21.7 (20)
-2.60 (-2.80) -26.7 (-28.2)
2.65 (2.50) 0.505 27.3 (25.8)
-3.35 (-3.40) -31.6 (-32.9)
3.40 (3.30) 32.1 (30.9)
0.504 -35.8 (-36.8)
-4.10 (-4.30)
4.15 (4.00) 0.503 36.2 (35.2)
-4.84 (-5.00) -39.1 (-40)
4.91 (4.70) 0.504 39.7 (38.7)
-5.58 (-5.80) -41.7 (-42.6)
5.67 (5.40) 0.504 42.3 (41.4)
-6.32 (-6.80) -43 (-45.5)
6.58 (6.00) 0.510 44.7 (42.2)
-10.67 (-9.80) -65.6 (-59.9)
0.514

M M

Figure 1.12
14 Multi-storey building frames

4·5 3·0
.xB = 27 +-2 = 29·25m·' .xT = 27 - -
2 = 25'5m

0'333)
C holt = ( - -2 X
.
{1 x [sinh(l x 29'25) - smh(l x 25'5)J
2x 1

0'333(29.25 2 - 25'5 2 )
-(cosh(l x 29'25)-cosh(l x 25'5)} + 4 ~O+ 17·10

= 17·lOkNm

From the diagram Mi we deduce the ratios f..

By considering the same frame loaded with inverted triangular forces and taking
into account the ratios £, we determine the final moments (Mi = Lil Ff:h/2). The
accurate values are given within the brackets and the approximate ones outside the
brackets. 0

1.2.4 PIN-SUPPORTED FRAMES

Multi-storey building frames with pinned supports are seldom used in design,
especially if the columns are of reinforced concrete. The actual supports are
elastic, but designers usually consider them as fixed. If we want to take into
account the elasticity of the supports against rotation in the frame of an
approximate analysis we have to perform two separate analyses, the first by
considering fixed supports (Figure 1.13a) and the second by considering pin-
ned supports (Figure 1.13b), and then interpolate the results. In the case of
slender frames, the effect of the high vertical reactions will further increase the
deflections (see section 1.3.3).
The pin-supported frame has two important features .

• Its reactions are statically determinate (the assumption of the axial in-
deformability of the beams allows the replacement of the actual loads F
with pairs of anti-symmetrical loads F /2 and leads to anti-symmetrical
reactions). As a result (Figure 1.14):

MG ="1... F·hG2
• Very flexible beams (v = kb/kc->O) bring the structure close to a mechanism
and excessive deflections are to be expected. The corresponding moment
diagram is shown in Figure 1.1Sa; it is noteworthy that its shape is similar
to the shape of a moment diagram of a frame obtained by superimposing
a series of three-hinged frames (Figure 1.1Sb).
 Multi-storey, one-bay frames 15

(a) (b) (c)

Elastically fixed supports

Figure 1.13

Figure 1.14

(a) (b)

Figure 1.15
16 Multi-storey building frames

The approximate analysis of pin-supported frames may be performed by

the ZMP procedure. Computations of frames with 6,10 and 15 storeys
subjected to uniform and inverted triangular distributed loads have yielded
curves of 8 2 and 8 m as functions of v, shown in Figure 1.16. The curve f: 2
refers to the columns above ground floor and the curve 8 m refers to floors at
mid-height.

o Numerical example 1.4

For the frame shown in Figure 1.17:

kb 1·33/4·0
V=-= = 1
kc 1/3.0 1

r-Em
0.5
--- ...J.2 I
0.4
0.3
0.2
- 1
I
I
0.1 I V

0.2 0.3 0.4 0.5 2 345 10

Figure 1.16

-0.75 (-0.97)

1
0.75 (0.53)
-1.50 (-1.75)
1.50 (1.25)
-2.25 (-2.50)
2.25 (2.00)
Z -3.00 (-3.25)
.><
3.00 (2.75)
x o -3.75 (-4.00)
o g 3.75 (3.50)
~ -4.50 (-4.75)
t"i 4.50 (4.25)
x -5.25 (-5.51)
o....
5.25 (4.99)
-6.00 (-6.36)
6.00 (5.64)
-7.83 (-7.92)
___• 5.67 (5.58)
-15(-15)

"J!~ ~~
.J.- .J. M
4.0 m

Figure 1.17
 Multi-storey, one-bay frames 17

From the curves of Figure 1.16: £2 ~ 0-42: Cm ~ 0·50

M 2h ,." = 9 x 3·0 x 0'42/2 = 5·67kNm (accurate: 5'58)
M 4h "" = 7 x 3·0 x O' 5/2 = 5·25 kN m (accurate: 4,99). D

1.2.5 APPROXIMATE ASSESSMENT OF DEFLECTIONS

It is of practical interest to assess the horizontal deflections of multi-storey
buildings subjected to lateral loads, with special reference to the following:
• the general deformed shape, in order to compute the dynamic characteristics
of the structure and to identify the storeys most exposed to excessive drift;
• the maximum deflection (at the top floor), in order to check the general
flexibility of the structure;
• the maximum storey drifts (usually limited by codes: e.g. the model code
CEB-85).
The deformed shape of the structure depends on the ratio v = kb/kc and, to a
lesser extent, on the type of loading.
Typical deformed shapes are shown in Figure 1.18 for three different ranges
of the ratio v, related to uniform fixed frames subjected to distributed loads. We
note that in the usual range (v = 0'1-5) the deformed shapes are rather close to
straight lines (the maximum deviation is Ilu/u max < 20%); this is in agreement
with the assumption adopted in most earthquake regulations that, for regular
structures, a straight line may be used in assessing the distribution of seismic
loads corresponding to the fundamental mode of vibration (see e.g. CEB-85).
Typical deformed shapes for uniform pinned frames are shown in Figure 1.19.
We note that, for very flexible beams (v < 0'01), the deformed shape tends to a
straight line. For v --+ 0, it indicates the vicinity of instability: the structure
becomes a mechanism and the columns rotate without deformations. With stiff
beams (v> 2), the deformed shape tends to two straight lines joining some-
where between the first and the second floor. In the usual range (v = 0.1-5),

v<O.1 v =0.1-1 v>2

Figure 1.18
18 Multi-storey building frames

U max U max

I
I
I
I

v<O.1 v=O.1-1 v>2

Figure 1.19

the deviation from the straight line is much greater than the deviation of
fix-supported frames; the straight-line assumption in this instance is no longer
justified.
The deflections are very sensitive to changes in geometry and rigidity, and
therefore it is rather risky to make an approximate assessment; we therefore
only give indications as to the order of magnitude.
We shall represent Urnax in the form
( 1.13)
where we define by UR the maximum deflection computed on the assumption
that the beams are rigid (v = kb/kc -> oc) and by Urnax the actual maximum
deflection (Figure 1.20).
Note that all computations performed in this chapter are based on the
assumption that the effect of axial forces and shear forces on the deformations
is neglected (see section 1.1).
Figure 1.20 presents four cases:
• a uniform frame subjected to a concentrated force F at the top (Figure 1.20a):
nF h 3
UR = 24 E I (1.14)
c

where n denotes the number of storeys;

• the same frame subjected to equal concentrated loads (Figure 1.20b):
n(n + 1) F·h 3
uR = 48 E I (1.15)
c

• the same frame subjected to inverted triangular concentrated loads (Fig-

ure 1.20c):

(1.16)
 Multi-storey, one-bay frames 19

(a) (b)

4_-..-"
,I

,
f
-+ t-fo-I-

(c) (d)

Figure 1.20

• a general case: a frame where I bi , lei and Fi are variable (Figure 1.20d):
n
Urnax = L l1i'
i= 1
URi (1.17)

where

URi=
1) [(F
(24' 1 +F 2 +···+Fn)El ,
e
hi
+(F +F
2 3
+"'+F)~+
n EI
... +Fnh~J
EI
C2 en

l1i: according to the ratio Vi = kb,/ k ei

An average curve of 11 versus V is displayed in Figure 1.21, which may be used
for any laterally distributed load.
For ratios v < 0.1, we have to compare the maximum actual deflection Urnax
with the maximum cantilever deflection Ueant ; for constant moments of inertia
of cantilevers:
20 Multi-storey building frames

/l=UmaxiUR

~
15

.~
10 -.....
"- ......

0.1 0.2 0.3

.........
......

0.4 0.5
- --..
2 3 4 5 10
v

Figure 1.21

I Hi

11
1 (a) (b) (e) (d)

Figure 1.22

• a concentrated force at the top (Figure 1.22a):

FH3
( 1.18)
U cant = 6 E Ic
• equal concentrated loads F (Figure 1.22b):
nFH 3 (1.19)
Ucant = 16E I c

• inverted triangular loads (Figure 1.22c):

II(n+ 1) FmaxH4 (1.20)
U
cant
= 240 E Ie

Numerical examples indicate the following ratios:

v = 0·001 Umax/Ucant ~ 0'8-0·9
0·01 0,3-0,4
0·1 0·04 -0·08 (1.21)
 Multi-storey, one-bay frames 21

In the case of nearly uniform multi-storey frames, we may determine the

columns moment diagram by using the continuum approach and then com-
pute the maximum deflection (Figure 1.22d):

U rna • = t(2;iIJ{ Mb i ( Hi -j) + Mt i ( Hi - 2:) J (1.22)

As mentioned above, we have to check the maximum drift. For regular

uniform frames, the deformed shape is almost a straight line, and consequently
we may assume

(1.23)

As may be observed in Figure 1.18, the location of the maximum drift depends
on the ratio v = kb/kc: for flexible beams (v < 0.1), the maximum drift occurs at
the top floor, while for stiff beams (v> 2) it occurs at ground floor. When
hG> h, the maximum drift occurs usua"Ily at the ground floor. In this latter
case:

UG ::::: (3 _ 1) (L F) h~ ( 1.24)
hG - CG 12 E I G

will be taken from the curves given in Figure 1.5.

I: G
The maximum deflections of pin-supported frames (up) may be assessed as a
percentage of the corresponding maximum deflection of the fixed end frame
(up/u r)·
Referring to uniform frames with 6-15 storeys sUbjected to laterally distrib-
uted loads, the ratio up / U r varies as follows:
v = 0·001 Up/U r = 7-10
0·01 2-3
0·1 1-4-1'6
For the usual range v = 0'1-5, we may consider

(1.26)

As mentioned above, in the case of pin-supported frames with very flexible

beams the deflections tend to infinity (the structure becomes unstable).

o Numerical example 1.5

Consider the uniform, fix-supported frame shown in Figure 1.23a:

22 Multi-storey building frames
220lEIc 3781Elc

-
o

Ie

4.5
--.f
J.-4.0 flJ,j,.
(a) (b) (c)

Figure 1.23

From the curve J.1 = urnaxl uR (Figure 1.21): J.1 ~ 2·7; J.1G ~ 2·2 According to equation (1.17):

nFh 3 IOxlx4·5 3 37·97

U ----
RG - 24 E I c - 24 E I c E Ie

Since

n(n+l) n n(n-I)
----- ---
48 24 48

n(n-l)Fh 3 10 x 9 x 1 X 33 50·62
48El c 48E Ie E Ie

(2·2 x 37·97 + 2·7 x 50·62)

U max =

(accurate: 2 !OlE Ie!.

The deformed shape is shown in Figure 1.23b. The maximum drift is to be expected at
the ground floor. According to equation (1.24):

uG (3s G - I) (L.F) hb
hG 12 E IG
 The substitute (equivalent) frame method 23

~
'n' ..... -+
..... ..... ...
';'~ -+ ..... ... ~'/2

~
kb ~c ~
-+
~ [.rkb rkc
..... I~ -+

'1'-+ .....
, 1- ....
.,,. ,,. "',. ,,~ ,"
,;- ~ " '2 -,f
-,r- "l
~~ L ------*'1
(a) (b)

Figure 1.24

From the curve f./v (Figure 1.5), for vG = 1·5, we obtain EG ~ 0·61.

!lUG (3 x 0.61 -1) x to x 4.5 2 14.0

--,;;:~ 12EIe = EIe

(accurate: 5·6/4·5EI e = 12·4/EIcl-

By considering the same frame, but pin-supported, according to equation (1.26):

1.5 x 220 330

~~
- 1·5·, up ~ = - (accurate: 378/ EIcl
uf E Ie E Ie

The deformed shape is shown in Figure 1.23c. The actual drift at ground floor is much
greater than the precedent: uG / hG = 213 / 4· 5 E Ie = 47·3/ E leas against: 63/4· 5E Ie =
14/ E Ie (for the fix-supported frame). D

For computation of Urn • ., in the case of variable heights and moments of

inertia, see numerical example 1.6 in section 1.3.2.

1.3 The substitute (equivalent) frame method

1.3.1 GENERAL APPROACH

The method of the substitute (equivalent) frame is an approximate method of

analysing a multi-storey, multi-bay frame by replacing it by a multi-storey,
one-bay frame.
In the following we adopt the same assumptions that have been formulated
in section 1.1.
Let us consider the building frame shown in Figure 1.24a. The substitute
frame is defined in Figure 1.24b. Its storey heights are the same as the storey
24 Multi-storey building frames

heights of the given frame, but the span L * may be chosen arbitrarily. Its
moments of inertia are chosen so that

kc
k*c = "L2 }
( 1.27)

where

The total loads acting at each floor are the same for the given and the substi-
tute frames:

F* = IJ (1.28)

Numerous computations have shown that the substitute frame exhibits two
important features:

u* ~ u (1.29)

where u is the deflection of the given frame and u* is the deflection of the
substitute frame, and

M*~"LMc (1.30)
c - 2

where M c is the bending moment in the columns of the given frame and M: is
the bending moment in the columns of the substitute frame. We shall admit
the distribution of the total moment "L =2 M: M:
at any storey among the
columns of the given frame, in proportion to their moments of inertia:

M
Cj
=2M*~
"LI C
( 1.31)
Cj

The validity of this assumption is considered in the following.

The method of the substitute frame is an approximate one. In a single case,
it becomes an accurate method (within the limits of the assumptions defined in
section 1.1), if the given frame is a 'proportioned' one (see Figure 1.25). This
frame may be considered as consisting of a number of identical one-bay
frames, each one loaded with the same loads (Figure 1.25a, b). As their de-
formed shapes are identical, when connected in order to obtain the original
multi-bay frame, no contact forces develop in these connections. Consequently:

• the deformed shape of the original multi-storey frame is identical to the

deformed shape of the elementary one-bay frame;
 The substitute (equivalent) frame method 25

• the sum of the bending moments in the columns at any level of the given
frame is equal to the sum of the corresponding bending moments in all
the elementary one-bay frames.
We may now replace the elementary frames by the substitute frame shown
in Figure 1.25c, where

'\' k* = 2 k* = 2 I: = "L.I c
~ c c h h
or

loaded with F* = 'LF. This will lead to the accurate relationships

u* = u; M* ='LMc
c 2
Note that the effect of axial and shear forces on the deformations has been
neglected (see section 1.1).
Obviously, the 'proportioned frame' is a theoretical concept, one that is
most unlikely to be encountered in actual design.

1.3.2 REMARKS DEALING WITH THE APPLICATION OF THE SUBSTITUTE

FRAME METHOD
,
The distribution of the total moment 'L Me = 2 M: among the columns of the
given frame according to their moments of inertia is accurate only in two
cases: when the beams are very stiff (v = kb/ kc -+ CXl) or when they are very
flexible (v -+ 0). In the intermediate range (most of the actual cases), the relative
rigidities of the internal columns are higher and the relative rigidities of the
end columns are lower than those yielded by the ratios k~nt / k~ide; as we see
from Figure 1.26 the end column deformation involves a beam segment ~ 1/2,
while the deformations of the internal columns involve a beam segment
~ 21/2 = I. This property is less significant at the ground floor, where the
bottom supports are identical for both types of columns.
From the numerical examples, we conclude that:
• we may neglect this effect for v < 0.05 and v> 10;
• we have to take it into account in the domain 0.05 < v < 10.
Several procedures have been proposed in the technical literature, aimed at
improving the accuracy of distribution of the total moment 2 M~ = 'LM c be-
tween the columns [see e.g. R. Smith, cited by Rutenberg (1966), Muto (1974)],
but these techniques require excessive computation, which is incompatible
with the purpose of using approximate methods.
26 Multi-storey building frames

We propose therefore:
• To increase the moments acting on the internal columns, computed ac-
cording to equation (1.31), by multiplying them by the following corrective
coefficients:
for the ground floor: M~orrec!ed = (1 ·1 ... 1· 2)M c
for the other floors: M~orreC!ed = (1, 2 ... 1· 3) M

• to leave the moments acting on the end columns, computed according to

equation (1.31), undiminished, so as to remain on the safe side.
After distributing the total moment 2 M: = I,M c among columns, we may
deduct the end moments acting on the beams.
For an end column (Figure 1.27), the equation of equilibrium suffices:
( 1.32)
For an internal column, we shall distribute the sum of moments acting on
the column in proportion to the relative rigidities of the beams:

L (M~ + M~) kt
Mb= kL+kR
b b

( 1.33)

In assessing the maximum vertical reactions due to horizontal loading, we

note that in most cases the reactions occurring at the extremities are much
greater than the intermediate ones (Figure 1.28). Consequently, so as to remain
on the safe side:

Rmax = M /L = Mex! - I,Mr = Mex! - FTOTGGhG (1.34)

v TOT L L

where Mex! is the moment of the horizontal loads with respect to any point
lying along the support line (the overturning moment); I,M r is the sum of the
support moments.
We note that the vertical reactions due to horizontal loads are usually
smaller than the reactions due to vertical loads, especially for relatively low
frames. Bearing in mind that currently accepted codes allow for an increase in
the stresses when horizontal loads are involved, we often can neglect the
additional vertical reactions due to horizontal loads, but it is desirable to take
these additional reactions into account for frames where H> L.
As already emphasized, we may arbitrarily choose the span of the substitute
frame (L *). To obtain the vertical reactions directly from the analysis of the
substitute frame, we shall choose a span equal to the total width of the given
multi-bay frame (L * = L).
--+ --+
F Ib; Ib, Ib, Ib, F;=F; rb,
....:..... -+ -+
Ie, 2/e, 2/e, I'ei hi l"e, l"e, hi
--+ -.r- --+ -i'-
--+ --+

'"
L
9 ,. "11
L=41
, , "
J- J
I ~/-J. L/J J-/j
j,

lej"
"
(a) (b) (c)

Figure 1.25

,..-,
I '
M TOT = Mext-Mfl ~ max
Rv

M~ M~'
.> ~ M" r. .... max
Rv
~
l·l~b(-t7
ML "-.:/
. ~i+J··-:-1
,........ __
Actual distribution

J.- ---J- --L J..

I I I M'e b M'e ~ L

Figure 1.26 Figure 1.27 Figure 1.28

28 Multi-storey building frames

The horizontal reactions are more or less uniformly distributed between the
supports:
'L.F
H ~ constant = - (1.35)

where ns is the number of supports.

D Numerical example 1.6

The to-storey symmetrical frame shown in Figure 1.29a may be replaced by the half
frame shown in Figure 1.29b (by assuming axially underformable beams). The substi-
tute frame is shown in Figure 1.29c.

k* = 2 I~ = '\' k = LIe
C h ~ C h

As h = constant, I~ = L Ij2.
t x (4'17 + 2 x 7· 2 + II -4 + 5· 7) = 17·835;

W'6+3 x 3'12+0-5)=6'26; ...

k: = Ikb

€I
200~

18~

16~

14~

12~
100
~

8~

J
60-.
40
~

20...
4.5
,~
" "
,,. ,,," ,'" ,'" ",. ,,,, ----r
l L=8 x 4.25=3.40 m
"(a)
Figure 1.29
 r r r r !~
0> '"'-I (Xl co
I\) VJ .j>. 01
o o 0 c 9
it ~ -I- + +
~r ...
~

- ..... 11

I\)
a
ioU

0
en--.J
I.U

0
en--.J
I.U

0
en--.J
v ..... ;J

0
en--.J
V.""t;J

0
en--.J
u ..... ..., u .....

0
en--.J
0
en--.J
0)
Ix
""x
.... 7.21\) 3.12 3.12 3.12 0.72 0.72 0.72 0.72 10",
00) 0 0 0 0 0 0 0
""
N
(J1
... 7.2 ~ en ~ ~ ~ ~ ~
!!. 'II I\) 3.12 3.12 3.12 0.72 0.72 0.72 0.72 0.72 0.72
--.J a 0 0 0 0 0 0 0 0 0
0 ".
0)
~ ~ ~ ~ ~ ~ ~ en ~

L
'11.41\) 3.1~ 3.1~ 3.1
~
5 1.07 1.07 1.0
0 0 6en 1.070 0·16 0.110
en
~ ~ ~ ~ .~
en
,.. " I~ ~ ~ ~ ~

!!
CO
r:: ,.. VJ
~ J.
I'
; 01
.....
N ....
! 1! r ~
CO 01
0 I\) VJ
o
.j>.
o o

-
0
::J
!=l-
:§:
o
it
0
... it

1117.84 6.26 6.26 6.26 ~.75 1.75 1.75 1.75 0.6/: 0.68
~ ~

T
r;
II
-
....x
0
VJ .... 0.... 0.... .... .... .... ....
0 9 9 9 9
~ 01
~.
~ VJ VJ
'"'-I '" N :..., :..., :..., '"'-I '"'-I '"'-I '"'-I 0>
(Xl I\) I\) I\) I\) I\) I\) I\)

"
1l.~Jf- :-.I

9x3.0 J It"
 ?(67)
Y
84(129)

~i,"5(121)
66(68) 105(128) 12 156(149) 78(74)
(101) 123(148) (89
77(81) 123(147) 156 183(167) 92(84)
105(113)
(127
86(91 ) 1 ?<:I/1 <:IQ\ 168( 158) 168(159) 183 168(191) 84(94)
(148)
183(202) 168 183(168) 92(88)
100(89) .<' I 194(172) \1lU
, .......V I 7194(172) (189
183 194(181) 97(89)

:.2t ()
<nn/<M' :;> n'l/DO\

74(53) 194(212 143(110) 194(212 .~143110 194

(175
(218r
69(74) 'l'lAI'lnl:\ 120(140) ""AI'''''''' 120(141) 234 b 190(176) 117fu'l\ I ::;::>" 95(92)
.--- J,
(288
~~
301(292) x2
4

94 183

~
J.;194 (~189
189 ~
100 194 (e)

Figure 1.29 Contd.

 The substitute (equivalent) frame method 31

We choose L* = L = 17·00.
l~ 4x2'08 l~ 4xO'67
l~ = 33'28; l~ = \0·72 ...
17 4'25' 17 4'25'
Bending moments in columns (Figure 1.29d):
33'28/17
At ground floor: v = ~ O· 5
G 17'835/4'5
From the curves given in Figure 1.5: cG = O' 76.

('LF) h = (550)
M*bolt = - 2 I; - 2 xO'76x4'5=941
G G

M*top 'LF)
=- (- (550)
2- (I-I;)h = - - 2 xO'24x4'5=-297
G G

We distribute the total moments 'LMc between the columns of the given frame
according to their moments of inertia (Figur.e 1.29d):
2 x 941 x 4'17
Column I: M = ~ 220
bot! 35.67 -
- 2 x 297 x 4·17
M = -69,4
top 35.67
2 x 941 x 7·2
Column 2: M bot! -- 35.67 - 380
=
Mtop~ -120

The accurate results are shown in Figure 1.29d within the brackets, and the approxi-
mate ones outside the brackets. We have to multiply the approximate moments of the
internal columns by 1.\-1.2.
As mentioned abvoe, the results obtained for the second storey are much less
reliable:
\0'72/17
v = = 0·3
2 6'26/3
Curve of Figure 1.5: 82 ~ 0'58.
540
M bot! =-xO'58x3'0=470
2

M top = - 540)
( - 2 x 0·42 x 3·0 = - 288

2 x 470 x 1·6
Column I: M = = 120
bot! 12· 52

-2x288x 1·6
-----~-74
12·52
32 Multi-storey building frames

At the intermediate floors: v = O' 30 ... 1·08.

Curve of Figure 1.5: 8 m ~ 0.5 (ZMP at mid-heights of storeys).

"LFh
M bolt = - M top =2-

The approximate results at the top storeys are unreliable, owing to the change in the
geometrical configuration and the resulting discontinuity in rigidity. As mentioned
above, they are not essential to the design.
The end moments acting on the beams of the third floor are shown in Figure 1.2ge
(within brackets: accurate results).
The maximum vertical reactions may be computed from Figure 1.28:

Rmax =
M TOT = M ext -"LM f
v L* L

L* = 17m; M ext = 10 x 4· 5 + 20 x 7·5 + ... + 100 x 31 ·5

=12375kNm

LM f = 2 Mi = 2 x 924 = 1848 kN m;

R~ax ~ (12 375 - 1848)/17 = 619kN

(accurate: 566 kN).

The reaction on the axis of symmetry is zero:

Rsym
v
= 619 - 619 = 0

The maximum horizontal deflection will be computed according to equation (1.17):

"LFih;
U -"/I'U'
max - ~ri Ri' URi = 24£ I
Ci

fLi is a coefficient taken from the curve shown in Figure 1.21.

Ground floor:
33·28/17
17'835/4'5 ~ 0·5 ···fl ~ 4·3
Storeys 2,3,4:
kb 10'72/17
V=-= ~0·3 ... fl~6·5
kc 6· 26/3
Storeys 5,6,7,8:
kb 10'72/17
V=-= ~ 1·1 ... fl~2·5
kc 1'75/3
Storeys 9, 10:
kb 5'36/17
V=-=---~ 1·4 ... fl~2·2
kc 0'68/3
 The substitute (equivalent) frame method 33

Eu 4·3
~ = 550 x 4· 53 X + (540 + 520 + 490) x 3 3
1000 (24 x 17·84)
6·5
x + (450 + 400 + 340 + 270) x 3 3(2.5)
24 x 6·26
I 22
x +(190+ 100) x 3 3 X = 5715·890
24 x 1·75 24 x 0·68
E=3xI0 7 kNm- 2
5715890
U
max
----:7=- = 0.190 m (accurate: urnax = 0·187 m).
3 x 10
The drift at ground-floor level (equation (1.24)):

dUG = (3 SG -I)L:F h~
hG 12E fG
v = 0.5 ... SG ~ 0.76

duG/hG (3 x 0·76-1) x 550 x 4.5 2

0·00222 m
1000 12 x 3 x 10 7 x 17·84/1000

(accurate: 0.00216 m).

1.3.3 EFFECT OF SOIL DEFORMABILITY

When taking soil deformability into account, we have to replace the 'perfect
supports' (fixed ends or pinned supports) by springs, i.e. horizontal, vertical
and sometimes rotational springs. The spring constants are determined in
agreement with the given subgrade moduli of the soil; they usually vary be-
tween 20000kNm- 3 (for soft soils) and 100000kNm- 3 (for hard soils).
Soil deformability leads to an increase in the horizontal deflections and a
corresponding decrease of the frame's rigidity compared with the same struc-
ture on perfect supports.
Results of numerical examples (Figure 1.30) suggest increases of horizontal
deflections up to 30-40% for normal frames (H / L < 2). For slender frames
(H/L > 2), the increase may exceed 60-70%.
The additional vertical reactions due to the overturning moment lead to an
increase in the general rotation of the structure and subsequently also to an
increase in the horizontal deflections.
The method of the substitute frame in structural analysis (statics, dynamics and
stability) was developed in the years 1945-1970 by Kornouhov (1949), Kloucek
(1950), Melnikov and Braude (1952), Lightfoot (1956), Ehlers (1957), Rutenberg
(1966), Murashev (1971), Williams (1977), and Allen and Darvall (1977).
During the years 1930-1940, Muto developed the procedure of approxi-
mate analysis of multi-storey frames by using the zero moment procedure
(M uto, 1974).
34 Multi-storey building frames

---..

Elastic
supports
(springs) L
-If'l'------ L

Figure 1.30

The author has proposed a computer program for the approximate analysis
of multi-storey structures based on the substitute frame (Scariat, 1986).
The main references for the analysis of one-bay frames by the continuum
approach are Beck (1956), Merchant (1955), and Csonka (1962b), although the
basic concepts of this method are much older (Bleich and Melan, 1927).

1.4 First screening of existing moment-resisting

frame structures
1.4.1 PURPOSE OF FIRST SCREENING METHODS
In many cases, we need a rapid evaluation of the order of magnitude of a
structure's capacity to resist horizontal, mainly seismic, forces. It is useful for a
preliminary design, as a check of a project, and especially as a tool to check
the probable seismic resistance of existing buildings.
A review of recent developments in this field, as well as a technique pro-
posed by the author, is given in Chapter 7.
As an integral part of these techniques, we have to determine the order of
magnitude of the total horizontal resisting force (the allowable force when
using the terminology of working-stress design) for different types of structure.
In the following section we shall deal with moment-resisting frames.

1.4.2 REINFORCED CONCRETE FRAMES

Let us consider a reinforced concrete moment-resisting frame. The columns are

rectangular with cross-sections be' he (he is parallel to the given forces) (Figure
1.31). We assume that:
 First screening of moment-resisting frames 35

0.20/0.30 m

0.20/0.40 0.25/0.50

Figure 1.31

• the characteristic (cube) strength of the concrete,fek varies between 15 and

25 MPa;
• the characteristic strength of the steel, /y is 300-400 MPa;
• the reinforcement ratio is p = 0·01;
• the axial (compressive) stresses in the columns vary between 3 and 8 MPa.
By using the graphs prepared for CEB/FIP (1982), we obtain the approxi-
mate evaluation of the resisting moment Ma due to the horizontal forces Fa in
the form (Scarlat, 1993).

(1.36)

The sum L refers to all the columns; he is parallel to the horizontal forces. The
total seismic moment acting on the given storey can be expressed in the form

M=Fsh (1.37)

where h is the storey height; s defines the position of the zero moment point.
We admit s ~ 0·7 for regular beams and s~ 1 for slab-beams. By equating
(1.36) and (1.37), we obtain the total resisting force acting upon the given
frame:

(1.38)

D Numerical example 1.7

Determine the total resisting force (allowable force) for the reinforced concrete columns
of the moment-resisting frame shown in Figure 1.31. ick = 20 MPa = 20000 kN m - 2;
h = 3·20m; regular beams.
36 Multi-storey building frames

5x/180

F/r------~-·------~------~r_----_,~7

Figure 1.32

0-10 x 0-406 x 20000

Fa=
0-7 x 3-20
=362kN o

1.4.3 STEEL FRAMES

We assume that:
• the steel profiles are St37 (yield strength: 225~235 MPa);
• the assumed resisting stress, (Ja = 220 MPa = 220000 kN m - 2;
• the assumed average axial (compressive) stress due to vertical forces is
(IN = N/A = 70~100MPa = 70000~100000kN m- .
2

This yields the resisting stress due to bending moments:

(J~= 125~150MPa; we admit (J~ ~ 130 MPa
By assuming M a ~ 0- 7 F h (see section 1.4.2), the total resisting horizontal seis-
mic force will be

Ma (J~ Lw"
F =--
a 0-7 h 0-7 h
l'4(JM",W
F= aL.. x (1.39)
a h
Fa results are in kN when units of kN and m are used and in N when units of
Nand mm (MPa) are used.
Lw" is the sum of the moduli of resistance of the steel profiles and
(J~ = 130MPa = 130000kNm- 2 .
 Bibliography 37

D Numerical example 1.8

Determine the total resisting horizontal force for the steel moment-resisting frame
shown in Figure 1.32.
LW, = 5 x 19,8+3 x 278+ 5 x 19·8 = 1588cm 3
~ 0'0016m 3

.
aM = 130 MPa = 130000kNm- 2 •
1·4 x 130000 x 0·0016
F= ~91kN
• 3·20

Bibliography
Allen, F. and Darvall, P. (1977) Lateral load equivalent frame. Journal of the American
Concrete Institute, 74, 294-299.
Beck, H. (1956) Ein neues Berechnungsverfahren fur gegliederte Scheiben dargestellt am
Beispiel des Vierendeeltragers. Der Bauingenieur, 31, 436-443.
Bleich, F. and Melan, E. (1927) Die gewohnlichen und partiellen Dijferenzen Gleichungen
der Baustatik, Springer, Berlin.
CEB/FIP (1982) Manual on Bending and Compression, Construction Press, London.
Csonka, P. (I 962a) Beitrag zur Berechnung Waagerecht belasteter Stockrahmen.
Bautechnik, 7, 237-240.
Csonka, P. (l962b) Die Windberechnung von Rahmentragwerken mit hilfe von Dif-
ferenzengleichungen. Bautechnik, 10, 349-352.
Ehlers, G. (1957) Die Berechnung von Stockwerkrahmen fur Windlast. Beton- und Stah-
lbetonbau, 52 (l), 11-·16.
Kloucek, C. (1950) Distribution of Deformations, Orbis, Prague.
Kornouhov, N. (1949) Prochnosti i U stoichivosti Sterjnevih System, Stroiizdat, Moscow.
Lightfoot, E. (1956) The analysis for wind loading of rigid-jointed multi-storey building
frames. Civil Engineering, 51(601), 757-759; 51(602),887-889.
Melnikov, N. and Braude, Z. (1952) Opit Proektirovania Stalnih Karkasov visotnih zdani,
GIPA, Moscow.
Murashev, V. (1971) Design of Reinforced Concrete Structures, MIR, Moscow.
Muto, K. (1974) Aseismic Design Analysis of Buildings, Maruzen, Tokyo.
Rutenberg, A. (1966) Muitistorey frames and the interaction of rigid elements under
horizontal loads. Master Thesis, Technion, Haifa (in Hebrew).
Searl at, A. (1982) Earthquake resistant design. Lectures, Israel Association of Engineers
and Architects (in Hebrew).
Scarlat, A. (1986) Approximate analysis of multistorey buildings in seismic zones. Pre-
mier colloque national de genie parasismique, St Remy, pp. 6.1-6.10.
Scarlat, A. (1993) First screening of aseismic resistance of existing buildings in Israel.
H andassa Ezrahit iu Binian, no. 2, 8-15 (in Hebrew).
Williams, F. and Howson, W. (1977) Accuracy of critical loads obtained using substi-
tute frames, in Proc. EECS, Stability of steel structures, Liege, April, pp. 511- 515.
2 Structural walls

2.1 Introduction

Structural walls represent the most efficient structural element to take lateral
forces acting on a multi-storey building and to transfer them to the founda-
tions.
This is borne out by experience of recent earthquakes, which has clearly
shown that structures relying on structural walls have been much more suc-
cessful in resisting seismic forces than structures relying on moment-resisting
frames.
Aoyama (1981) notes with reference to the effects of recent earthquakes in
Japan that
... the amount of shear walls used played the most important role. This was also the
case in previous earthquakes ... As to frame buildings or frame portions of buildings,
shear failure was the most prevalent type of serious damage or even failure.
Wood (1991), referring to the 1985 earthquake in Vida de Mar, Chile (mag-
nitude 7.8), notes that ' ... most of these buildings [relying on structural walls]
sustained no structural damage.' Fintel (1991) has summarized the effects of 12
strong recent earthquakes and stated that: The author is not aware of a single
concrete building containing shear walls that has collapsed.'
Structural walls may be either isolated (Figure 2.1a) or form a component of
a core (Figure 2.1b). They are either without openings (uniform) or with open-
ings (coupled structural walls).

(a)

(b)

Figure 2.1
 Structural walls without openings 39

The main feature of structural walls, as compared with columns of moment-

resisting frames, relates to the fact that their rigidities are much more affected by
shear forces and soil deformability than are the rigidities of the columns of moment-
resisting frames. Both effects will be studied in some detail in the following.

2.2 Structural walls without openings

2.2.1 EFFECT OF SHEAR FORCES

Structural walls without openings may be treated as cantilever beams, but, in
contrast with usual cantilevers, where the effect of the bending moments on the
deformations is overwhelming (and subsequently the effect of the shear forces
may be neglected), the deformations of the structural walls may be strongly
influenced by the shear forces; consequently, we have to consider this effect on
the deflections, as well as on the rigidities.
In order to quantify this effect, we shall refer to the structural wall shown in
Figure 2.2, and compute successively the maximum deflections U M (by con-
sidering only the effect of bending moments) and U = U M + U v (by also adding
the effect of the shear forces):

uM=fmMdX
EI

_ _fmMdX
u-uM+U V - EI +
ffGA
VVdX
(2.l)

which give

U = UM
2S)
(1 +3 (2.2)
6fEI
s---
. - GH 2 A

I
I
I
,..-
,
H
u
.,,
,
,,,
-....---
1

H
I
I
,
l 2
pH /2 pH H
L M V m v
Figure 2.2
40 Structural walls

f is the shape factor depending on the form of the cross-section (f = 1·2 for
rectangles and 2.1 ~2.5 for I and C sections).
In the case of reinforced concrete rectangular sections and admitting
G = 0·425 E, we obtain

(2.2a)

The shear forces lead to a decrease in the rigidity of the structural wall.
The graph shown in Figure 2.3 displays the variation of the ratio u/u M versus
1/ H for a rectangular cross-section. We note that for 1/ H < ~ the effect of the
shear forces may be neglected (it represents less than 4%); for 1/ H > 1, this effect
is important (more than 10%); while for 1/ H > 1 it becomes predominant.
Table 2.1 gives:

• the ratios of maximum deflections u/u M ;

• the ratios of fixed end moments .~ / AtM;
• the ratios of rigidities K/ KM and carry-over factors t/t M •
The index (M) signifies that only the effect of bending moments has been taken
into consideration. Note that the effect of shear forces appreciably modifies the
carry-over factor (Figure 2.4).
In Figure 2.5, we can follow the change of shape of the elastic line of a
cantilever subjected to a uniformly distributed load, due to the effect of shear
forces.
Note: If we intend to assess the global effect of the shear forces on the stresses and the
deflections of a multi-storey building with structural walls, we have to refer to the ratio
1/ H, where H is the total height of the building, and not to the storey height.

uAlM

9
8
f't

~DJ
7
6

5
4

3 J.-/:J
2

1/51/3 2 3 IIH
Figure 2.3
 Structural walls without openings 41

Table 2.1 Effect of shear forces on deflections, fixed end moments and
rigidities

p-Y"1 :
, j
H
PH 3
UM= 3EI
U
- = 1 +O'5s
uM

pif" PH 4
UM = gEl
U
- = 1 +-s
UM 3
2

p1 I A
PH 2
I'~AMI =-g-
'~A
AAM 1 + O'5s

PH 2 .A

P1 1'~MI=12 -=1
AM

3EI K

1
KM=-
H KM 1 + O'5s

uo1 , 1
I.HA)=li
3EI '~A
.H AM 1 + O'5s

4EI K 1 + O'5s
KM= -

1
H KM 1 + 2s

1 tAB 1- S

"or
tABM ="2
tABM 1 + O'5s

6EI .H
I'#MI=-
H ·Jl M 1 + 2s
42 Structural walls

Mo
~~~---4~Ir----+~~--~~1

O.SMo o O.SMo Mo
5=0 4 +00
0.84 1.68 ~O

Figure 2.4

UM Uv
4th degree 2nd degree
parabola parabola

Figure 2.5

When // H < t, the effect of shear forces becomes negligible, and the analysis
can be performed by referring only to the moments of inertia.
When the structural wall is part of a dual structure, we can evaluate its
rigidity approximately, by taking into account its moment of inertia divided by
a correction factor equal to (1 + 2s), where s is computed for the total height of
the structural wall: in the case of rectangular sections, s = 1·41 /2/ H2.

2.2.2 EFFECT OF SOIL DEFORMABILITY

Design analyses are generally based on the assumption of fixed (mds supports:
i.e. the effect of soil deformability is neglected. In fact, this effect is important,
and overlooking it may lead to significant errors. Soil deformation strongly
 Structural walls without openings 43

affects the rigidity of structural walls and, as such, the magnitude of seismic
forces and their distribution between structural walls and frames, as well as
the magnitude and distribution of stresses due to change of temperature.
Detailed methods to quantify the soil deformability are given in Appendix
C. It was found advisable to use elastic models: a set of discrete springs for
large foundations and a 'global' central spring for small foundations.
The basic parameter in defining the soil deformability is the sub-
grade modulus ks ; it usually varies between 20000 kN m - 3 (soft soils) arid
100000 kN m - 3 (hard soils).
In the following, we shall deal with the effect of soil deformability on the
rigidity of structural walls. We shall check the rigidity of the structural walls
without openings, by referring to Figure 2.6a.
A fixed supports analysis yields the maximum deflection UO and the corre-
sponding rigidity (Figure 2.6b):

u =uo; (2.3)

Note: The rigidity of a structure is usually defined as 1/ u, where u is the maximum

deflection due to a given lateral load. Here, we considered a concentrated lateral force
acting at the top.
An elastic supports analysis includes three components of the total displace-
ment uT (Figure 2.6c, d):
1
KSW=T (2.4)
u
where uH is the supplementary maximum deflection due to horizontal soil
deformation and u'i' is the supplementary maximum deflection due to the
foundation's rotation (usually u'i'» uH ).

u· u H
f4u.
--- .7
-- ? t4
-7
I
I
I
.rf,
I
I
I I
I
I
-1

I
I
I

I I I
I I I
I I I
I
I ,
I I
I
I

,
I
J
(a) (b) (c)

Figure 2.6
44 Structural walls

Computations of 0·20 m thick structural walls with heights of 20-30 m on

spread footings of 1·50/6·00-2·50/10·00 m have yielded the following results:

for soft soils (k s = 20000- 30000 kN m - 3): ~~,w = 0·05-0·10

sw

for hard soils (k s = 100000 kNm- 3 ): Ksw =0·20-0·30

. K:w
If we replace the elastic supports analysis by a fixed supports analysis,
considering structural walls with an equivalent length leq (Figure 2.7), we can
obtain the same results:

for soft soils: li q

= 0-4-0·6

for hard soils: li q

= 0·5-0·7

The decrease in the structural wall rigidities involves an increase in the funda-
mental period T and a corresponding decrease in the seismic forces acting
upon the structural walls. Accepting the provisions of the SEAOC 88 code,
which gives a decrease proportional to 1/T2f3, a decrease in the seismic forces
acting on the structure of 35% (hard soils) and 65% (soft soils) is obtained
when the structural walls are the only resisting elements. For usual, dual
structures (structural walls plus moment-resisting frames), the extent of the
decrease is less, according to the rigidity of each type of structural element (see
Chapter 3).
From the above-mentioned results it follows that neglecting soil deformabil-
ity leads to significant errors in the evaluation of rigidities of the structural
walls, unacceptable in design. An approximate evaluation of the fundamental
period of a structural wall on deformable soil may be obtained as follows.

1. Compute the period To of the structural wall, based on the usual assump-
tion of elastic wall on rigid soil.
2. Compute the period T. by assuming a rigid wall on deformable soil
(subgrade modulus k.); 'rocking vibrations' (rotations with respect to a
horizontal axis) only are taken into account.

~
~ leq---J.
.., l l
)\

Figure 2.7
 Structural walls without openings 45

3. Evaluate the period T of the elastic wall on deformable soil using the
formula [due to Dunkerley (1895)]:

T~J(T6 + Tn (2.5)
The moment equation for a rigid body in plane motion is:
(2.6)
M 0 denotes the moments of forces about the axis y, and 10 denotes the polar
moment of inertia about the same axis.
For the sake of simplicity, we assume that the structure is symmetrical
about the plane yz (Figure 2.8) and uniform. After omitting negligible terms,
equation (2.6) becomes

where

yielding

T, = 2n
(J)
= 3·63 J(W9 K,pHi) (2.7)

where W is the total weight of the area tributary to the given structural wall
(kN); K<r = k,If = ks t f IN 12 (kN m rad - 1); ks is the subgrade modulus of the
soil (kN m - 3), and 9 is gravitational acceleration (m s - 2).
A similar expression is recommended in FEMA 95 (1988). We note that the
same expression can also be used in the case of asymmetrical and non-uniform

t-,

1
-f' -x

Figure 2.8
46 Structural walls

structures, without significant errors. In the case of uniform or nearly uniform

structural walls with a uniform or nearly uniform distribution of the weights
along the height (see section 6.1.4):

To ~ J(;:1 1·787
3

D Numerical example 2.1

Consider the eight storey structural wall shown in Figure 2.9. The sum of the weights of
the tributary areas is W = L It; ~ 17000 kN.

(a) Soft soil: ks = 20000 kN m- 3

The structural wall is uniform. We assume that the weights It; are uniformly distributed
along the height:

As the cross-section of the foundation is much bigger thim the cross-section of the
structural wall, we let H ~ 24 m.

3 X 10 7 x 0·2 X 5'0 3
El= =62'5 x 10 6 kNm 2
12

g=9·81 ms- 2

W=1000 kN
r 50
.,
' -t
2000
2000
2000
2000
"
J.

...
l
8x3.0=24.0
2000 ..
2000 .

Wt=50
2000
2000

~ 8.0
J.

...
---J,.
Jo
tw=0.20 m
tt =2.50 m
Figure 2.9
 Structural walls with openings 47

3
17000 X 24 )
To ~ 1·787 ( 9.81 x 62.5 x 106 = 1·11 s (accurate: 1·18 s)

I; ~ 3·63 J(;::)
20000 x 2·5 X 8.0 3
K",=k,fr = 12 =2133333kNmrad- 1

17000x252 )
T, ~ 3·63 ( 9.81 x 2.133333 = 2·59 s (accurate: 2·6 s)

T ~ )(T~ + T;) = )(1.11 2 + 2.59 2) = 2·82 s (accurate: 2·89 s).

(b) Hard soil (k, = 100000 kNm- 3 ).

To ~ 1·11 s
100000x2·5x8·0 3 -I
K",=k2Ir= 12 =10666666kNmrad

17000 X 25 2 )
I; ~ 3·63 ( 9.81 x 10 666 666 = 1·16 s (accurate: 1·17 s).

T~ )(T~ + T;) = )(1.11 2 + 1.16 2) = 1·61 s (accurate: 1·71 s).

The 'accurate results' were obtained by finite element technique, by considering the
structural wall supported on springs. 0

2.3 Structural walls with openings (coupled walls)

Most structural walls or cores have openings for windows and doors (Fig-
ure 2.10). This is an intermediate type of structure, one that is between a

D D
CJ CJ
D D
CJ CJ
0 0
r- -
,
"
Figure 2.10
48 Structural walls

structural wall without openings and a moment-resisting frame. The effect of

the shear deformations is still important, but it is less than the effect observed
in the case of structural walls without openings; it is especially significant in
the deformation of lintels, which are usually short and deep. The lintels are the
sensitive elements of the structure; they are generally the first to crack during
an earthquake, and their proper design is therefore essential for the behaviour
of the whole structure.
Moreover, their role in dissipating energy is very important, even after
cracking, as they ensure a high ductility of the structure (Paulay and
Taylor, 1981). The commentaries to the New Zealand code (1976) point out
that

well-proportioned ductile coupled cantilevered shear walls could well be the best earth-
quake-resisting structural systems available in RC. The overall behaviour is similar to
that of a moment-resisting frame but with the advantage that, because of its stiffness,
the system affords a high degree of protection against non-structural damage, even after
yielding in the coupling beams.

2.3.1 SLAB COUPLING: EFFECTIVE WIDTH OF LINTELS

Structural walls are in some cases coupled by slab coupling (the beams are
included in the slab's thickness) (Figure 2.11). In such cases, we have to assess
the effective width of the lintels to be taken into account in the structural
analysis.
According to our computations performed in the elastic range, an effective
zone with a width heff = 12 ts + tw (where ts denotes the slab's thickness and tw
the wall's thickness) can be admitted for the usual range of lintels net spans
(1'00-2'50 m), from the points of view of both the deformations and the
magnitude of vertical shear forces developed in the lintels.

2.3.2 APPROXIMATE ANALYSIS BY FRAMES WITH FINITE JOINTS

A satisfactory approximation may be obtained by replacing the coupled struc-
tural wall by frames with finite joints [the term has been coined by I. McLeod
(1971)], as shown in Figure 2.12. In this computation we have to consider the
effects of the bending moments, axial and shear forces.
The analysis is usually performed by computer. The basic stiffness matrix of
members with offset connections is displayed in the technical literature; see
Weaver and Gere (1980). This method has two important advantages: it re-
quires a rather simple program (a system of bars instead of finite elements),
and allows much simpler interpretation and application of the results than
nnite elements (we obtain directly the bending moments, the axial and the
shear forces acting on the members that we have to proportion).
 Structural walls with openings 49

Figure 2.11

o Numerical example 2.2

The coupled structural wall shown in Figure 2.13a has been replaced by the finite joint
frame displayed in Figure 2.13b.
The ratios of the approximate vertical shear forces (V~Pp) to the accurate forces
(v~CC), V~Pp/V~cc, are also displayed in Figure 2.13. At most levels V~Pp/ V~cc~0·9. The
maximum displacements are nearly equal:
app......... ace
U max = U max

Considering different component walls, we obtain

v~pp

for 12 = 2 II: ~cc = 0'85-0'90;

vapp
for 12 = 8I 1: ~CC = 0'80-0'95;
L

In all cases, :<0/ b ~ 0·5 (zero moment point in lintels). o

50 Structural walls

-
DO-
DO -
DO -
DO -
DO
"" "~ ~"

Figure 2.12

2.3.3 APPROXIMATE ANALYSIS OF UNIFORM COUPLED STRUCTURAL

WALLS BY THE CONTINUUM APPROACH

Several alternative methods dealing with the analysis of uniform coupled struc-
tural walls were developed in the 1960s (Beck, 1962; Arcan, 1964; Rosman, 1966;
Penelis, 1969). In the following, we shall refer to the Rosman method, the one
with which designers are best acquainted.
The approach is similar to the approach used for uniform one-bay frames
(section 1.2.3).
Let us consider a uniform coupled structural wall (all the floors are ident-
ical). We replace the actual structure (Figure 2.14a) by a finite joint frame
(Figure 2.14b). We then replace the lintels, positioned at distances h, by a
continuous medium formed by an infinite number of very thin horizontal
laminae at distances dx, with the moment of inertia (Figure 2.14c)

The laminae are acted upon by infinitesimal vertical shear forces X dx. Their
sum between level 0 (top) and x equals the axial force:

(2.8)

The equation of compatibility at level x (flexibility method) leads to the differ-

ential equation

Nil - rJ.2 N x = t/J M p, (2.9)

 Structural walls with openings 51

1500
-
1500

-
"+
0
-- D
3OO~

3 OO~
3 00o

~o
0.S3

0.S7
D
- - 0j/O,S

- 0.90

-- D
--
0.2/4.5
V
0.2/4 ~ 0.91
2000 kN m- 1

-
0
D
j
O'S~
2.2g
O.S
-- - o
M 0.92
x
C1l

0.92
D
- -
2.2

-
VL
.- O.

-
-
D
0 - -
3 00].
"VL

3~ o 0.S3
0.S9

n , ~
2.6
---".
"",
"
,

~- 4.5-$~- 4.5-+
.,f-- 11 ,0 ----.J<
(a) (b)

Figure 2.13

where

2 _ 12Ib(1l ~)
!Y. - hb3 21 c + A c

./, = 6/ o Ib ( -3) (2.10)

'I' hb3 I c m
Mpx is the bending moment acting upon the cantilever wall subjected to the
lateral load PX'
Solving the differential equation (2.9) yields
N x = N~om + N~ar\ N~om = C sinh!Y.x + D cosh!Y.x
N~art depends on the type of loading.
After determining N x we obtain the shear force X:

X=dN x (2.11)
dx
The shear force Xi acting upon the lintel i is shown in Figure 2.14d:
Xi ~ Xh (2.12)
 P I
0
D
I~oo
Ab\,
1 x

~
n
D
-t I
+..... t
Nx
Xdx

tl
H ) Ib.~
h
4 c W.
D +
Nx+dNx

101 Ib
thg/12
1 I I 1 1+2.a(lEf
~ 7l~ .,.100

(a) • ,- v " 1 (b) (c)

1
~
10 )
"
r t t
h

t- +Xh t-
,;
txVi;';')
h

t- ~
t- f
(d) (e)

Fiigure 2.14
 Structural walls with openings 53

When the distances between lintels are variable (Figure 2.14e) we can use the
same expression (2.12) by replacing:

h' +h"
h=-- (2.13)
2

Computation of the shear forces Xi reduces the problem to a statically deter-

minate one, and we can evaluate any stress resultant.
Rosman (1966) has prepared tables for the computation of the shear forces
X in the form

X -lJx
_ ,(tj; rxFTh)
2 (2.14)

where F T is the total lateral force acting upon the structural wall. The coefficient
IJ~ depends on the type of lateral loading, the coefficient rxH and the height x.
The tables prepared by Rosman (1966) include data for uniformly distrib-
uted loads, trapezoidal loads and concentrated forces at the top. They have
been prepared forrxH=0'5-25; when rxH>25 we may use rxH=25 (IJ~ re-
mains practically constant). IJ~ is given for the sections xl H = 0; 0·1; ... ; 0·9; 1;
when the number of storeys n # 10 we have to draw the diagram VL through
the points x I H = 0; 0'1; ... ; 1 and find VL at the levels of the lintels by interpo-
lation.

o Numerical example 2.3

The coupled structural wall shown in Figure 2.15a is subjected to a uniform lateral load
of intensity

p = 20 kN m - 1 (total load: FT = 600 kN).

th~/12 5·893 4
1= =···=--m· 10 = 6.5m.
b 1+ 2'8h 2 /h 2 10 3 '

The term 2·8 h 2 / h 2 is intended to take into account the effect of shear forces VL on the
deformations.

0·2 X 4.5 3
L= =1·519m 4 .
'12 '

!X
2 Ib(l~
= 12-
hh 3 2/, Ac
2)
- + - = ... =0'0475m- 2

!XH~6'5
 t-'0=6.5-1'
1=0.2 mt

-, I
_, i
-, i
D I
0 i
D i
I
I
1
P=20 kN m-1 _, D I
-, D_ jO.. H hb=
""3.0=30.0hn

-, D~ 2.2

-, D- 3.0

-, D
~I I D!
i D iI -J. L-
32kN

~
M

1600kN.m

(a)
.5 ~~) 4.5 .,I..
(b) (c) (d)

Figure 2.15
 Structural walls with openings 55

The vertical shear forces acting on the lintels are

VL , = 17~(~ )FTh = 17~ x 0'133 x 3 x 600= 239-417~

17~ is taken from the tables for uniformly distributed loads and rxH = 6'5. Then we
compute

and
2 N)o
M x =-(p/2)x(x /2)+-= .. ·=-5x+3·25Nx
2
The results are shown in Figure 2.15b, c, d. o
2.3.4 ASSESSMENT OF VERTICAL SHEAR FORCES IN LINTELS

In order to assess the order of magnitude of the vertical shear forces in the
lintels (VL ) and their distribution we shall refer to three types of lintel:
• flexible lintels - usually a strip of the slab (slab coupling);
• medium lintels-with spans of 2-3 m and heights of 0·50-0·80m;
• stiff lintels - with spans of 1-2 m and heights of more than 1.00 m (above
windows)
Figure 2.16 displays the diagrams of the vertical shear forces VL developed in
the lintels of the coupled structural wall shown in Figure 2.13 along the height,
for two types of lateral load: uniformly distributed and inverted triangular.
According to luravski's formula (by assuming rigid lintels):

Vu = V:h = ~h (2.15)

where V denotes the horizontal shear force acting upon the cantilever struc-
tural wall; S is the first moment of area of the wall with respect to the neutral
axis; I is the moment of inertia of the horizontal cross-section with respect to
the same axis; and h is the storey height.
I
Z = -S-
~ 0·71 for structural walls

(0'8-0'9)1 for cores

1 denotes the length of the structural wall/core, parallel to the given forces.
V L increases with the lintel rigidity; VL,J represents their upper bound. For
medium and stiff lintels, the maximum vertical shear forces occur at 0,2-0,3 of
the total height measured from the base. For flexible lintels, the vertical shear
forces are nearly constant on the upper zone of the shear wall.
The results of several computations are given in the following: the computa-
tions were performed for coupled structural walls with 8-12 storeys in order
to assess the type of variation and the order of magnitude of the vertical shear
 56 Structural walls

Uniform loading

T
,
~

I
~

----- --
Flexible lintels Stiff lintels According to Juravski

Inverted triangular loading

H
r
--- ----
--
I I I
Flexible lintels Stiff lintels According to Juravski

Figure 2.16

forces acting on the lintels. The results are given for fixed ends base (rigid soil).
They are compared with the diagram of vertical shear forces VL.J obtained by
luravski's formula. Similar results are given in section 2.3.7 for deformable
soils.
 Structural walls with openings 57

(a) Position of the maximum vertical shear force (HMI H):

Loading Flexible lintels Medium lintels Stiff lintels

type

Uniform 0·5-0·7 0'3-0'5 0,2-0-4

Inverted 0'6~0'8 0-4-0'6 0'2-0-4
triangular

(b) The sum of the vertical shear forces LVL equals the vertical reaction acting
upon each component wall (Figure 2.17a). This can be expressed as a function
of the total lateral force FT' In the following table, the ratios L VLI F T are given
for a coupled structural wall with HI L = 3:

Loading Flexible lintels Medium lintels Stiff lintels

type

Uniform 0'5- 0·9 1,1-1,7 1'5-3

Inverted 0,7-1,5 1,7-2,5 2-4
triangular

o D

(a) (b)

Figure 2.17
58 Structural walls

(c) The intensity of the maximum vertical shear forces V~ax can be expressed as
a function of the same shear force computed according to luravski (V~ax/ V~.n

Loading Flexible lintels Medium lintels Stiff lintels

type

Uniform 0,3-0,6 0,6-0,85 0,7-0,9

Inverted 0,2-0,5 0,7-0,85 0,8-0,9
triangular

(d) As an alternative, we may assess the same maximum vertical shear force
V~axas a function of the average lateral force Fav = FT/n, where n is the number
of storeys (V~ax/ Fav):

Loading Flexible lintels Medium lintels StifT lintels

type

Uniform 0'6-\ \·2-2·5 \,5-3

Inverted 0,8-\,5 \·8-3·5 2-4
triangular

The data given above relate to the maximum vertical shear force at the axis of
symmetry of each section. When the lintels are positioned at a certain distance
from the axis of symmetry (Figure 2.17b), we can compute the corresponding
vertical shear forces as a function of the shear forces in the axis of symmetry:

, _TI A!d!
V L - VL (2.16)
A!d! + A 2 d 2
After determining the vertical shear force Vv we may deduct the maximum
bending moments acting on the lintel (Figure 2.18):

~ VLb
Mmax=T (2.17)

(equal slopes at both extremities of the lintel have been assumed: f{JL ~ f{JR)'

I t4Mmax"VL~
[;:? ~VL
k b ---,/'

FiQure2.18
 Structural walls with openings 59

Note: Slab coupling is a rather dangerous solution for coupled walls: the early cracking
of the lintels significantly decreases the effect of coupling, leading to an important
increase in the bending moments acting on each component wall. The results of elastic
analyses based on uncracked members are no more valid.

2.3.5 ASSESSMENT OF MAXIMUM HORIZONTAL DEFLECTIONS

If the lintels are considered as infinitely rigid we can determine the maximum
deflections (u~ax) as for a cantilever with constant moment of inertia
(Figure 2.19):

where

Owing to the deformability of the lintels:

The ratio urnax/ u~axdepends on the actual rigidity of the lintels, as displayed in
the following table (average values for uniformly distributed and inverted
triangular loads):

D
".
h
,f-. .-
0
0
~ ~ =+t
-J--+b~I

Figure 2.19
60 Structural walls

Flexible lintels Medium lintels Very stiff lintels

3-6 2-4 1.5-2.5

For the numerical example 2.3 (section 2.3.3):

Loading Lintels
type
0.5/0.2 1.0/0.2 0.2/0.5 0.2/0.8 0.3/1.0 m
Uniform 5.4 4.2 2.9 2.0 1.7
Inverted
triangular 5.5 4.3 2.9 2.0 1.7

2.3.6 DISTRIBUTION OF VERTICAL REACTIONS

Consider a structural wall on rigid supports (Figure 2.20a), subjected to lateral

loads.
Usually one assumes a linear distribution of vertical reactions, as a result of
the rigid rotation of the structural wall. This leads to the maximum reactions:

RO _ 6 M (ns - 1)
(2.18)
V m" - -1- ns(n s + 1)

where M is the overturning moment (the moment of the resultant lateral loads)
at base, and ns is the number of supports.
In the case of structural walls without openings the distribution of the
vertical reactions yielded by an accurate analysis is shown in Figure 2.20b,
where R vm" ~ R~m",. The different distribution of the reactions is due to the
shear deformation of the wall (more significant for squat walls). In the case of
structural walls with openings, local reactions of opposite sign may occur close
to the openings (Figure 2.20c). In the case of cores on rigid supports a signifi-
cant concentration of reactions occurs at corners (Figure 2.20d):
(2.19)

2.3.7 EFFECT OF SOIL DEFORMABILITY

(a) Problems of rigidity

The rigidity of coupled structural walls on deformable soil is affected mainly
by two factors: the deformability of the lintels and soil deformability. As such,
large horizontal deflections are to be expected.
Let us denote by Urnax the maximum deflection of a coupled structural wall,
and by Urnax the corresponding maximum deflection by assuming rigid lintels
 Structural walls with openings 61

(a){~~
RVmax

(b) !f>"ZV
Vrnax

D
(c) I
I
I I
I
L/f\
~

(d)

Figure 2.20
62 Structural walls

and fixed supports (rigid soil). The ratio 13 = umax/u~ax gives the order of mag-
nitude of the expected horizontal deflections of coupled structural walls. The
results of finite elements analyses carried out for 8 and 12 storeys of coupled
walls with various lintels and soil deformability, subjected to uniform and
inverted triangular loadings, point to the following limit values of the coeffi-
cient 13:
Deformable, hard soil (k s = 100000kNm- 3 ):

Flexible lintels Medium lintels Stiff lintels

6-10 5-9 4-7

Deformable, soft soil (k s = 20000 kN m - 3):

Flexible lintels Medium lintels Stiff lintels

12-16 11-14 10-12

(h) Distrihution and order of magnitude of vertical shear forces

Figure 2.21 shows the distribution of vertical shear forces for the coupled
structural wall displayed in Figure 2.15, when a deformable weak soil is taken
into consideration.
By comparing these diagrams with those obtained for fixed ends (Fig-
ure 2.15), we note an increase in the maximum vertical shear forces of 10-15%
in the case of stiff lintels and 50- 70% in the case of flexible lintels. We have to
point out that, as the effects of cracking of reinforced concrete elements and of
soil deformability are interdependent, a direct superposition of these effects
may lead to excessive values of the coefficient 13 (see Appendix C).

( c) Distrihution of vertical reactions

By referring to Figure 2.20 and considering the soil deformability, we obtain a

distribution of reactions close to a linear one.

( d) Approximate evaluation of the fundamental period T

An approximate evaluation of the fundamental period T of coupled structural
walls on deformable soil may be obtained by the procedure displayed in
section 2.2.2, according to the formula

(2.20)
where To is the period computed for the elastic wall on rigid soil and ~ is the
period computed for a rigid coupled structural wall on deformable soil.
 Structural walls with openings 63

o Numerical example 2.4

Let us consider the coupled structural wall shown in Figure 2.22.

I = II m; t = O' 2 m; H = 30 m;
If = 14m; t = 2m; H f = 1m;
E = 3 x 107kNm- 2 (Young modulus).

Uniform loading

It'ax
L,J

...... ....
..... .... .....
II.t=O.83II.t,j
----- ---------
I~=O.93II.t,j
-_ .. ------ ---------
Flexible lintels According to Juravski
Stiff lintels

Inverted triangular loading

\,(']ax
,J

_.. .. - ....
II.t=O.84I I.t.j II.t=O,93II.t,j According to Juravski
Flexible lintels Stiff lintels

Figure 2.21
64 Structural walls

Total tributary weight,W= 5700kN.

Subgrade moduli to be accounted for: ks = \00000 kN m - 3 and k, = 20000 kN m - 3
A computerized analysis yields To = O' 35 s.
(a) ks = 100000kNm- 3
\00000 x 2 X 14 3
Kp,=.
kslr = 12 =45733333kNmrad- 1

T= 3'63J(WH~)
2
= 3'63J( 5700 x 31 ) = OAOs
r 9 K<p 9·81 x 45733333

T= J(T~ + T;) = J(0'35 + 0040 2 2

) = 0'53s (accurate: 0·50s).
In order to avoid computerized analysis, we may assess the fundamental period To
according to the data given in section 2.3.5, by relating to the period of the coupled
structural wall with rigid lintels:

0.8
D 2.2-

D
D
0
D E
0

D c:i
C')

D
D
D
o
J.- ¥ 4.5 4.5 J
-,,~,-.5:,j'L~_1
>( ~ =11.01---,I'_O..:.::.~iI'
- .. :;j<

"';."----/1=14.0----,,,"'''

Figure 2.22
 First screening of existing structures 65

For a uniform wall (see Appendix A):

To =I'78
J( WH
-
3
- ) =1·78
gEl
J( 5700 X 30
3

9·81 x 3 x 10 7 x 22·05
)
=0'27s

As the lintels are stiff, an increase of 1·3 ... 2·5 in the maximum horizontal deflection is
to be expected, yielding an increase of the fundamental period To in proportion to the
square root of the deflections: 1·1 ... 1'6; let To = !(0'30 + 0-40) = 0·35 s, resulting in the
final period of T~ J(0'35 2 + 0-40 2 ) = 0.53 s (accurate: 0.50 s).
(b) k, = 20000kNm- 3

J(
2
5700 X 31 )
K,p = 9146667 kNmrad -1; T. = 3·63 = 0·90 s
r 9.81 x 9146667
When using the approximate assessment of the period To we obtain:
T ~ (0'35 2 + 0'90 2 ) = 0'97s (accurate: 0'89s) D

2.4 First screening of existing structures laterally

supported by RC structural walls
2.4.1 INTRODUCTION

As was shown in section 1.4,1 we often need a rapid evaluation of the order of
magnitude of the seismic capacity of a structure (a detailed review of recent
developments in this field is given in Chapter 7).
As a part of this problem we shall now deal with the first screening of the
seismic capacity of a building laterally supported by reinforced concrete struc-
tural walls.
The technique we use is based on concepts set forth by Japanese engineers,
derived from statistical studies of the behaviour of numerous buildings during
strong earthquakes, especially in the Tokachi area during 1968 and 1969
(Shiga, 1977; Aoyama, 1981). A relationship has been proposed in the form

>-: 0.3" (2.21 )

Asw? 100L..A

where Asw is the total horizontal area of the reinforced concrete structural walls
in a given direction (parallel to the considered seismic forces) at the ground floor
level, and LA is the sum of the areas of the slabs above the ground floor.
Vertical loads of p = 10 kN m - 2 have been assumed. Equation (2.21) was subse-
quently checked for various earthquakes (Glogau, 1980; Wood, 1991).
Similar relationships are included in the Japanese seismic code (1987) for
structures with structural walls and columns. After adjusting the units, they
take the form (we refer to ground floor walls):
2500 Asw + 700 Ac + 1000 AS! > 0·75 Zj W (kN m) (2.22)
66 Structural walls

and
(2.23)
where Asw was defined above (m2); Ac is the total area of reinforced concrete
columns (m2); Ast is the total area of steel columns (m2); W = P L A is the total
weight of the slabs above ground floor; and Zj is the hazard seismic coefficient,
admitted in the Japanese code (1987), varying between 0.7 and 1. A rough
equivalence with the relative maximum acceleration Z used in the USA may
be expressed as Zj ~ 2·5Z (Paulay and Priestley, 1992).
When structures rely on reinforced concrete walls only, equations (2.22) and
(2.23) yield
3333Asw ~ Zj W (2.24)
1800A sw ~ Zj W (2.25)
(units: kN m).
In this case, only equation (2.25) is significant.

2.4.2 PROPOSED RELATIONSHIPS

Based on the above-mentioned data, the author has proposed the relationship
1·2 P
Asw ~ 100 Z PoLA (2.26)

where Z = amaxi g is the relative peak ground acceleration, given in seismic

maps; Po = 10 kN m - 2 is the reference vertical load; P (kN m - 2) is the consider-
ed vertical load (unfactored). Currently, P ~ Po, leading to

(2.27)

Z = 0·2 ... 0·3 yields

close to the basic relationship (equation (2.21)). It is consistent with data

obtained from the 1985 earthquake in Vifia del Mar, Chile (Wood, 1991).
We now consider the approximate total shear force derived from the SEA-
OC-88 code by considering an average soil, normal importance of the building
and referring to unfactored seismic forces:
~ 3ZW _ 3ZPL A
V= fl- fo (2.28)

where H is the total height of the structure (m).

 First screening of existing structures 67

By considering an average storey height of 3 m and n identical storeys:

H=3n;
yielding

Vc= 1·73JnZpA (2.29)

The resisting (allowable) base shear V. results in

(2.30)

where Ta is the allowable shear stress of the structural walls.

By equating (2.29) and (2.30) we obtain
144
Ta = JnP (2.31 )

Ta decreases with the number of storeys.

When a load p = 10 kN m - 2 is assumed:
n= 2 Ta = 1020kNm- 2 = 1·0MPa
4 = 720 =0·7
6\ = 588 =0·6
8 = 509 =0·5
10 = 455 =0·45
For n ~ 12 a constant value Ta = 0.4 MPa is admitted.
Note: When computing the total horizontal area of reinforced concrete structural walls
Asw we take into account walls at least 0·12 m thick and at least 1·00 m long (parallel to
the given forces). Walls with less than 1 m length are considered as columns. In the case
of coupled walls we count for the whole length Lsw (Figure 2.23) when the height of the
lintels hh (including the slab) exceeds 0·70 m; otherwise the length 2a is considered.

A-A

Figure 2.23
68 Structural walls

In the case of cores, we take into account in each direction the total length of the
walls included in the core (parallel, as well as normal, to the given forces).

In order to compare the results of the proposed relationships with the

results given by the Japanese code, we consider equation (2.25), which yields
(2.32)
As Zj~ 2·5Z
1880 Asw > 2·5 Z p LA (2.33)
2
When p= 10kNm-
1·3 1·3
Asw ~ 100 Z LA = 100 Z LA (2.34)

(compared with the proposed value Asw ~ (1·2/100) ZL: A).

The allowable shear stress yielded by equation (2.32) is:
,~= 1880kN m- 2 = 1·88 MPa
for a conventional force V' = Zj npA = 2.5 ZnpA, much higher than the force
we considered: V = l·73JnZpA (equation (2.22)).
By adjusting ,~ accordingly:

_. l·73JnZpA ~ 1·3
'a - 1 88 2
·5nZpA
= r:. MPa
....;n

leading to:
n = 4: 'a = 0·65 MPa (compared with 0·7 MPa)
n = 10: 'a = 0.41 MPa (compared with 0·45 MPa).
We note that the seismic coefficients resulting from the Japanese formulae do
not vary with the building's height as in the SEAOC-88 code used in our
proposed relationships.

2.5 Masonry structural walls

2.5.1 INTRODUCTION
Traditionally, masonry has not been accepted as a structural material fit for
modern multi-storey buildings subject to significant horizontal forces, al-
though up to the turn of the century most building structures, including tall
buildings, relied on masonry walls for lateral support.
Relatively recent experimental and theoretical research has cleared the way
for considering masonry walls as a recognized material, alongside steel and
reinforced concrete.
 Masonry structural walls 69

In the case of usual structures, i.e. in reinforced concrete or steel, the ma-
sonry walls, either facades or partitions, are not considered by most designers
as structural elements. Neglecting masonry walls has three main effects.

• It lowers the actual rigidity of the structure and subsequently leads to

underestimation of seismic forces.
• It significantly modifies the position of the centre of rigidity (see Chap-
ter 4) and subsequently the magnitude of the general moment of torsion.
• It neglects important reserves of resistance. Numerous cases are cited in
the technical literature where brick walls acting together with RC el-
ements have saved buildings from collapse during seismic attacks (Anon,
1982; Clough et al. 1990; Gulkan et at. 1990).

Several authors, as well as codes (e.g. CEB, 1985) recommend separating ma-
sonry walls from the main structural elements, in order to achieve a clear
statical scheme and to avoid undesirable eccentricity. However, we do not
agree with such proposals, as separating these walls leads to a significant
decrease in the capacity of the structure and to possible collapse due to out-
of-plane forces. We also have to recognize that the means of separation
(usually by introduction of deformable materials such as polystyrene) are not
sufficient to ensure a clear separation.
In the following, we shall refer to three main types of masonry walls: plain
masonry (unreinforced and not connected to a grid of reinforced concrete or
steel columns and beams); infilled frames (the infill masonry is not reinforced,
but is connected to a reinforced concrete grid of columns and beams); rein-
forced masonry.
The main data for determining the strength and the deformability of brick
masonry can be found in specific regulations, e.g. in UBC-88. They are as
follows.

( a) Strength data
The specified compressive strength (f~) varies as a function of the quality of
the mortar and the level of inspection, between 10 and 40 MPa for clay ma-
sonry units, and between 6 and 20 MPa for concrete masonry units. The
recommended limits for seismic zones where Z = 0·3 aref~ = 18 MPa for clay
units andf~ = 10·5 MPa for concrete units.
The allowable compressive stress (flexural) fm = 0·33f~.
The allowable shear stress 'am (Fv in UBC-88 notation) In unreinforced
masonry:

• clay units: 'am = 0·025 jJ: and 'am = 0·56 MPa

• concrete units: 'a = 0·24 MPa;
• stone with cement mortar: 'a,sm = 0,03-0,06 MPa, as a function of the
level of inspection,
70 Structural walls

(b) Deformability data

Modulus of elasticity for uncracked masonry, Em = 750 f'm for concrete ma-
sonry and Em = 1000,f'm for clay brick masonry. Shear modulus G = 0·4 E.
For cracked masonry, Hart (1989) suggested multiplying the modulus of
elasticity by 0·4.

2.5.2 PLAIN MASONRY (IN-PLANE BENDING)

The use of plain masonry as a structural element in seismic areas is not

recommended. Consequently, we have to deal only with existing plain ma-
sonry structures, in order to evaluate their seismic resistance.
We have to point out the uncertainty stemming from the questionable
quality of existing plain masonry structures (chiefly the mortar strength), and
from the conflicting data cited in the technical literature dealing with the
allowable stresses of plain masonry walls. It is therefore advisable to adopt
very approximate-and usually very conservative-methods of analysis.
By referring to data used for 'first screening' techniques (see section 2.7 and
Chapter 7) we recommend resisting (allowable) base shear v..m in a given direc-
tion in the form

where Am is the total horizontal area of the structural masonry walls in the
given direction. Several limitations in defining the area Am are detailed in
section 2.7.1.
The allowable shear stress ram can be taken between 0·03 and 0·05 MPa for
solid unit masonry, and between 0·015 and 0·03 MPa for hollow unit masonry,
as a function of the present condition of the masonry.

2.5.3 INFILLED FRAMES (IN-PLANE BENDING)

Several techniques have been proposed to evaluate the allowable horizontal

force of an infilled frames masonry wall subject to in-plane bending and axial
force. We have chosen the procedure proposed by Stafford Smith and Coull (1991).
Three failure modes are studied:
• tension failure mode, involving yield of the steel in the tension column;
• compression failure mode, involving failure dictated by the compression
strength of the diagonal strut of the panel (Figure 2.24);
• shear failure of the masonry infill, leading to the onset of flexural plastic
hinges in the RC columns (Figure 2.25).
As the second and the third failure modes are usually more dangerous, we
shall deal only with these modes.
The resisting (allowable) base shear for compression failure may be ex-
pressed in the form
(2.35)
 Masonry structural walls 71

for RC frames, and

v..m = 3fmL [cos 2 e(lchmt3)1/4] (MPa,mm) (2.36)

e
for steel frames; tan = hm/ Lm.
fm is the allowable compressive stress (MPa). The sum L refers to all the
infilled panels included in the considered wall; Lm is the length of the infill; t
and hm denote the thickness and the height of a masonry panel; Ie denotes the
moment of inertia of each column bounding a masonry panel. The sum L
extends to all the panels included in the wall.
The resisting (allowable) base shear v..m for shear failure may be expressed
in the form

(2.37)

ram is the allowable shear stress and Lm is the length of the panel. In evaluating
v;..m' an internal friction coefficient of 0·2 was assumed.

o Numerical example 2.5

Check the infilled RC frame wall shown in Figure 2.26, at the ground floor.

t = 200mm; hm = 2·80m
Allowable shear stress Tam = 0·24 MPa
Allowable compressive stressfm = 7 MPa
Base shear V = 400 kN

We shall compute the resisting base shear v;..m' D

( a) Compressive failure
Zones (1) and (2): Lm = 4700mm; hm = 2800mm; tan e = 2800/4700 = 0·596;
cos e = 0·859.

Zones (3): Lm = 3800mm; hm = 2800; tan e= 0'737; cos 0 = 0·805.

200 4
I =~-= 1·33 x 10 8 mm 4
c 12
v..m = 2 x 7 [2 X 0.859 2 x (4.5 X 108 x 2800 X 200 3 )1/4

+0.805 2 x (1,33 X 10 8 x 2800 X 200 3 )1/4] = 1541 072 N

~ 1541 kN
t )<
. _' ~iagonal
v---~ cracks Flexural
plastic
hm hinge
Shear
-t + cracking

'i<

Figure 2.24 Figure 2.25

71.4 kN
--+

57.2
-+

42.8
-+

28.6
-+

1--
t-- 5.0 L
.. 5.0 2.5~4.0=i
(a) 16.5m .., (b)

Figure 2.26
 Masonry structural walls 73

(b) Shear failure

200
v..m = 0·24 x [ 2 x 4700 x (1'39-0.16 x 2800/4700)
+ 3800 x (1'39-0'16200
x 2800/3800)
] - 49
-
~ k
1887N=492 N

v..m = 492 kN > V= 400 kN

The structural wall can resist the given lateral forces. D
Note: Deflections can be computed by replacing the infilled panels by trusses formed
by RC members (columns and beams) and masonry diagonal struts (Figure 2.27) with a
width dm=Ld/4 (Wakabayashi, 1986). Their moments of inertia are: td!/12; the
modulus of elasticity Em = 750f~.

D Numerical example 2.6

By referring to the infilled frame of Figure 2.26 we obtain the truss shown in Fig-
ure 2.28, intended to permit the computation of horizontal deflections.
RC beams: 1200/200mm, columns: 200/300 (left zone), 200/200 (right zone).
Masonry diagonal struts: 1460/200mm (left zone), 1250/200 (right zone).
RC elements: E = 30000 MPa; masonry: Em = 16000 MPa.
Computer analysis yields the elastic line displayed in Figure 2.26b. D
As is customary in design analyses we assume elastic, uncracked members for
masonry as well as for reinforced concrete, but we must be aware that, in the
case of strong earthquakes, the rigidity of infilled frames significantly decreases
owing to cracks in the RC elements, as well as in the masonry infill.
In the following, we shall refer to two tests performed in order to determine
the effect of the brick infill on the stiffness and the failure load of an infilled
frame.

Figure 2.27
74 Structural walls

Tests with statically applied forces have been performed in Johannesburg

on a three-storey RC frame (cited by Stafford Smith and-Carter, 1969) (Fig-
ure 2.29), yielding:
u/H= 1/1000 Finfill/FRC=7'4
1/500 =5·1
1/300 = 4·7 (near failure of the infilled frame)

Figure 2.28

F(kN)

80
nfilled frame Masonry infill

60 ..... -t
3.65m
0.28/0.2

40
3.65

20
4.00

~~~--~--~--~--+---~--~-+
u(mm)
. .=-..... -.k
40 50 60 70 *2.80*

Figure 2.29
 Masonry structural walls 75

where H denotes the total height and u the maximum horizontal deflection.
Note that the infilled frame exhibits a remarkably high strength up to near
failure.
Cyclic load tests have been performed on a 1/4 scale model of a seven-
storey building in two alternatives: a RC frame and an infilled frame
(Govindan et al., 1986). The effect of infill on rigidity is significant only during
the first earthquake shocks, and subsequently nearly vanishes, in spite of the
fact that the failure load of the infill frame remains much higher than the
failure load of the RC frame (approximately 2.7 times higher).
This is interpreted as indicating that the parasitic torsional moments due to
the collaboration of the masonry wall with the RC structure may be significant
only during the first earthquake shocks, when the structural rigidity is high
due to the in filled frames. When the first cracks set in, the increase in the
rigidity due to th~ masonry walls is less important (and the parasitic torsional
moments decrease as well), but the increase in the strength of the masonry
remains important and beneficial to the structure as a whole. This confirms
our view that the interaction between the main structure (steel or RC) and
masonry walls should be maintained.

2.5.4 REINFORCED MASONRY (IN-PLANE BENDING)

( a) Design for bending moments and axial forces
Proportioning of reinforced masonry is usually performed by an elastic ap-
proach. Paulay and Priestley (1992) proposed that the proportioning be based
on the flexural strength of the wall, similar to that of proportioning of RC
sections. They showed that such an approach is also reasonable for reinforcing
bars uniformly distributed along the wall. In the following, we shall adopt this
approach; we shall determine the resisting (allowable) bending moment M by
considering a constant axial force N. _.m
Referring to Figure 2.30 it may be assumed that, at failure, most of the
reinforcing bars reaching the yield point (fy) are in tension, while only a few
are in compression.
We denote:

(2.38)

(when the masonry IS confined by steel plates, we let a = 0·96 c instead of

a = 0·85 c).

(2.39)

Ts=n,AJy

where As denotes the area of one reinforcing bar.

The total number of bars n = n, (in tension) + nc (in compression).
76 Structural walls

Equivalence equations of the resultant stresses M a.m (allowable moment),

axial force N and the forces C b , C s ' Ts yield
N= C b + Cs - T,
= 0·85 C tf~ + nc AJy - nt AJy (2.40)
Ma,m = Cb(c - 0,5 a) + Ash ~)c - dJ + N (0'5 Lm - c) (2.41)
Equation (2.40) yields
c = N + (n t - nJ AJy (2.42)
0·85 tf'm
An iterative computation is performed by initially choosing a given number of
compressed bars (n c ); usually nc = 1 or 2.

D Numerical example 2.7

See Figure 2.30. Concrete masonry blocks, t = 200 mm thick; Lm = 15 x 0·20 = 3·0 m.
Specified compressive strengthf'm = 8 MPa. n = 8 steel bars DI6mm,j~ = 400 MPa; for
each bar AJy = 200 x 400 = 80000 N = 80 kN. We shall compute the resisting moment
M a .m ·
We assume nc = 2; then nt = 8 - 2 = 6. According to equation (2.42):
c = 20000 + (6 - 2) 200 x 400 = 250 mm
0·85 x 200 x 8
a=0'85c=213 mm
We assume that the compressive zone includes only one steel bar: nc = 1;
nt = 8 -1 = 7.
c = 20000 + (7 - 1) 200 x 400 = 368 mm
0·85 x 200 x 8
a = 0'85c = 313 mm
The compressive zone includes one steel bar, as assumed; no additional iter-
ation is needed.
C b = a t f 'm = 313 x 200 x 8 = 500 800 N ~ 501 k N
As equations (2.40) and (2.42) are equivalent, equation (2.40) is satisfied. Ac-
cording to equation (2.41) the resisting moment results in
M a.m = 501 (0'37 - 0·5 x 0'31) + 80 x [(0'37 - 0'10) + (0·50 - 0'37)
+ ... + (2'90 - 0'37)] + 20(0'5 x 3·00 - 0'37) = 897 kN m D

(b) Design .f(Jr shear forces

The total resisting shear force:
(2.43)
 Out-of-plane bending of masonry walls 77

where Va.m(V..sh) denote the resisting shear forces taken by the masonry (hori-
zontal steel bars).
According to Shing et al. (1989), cited by Paulay and Priestley (1992):
(2.44)
where

La ~ 0·04 J7: + -0·16A N-

m
(MPa)

0·16 N
L ;::; 0·20 + - - (MPa)
a --.;: Am

La ~ 0·52 (MPa)
Am = Lm t
v..s = 0.8j~· ASh (2.45)
where ASh denotes the total area of the horizontal steel bars of the considered
wall.

2.6 Out-of-plane bending of masonry walls

In evaluating the out-of-plane seismic forces (normal to the wall plane) we can
use the formula given by the SEAOC-88 code in the form
ZIC
F=-W (2.46)
Rw m
where Wm denotes the weight of the masonry wall and I is the importance
factor (l = 1 ... 1'25).
The maximum forces will develop at the top storeys. We admit, conserva-
tively:
C = C max = 2·75

Rw = R wm" = 1
resulting in
F=2.75ZIWm (2.47)
The wind pressure on masonry partition walls varies, according to national
codes, between 0·3 and 0·5 kN m - 2.
When plain masonry walls of existing buildings are checked for out-of-plane
forces, and we cannot ascertain the existence of an accepted type of connection
to the main structural elements (steel or RC columns), the situation may be
very dangerous; such masonry walls sustaining vertical loads must be
78 Structural walls

strengthened. In the case of infilled frames, where a connection exists between

the in fill and the structural elements, we check the infill acted upon by out-
of-plane forces, by taking into account the existing supports.
It should be pointed out that the actual resistance exhibited by the infill
masonry connected to structural elements is significantly higher than the resis-
tance resulting from usual design. This can be explained by the arching effect:
i.e. the masonry behaves like a compressed arch fixed at the extremities (Fig-
ure 2.31).
In the case of reinforced masonry, the existing reinforcing bars, even when
corresponding to minimum reinforcement ratio, usually suffice to ensure resis-
tance to out-of-plane forces.

2.7 First screening of existing buildings laterally

supported by masonry walls
In the case of masonry walls, we shall compute the resisting (allowable) base
shear v..m
in the form

(2.48)

As usual in 'first screening' procedures, we refer to existing masonry walls at

the ground floor only.
Am = LLm t denotes the total horizontal area of the brick masonry at the
ground floor in the given direction. Its computation is subject to several
restrictions (Figure 2.32): masonry walls with at least 1 m length and 150 mm
thick are taken into consideration; the masonry above the doors is neglected;
windows are included in the area Am. When stone masonry is checked we take
into consideration only walls at least 200 mm thick.
Different allowable shear stresses 'am ('a,sm) are prescribed for each type of
masonry wall (stone masonry wall).

2.7.1 PLAIN MASONRY

As stated earlier (section 2.5.3), the quality of existing plain masonry and,
especially, the quality of the mortar are very uncertain, and we must therefore
be cautious in assessing its seismic resistance.
The allowable shear stress 'a.m will be chosen according to data provided by
the codes and handbooks. FEMA-178 (1989) recommends values of 0.07 MPa
for solid concrete masonry and 0·04 MPa for hollow unit masonry. When
vertical compressive stresses of the masonry are neglected, Hendry (1990) rec-
ommends 'u,m = 0·3 MPa (ultimate strength); by letting 'a,m = 0·1 'u.m [accord-
ing to UCBC (1992)], we obtain 'a.m = 0·03 MPa. UBC-88 recommends
 /1
~i I I. 1 1.1 '+"
Reinforcement
1~
Ma,m
--
t,
bars As --r
tI· I I· I I· I I·I Wv
d,
u?1_ +
~

L
-r Lm=1SXO.20 m 'I
-~F
~
E~

~.1 -+-
II c;,I t t
a-0.37
I t
tf Asfy-BOkN
t t 1
J -
~
+-

+-1+ a =0.31
Figure 2.30 Figure 2.31

·
.
nm r··H··· . · · · .
........
::::::::
. ...
....
••••
···· . h
'::.'.
\::::
....
••••

.....
········0······0·····0···.....: .....
::::»:::::::::::::::>:::: ':>:::\
..

.. ..........................
........................ ..
............ 0 . . . . . . . . . . . . . . ..
I
~
O.BO
I 1.10 ~ I I
-hos+. M ~ +. .- t _1
Figure 2.32
80 Structural walls

= 0,04-0,08 MPa for solid unit masonry and ' •. m = 0,035-0,07 MPa for
' •• m
hollow unit masonry.
We have chosen the following values:
' •• m = 0,03-0,05 MPa for solid unit masonry
0,015-0,03 MPa for hollow unit masonry
as a function of the present condition of the masonry.
As stated in section 2.6, plain masonry walls that are not connected to
surrounding structural members and sustaining vertical loads must be
strengthened.
For stone masonry, we let ' •. sm = 0,05-0,1 MPa, as a function of the present
condition of the masonry. These values are based on the results of tests per-
formed at the University of Edinburgh, cited by Hendry (1990), and the values
specified by UBC-88:

' •• sm = 0,03-0,06 MPa.

2.7.2 INFILLED FRAMES

The allowable shear stresses are much higher than those admitted for plain
masonry: ' •. m = 0'1-0,2 MPa for solid unit masonry and 'a.m = 0,05-0,1 MPa
for hollow unit masonry, depending on the present condition of the masonry.
These values were determined by using results of computations based on the
procedure proposed by Stafford Smith and Coull (1991) (see section 2.5.3) and
on the empirical rule used in Japan (Wakabayashi, 1986): a length of 0,15-0,20
m of masonry at ground floor is required for each square metre of slab above
the ground floor.

2.7.3 REINFORCED MASONRY

In order to assess the allowable shear stress ' •. m we shall use the procedure
described in section 2.5.4, based on strength design, by assuming a minimum
steel ratio P = Pmin = 0'07%, and by conservatively neglecting the effect of the
axial force. The allowable bending moment results in

(2.49)

where

Lm denotes the length of the wall, and H is the total height of the masonry
w,all.
 Connection forces between facade panels 81

The allowable horizontal force takes the form

(2.50)

The allowable shear stress results in

(2.51 )

2.8 Connection forces between facade panels in

one-storey industrial buildings
In the case of one-storey industrial buildings, the horizontal forces are often
carried to the foundations by precast facade panels. It is important to assess
the order of magnitude of the connection forces that develop between these
panels subjected to horizontal loads (Scarlat, 1987).
Let us consider a wall of this type supported by a number of isolated
foundations (spread footings or piles); they are replaced by a continuous foun-
dation (Figure 2.33). The top face of the foundation is acted upon by horizon-
tally distributed shear forces PH = F / L and distributed moments m = PH h. The
bottom face is acted upon by the vertical forces Pv = (J b (b denotes the width of
the foundation beam).
The moment equilibrium yields

Mb h
Pmax=(Jmax b =bL 2/6=6F L2

Fh Fhx 2
Vx = VS -px = 1'5
L - - 6L3
--

M x = Px ~ + m x - v: x = F hi (2 ~: - O' 5)

Mmax = 0·1 Fh (at .x = 0·29 L)}

Fh (2.52)
Vcorresp = L
82 Structural walls

-
F= Fr/2

(a) (b)

-...,. -+
- --f h

A ~
- - - 1-- -

L
--"
- B-+

Figure 2.33

D Numerical example 2.8

F = 1120 kN. Wall length 26 m; height 12 m; thickness 0.16 m.

M max = 0·1 X 1120 X 12 = 1344 k N m

 Connection forces between facade panels 83

In section:

.X = 0·29 x 26 = 7·54 m

In the same section:

Fh 1120 x 12
Vcorresp ---=517kN
L 26

Section modulus:
0·16 X 122
----=3·8m 3
6

Maximum stress:

1344
a =--=354 kN m-2~0.35 MPa
max 3.8

Let us assume that the panels are connected at distances of a = 2 m. Each connection is
subjected to:

Horizontal forces:

PH ~ a max ba = 354 x 0·16 x 2·00 = 112 kN

Vertical shear stresses:

V 517
Tmax ~ = 317 kN m- 2 ~ 0·32 MPa
bz 0·16 x (0'85 x 12)

Vertical shear forces:

P~ 317 x 0'16 x 2·00 = 101 kN

On the axis of symmetry:

h 12
M =0' V=I'5F-=15xI120x-=775kN
, " L' 26

775
Tmax
- - - - - = 4 7 5 kN m-2~0'48 MPa
0'16(0'85 x 12)

PH = 0; Pv = 475 x 0·16 x 2·00 = 152 kN

Assuming foundations at distances of 5.2 m, these are acted upon by the forces:

6 F h 6 x 1120 x 12 1
P =-= = 119·3 kN m-
max L2 26 2

Reactions:

1120
Rv ~ 119·3 x 5·2 = 620 kN; RH~--=190kN
6
o
84 Structural walls

Bibliography
Anon (1982) The 1977 March 4 Earthquake in Romania, Ed. Acad. RSR (in Romanian).
Aoyama, H. (1981) A method for the evaluation of the seismic capacity of existing RC
buildings in Japan. Bulletin of the NZ National Society for Earthquake Engineering
14 (3), 105-130.
Arcan, M. (1964) Berechnungsverfahren fur Wandscheiben mit einer Reihe Offnungen.
Die Bautechnik, 41 (3), 95- 100.
Beck, H. (1962) Contribution to the analysis of coupled shear walls. Journal of the
American Concrete Institute, 59, 1055- 1070.
Clough, R., Gulkan, P., Mayers, R. and Carter, E. (1990) Seismic testing of single-
story masonry houses, Pt 2. Journal of Structural Engineering ASC E, 116, 257 - 274.
Dunkerley, S. (1895) On the whirling and vibration of shafts. Philosophical Transactions
of the Royal Society, 185, 269-360.
Fintel, M. (1991) Shear walls-an answer for seismic resistance? Construction Inter-
national, 13 (July), 48-53.
Glogau, O. (1980) Low rise RC buildings of limited ductility, Bulletin of the NZ Nation-
al Society for Earthquake Engineering, 13 (2), 182-190.
Govindan, P., Lakshmipaty, M. and Santhakumar, A. (1986) Ductility of infilled frames.
Journal of the American Concrete Institute, 83,567-576.
Gulkan, P., Clough, R., Mages, R. and Carter, E. (1990) Seismic testing of single-storey
masonry houses, Pt 1. Journal of Structural Engineering ASC E, 1 (116), 235- 256.
Hart, G. (1989) Seismic design of masonry structures, in The Seismic Design Handbook
(ed F. Naeim), Van Nostrand, New York, Ch. 10.
Hendry, A. (1990) Structural masonry, Macmillan.
McLeod, 1. (1971) Shear Wall-Frame Interaction, Portland Cement Association, Skokie,
IL.
Paulay, T. and Priestley, M. (1992) Seismic Design of Reinforced Concrete and Masonry
Buildings, 1. Wiley & Sons, New York.
Paulay, T. and Taylor, R. (1981) Slab coupling of earthquake-resisting shear walls,
Journal of the American Concrete Institute, 83, 130-140.
Penelis, G. (1969) Eine Verbesserung der R. Rosman-H. Beck Methode. Der Bauin-
genieur, 44 (12), 454-458.
Rosman, R. (1966) Tablesfor Internal Forces of Pierced Shear Walls Subject to Lateral
Loads, W. Ernst & Sohn, Berlin.
Scarlat, A. (1987) Precast Facade Panels Subject to Seismic forces. H andasah, December,
11-13 (in Hebrew).
Shiga, T. (1977) Earthquake damage and the amount of walls in RC buildings, in
Proceedings of 6th WCEE, New Delhi, pp. 2467-2472.
Stafford Smith, B. and Carter, C. (1969) A method of analysis for infilled frames,
Proceedings of the Institution of Civil Engineers. London, 44, 31·48.
Stafford Smith, B. and Coull, A. (1991) Tall Building Structures: Analysis and Design, J.
Wiley & Sons, New York.
Wakabayashi, M. (1986) Design of Earthquake Resistant Buildings, McGraw-Hill, New
York.
Weaver, W. Jr and Gere, J. (1980) Matrix Analysis of Framed Structures, Van Nostrand,
New York.
 Bibliography 85

Wood, S. (1991) Performance of RC buildings during the 1985 Chile earthquake: impli-
cations for the design of structural walls. Earthquake Spectra, 7 (4),607-638.
Uniform Building Code (UBC) (1991) International Conference of Building Officials,
Whittier.
Uniform Code for Building Construction (UCBC) (1992) Appendix, Ch. 1: Seismic provi-
sions for unreinforced masonry bearing walls and buildings (Advance version).
3 Dual systems (structural
walls and frames)

3.1 Introduction

Most structures are dual, and we only conventionally refer either to building
frame systems (by neglecting the structural walls), or to structural walls sys-
tems (by neglecting the columns). Dual systems are mostly taken into consider-
ation up to 30-40 storeys. Above this limit either RC tube structures (see
section 4.2) or steel structures are commonly used.
Dual systems are generally space structures but, as can be seen in Fig-
ure.3.1, we may perform the analysis in two stages: a planar and a three-
dimensional (torsion) problem. Instead of the resultant lateral force F passing
through the centre of forces (centre of masses when we deal with seismic
forces), denoted as CM in Figure 3.1, we consider an equivalent system of
forces: the resultant force F passing through the centre of rigidity CR, chosen
so that it involves mostly translations and corresponding translational forces
(i.e. a planar problem), plus the storey moment of torsion MT = F·e (i.e. a
spatial problem), yielding additional torsional forces (see section 4.1).
Generally, we have to determine the centres of masses as well as the centres
of rigidity for each storey separately. In the case of uniform or nearly uniform
structures we can assume identical centres of masses and identical centres of
rigidity.
In the following we shall assume the slabs are rigid in their planes as
compared with the vertical structural walls (the assumption is valid in most
cases, although not always).

3.2 Analysis of dual systems: classical approach

3.2.1 ANALYSIS OF DUAL SYSTEMS ON PLANAR SCHEMES

The main problem to be dealt with is the correct distribution of the lateral
forces between frames and structural walls, as a result of their interaction.
After solving this problem we may return to the approximate methods pres-
ented in Chapter 1 (for frames) and Chapter 2 (for structural walls).
Consider the structure shown in Figure 3.2a. As the effect of torsion is
neglected at this stage the deflections of each substructure 1,2, ... parallel to
the resultant force F are equal. Consequently, the analysis may be performed
$.3-';"_' Uk_Z-!i¥L,k·$.· jU2QZWJi$2S
iA
ii*£"Z&Ai4i: -Z" .-'"""-~> _ .
·CR ·CM·
•• ~ ..
_ . .,.J .CM I O-R-·
- • .CM I +
• I _I • .:CPw • .!
, . •
I :
I I

UF tF' 'iF
Figure 3.1

. --.
2 345
-
0· · • •
• • • I
• •

t t t t t 1,,1' tF2 t
F1 F2 F3 F4 Fs F1 F3 F4 Fs
(a) (e)
A-.oo
r4
F1 F2
! t-3
----.. Fs
---+

!!ill'
~

I
~ ~
I:I~ I ,
~~I~ . lifllH
~
~ I

~:I t-

~~I!I!I1111
111111/
1II i t-
~ ~ l. ,. ,,, ,Ir ,"
"
ll,o ,,,, ,. ,Ir
"
," ,1,0 JlP ,,,, ,
1,0 "
\>. ~~ ,
'---y---J , v v v " \i
2 3 4 5
(b)

Figure 3.2
88 Dual systems (structural walls and frames)

on a general planar structure consisting of the parallel substructures 1,2, ...

connected by rigid bars pinned at their ends, to ensure equal deflections (Fig-
ure 3.2b).
The analysis of this general planar structure is usually performed by
computer. The results obtained are very close to those yielded by an 'accu-
rate analysis', where torsional effects were intentionally disregarded (for
instance: preventing them by adequate horizontal supports, as shown in
Figure 3.2c).

3.2.2 APPROXIMATE METHODS FOR THE DISTRIBUTION OF LATERAL

FORCES

Consider several vertical elements of the same type, e.g. structural walls with
similar stiffness variation along the height, subjected to a given pattern of
loads (Figure 3.3). The magnitude of their horizontal deflections depends on
their rigidities, but their deformed shapes are similar. Consequently we may
distribute the total lateral forces according to the rigidities of each component
and we shall obtain nearly identical deformed shapes: in other words, the
condition of compatibility of the elastic lines is nearly satisfied.
The rigidities of the vertical elements may be defined in several ways. We
shall define them as follows:
1
ki = -
u·lmax

where ui is the maximum horizontal deflection due to the same type of

lateral lo"ict (for instance an inverted triangular load); the index i denotes the
number of the element considered.

U2max u3max u·
/..Imax
J /
I
.1-..., .J _,.
I I I
I
I
I I
I I
I
I
, I
I

I
I
, I
I I

I ,I
I

,
I
I I
I

,,
I I

,
,,,
I I
, I
I

,, I
,
I
I
, ,I
I

,,~

SW,
., ,. , ~

SWj
(k,) (k;)

Figure 3.3
 Analysis of dual systems: classical approach 89

The coefficient of distribution is defined as

d.=~ (3.1 )
I 'Lk
We point out that the coefficient of distribution di does not significantly differ
if we take into consideration another type of lateral loading in defining the
rigidity k i (obviously, on condition that we use the same loading for all the
vertical elements).
Each substructure is acted upon by the lateral forces:

PTki
Pi=PTdi = 'Lk (3.2)

In the case of dual systems (Figure 3.4) the problem of distribution of lateral
loads is more complex. We shall distinguish between two types of dual system:
structural walls with openings and frames and structural walls without open-
ings and frames.

( a) Structural walls with openings and frames

It will be recalled that coupled structural walls are an intermediate structure
between structural walls without openings and frames. When the openings are
large (doors or large windows) their deformed shape is close to that of frames.
In such cases we can distribute the lateral loads in proportion to the rigidities
of each vertical element (equation (3.2)).

(b) Structural walls without openings and frames

When the structural walls do not have openings or when the openings are
small, then their deformed shape significantly differs from that of frames, and
we have to resort to more sophisticated techniques (Figure 3.5). We shall use the
concept of equivalent dual structure proposed by Khan and Sbarounis (1964).
We replace the frames by an equivalent one-bay frame, as defined in
section 1.2:

1* = 'LIe
c 2 (columns); I~ = IIb (beams)

In the same way, we replace the structural walls by an equivalent structural

wall, defined as

The equivalent mixed structure is made up by the equivalent frame connected

to the equivalent structural wall by the usual connection bars meant to ensure
equal horizontal deflections (Figure 3.6).
 ::=f~
~-~~~
r- -_

...,"
"'
-t
-l:~

I
~
0 E
~ Lti o
x o
"' 0 «)
.J:
r
,..t;.

...I:,

@ .J:.

§§
E=
~ r
0
Lti
 Analysis of dual systems: classical approach 91

We let, based on numerical examples:

u* ~u; (3.3)

Finally, we distribute M~ = 2M~ (M:w) to each column (structural wall)

according to their rigidities.
It is usual to distribute the resultant stresses between columns and struc-
tural walls according to their rigidities, computed for each storey. This pro-
cedure may lead to significant errors due to the different effects of shear forces
on the considered rigidities. For squat structural walls (H < liS) we have to
consider the effect of the shear forces on the rigidity by using the data given in
section 2.2.1. Numerical computations show that an acceptable decrease in the
rigidities of the structural walls can be obtained by multiplying them by the
coefficien t:

(3.4)
1 + 2s

where oS is defined by equation (2.2a):

6fEI
s = - -2
GAH

f is the shape factor of the transversal section.

In the case of rectangular sections:

We point out that H denotes the total height of the structure. We have
computed uniform equivalent structures with 8, 12 and 20 storeys with 3·0m
storey height by taking into account three rigidity ratios (kswl kFr =;j:; 1; 4); for
each case, lintels with sections o· 20 x O· 50 m and 1·00 x O· 20 m (slab beams)
have been considered.
The structures have been loaded with inverted triangular and uniformly
distributed loads. Full fixity at base has been assumed (Figure 3.6).
A typical variation of the interaction forces Fe between the equivalent frame
and the equivalent structural wall is displayed in Figure 3.7; it has been drawn
assuming that the given loads have been distributed to the component struc-
tures in proportion to their rigidities.
The curves Fswl FT and Mswl MT are displayed in Figure 3.8, where Fsw =
F:w + Fe and Msw are the resultant forces and the overturning moment acting
on the equivalent structural wall; F T and MT are the resultant force and the
overturning moment of the given loads acting on the entire structure.
The curves represent average values obtained for the two types of loading
mentioned above.
92 Dual systems (structural walls and frames)

Notes: (a) Most of the lateral loads FT are taken by the structural wall, even when its
rigidity is relatively low. Consequently the bending moments occurring in the frames
are rather small.
When the structural walls are much more rigid than the frames (ksj kFr ~ 6) we may
consider that the horizontal forces are taken by structural walls only, and the frames
are 'braced'.
(b) The overturning moment acting on the structural walls (MsJ varies between 30%
of the total overturning moment M T (for ksj kFr = ±)
and 80% (for ksj kFr = 4).
Note that these results are based on the fixed supports assumption. The results
obtained by considering elastic supports, as shown in section 3.3, are quite different.
(c) Consider a structure made up of several complex substructures S l' SZ' .... Sn; we
assume that the total number of unknowns is excessive.

Figure 3.7

1
Figure 3.8
 Analysis of dual systems: classical approach 93

o storeys

0.5

kswlkFr
~----------------~----------------~--~
1/4 4

0.5

-t-________~I_--------!_....kswlkFr
1/4 4

Figure 3.8 Contd.

94 Dual systems (structural walls and frames)

We can solve the problem of the substructures interaction by using elastic models.
Each substructure Sj is replaced by a very simple equivalent model, S;', involving a
small number of unknowns. We load the complex substructure Sj (Figure 3.9a) by
lateral distributed forces - preferably of the same type as the given external forces .- and
we determine accordingly the horizontal deflections U j (by computer). The elastic model
S; is chosen as a one-bay symmetrical frame with rigid beams and elastic columns
(moments of inertia (1' I~z' ... )(Figure 3.9b). We determine these moments of inertia
from the condition of equality of the corresponding deflections (u; = u;):

I' = CI.J) h; (3.5)

c, 24£ ~ uj

where Lj F is the sum of the lateral forces above storey i and ~ U j is the drift of the
same storey. Computation of the moments of inertia (; is performed downwards,
starting from the top storey.
As a result, we replace all the given substructures SI'SZ' ... by their equivalent
models S'I' S'z . ... and then compute the deflections u; of the complete elastic model; we
load each real substructure with these given deflections and obtain the final stresses.
It has been found from numerical examples that statisfactory results (errors of less
than 10%) are obtained for:
• the computation of the horizontal deflections of the structure;
• the first mode of vibration and the corresponding fundamental period.
The resultant stresses are satisfactory only when the structural walls are more rigid
than the frames (k sw ~ k Fr ).

3.3 Effect of soil deformability and cracking of RC

elements on the distribution of lateral forces
Soil deformability leads to decreases in the rigidities of the vertical resistant
elements; the effect is much more important for structural walls and cores than
for moment-resisting frames, and subsequently leads to significant changes in

---..,
--- ... I
I
--- ... I
---~I
---1
I
I

1
I
u
/

-- llh
I I

(a) (b)

Figure 3.9
 Effect of soil deformability and cracking of RC elements 95

the distribution of lateral forces with respect to the results yielded by a classi-
cal analysis, where this effect is disregarded (Scarlat, 1993).
In order to assess the order of magnitude of this effect we have analysed the
dual systems shown in Figure 3.6, on the assumption that the supports are
elastic, and we have compared the results with those obtained by considering
fixed supports (Figure 3.10: bending moments acting on the shear wall).
The bending moment diagram acting on the structural wall yielded by a
fixed supports analysis is displayed in Figure 3.10a; it is close to a cantilever
diagram (the positive moments usually remain less than 10% of the base
moments MD.
The diagram displayed in Figure 3.1Ob is based on the assumption of a
deformable soil with a high subgrade modulus (k s = 100000 kNm ~ 3); the base
moment M~ decreases to nearly half of the fixed support moment M~ and the
positive moment increases correspondingly.
The diagram displayed in Figure 3.1Oc corresponds to a low subgrade
modulus (k, = 20000 kNm ~ 3). The base' moment (MD decreases to {-j of the
fixed end moment Mi, leading to significant positive moments; the moment
diagram is close to that of a simply supported column.
In order to assess the order of magnitude of the changes involved by
considering the soil deformability we shall refer to the eight-storey dual struc-
ture shown in Figure 3.11. Four analyses have been performed, as follows.

(a) Fixed supports

The lateral forces are distributed between the structural wall and the frame in
proportion with their rigidities. By loading them separately with the same
uniformly distributed loads we obtain

(O-O.1)M~
O.3-D.5)M;
O.9-1.2)M;'

M;=(0.4-D.6)M? M;'=(O.2-D.3)M?
(a) (b) (c)

Figure 3.10
 --+
-+
--+
r Frame:
Columns 0.4010.40m
Beams: 0.20/0.50
SW: t=0.20

---. I
8x200 kN
---.
---.
--+
--+

..,ho-J..

8x152 kN F..=~'F..
sw ksw+kF' TOT

1216kN 16416kNm(0.76xMT )
(a) Vsw Msw

-150

Fixed supports

Elastic stiff soil

-2406

-1935

876

(c) \1.1476 7329 (0.34M,)

sw Msw

-785

-3267

-3579 Elastic soft soil

398
(d) Msw

Figure 3.11
 Effect of soil deformability and cracking of RC elements 97

1
ksw = 0.845 = 1·18 = 3.15k Fr

(no interaction forces were taken into consideration).

The corresponding diagrams of the shear forces V.W and bending moments
Msw acting on the structural wall are displayed in Figure 3.11a.

( h) Fixed supports
The lateral loads act on the dual system and an accurate interaction analysis is
performed. The total forces Fsw acting on the structural wall (given loads +
interaction forces) and the corresponding diagrams V.W and Msw are displayed
in Figure 3.11 b. The maximum deflection U o = 1·04.

(c) Elastic supports

The analysis of point (b) has been repeated by assuming a stiff deformable soil
(subgrade modulus ks = 100000 kN m - 3). The results are displayed in Figure
3.11c. The maximum deflection u' = 3·24.

( d) Elastic supports
By assuming a soft deformable soil (k s = 2000 kN m - 3); we obtain the results
displayed in Figure 3.11d. The maximum deflection un = 4.03.

The moments acting on the foundations-points (c) and (d)- are compared
with the total overturning moment due to the lateral loads (M T = 19200 kN m).
We see that the results (diagrams v'w, Msw) and the maximum deflections
obtained by taking into account the soil deformability differ sharply from the
corresponding results yielded by the usual analysis, where fixed supports are
assumed. As the data dealing with soil deformability are often not reliable,
particularly when cyclic dynamic loads occur (as during earthquake attacks),
the reliability of the results based on analyses ignoring this effect is question-
able, too.
Additional doubts arise when we refer to the effect of cracking and inelastic
behaviour of RC elements. As shown in Appendix A3, the order of magnitude
of the decrease in rigidity of stnJctural elements due to this effect varies be-
tween 30% and 80%, depending on the type of element, the existing reinforce-
ment, and the intensity of compressive stresses.
It is therefore legitimate to question the validity of the 'classical analysis' of
dual systems we perform. The author's opinion is that the picture provided by
98 Dual systems (structural walls and frames)

such analyses is strongly distorted, and we have to look to other ways of

analysing dual systems.
In the following we shall refer to two approaches when dealing with the
structural analysis of dual systems: a classical approach and a limit design
approach.
The classical approach is based on two main simplifying assumptions:

• the structural elements are considered in the elastic range and the rein-
forced concrete elements as uncracked;
• the foundation soil is rigid (we neglect the effect of the soil's deformabil-
ity).

The distribution of the horizontal forces between structural walls and columns
is performed in agreement with an elastic analysis, either spatial or planar or a
combination of both. A 'complete spatial analysis' gives sometimes the agree-
able illusion of an 'accurate analysis'.
This approach is recommended in most modern codes, although many
engineers are well acquainted with its shortcomings. Both assumptions disre-
gard the significant decrease in rigidity as a result of each of the aforemen-
tioned assumptions. What is more, the decrease in rigidity sharply differs as a
function of type of the structural members (see Appendix C). However, the
temptation of the simplicity is yet too great to be resisted. Several amendments
are recommended in order to take into account the temporary excursions of at
least several structural members into the inelastic range (see Chapter 6).
In the limit design approach, the total resisting force results as a sum of the
resisting forces of the component elements in each direction; in order to take
into account the different ductilities of each type of structural element the
forces are multiplied by various participation factors. Such an approach is
presently used in the frame of 'first screening' procedures. The author is aware
of only one seismic code permitting such a technique in structural analysis: the
Japanese code (1987).

Bibliography
Anastasiadis, K. and A vramidis, I. (1988) Einheitliche Methode fur die Berechnung
gekoppelter Rahmen-Scheiben System auf elastischer Grundung. Bautechnik, 65,
(4).
Khan, F. and Sbarounis, G. (1964) Interaction of shear walls and frames. Journal of'
Structural Division ASC E, ST3, 285 336.
McLeod, I. (1971) Shear Wall-Frame Interaction, Portland Cement Association, Skokie,
IL.
Paulay, T. and Priestley, M. (1992) Seismic Design of'Reinforced Concrete and Masonry
Buildings, J. Wiley & Sons, New York.
 Bibliography 99

Scarlat A. (1989) Criteres sismiques dans Ie project du biitiment Shikmona, Haifa, in

2eme Colloque National AFPS, St Remy, pp. CB/2 7-15.
Scar/at, A. (1993) Soil deformability effect on rigidity-related aspects of multistory
buildings analysis, ACI Str. Journal, 90 (2) 156-162.
Taranath, B. (1988) Structural Analysis and Design of Tall Buildings, McGraw-Hill, New
York.
4 Space structures

4.1 Torsional forces: Lin's theory

Consider an asymmetric multi-storey structure subjected to lateral loads due

to earthquake or wind; the forces are parallel to a given direction, for instance
to the axis Y (Figure 4.1). We assume that the slabs are rigid in both the
horizontal and vertical planes; we neglect the axial deformations of the vertical
resisting elements. We have to point out that the assumption of the rigidity of
the slabs in the vertical plane is obviously untrue in most cases; it entails
'anomalies', which are further detailed.
The aforementioned assumptions enable us to check the displacements and
the stresses of columns and structural walls separately for each storey (by
'decoupling' the unknowns of the problem).
At first we assume that the storeys are identical (Figure 4.2). Consider the
storey j of the structure. Vj = 'I.:;
F j denotes the sum of the forces above the
considered storey (the storey shear force).
The relative displacement of the slabs} + 1 and} can be resolved into two
components:
• a translation parallel to the given forces, accompanied by 'translational
forces and stresses';
• a rotation about a vertical axis passing through the centre of rigidity CR,
accompanied by additional 'torsional forces and stresses'.
The position of the resultant force Vj depends on the type of forces: in the
case of wind pressure, at mid-length (Lj2); in the case of seismic forces, Vj
passes through the centre of mass CM. In determining the position of the
centre of mass we have to take into account all existing elements of the storey
}, including non-structural ones. In most cases the centre of mass is close to the
centre of gravity of the slab, considered as a geometrical area, and we can
replace the centre of mass by the centre of gravity of the slab's area.
As the storeys are identical, the positions of the centres of rigidity CR are
the same; also, the positions of the centres of mass are the same for all storeys.
The distance e between the centre of rigidity and the direction of the result-
ant force Vj is referred to as the eccentricity of the resultant force Vj. The storey
torsion moment acting on the storey} results in
MT=Vje (4.1)
In the case of non-uniform storeys (Figure 4.3) the position of the centres of
rigidity CR, as well as the positions of the centres of mass CM, differ from
storey to storey.
 '+
......

><t
~t ,..
.,;
I
, ...
Q)

::l
I• · CI
u::

· . .-
I
. I,. •
·•
:!:I
•
.0,. • • ::s.....
~, -t~

· ·1]].°
I Q)

~- - - -
• -+
r •
• •
II:
~
II:
X
<
• 4

>-
~ -t
+--g: -t
102 Space structur~s

F"
n

1=GF
",. ,,. " ,,, ~,. ,~
Figure 4.2

When determining the position of the centre of mass of a storey j, we have

to take into account the weights and the positions of the centres of mass of all
the storeys above storey j. Referring to Figure 4.3, the coordinate X CM of the
centre of mass of storey n - 2 results in ,, "

X = w"X~ + w,,-l X~-l + w,,-2X~-2 (4.2)

CM,,, 2< w" + w" _1 + w" - 2 .

where X~, X~-l' X~-2 define the centres of mass of storeys n, n-l, n-2,
considered as isolated (denoted cm in Figure 4.3). Equation (4.1) holds true
also in the case of non-uniform structures.
In the following we shall refer to any storey j without the index j. The storey
torsional moment MT acting on storey j will be defined as

(4.1a)

where V denotes the shear force of storey j. In most cases we deal with the
torsion at the ground floor, so that V denotes the base shear force.
In order to define the position of the centre of rigidity CR we choose a
temporary system of coordinates X, Y with an arbitrary origin. It can be
shown that the coordinates of the centre of rigidity CR result in (Figure 4.1):

(4.2)
YR = I,Ul', 1';) + I,.; Y a + Iy, Yb
I,I,., + I)" + I)., 1
 Torsional forces: Lin's theory 103

-- n
n-1
n-2

-- j t-I-
-
F,-... 2

~1
=GF
'!!~ , II , " ," ..

Storey n
Weight Wn
em
'i
Y~
,.Ie--

+-x~+
yt
'!'- - --
Yn-1
,I<-

yt
rn
+X'n-1.J.
I
em

~x
Storey n-1
Weight Wn -1

Pm
Storey n-2
Weight Wn-2
-t- _ __ .em
Y~-2 I
,Ie- : --+X
I

f- X~-2-+

Figure 4.3

where I x' I r are the moments of inertia of column i. For the definition of the
moments of'inertia of the structural walls (Ix., I y ., ••• ), see Note (a) below.
After determining the position of CR, we use the final system of coordinates
x, y with the origin in CR.
104 Space structures

The centre of rigidity CR is 'attracted' by vertical elements with high mo-

ments of inertia (structural walls parallel to F or cores).
Lin (1951) showed that when the principal axes of the vertical resisting
elements are parallel, then the additional torsional shear forces acting on the
element i are

(4.3)

The torsional moment acting on the same element is

2
T T Gh
M; =M 12EC It; (4.4)
o
The slab describes a rotation in its plane with the angle

T_ M T h3
qJ - -12-E-C- (4.5)
o
where h is the storey height (m); It is the torsional moment of inertia of the
element i (m4); E is the Young modulus of elasticity (kN m - 2); G is the shear
modulus of elasticity (kN m - 2); Co is the torsion constant (m6); and

where
Ceo I = L (I x, x?) + L (Ir, yn
. Gh 2 I t
Ca=Ix,xa+IrJa+ 12E ,
(4.6)

According to equations (4.3), the torsional shear forces increase with the dis-
tance to the centre of rigidity CR. Figure 4.4 shows the torsional moments of
inertia for several sections:
• Cases (b), (c), (d), (e): Thin-walled sections are assumed.
• Cases (c), (e): Ao denotes the area inside the axes of the wall.
• Cases (d): In this case the presence of the interior walls may be neglected.
• Cases (e): In the case of cores with openings an equivalent thin wall with
a thickness teq may be considered; the torsional moment of
inertia It will be computed as in case (c); f denotes the shape
factor (see section 2.2.1). The expression of the equivalent thick-
ness teq has been evaluated by Khan and Stafford Smith (1975).
Usually, the thickness teq is in the range 5-30mm. It leads to a
very significant decrease in the torsional moment of inertia,
 Torsional forces: Lin's theory 105

It=k;b t3
If - 0.3
~ = 1+(t/b)2

~t
~L ---./c
(c) I

I
-- ..
.., '4
", I ,
-~

...
,.--.
,,
I
,,
.. -
~
------.

......---
(d~-- +-- ..

r:~-
(e)

t _ 1
8--- eq - h (C 2 G + f )
12/b E Ab
h

Q--- -..I:.

Figure 4.4

which may reach 50-80% for small cores and 30- 70% for large
cores.
Notes: (a) The deformation of the columns is mainly moment dependent (Figure 4.5);
consequently their lateral rigidity may be defined as a function of their moments of
106 Space structures

(a)

Figure 4.5

inertia without correction due to the effect of shear forces:

12E Ie
K e =-h-3- (4.7)

On the other hand, the deformation of structural walls and cores depends, in most cases
on the effect of both bending moments and shear forces; consequently we have to
correct the usual moment of inertia I~w in order to account for the effect of the shear
deformations (see Chapter 2, Table 2.1):

12EI
_ _s_w.
["
K sw I ~~ (4.8)
h3 sw 1+ 2s

I~w denotes the usual moment of inertia (for rectangular sections, I:w = hl 3 112). I sw
denotes the corrected moment of inertia, by taking into account the effect of the shear
forces, too (Figure 4.5b).
The 'correction factor' 11(1 + 2s) depends on the coefficients s (see section 2.1.1):
s = 6f E r I(G A h2), where f denotes the 'shape factor'. In the case of rectangular
sections,;: 1·41/ 2 IH 2 . We point out that H denotes the total height of the structure.
When HII > 5 the effect of shear forces on the deformations can be neglected (lsw ~ I:J.
In equations (4.2) and (4.6) we have to consider the corrected moments of inertia Isw for
structural walls and cores.
(b) Lin's theory gives more accurate results when all the vertical resisting elements are
of the same type: either columns, or structural walls without openings, or structural
walls with openings.

(c) The torsional rigidity of usual columns is negligible.

(d) The torsional rigidity of thin-walled closed sections is much higher than that of
thin-walled open sections (we define a thin-walled section when the thickness I of any
component wall is less than 1/10 of its length). In the case of dual structures when open
section resisting elements are predominant, very high shear and additional (warping)
normal stresses may develop. The coordinates X, Y and x, y of thin-walled open
sections refer to the shear centres SC of these sections rather than to their centres of
 Torsional forces: Lin's theory 107

rb-,j<

L set-- sTf~;--r-
sc
(a) (b)
I
I
-,f-df- b
tf

-J.
+-
(c)
f
t-b~ -:t d _ 3tftf

-1 -6 tfb+htw
When tf=tw: d= 2+b h

'
t1
SC 3b
tw
h

-.
SC

(I)
E ~v
I

J.-- d4
"
I
d=2r sina.-acosa
a-Slnacosa

d=hl2

Figure 4.6

gravity. Figure 4.6 gives the positions of the shear centres for several thin-walled open
sections.
(e) Owing to the uncertainty as to the mass distribution, a minimum eccentricity must
be considered even for symmetric structures. Most codes (e.g. the UBC-~2 code) recom-
mend a value of O'OS L (L denotes the length of the slab, normal to the force V).
(I) Owing to its basic assumptions, Lin's theory is essentially an approximate one.
When we relinquish these assumptions, the analysis becomes an 'accurate' one; in this
case several definitions are possible for the centre of rigidity of a given storey (Jiang
et ai., 1993), but their practical interest is limited.

o Numerical example 4.1

Consider the eight-storey building with identical slabs, shown in Figure 4.7 (total height
H=8x3m=24m). Zone I: columns 0·80jO·30m. Zone II: columns O·SOjO·SOm. RC
structural walls a and b: 0·20jS.00m.
 f fr
r=-
14.92

~I
XR=20.08
t-
S.O XM=18.75 • •
t T - - - I-I
_ Jy ~ _. =
3xS.0 I- a0.2/S.0_ _
I
CM~=_~
•
CR •
b
.0.2/S.0
YM=8.70

i
t--

-l- 4xS.0 m
(I)
3xS.0
(II) -+
Figure 4.7
 Torsional forces: Lin's theory 109

We assume that the centre of mass CM coincides with the centre of gravity of the
slab's area:
XM = 18·75m; YM =8·70m
We shall determine the torsional forces in the vertical elements of the ground floor
according to Lin's theory, by assuming non-deformable soil (fixed ends).
We consider a concentrated force F = Fy = 1000 kN acting in the centre of mass of
the top slab. The shear force is constant along the height. The base shear (at ground
floor) is V= 1000kN. Columns in zone I: I x, =0'0018m 4 ; I y, =0·0128m 4 . Columns in
zone II: I x, = I y, = 0·00521 m4.
IJx, = 14 x 0·0018 + 18 x 0·00521 = 0'119m 4
I II" = 14 x 0·0128 + 18 x 0·00521 = 0·273 m 4

I(/x,X;) =0'0018(4 x 0+2 x 5+4 x 10+4+ 15)

+ 0·00521 [5(20+ 25 + 30)+ 3 x 35] = 2·698

I(/y, 1;) = ... = 2·307

Structural walls a and b:

0·2 X 53
Moments of inertia about x:I = - - - = 2·083
x 12

As HI/ = 24/5 = 4·80 < 5 we shall consider the effect of shear forces on the deforma-
tions. The correction factor:
I -41/2 I -41 X 52 I
S = ~ = -2-4-=2- = 0·0612; --=0·891
1+28
I x =2'083xO'891 = 1'856m 4
Moments of inertia about y: As HI/ = 24/0' 2 = 120> 5, we shall neglect the effect of
shear deformations:
5 X 0.2 3 4
I y --:::.1
.\'
=---=0'0033m
12

The position of the centre of rigidity CR:

2·698 + 1·856 x (5+35)

= = 20'08m
0'119+2xl'856
YR= ... =8-43m
The torsional moment of inertia:
5·0xO·2 3
I =1 = =0'0133m 4
1" '" 3
110 Space structures

We let G/E = 0-425.

The torsion constant:

Co = Ceol + C a + C b
Cco1="(/ X2)
~ x, I
+ "(1,
~},
y2)
I

I(1 x,x;) = 0·0018 [4(20'08 - W + 2(20'08 - 5f

+ 4(20'08-IW + 4(20'08-IWJ + 0·00521 [5(20'08 -2W
+ 5(20'08-2W + 5(20'08-3W + 3(20'08-3WJ = 11·31 m 6
I(/yX) = ... =11'79m 6
Ceol = 11·31 + 11·79 = 23'lOm 6
I
x 2 + I y, y2a + Gh 2 --1...-
Ca = I x,'a 12E = 1'856(20'08 - W

0-425 X 32 X 0·0133
+0'0033(8-43-7'W+ =422'lm 6
12
Cb = ... = 413.2 m 6
C o = 23'10+422·1 +413'2~858m6

The storey torsional moment results in

MT = 1000(20'08-18'75)= 1330kNm
The torsional shear forces resisted by the structural wall a are
MT IT xa 1330 X 1·856(20'08 - 5)
~ VT = x, = = 43-4 kN
y, Co' 858
~ VTx, ~ O.
The torsional moment resisted by the structural wall a is:
MT Gh 2 I
M = t,
t, 12E Co
0·1333
= 1330 x 0-425 x 32 x = 0·007 ~ 0
12 x 858
The torsional forces taken by the columns are negligible.
An accurate analysis (as a tri-dimensional structure) yields:
Eccentricity e = (XR -XM)= 20'11 -18·75 = 1·36m
The storey torsional moment MT = 1000 x 1·36= 1360kNm o

4.2 Approximate torsional analysis of dual structures

based on Lin's theory
Dual structures with eccentric structural walls/cores are subjected to high
supplementary torsional shear forces, When determining the position of the
 Approximate torsional analysis of dual structures 111

centre of rigidity CR (the coordinates XR, YR) and the torsional shear forces,
the contribution of columns is usually much smaller than the contribution of
structural walls/cores. Consequently, we may admit some simplifications re-
garding the columns without adversely affecting the results (Scarlat, 1986).
These simplifications are intended to avoid the cumbersome computations
resulting from Lin's theory.
Consider a slab supported on the columns i and the structural walls/cores
a, b, ... (Figure 4.8a).
We denote

Icol
x
= "I .
Lx,'
Icol
y
="
L..I y, (4.9)

We replace the columns i (area Ai' moments of inertia lx" Iy) by an infinite
number of identical infinitesimal columns (dA, dl x , dl y ) covering the entire
area of the slab (Figure 4.8b). Assuming the area of the slab to be

we admit:
col
dl x~ I
dA = cons t an t -- Ax *
dl Icol
-y ~ constant = -Y-
dA A*

It follows that

~ (I~ol/ A*) JX dA + Ix. Xa + IXbXb

Icol
x
+I +I x~ Xh

The centre of gravity of the area A* is defined by

 • ><

~ .Q

•
1\
0:
+-'"
,.....'L
oJ
() -;.-><
..........
.8 w.

+~
... ...!<
II
::0.,"

.CIl

I':

1 1
j

.~
.Q

\
-\
,
f
c..c:..-=.=~ _-.=-=.:1

::..,.
~
•
•
-• • • •
.0:- •
~
0:
()
-+
()
• • •
-• ••
•
•
-t i
• • •
p t:::==-:==~

>- _\ >-
• \
1l1li(
 Approximate torsional analysis of dual structures 113

yielding (Figure 4.8c)

(4.10)

We may interpret these formulae geometrically as follows: in order to com-

pute the position of the centre of rigidity CR we can replace the columns by a
single resultant column having the moments of inertia I~ol = 'LJx. and I~ol =
'LJy .' positioned in the centre of gravity G of the area of the slab.
We recall that in most cases the centre of mass CM coincides with the centre
of gravity G of the slab's area.
Similarly

'\'(1 X2) ~ fX 2dA (dl x ) ~ fC ol fX 2dA

L.. x, ' - dA - x A*
y2
'\'(1
L.. y,
y2) ~ f},2 dA (dl
' -
y)
dA -
~ leol f dA
A* Y

but Jx 2 dA and Jy 2 dA represent the moments of inertia of the area A * with

respect to axes passing through the centre of rigidity CR:

fX2 dA = I;; fY2 dA = I;

This yields

lcol/*
'\'(1
L.. y,
y2) ~ ~
, - A*
= lcol i*2
Y x

The torsion constant Co becomes

Co = C eol + C a + C b
where

(4.11)

Xe-ly-
_ .* _
I* )
JrtJ
A*'
Y. _
Ye-lx-
.* _
JrtJI*
A* )
x
114 Space structures

We may interpret these formulae geometrically as follows: in order to com-

pute the torsion constant Co we can replace the columns by a single resultant
column (I~ol, I~ol) positioned in a point CC (Figure 4.8d), at a distance defined
by the coordinates

Yc = i~
from the centre of rigidity CR.
We note that the sign of the coordinates XC' Yc has no practical significance.
When the slab is divided into several zones with significant differences be-
tween the rigidities of the columns, we have to consider separately each zone
(I, II) and compute the geometrical data accordingly:

l 1': col 1': + I Y + I Y. (4.12)

YR. ~
Ico
YU! G m
+ I Yon G y, a
OIl}\ b
- I~ol + I~ol + I" + I,.
J(I) _ (111 J ~ J I>

Note that when the centre of rigidity is close to the centre of mass, the analysis
is very sensitive to inaccuracies occurring in the computation of the eccentric-
ity, but then the whole problem of torsion is not important.
When structural walls or cores with high rigidities are present, we may
compute the position of the centre of rigidity CR and the torsion constant Co
by neglecting the presence of the columns:

(4.13)

o Numerical example 4.2

Consider the slab shown in Figure 4.7. The coordinates of the centre of rigidity and the
torsional shear forces will be computed according to the approximate method. As the
columns in zones I and II are very different, we shall divide the area of the slab into two
zones (Figure 4.9) and compute accordingly the coordinates XR, YR (equation (4.12)).
The areas of the slabs:

Centres of gravity:

x GUll
= 27-Sm''
 Approximate torsional analysis of dual structures 115

Y
t
~
1 XR=20.76
1
~
I
II
1
1'"
5.0

t - - 4_~·GIII
• •
ecc
3x5.0

- - • •
-GI CMCR- - - b-

"I
I
l- l ~X
1

I I
:f- 4x5.0 -----t-- 3x5.0

Figure 4.9

Total moments of inertia of columns:

The centre of rigidity CR:

0·0252 x \0 + 0·0938 x 27· 5 + 1·856 x (5 + 35)

= = 20.12m
0·0252 + 0·0938 + 2 x 1·856
(accurate result: 20·08 m)
YR = ... = 8·34 m (accurate result: 8-43 m)

Area of the slab: A * = 600 m 2

The moments of inertia of the slab with respect to axes passing through CR:
J: =61562m 4

Xc = i: = 10·12; Yc = i: = 5·26m
116 Space structures

+ I }\
col
Ccol = I col
x Xc
2
Yc
2

= 0·119 X 10'12 2 +0·273 X 5'25 2

= 19'7m 6
2
2 2 G h It
Ca=Ixxa +I .. Ya
, f.
+-_.
12£

OA25 X 3'0 2 x 0·0\ 33

= 1·856 x 15'12 2 +0'0033 X 0'84 2 + - - - - - - -
12

C b = .. ·=41Im 6

Co = 19·7 +424·3 +411 = 855 m 6

MT = 1000(20'12 -18'75) = 1370kNm
(accurate result: 1330 kNm).
T 1370 x 1'856(20'12 - 5)
~V = =45kN
h 855

(accurate result: 43 kN)

~ V;, = 0·005 kN
1370 x OA25 x 3.0 2 x 0'0133
M =
r. 12 x 855
=0'007kNm o

4.3 A critical review of Lin's theory

4.3.1 EFFECT OF VERTICAL DEFORMABILITY OF SLABS

A basic assumption of Lin's theory implies the neglect of the vertical deforma-
bility of the slabs. This leads to two anomalies.
First, according to Lin's theory, torsional shear forces increase in propor-
tion with the distance from the centre of rigidity. An accurate analysis displays
a significant decrease of these forces close to the facades. In order to assess the
order of magnitude of this effect, we shall refer to the structure shown in
Figure4.10a (12 storeys, slabs 60x32xO'20m, columns O'40xO'40m on a
mesh of 5 x 4m and a central core; shallow beams have been assumed in order
to emphasize the effect of the vertical deformability of the slabs). The distribu-
tion of the torsional shear forces along the radii at each floor is displayed in
Figure 4.10b. The maximum forces do not occur along the facades.
This phenomenon is due to the fact that the effective width of the slab strip
acting together with the facade columns is much smaller than the correspond-
ing strip acting together with the central columns; this leads to significant
decreases in the rigidities of the facade columns.
 A critical review of Lin's theory 117

f
013 007 !c'
0.10
0~5 0.15
0.20
0.q7 0.21
, Central core C
/ 0.28
0.27
C 0.35 .4t

0.41 0.55 O.;j; Central

~[
/ core
~
0.46 0.63 0.37

~
0.52 0.69 0.4
1q,"'i: (a)
0.55 0.74 ro:« r--
12x5.0 m
·0.58 iQ.79 roM r--
OM r-- r--
r;;FFr IMTFr
0.61 0.83

0.96
r--
0.63 0.88

1
-0.51 r---
r--
0.62 I---
r---
r---
0.80
(c) 6x5.0 (b)

Figure 4.10

Mazilu (1989) noted that, in the 1977 and 1986 earthquakes in Romania,
relatively small torsional forces acted on the corner columns of structures
subjected to general torsion. This may have been due to the vertical deforma-
bility of the slabs and the subsequent low rigidity of the facade columns.
Second, the vertical deformations of the slabs lead to a decrease in the
rigidity of the columns with respect to the rigidity assessed by Lin's theory; this
means that the ratio of the moment of torsion taken by the columns is smaller
than the ratio assumed by Lin's theory. By referring to the model shown in
Figure 4.10, we obtain according to Lin's theory

MT = M~w (taken by the core) + Mi:, (taken by the columns).

An accurate analysis yields

where !Wi:, = (0.70 ... 0.80) Mi:,. This effect increases along the height of the
structure (Figure 4.1 Oc).
118 Space structures

4.3.2 EFFECT OF SOIL DEFORMABILITY ON THE DISTRIBUTION

OF TORSIONAL FORCES

We have checked the effect of soil deformability on the distribution of lateral

forces in the case of dual structures (see section 3.3). We reached the conclusion
that this effect is important, and to overlook it may lead to significant errors.
As the problem of the distribution of torsional forces is very similar to the
aforementioned problem, we may expect that the effect of neglecting soil defor-
mability will be similar.
In order to check the order of magnitude of this effect, we have considered
two different structures. Referring to the structure shown in Figure 4.11, we
considered elastic supports corresponding to: (a) average-stiff soils, with a
subgrade modulus of 60000 kN m - 3; (b) soft soils, with a subgrade modulus of
20000kNm- 3 . In both situations we have compared the torsional moments
taken by the structural walls assuming fixed supports (M;w.rl and elastic sup-
ports (M;w,el)'
We have obtained:

(a) For average-stiff soils:

~=0'53
(Msw,r)

(b) For soft soils:

~=0'34
(Msw,r)

We see, in the considered structure, that the structural walls take only i-i of
the torsional forces yielded by a 'classical solution' (where we overlook soil
deformability).
In Figure 4.12, the structure has two basement floors, surrounded by RC
structural walls and eight storeys above the ground floor. An RC core
10 x 5 x 0·2 m is placed eccentrically. Columns 0·5 x 0·5 m are positioned on a
grid 5 x 5 m; the slabs are 0·2 m thick. A concentrated force normal to the
longitudinal axis of symmetry acts in the centre of the roof slab.
We considered elastic supports corresponding to: (a) stiff soils
(k s = 100000kNm- 3 ); (b) soft soils (k s =20000kNm- 3 ). By comparing the
torsional moments with those yielded by an analysis performed on the
assumption of fixed ends we obtained:

(a) For stiff soils:

 A critical review of Lin's theory 119

1
8x3.0=24.0

d c

1
8x5.0

8x5.0

Figure 4.11
b L
A
J
(b) For soft soils:

We have to point out that in both examples we also considered given loads
for fixed ends and elastic supports. Actually, the seismic forces are smaller
120 Space structures

---------- -------.-- - -.- --1

r+ 0
t
• • • • • •
• • • • ,t
0
10.0 • • • • f
ci
C')
+- i
• • • •
t
Lr I

L •
+5.0.,1<.-
• • • •
• • - ___• -----.--------.-...-J
5x5.0

t--- 30.0
Figure 4.12

where elastic supports are taken into account; also, the bending moments will
differ due to the change in the positions of the zero moment points.
In conclusion: considering the soil deformability leads to significantly small-
er torsional forces than those yielded by 'classical analysis', based on the
assumption of fixed ends. The decrease can be quantified by referring to the
ratio eel/ e, where e denotes the eccentricity of the resultant forces yielded by
 First screening of existing dual structures 121

1
..;-1 j
!
Figure 4.13

assuming fixed ends and eel the eccentricity yielded by assuming a deformable
soil.
By admitting conservatively ee/ e = 0·6 for stiff soils (ks = 100000 kN m ~ 3)
and eel/e=0'4 for soft soils (ks=20000kNm~3) and considering a linear
variation of the ratio eel/ e between these limits, we obtain
eel ~ 4 0.2 X (ks - 20000)
e - O. + 80000 (4.14)

where ks is expressed in kN m ~ 3.
In the case of foundations on piles we can consider (see Appendix C):
• for a small pile diameter (Dp = 0-40 m), an equivalent spread foundation
on stiff and very stiff soils; by considering ks = 100000- 200000 k N m ~ 3,
we obtain eel/ e = 0,6-0,85 .
• for a large pile diameter (Dp = 1·50 m), an equivalent spread foundation on
soft and regular soils; by considering ks = 20000-60000 kN m ~ 3, we ob-
tain ecI / e = 0-40-0' 50.

4.4 First screening of existing dual structures

subject to torsion
First screening of dual structures implies their classification from the point of
view of their regularity in the horizontal plane (see Chapter 7).
At this end we define the torsional index (Figure 4.13):

TI=S L(Asw d ) (4.15)

LavLA sw
where Asw denotes the area of an RC structural wall, belonging to a pair of
parallel walls (where the walls are not identical we choose the wall with mini-
mum area), d is the distance between the pair of walls, and Lav is the average
horizontal dimension of the slab; in the case of irregular slabs La!' = perimeter/4.
122 Space structures

Where masonry walls are present, the areas Asw are replaced by the areas
Am of the masonry walls, multiplied by the following correction factors:
• reinforced masonry: 0'6;
• infilled frames: 0-4;
• plain bricks or stone masonry: 0·3.
We have checked structures with 8, 10 and 12 storeys with columns 0.50 x 0.50,
slabs 0.20 m thick and various RC structural walls with various deformable
soils, and we have compared the ratios torsional shear forces/translational
shear force. According to the results of these computations, we propose tenta-
tively the following classification.
A regular structure is defined as either a symmetrical or a nearly symmetri-
cal structure in both main directions, or as an asymmetrical one, but with a
torsional index TI > 2 (when the pairs of parallel walls are positioned in one
direction only, Figure 4.14a) or TI> 1 (when pairs of parallel walls are posi-
tioned in both main directions, Figure 4.l4b). Note that in the case shown in
Figure 4.l4b, at least two parallel walls are subjected to torsional forces only,
whereas in the case shown in Figure 4.l4a the existing parallel walls are loaded
simultaneously by translational and torsional forces (Paulay and Priestley,
1992).
A moderately irregular structure is an asymmetrical one with a torsional
index 1 ~ TI ~ 2 (Figure 4.l4a) or 0.5 ~ TI ~ 1 (Figure 4.l4b).
All other structures are defined as significantly irregular.
Taking into account the results of comparative computations, we proposed
an increase of the translational forces in the columns close to the perimeter,
due to the presence of torsional forces (Scarlat, 1993) as follows:
15% for irregular structures;
25% for moderately irregular structures;
40% for significantly irregular structures.
As noted above, most modern structures require a mInImum accidental
eccentricity e = 0.05 L to be taken into account for symmetrical structures

1
.. d1 1 ..,f- d1 ----f
-
+-d2~ I I

j
I I

I I I I -
(a) (b)

Figure 4.14
 Vertical grids composed of structural walls and slabs 123

(L is the length of the building, normal to the given forces). The CEB-85 code
proposes, as an alternative, multiplying the translational forces by the factor
(Figure 4.15a)

1 +0.6x
(4.16)
L

This leads to a maximum increase in the translational forces of 30%.

Accurate analyses performed by taking into account the soil deformability
show that such an increase is equivalent to the presence of a quite strong
eccentric core at a distance of about L/4 from the centre of masses (Fig-
ure 4.15b). This is obviously an exaggeration; we consider that the proposed
increase in the translational forces of maximum 15% is more realistic.

4.5 Vertical grids composed of structural walls and slabs

In order to assess the stresses that develop in the slabs due to horizontal
forces, we have to refer to a vertical grid formed by the structural walls and the
slabs (Figure 4.16a,b). Obviously, we have to take into account the deformabil-
ity of both types of elements: structural walls and slabs.
For an approximate analysis, we can refer to each slab separately, as beams
on elastic supports (Figure 4.16c). The rigidity of the supports (k) is determined
by loading each wall at the considered level by a unit horizontal force (Fig-
ure 4.16b). This procedure neglects the interaction of the structural walls.
When only two structural walls are present, the simplified procedure leads
to statically determined structures; in this case the rigidity of the supports, as
well as the effect of the shear forces, are not relevant (Figure 4.16d); we can
consider the beams as resting on simple supports.
A procedure for assessing the distribution of the horizontal forces among
structural walls by using an approach based on the use of difference equations
has been proposed by Rutenberg and Dickman (1993).

•
•
•
•
•
•
•
·0
•
• •
• •
• •i
-J,.
[Q. • •
·0 •
• •
•
•
.i

,f-L/2 -+ x ,f-Ll4 + L/4-+

t- L -J- .,
~ L 1
(a) (b)

Figure 4.15
124 Space structures

(a)

-,

t t
(e)

Two supports:

t t
(d)

Figure 4.16

Notes: (a) As the depth of the slabs is very large. the depth/span ratios are also high
(usually 2/1 ... 1/1); consequently, the effect of shear forces on deformations is import-
ant. and we have to take it into account. The pattern and the values of the diagrams of
resultant stresses differ from the classical pattern and values (resulting from an analysis
where we consider only the effect of bending moments on the deformations). The effect
of the shear forces increases with the rigidity of the supports. and reaches a maximum
when the supports are rigid.
Figure 4.17 presents diagrams of bending moments that have been drawn on the
basis of two assumptions: by taking into account the effect of shear forces on deforma-
3
tions and by neglecting this effect. Rigid supports and elastic supports (k1 / EI = 20 and
kl 3 / EI = 2) have been checked (/: moment of inertia of the beam). When elastic
 Vertical grids composed of structural walls and slabs 125

P=1 kN m- 1
JlXJ( 1

1=10.0 + 1=10.0 m + 1=10.0 -t

Neglecting ~.O kN m ~O.O
~~~~~
shear
deformations t
4.0 8.0
t
11.0kN
2.5 t
11.0
~
8.0 14 .0
Including 4.8 N 4.8
KO$;?M<:::$;?>_~
shear
deformations t4.5 .l
10.5 110 .5 7.7
110.5 t 4.5

(2) Elastic .;y----T"'""----r---'"""'I

supports .it. 10 " 1.
kI 3/EI=20
7.1 7.1
Neglecting
shear
deformations t:::3 .
.l
19.3 19.6
t9.3 16.3
t5.7
'oo'"d',. ~
shear
t55
deformations t • t5.5
. 14.8 9.5 17.1 19.5 14.8

(3) Elastic
supports
kI 3/EI=2
I 1 I 1
21.7 21.7
N""ed',.
shear ~
deformations 7.2
25.7f7.8 34.2 17.8 25.7

~~~~~". ~
deformations

t
7.0
24.3 8.0
32.2
8.0 24.3 t. 70

Figure 4.17

supports are considered, the negative moments decrease significantly, and usually only
positive moments need to be taken into account. These are much greater than the
moments resulting from an analysis where rigid supports have been assumed and reach,
in the case of equal spans, values of pl21S ... p12/3.
126 Space structures

(b) The resisting elements sometimes present discontinuities in the vertical plane. We
then have to transfer the horizontal forces from one group of elements to another
group through the slabs. Referring to Figure 4.18, above the level 3'00, the resisting
elements are 2a3a-Sa6a and 2b3b-Sb6b. Below this level, the resisting elements are
I a2a -4aSa, I b2b-4bSb and I c2c- SbSc.
Obviously, the storey torsional moment MT and its distribution have to be modified
accordingly. Moreover, we have to ensure that we transmit the 'local moments' M' and
M" to the corresponding resisting elements beneath level 3.00.
As shown in section 6.3, such discontinuities lead to concentrations of seismic stresses
and possible local distress.

o Numerical example 4.3

Let us consider the space structure shown in Figure 4.19a subjected to a uniform
horizontal load p = I kN m - I at each floor. We shall check the diagram M of the slab
at level 6·00 (second floor). An accurate analysis, based on a grid analysis, leads to the
diagram shown in Figure 4.19b.
An approximate analysis of the same slab, based on a continuous beam on elastic
supports, must be preceded by an assessment of the rigidity of the supports. We then
load each structural wall with a unit force at level 6·00 (Figure 4.19c) to obtain the
deflections at the same level: it = I'S3 x 10- 6 m kN - I (for the end supports) and

Figure 4.18
 One-storey industrial buildings 127

ii = 1.02 x 10- 6 m kN - 1 (for the intermediate support), and the corresponding rigidities
(spring constants) k = l/ii = 654000 kN m - I and 980000 kNm - I respectively. The re-
sulting diagrams M are displayed in Figure 4.19d. D

4.6 One-storey industrial buildings: bracing conditions

Consider a fragment of a one-storey industrial building, between a gable wall

and an expansion joint (Figure 4.20a), on rigid supports; the structure is com-
pletel~ braced if no side sways are possible when rigid walls are assumed. In
order to check the condition of complete bracing, we replace the structure by a
space pin-jointed truss (Figure 4.20b); a minimum number of bars are consider-
ed for each wall, sufficient to ensure the bracing. The structure is completely
braced if the pin-jointed truss satisfies the condition of geometrical fixity:
(4.17)

Wallt=0.30 Wallt=0.20

slabs
t=0.20

:; ~/
7,-~.

(a) -,1.--- 25.0 - - . . . : : } /

~~
(b) 78.04 kN m

6.ot ~~
if
1 U=1.53X10 -6 fU = 1.02x10-6 m KW1
j
_
i
1.w I
1U =1.53x10-6

-.r-
Section: 0.20x20.00 0.30x20.00 0.20x20.00
(c)

(kN km_1) 1. 654000 L980000 6540001

~.OO
<:::::!? ~
(d) 20.20 69.60 kN m

Figure 4.19
128 Space structures

A B
(a)

(b)
c,

Figure 4.20

In our case:
Number of bars b = 12
Number of joints j = 8
Number of restraints due to supports s = 4 x 3 = 12
12+12=3x8
 Framed tube structures 129

I~
/l I~~~~
.IVVV.VV
IVv

r
I.lVVI'V
~/' ~

;'
1/

I' ;I
~~V ~
I' (W) ~

~~ /~~
lI'f
/ IJI ;l/I'V/~
H ~ (t) 17 ~VI'
I'I'~I'~
I'.
~
r--~~ I~~~~~?
~V~~,
~ ,~
1~~fiT
~~
'I . B

Figure 4.21

As a result, only axial forces have to be taken into account. When a significant
settlement of the supports occurs, the bracing effect vanishes and bending
moments develop along the edges (Figure 4.24c).

4.7 Framed tube structures: approximate assessment

of stresses

4.7.1 INTRODUCTION

A new type of structure intended for very high-rise buildings (in excess of
30-40 stories) was developed in the 1970s, namely framed tube structures.
In this structural system, the horizontal forces are taken by a dense perimet-
ral grid of columns and spandrels (Figure 4.21). The interior elements are
designed to take only vertical forces. Columns with depths of 1-3 m are placed
at 4-5 m distance. Spandrels 0.25-0.75 m thick and 1-1'5 m deep are placed at
each floor.
If we replace the exterior grid by a continuous wall, a shell is obtained with
a thickness-to-Iength ratio of 510-160 (Taranath, 1988). Consequently, the
structure behaves like a thin walled tube; a typical stress distribution is dis-
played in Figure 4.22, and is compared with a linear distribution yielded by the
classical analysis, based on the assumptions of strength of materials.
A significant stress concentration due to the 'shear lag effect' occurs around
the corners.
130 Space structures

In order to take into account the presence of the rigid slabs without adding
supplementary unknowns, we shall transfer the horizontal distributed loads to
the side planes (w) (Figure 4.23).
An accurate analysis of framed tube structures, by finite elements, usually
involves 5000-50000 unknowns. A significant reduction in the number of
unknowns can be achieved by considering a space system of bars; a high
degree of accuracy is obtained by using finite joints (see section 2.3.2).

4.7.2 APPROXIMATE ANALYSIS BY REDUCING THE SPACE STRUCTURE TO A

PLANAR STRUCTURE

The method was developed by Khan (1966, 1971) and completed by Coull and
Subedi (1971), by Rutenberg (1972, 1973, 1974), by Mazzeo and De Fries
(1972) and by Khan and Amin (1973). It is based on the fact that the displace-
ments normal to the flange planes and to the web planes are negligible (Fig-
ure 4.24a). By also neglecting the torsion stresses, the deformations of the
component plane substructures (f and w) are planar; subsequently, we may
consider the planar structure displayed in Figure 4.24b. The connections

Strength of
material~IIiJ'''''''''''''''''u.u~",

Thin-walled
structure

Figure 4.22

Figure 4.23
132 Space structures

between the substructures f and ware meant to transfer vertical shear forces
only (Figure 4.24c). These forces increase with the vertical rigidity of the flange
substructure f: a rigid flange substructure prevents the free vertical displace-
ments of the corner joints of the web substructure. The method was extensively
used in the 1970s. In order to facilitate the computations, influence lines were
drawn by Khan and Amin (1973) for basic 10-storey buildings, with possible
use for buildings with a different number of storeys.
Several noteworthy procedures were proposed, intended to model space
(tri-dimensional) problems by planar systems, based on various analogies
(Rutenberg and Eisenberger 1986).

4.7.3 APPROXIMATE ANALYSIS BASED ON EQUIVALENT FLANGES

This method is very simple, and gives the order of magnitude of the axial
forces in the corner columns and of the shear stresses in the spandrels of the
web substructure (Figure 4.25). In this analysis, we compute the geometrical
data of the section by admitting the classical assumptions of strength of ma-
terials, but considering a limited number of columns, in agreement with the
equivalent flange Leq < L.
According to Khan's proposal, we may choose Leq as the smallest of the
two lengths: Hjl0 or Bj2. Furthermore, we compute the moment of inertia Ieq
and the statical moment Seq of the active zone with respect to the axis x« and
deduct the axial forces in the columns (area AJ:

(4.18)

Leq/2

++ Active zone II

(a) (b)

Figure 4.25
 Framed tube structures 133

and the vertical shear forces V:

(4.19)

where h denotes the storey height.

The proposed method gives satisfactory results for the maximum axial forces
in the corner columns and for the maximum vertical shear forces in the bottom
beams of the web substructure on the neutral axis (x-x); however, because of the
shear lag effect, higher vertical shear forces may occur close to the corners.

o Numerical example 4.4

The framed tube shown in Figure 4.26a is subjected to a lateral uniformly distributed
load. Taking into account the in-plane rigidity of the slabs, we shall consider the lateral
loads as linear loads distributed along the edges: let be p = 4 x 3 kN m - 1. The total shear
force results in Vmax = 1080 kN, and the total overturning moment M TOT = 1080 x
90/2 = 48 600 kN m. We compute the statical data for half of the given structure (a
channel-type section). Vmax = 1080/2 = 540 kN; Mmax = 48600/2 = 24300 kN m.
The equivalent width of the flange (Figure 4.26b):
H 90 B 24
-=-=9 m· -=-= 12m· L eq =9m
10 10 ' 2 2 '

on each side, Leq/2 = 4.5 m includes 2.5 columns. Let us denote the area of a column by
Ac. The moment of inertia results in
Icq = 2 x 2.5 x Ac x 122 +4Ac(42 + 82) = 1040A c

with the corresponding modulus of resistance

1040A c
Weq =--=86.7
12 Ac

The static moment with respect to the axis x-x:

Seq=2.5 Ac x 12+A c(4+8)+0.5A c x 12=48A c
The maximum axial force (at corners):
M Ac 24300
N = ~ = - - = 280 kN (accurate analysis: N = 304 kN).
Weq 86.7

The maximum vertical shear force (in the spandrels of the web substructure):

I 1040
z =~=-- = 21.66m· h=3m
Seq 48 '
540 x 3
V. m =--=75kN
L " 21.66
 I
><i
u-J _ . _ .

o
o
CJ)
II
o~
Ml:
x~

r
oC')

I
I
..... I
-- ------ --------- ------~ \
\
\
\

~,\ ","
 Axi-symmetric system of bars 135

Accurate analysis (tri-dimensional finite element analysis):

VLm " = 46 kN (in the neutral axis) ... 75 kN (close to the corner)

The distribution of the vertical reactions on half of the wall yielded by the accurate
analysis is displayed in Figure 4.26c; a significant shear lag effect is visible. D

4.8 Axi-symmetric system of bars: evaluation

of maximum stresses and deflections
4.8.1 INTRODUCTION

Consider a water tank (Figure 4.27) supported on a space, axi-symmetric frame

(a regular polygon with n sides), subject to horizontal forces (either seismic
forces or wind pressure). The resultant force FT acts at height H T; it may be
replaced by the same resultant force F T acting above the columns plus a
resultant moment M k = F T H k/2. As the bending moments M, the shear forces
V and the horizontal deflections u due to the resultant moment M k are usually
negligible we may, when computing M, V and u, relate only to the effect of the
resultant force FT. The columns of the space frame are usually placed on a

r- FT
>

II
Hk

t- t
17

HF +
17

J-
17
I
I
i
+ I

17
I I
?I'- --.r- +.-
.,0'

Figure 4.27
136 Space structures

hexagon, an octagon or a dodecagon. They may be radial (Figure 4.28a),

square (Figure 4.28b) or tangential (Figure 4.28c).
In the following, we shall give approximate values of the fixed end bending
moments, at the base of the columns (radial moments M r and tangential
moments M\ Figure 4.28d), maximum deflections and axial forces due to the
resultant force F T"
We note that in most cases the fixed end moments also represent the
maximum moments
We assume that:
• the polygon beams are positioned at equal heights H = m h;
• the polygons maintain their initial form in the deflected shape (it has been
demonstrated that this assumption is theoretically valid for the considered
type of loading: Mutafolo, 1959);
• the columns are fixed at their base and at their top ends, in the tank (the
structure of the tank is considered as a rigid body).

4.8.2 FIXED END MOMENTS

The fixed end moments depend on the given resultant force F T' the type of
polygon (number of sides n), the direction of the columns (radial, square or
tangential), and the relative rigidities of the columns and beams.

a-a
~ 3 <- vertical

-f.t>t,+

(a) (b) (c)

~
adiUS ~ radius
M' ¥/ Mt

- taog.'" - ~ taoge'"

(d)

Figure 4.28
 Axi-symmetric system of bars 137

We have computed the fixed end moments Mr for three types of structure:
hexagonal, octagonal and dodecagonal. For each type, we considered (Fig-
ure 4.28): radial columns (he = 5b e), square columns (he = bJ and tangential col-
umns (he = 0·2 bJ. Beams of various rigidities were taken into account (,flexible
beams', 0·20 x 0·20 m; 'average beams', 0·32 x 0·32 m; and 'stiff beams',
0·50 x 0·50 m). Space frames ranging from three to five storeys were considered.
The results are given in the form of graphs: M~(n/FTh) and M}(n/FTh) as
functions of the relative rigidities kb/ ke' where

(see Figure 4.28).

The graphs are shown in Figure 4.29.

After computing the maximum moments M~ = M~t and M} = ML we can
obtain the corresponding moments in the other columns i according to the
expressions (Figure 4.30)

M~, = ML cos(2ni/n); M}, = ML sin (2 n i/n) (4.20)

4.8.3 MAXIMUM AXIAL FORCES

The axial forces are due to the effect of vertical loads and the effect of horizon-
tal forces. In the following we shall relate only to these latter forces. The
maximum axial forces are equal to the vertical reactions Rv: N max = Rv'
We shall assume that the reactions vary linearly, in proportion to the
distance di to the axis of symmetry S, normal to the resultant force FT'
Equilibrium by moments about axis S (Figure 4.31) yields

where Mext = FT H F ; M R , is the moment given by the vertical reactions Rv; MM

is the moment given by the fixed end moments acting on the columns, projec-
ted on the axis S; d i = r sin 'Xi = r sin 2n i/n; and M R , = L Rv di .
Examination of numerical examples indicate that we may accept
MM~0'1 M ex !

Accordingly:
R rn .
PHF=~+O'lFTHF (4.21 )
2
1·8 FT H
N max =R ~ F
Vm;l\ - nr

In the remainder columns:

Ni = N max cos 2n i/n (4.22)

138 Space structures

2 3 4
TC-~

(a) Hexagon

2
/RC-Mf
/ TC-Mi

\,,~ sc-~i ----

'-----r=~ML------
' - - - __ -*C RC-
M;
--..:..:: ---------------
",",TC-Mf
I I I ~
2 3 4 5 kb1kc

(b) Octagon

Figure 4.29
 Axi-symmetric system of bars 139

t ----

~-- - - - - - - -
" '---. ~SC-~~_M~ -~-==--.:::.-..:::-..:::-..:::-=-=-,,-,-=-=-=-=-=-=---
7-_
-------~Rc=_Mt --------
"eTC M[ f

2 3 4 5 6

(c) Dodecagon
RC- Radial columns, SC- Square columns, TC- Tangential columns,
M' - Radial moments, Mt_ Tangential moments
Figure 4.29 Contd.

4.8.4 MAXIMUM HORIZONTAL DEFLECTIONS

We shall first compute the maximum horizontal deflections by assuming that

the beams are rigid (u;;'aJ We shall then correct the deflections u;;'ax by taking
into account the deformability of the beams.
Referring to Figure 4.32 and considering only one storey and assuming rigid
beams (sliding fixed ends), we find:
n
FT = I (V~ cos ai +
1
v: sin aJ

(4.23)
12 E u;;'ax( It nl2 + Ir n12)
h3
12 E Is u:"ax
h3

where

I =bch~.
r 12'
140 Space structures

Figure 4.30

(5)
n-1

F.
~1

L d
1-~1---t

Figure 4.31

By considering m storeys (H = m h):

(4.24)

We take into account the deformability of the beams by multiplying the deflec~
tion u';'ax by the factor II:
 Axi-symmetric system of bars 141

-t
h

-+

,g Uj

\/,
v;r

o
12EIr·uma
h3 ~~-t
j' .
,XSinY

12EIt·uo
h3 max'
COs2 u
, __
~
-,Ie-
Figure 4.32

Graphs of the factor J1 are shown in Figure 4.33, for the same cases that we
considered previously.
Evaluation of maximum horizontal deflection also enables us to assess the
fundamental period of free vibrations (T). Let

Ii Wm h3
uw = (4.25)
12£ Is
142 Space structures

SC

2 3 4 5 kb/ke
(a) Hexagon

~,

SC

2 3 4 5 ""Ike
(b) Octagon

Figure 4.33
 Axi-symmetric system of bars 143

2 3 5

(c) Dodecagon
RC- Radial columns; SC- Square columns; TC- Tangential columns

Figure 4.33 Contd.

be the maximum horizontal deflection due to a horizontal force equal to the

total weight of the tank (W), including its content. According to Geiger's
formula (see Appendix A), T = 2 Fw,
where Uw is given in m.

o Numerical example 4.5

Consider the octagonal water tower shown in Figure 4.27 with the following data:
radial columns 0-20 x 0'50 m, beams 0-32 x 0-32 m (average beams), radius r = 4-0 m,
H = 4h = 4 x 4-0 = 16 m, H F = 18 m; FT = 800 kN; E = 3 X 10 7 kN m - 2_

I 0-32 4 2'85
kb=...!'.= =-m 3
I 12 x 3·06 104

Ie 0-20 X 0'50 3 5-21 3

k =-= =-m
e h 12 x 4·0 104

From Figure4_29b, we read M;n/(FTh);;;; 1·35 and M;n/(F T h);;;;0'37; n/(FTh)=

144 Space structures

8/(800 x 4) = 2'5/ \0 3 . This gives M; = 540 kN m and M~ = 148 kN m (accurate results.

yielded by a space frame analysis: M; = 530 kN m and M; = 150 kn m).
From Figure 4.33b we read fl ~ 5·3. This gives

n(lT + It) 9·67 4

I = =···=-m
s 2 10 3

F T mh 3 800 x 4 X 4.0 3
U =---= =0·0589 m
max (12 E I,l (12 x 3 x 10 7 X 9'67/10 3 )

U
max =5'3xO'0589=0'3Im (accurate result: urn"X = 0·28 m)

The maximum axial force due to horizontal forces is

1·8 x FTH 1·8 x 800 x 18·0

N ::= =8\OkN
max nr 8 X 4·0
(accurate resultN max = 632kN).

Fundamental period:

10000 x 0·31
W = 10000 kN; Uw = - - - - - = 3-875 m
800

T = 2 )3'875 = 3·94 s (accurate computation: T = 3'785 s)

In the case of tangential columns (0'50 x 0·20 m) we obtain

M; = 80 kN m (accurate: 75 kN m)

M; = 420 kN m (accurate: 454 kN m)

Urn"X = II urn"x = 2·2 x 0·0589 = O' \3 m (accurate: 0·16 m)

Nrnax = 8\0 kN (accurate: 675 kN)

T = 2·55 s (accurate: 2·82 s)

We note that the tangential orientation of the columns is preferable. in terms of

stiffness. to the radial orientation of columns; it also provides a higher torsional
stiffness. o
 Bibliography 145

Bibliography
Coull, A. and Subedi, N. (1971) Framed-tube structures for high-rise buildings. journal
of"Structural Division ASCE, 97,2097-2105.
Jiang, W., Hutchinson, G. and Chandler, A. (1993) Definitions of static eccentricity for
design of asymmetric shear buildings, Engineering Structures, 15 (3), 167-178.
Khan, F. (1966) Current trends in concrete high rise buildings, in Proceedings S ympo-
sium on Tall Buildings, University of Southampton, April, pp. 571-590.
Khan, F. (1971) Tendances actuelles dans la construction des immeubles de grande
hauteur a structure en beton arme et en acier. Annales Institut technique du Mtiment
et des trauvax puhlics, Supp1. au No. 281 (Mai), pp. 37-53.
Khan, F. and Amin, N. (1973) Analysis and design of framed tube structures for tall
concrete buildings. journal of the American Concrete Institute, 53 (SP36), 85-90.
Khan, F. and Stafford Smith, B. (1975) Restraining action of bracing in thin-walled
open section beams. Proceedings of the Institution of" Civil Engineers. Part 2. 59
(March),67-78.
Lin, T. Y. (1951) Lateral forces distribution in a concrete building story. journal of the
American Concrete Institute, 23 (4), 281-294.
Mazilu, P. (1989) Behaviour of buildings during the 1977 and 1986 earthquakes in
Romania, lecture at the Faculty of Civil Engineering, Technion, Haifa, November.
Mazzeo, A. and De Fries, A. (1972) Perimetral tube for 37-story steel building. journal
of" the Structure Division ASCE, 98 (ST6), 1255-1273.
Mutafolo, M. (1959) Contribution to the analysis of multi-storey, axi-symmetric towers.
Thesis for PhD, Institute of Civil Engineering, Bucarest (in Romanian).
Paulay, T. and Priestley, M. (1992) Seismic Design of RC and Masonry Buildings, J.
Wiley, New York.
Rutenberg, A. (1972) Discussion of paper by A. Coull, N. Subedi (Framed-tube struc-
tures for high-rise buildings), journal of the Structural Division Proceedings of the
ASCE, 98 (ST4) 942-943.
Rutenberg, A. (1972) Discussion of paper by A. Mazzeo, A. De Fries (Perimetral tube
for 37-story steel building), journal of the Structural Division Proceedings of the
ASCE, 99 (ST3) 586-588.
Rutenberg, A. (1974) Analysis of tube structures using plane programs, in Proceedings of
the Regional Conference on Tall Buildings, Bangkok. pp. 397-413.
Rutenberg, A. (1980) Laterally loaded flexible diaphragm buildings. journal of" the
Structural Division Proceedings of" the ASC E, 106 (ST9), 1969-1973.
Rutenberg, A. and Eisenberger, M. (1986) Simple planar modeling of asymmetric shear
buildings for lateral forces. Computers and Structures, 24 (6), 885-89\.
Rutenberg, A. and Dickman, Y. (1993) Lateral load response of setback shear wall
buildings. Engineering Structures, 15 (1),47-54.
Scariat, A. (1986) Approximate analysis of multistorey buildings, in 1er Colloque Na-
tional de Genie Parasismique, St Remy, pp. 6.1-6.10.
Scariat, A. (1993) Soil deformability effect on rigidity-related aspects of multistorey
buildings analysis. Structure journal of the American Concrete Institute, 90 (2),
156-162.
Taranath, B. (1988) Structural Analysis and Design of Tall Buildings, McGraw-Hill, New
York.
5 Pile foundations and
retaining walls

5.1 The 'equivalent pile length' concept

A quick approximate assessment of the maximum moment and the maximum

deflection of a pile subjected to a horizontal force can be performed by ap-
plying the equivalent pile length concept. In the following, we shall use the
results obtained by Kocsis (1968) in the form adopted by the Ministry of
Works and Development of New Zealand (1981) (cited by Dowrick, 1987). In
order to characterize the mechanical properties of soils we may classify them
as cohesive (e.g. clays) or cohesionless (e.g. sands).
For cohesive soils, we defined the horizontal soil deformability by using the
modulus of subgrade reaction k" i.e. the normal stress corresponding to a unit
horizontal displacement (see Appendix C). We may assume a constant value of
ks along the height of the pile (Figure 5.1). ks is usually in the range
5000- 70000 kN m - 3. We define the stiffness radius

R=(E kJ )1!4
C
p p
(m) (5.1)

where

(5.2)

Figure 5.1
 The 'equivalent pile length' concept 147

Dp is the diameter of the pile (m), Ip is the moment of inertia of the pile (m4),
and Ep is the modulus of elasticity of the pile (kN m - 2).
For cohesion less soils we assume that soil stiffness increases linearly with
depth z (Figure 5.2); we characterize it by using the unit subgrade reaction ns ,
defined as

(5.3)

ns is usually in the range 1000- 20000 kN m - 3.

The stiffness radius is

Rn=(E~:p)1/5 (m) (5.4)

In the following, the piles are assumed flexible, i.e.

(5.5)

where L denotes the length of the pile below the ground surface. We note that
the modulus of subgrade for groups of piles may significantly decrease with the
pile spacing (Davisson and Salley, 1970; Poulos, 1979).
We define the "equivalent pile lengths" in the following by referring to two
situations, free-headed piles and fixed-headed piles.
For free-headed piles (Figure 5.3), the equivalent length Lm for the evalu-
ation of the maximum moment is given by

(5.6)

The equivalent length Ld for the computations of the maximum deflections is

. . . . - - - -.... k

z
Figure 5.2
148 Pile foundations and retaining walls

F, a
a

t L Mmax=F{a+L m}

Figure 5.3

given by

(5.7)

Lrn and Ld are determined as follows.

For cohesive soils:

(5.8)

For cohesionless soils:

(5.9)

For fixed-headed piles (Figure 5.4), the equivalent length Lrn is given by

F Lrn
Mrnax =-2- (5.10)

The equivalent length Ld is given by

F L~
(5.11)
Urnax = 12 E I
p p
 The 'equivalent pile length' concept 149

.,f- Umax +-
JJ.L _~_
- ---r-
I
/
I
I
/
I
I
I
Mmax

Figure 5.4

Lm and Ld are determined as follows.

For cohesive soils:
Lm=I'5Rc;' L d =2'2R c
For cohesion less soils: (5.12)

The radii Rc and Rn were defined by equations (5.1) and (5.4).

o Numerical example 5.1

A free-headed pile with a diameter Dp = 0·80 m and a length L = 12· 5 m is acted upon
by a horizontal force F = \0 kN (Figure 5.5).

Ep lp = 3 X 10 7 x 0·0201 = 603000 kN m 2 .

(a) Cohesil'l' soil

603000) I /4
k = \0 000 k N m - 3. k = k D = 8000 k N m - 2 R.= -- = 2·94 m
, " p ' 8 (0 0 0 '

L 12
-=-=4·25>4;a=0·L =0·5x2·94=1·47m·M =IOx 1·47= 14·70kNm.
Rc 2· 94 ' m ' max

(An 'accurate analysis', performed by considering discrete horizontal springs, yields

Mmax= 13kNm.)

Ld = 1·6 x 2·94 =4·70m

IOx4'70 3
II = =0·00057m=0·57mm(accurate:0·60mm).
max 3 X 603000
150 Pile foundations and retaining walls
F=10 kN
~

1
L=12.S m

Figure 5.5

(h) Cohesion less soil

n, = 2000 k N m ~ 3

Rn = (603000/2000)1/5 = 3·13 m

L
- = 12· 5/3' 13 = 3·99
Rn

As LIR is very close to 4 we shall relate to the pile as to a flexible one.

M max = 10 X 2· 50 = 25 kN m (accurate: 24·1 kN m).

a
-=0<1' Ld = 2·2 x 3·\3 = 6·89m
Rn '
IOx6'89 3
U = =0·0018m=I·8mm(accurate:I·6mm). 0
max 3 X 603 000

o Numerical example 5.2

A fixed-headed pile with a diameter of 0-40 m and a length L = 10 m is acted upon by a

horizontal force F = 10 kN (Figure 5.6).
 The 'equivalent pile length' concept 151

F=10 kN

Dp=0.4 m
1 L =10 m

J
Figure 5.6

(a) Cohesive soil

k,=10000kNm- 3 ; k=k,D p =4000kNm- 2 ; E p l p =37700kNm 2

37700)1 /4 L 10
R = (-- = 1·75m -=-=5·71 >4
c 4000 Rc 1·75

Lm= 1·5 x 1·75=2·63m

10 x 2·63
Mmax. - - - = 13·15 kN m (accurate: 12·10 kN m)
2
L d =2·2 x 1·75=3·85m
lOx 3.85 3
U
max
----=O·OOl3m= 1·3mm (accurate: I mm).
12 x 37700
( h) Cohesionless soil

n, = 2000 k N m - 3

37700)1 /5
R = - - =1·80m
n ( 2000

L m =2·0R=2·0 x 1·80=3·60m
10 x 3·60
Mmax - - - = 18kNm (accurate: 20kNm)
2
L d =2·5R=4·50m
10 x 4.50 3
1/
max
----=0·002m=2mm
12 x 37700
(accurate: 2·5mm) o
152 Pile foundations and retaining walls

I 2
n

(a) 2 (b)

Figure 5.7

5.2 Distribution of lateral forces in a pile group

We often have to deal with the problem of distributing a given horizontal force
among piles with different diameters and different lengths.
In the case of a group of columns with various sections and lengths with
fixed base and top sliding fixed end. as in the case of rigid connecting beams
(Figure 5.7a). subjected to a given horizontal force FT. the solution is simple:
the force FT will be divided among the columns in proportion to their rigid-
ities:
IlL
F=F T -'--' (5.13)
, Ll;/ L j
The same formula governs the distribution of the force F T among columns
when they are pinned at their lower ends (Figure 5.7b). In the case of elastic
connecting beams equation (5.13) gives an approximate solution.
However. in the case of piles. the presence of soil resistance to horizontal
displacements completely changes the picture (Figure 5.8): the distribution of
the force F among piles according to their rigidities would be very far from the
reality. Hence. in order to determine an acceptable distribution of the given
horizontal force F T among the piles. we shall refer to two cases. as follows.
The first case is a pile group with constant length in an identical soil. but
with different diameters. For this purpose we consider a group of five piles.
10 m long. and diameters varying from 0·40 to 1· 20 m (Figure 5.8). The piles
are connected by a horizontal beam. The following alternatives are considered:
• a soft soil (k, = 20000 kN m - 3) and a stiff soil (k, = 100000 kN m· 3);
• a very stiff connecting beam (sliding fixed top ends) and a very flexible
connecting beam (pinned top ends);
 Distribution of lateral forces in a pile group 153

F,-

r
~

10.0 m

Dp=O.4 m 0.6 0.8

Figure 5.8
1.0 1.2
1
• soil resistance to horizontal displacements at top is neglected for the first
1 m and for the first 2 m.

The results are displayed graphically in Figure 5.9, as average curves Fj F T'
where Fi is the horizontal force taken by the pile i, versus the diameters' ratio
Dpj Dp , for soft and stiff soils.
'We 'c'im assume the average curve to be a straight line (see Figure 5.10);

Fi~~ __ (5.14)
F - 2D 10
T Pm",

The second case is a group of micropiles with identical diameters but of

different lengths.
We refer to a rather usual case, when a backfill (compacted or not) is placed
on a sloping hard soil (often a rock-type soil). The micropiles will have differ-
ent lengths, according to the slope of the hard soil.
In order to determine the distribution of the horizontal force F T among the
micropiles we consider (Figure 5.11) a group of five micro piles with a diameter
of O' 35 m, placed on a backfill of 2· 5 ~4' 5 m and penetrating into the hard soil
at depths of 1,5-3,5 m.
We consider the following alternatives:

• a very soft backfill (k s = 5000 k N m - 3) on a very hard soil (k s =

200000 kN m - 3); an average backfill (k s = 30000 kN m - 3) on a hard soil
(k s = 100000 kN m - 3);
• a very stiff connecting beam (sliding fixed top ends) and a very flexible
connecting beam (pinned top ends);
• the soil resistance to horizontal displacements at top is neglected for the
first 1 m and for the first 2 m.
154 Pile foundations and retaining walls

0.5

0.4

0.3

0.2

0.1

--1-------,r-----r------r------r--~Dp/Dpm'"'
1/3 1/2 2/3 5/6

FJF-r
0.5

0.4

0.3 Average beam

0.2

0.1

-+-------.--------,.------r------..---~Dp/Dpm'"
1/3 1/2 2/3 5/6

Figure 5.9

The results of the computations show that:

• In the case of average backfill (usually obtained by compaction) on hard

soil the forces are distributed nearly equally among the micropiles. As a
first approximation we can consider for any micropile a force equal to
Fi = 1·2 FT/n (n = the number of micropiles).
 Distribution of lateral forces in a pile group 155

F;lpt
0.5

0.4

Admitted straight line

0.3

0.2

0.1

~ D p,
Dpmax
1/3 1/2 2/3 5/6

Figure 5.10

Connecting
beam

t
2.5m

-t 4.5m
1.5 m
.J<..-

-+
Stiff base soil 3.5m

Figure 5.11

• In the case of very soft backfill on a very hard soil the shortest micropiles
take (1'6-1,9) Fr/n, while the longest micro piles take (0,6-0'9) Fr/n. As
a first approximation we can consider for the short micro piles a force
equal to 2 Ft/ n, whereas for the long micropiles we can consider a force
equal to Fr/n.
156 Pile foundations and retaining walls

5.3 Retaining walls: earth pressure during earthquakes

The computation of earth pressure on retaining walls during earthquakes is

usually based on the Mononobe-Okabe formulae, derived in the frame of
Coulomb's theory (Mononobe, 1926; Okabe, 1926).
The resultant of the earth pressure PT is resolved into two components:
PT = Pst + PDyn (5.15)
where Pst denotes the static component and PDyn the dynamic supplement due
to inertia forces developed during the earthquake. We define the intensity of
the ground acceleration by using the coefficients
Uv
Zv=- (5.16)
g
where U H and U v denote the peak horizontal and vertical accelerations charac-
teristic for the given area; g is the acceleration of gravity.
When no detailed data are available we may let
Zy 2
Z~3 (5.17)

We define 0 by
Z
tan()=-- (5.18)
l-Zy
According to Mononobe-Okabe, the active resultant force is
Ka }' H2(1- Zy)
P - ~.!-'. ' - - - . : . - - . : . : . (5.19)
T- 2
where }' is the unit weight of the soil and
cos 2 (ifJ - [3 - 0)
KaT = ----------=--'--i==========:=:::-:: (5.20)
COS ()
2

(for notations see Figure 5.12).

l (
cos [3 cos( i5 + [3 + &) 1 +
sin(ifJ + (5) sin(ifJ - i -
cos(i5 + [3 + O)cos({J - i
0))]2

By setting 0 = 0 (Z = 0, Zy = 0), we obtain

where
cos 2 (ifJ - (3)
K a . st = ------------.:,~~~=====~ (5.21)
i))]2
l (
cos 2 [3 cos (i5 + (3) 1 +
sin(ifJ + ())sin(ifJ -
cos(() + (3)cos([3 - i
 Retaining walls: earth pressure during earthquakes 157

r H

1 Figure 5.12

The dynamic supplement results in

(5.22)

The static component Pst acts at a height H /3. Experiments performed on

gravity (rigid) walls suggest that the dynamic component acts at a height 0·6 H
(Seed and Whitman, 1970). The resultant will act at the height

H =Pst(H/3)+PDyn(0·6H)
(5.23)
P PT
As no similar experiments are available for flexible retaining walls (Dowrick,
1987), we also have to use the equation (5.23) for such type of walls.
A height Hp ~ O' 5 H has been suggested for the resultant force PT (e.g.
AASHTO, 1983).
We note that the increase in the height of the resultant force PT leads to a
significant increase in the overturning moment.
In order to assess the order of magnitude of the increase in the earth
pressure due to earthquake, we refer to a simple case where Ii = i = (j = 0
(Figure 5.13) and we let Zv / Z ~ ~, resulting in

cos 2 (ip - 0)
(5.24)
158 Pile foundations and retaining walls

~
Hp

J
Figure 5.13

1.8

1.7

1.6

1.5
1.4

1.3

1.2
1.1

0.05 0.1 0.2 0.3 z

Figure 5.14

and
K a . st ~' H 2
Pst = 2

cos 2<p
K = =tan 2 (4Y-O·5<p) (5.25)
a" (I + sin <p)2

The graphs shown in Figures 5.14 and 5.15 display the curves PT / Pst and
Hp/ H versus Z.
These curves can be also applied to approximate PT / Pst for the more gen-
eral cases (when f3 of. 0, i of. 0, b of. 0).
More recent research suggest that a diminished value of the relative acceler-
ation (2' < Z) should be taken into account when using the Mononobe--Okabe
 Bibliography 159

Hp/H
0.6
--
r:: --
0.5
--:::;:L,-':;:~
_- -::::- ........... =25 0

0.4
_~-
__ ~~""""~ ----
oljl=30°
Ijl

0.3
-- -- -- -- -- -- --Ijl~S-- -- -- - -
1jl=40°
~----~--~----------T----------+----~
0.05 0.1 0.2 0.3
__ Z
Figure 5.15

formula. The reduced relative acceleration depends on the allowable maximum

displacement of the retaining wall; maximum values of Z' = (0·4 -0'5) Z have
been considered. The main sources for this approach are Richards and Elms
(1979), Elms and Richards (1990), the AASHTO (1983) code, the proposed new
French code (1990), and Frydman (1992).

Bibliography

Amir, Y. (1989) Foundation Desion in Seismic Areas, Lishkat Hamehandesim, Tel Aviv
(in Hebrew).
Davisson, M. and Salley, J. (1970) Model study of laterally loaded piles. journal of the
Soil Mechanics and Foundations Division of ASCE, 96 (SM5), 1605-- 1628.
Dowrick, D. (1987) Earthquake Resistant Desion, J. Wiley & Sons, Chichester.
Elms, D. and Richards, R. (1990) Seismic design of retaining walls, in Desion and
Perfimnance of Earth Retaining Structures, Proceedinos of Conference at Cornell
University, pp. 854-869.
Frydman, S. (1992) Design of retaining walls in seismic areas, in Proceedinys of the Fifth
Conference of the Israel Association of Earthquake Enoineering, Tel Aviv, pp.
97 102 (in Hebrew).
Kocsis, P. (1968) Lateral Loads on Piles, Bureau of Engineering, Chicago.
Matthewson, M., Wood, J. and Berril, J. (1980) Earth retaining structures. Bulletin of
the New Zealand Societyfiil' Earthquake Eni/ineeriniJ, 13 (3),280-293.
Mononobe, N. (1926) Earthquake-proof construction of masonry dams, in Proceedinys
of World Enoineeriny Conference 9.
Okabe, S. (1926) General theory of earth pressures. journal of the japanese Society of
Civil Enoineeres, 12 (I).
Poulos, H. (1979) Group factors for pile-deflection estimation. journal of the Geotechni-
cal Dil'ision of ASCE, 105 (GTI2), 1489-1509.
Richards, S. and Elms, D. (1979) Seismic behavior of gravity walls. journal of the
Geot.:cilnical Division of ASC E, 105 (GT4), 449-464.
Seed, H. and Whitman, R. (1970) Design of earth retaining structures for dynamic loads,
in ProceediniJs of' ASCE Specialty Conference on Lateral Stresses in the Ground and
the Desiyn of Earthqlwke Retainino Structures, New York, pp. 103-147.
6 Earthquake design:
basic concepts,
approximate methods

6.1 Evaluation of seismic forces: regular structures

6.1.1 GENERAL DATA. PERFORMANCE CRITERIA. SEISMIC COEFFICIENTS

The performance criteria in seismic design accepted by most of the present

codes follow the guidelines established by the Californian Code SEAOC-88.
According to these guidelines buildings should:

• resist earthquakes of minor intensity without damage (it is expected that

the structure remain in the elastic range);
• resist moderate earthquakes with minor structural and some non-struc-
tural damage;
• resist major catastrophic earthquakes without collapse.

For exceptionally important buildings (such as hospitals, fire stations, and

power plants) or exceptionally hazardous buildings (such as nuclear power
plants), the criteria should be more stringent.
Modern codes evaluate the total horizontal seismic force F T (equal to the
base shear V) acting on a building by taking into account the following factors:

1. the seismic intensity factor, i.e. the probable intensity of seismic motions,
depending on the seismic area in which the building is located (the seis-
mic areas are defined by seismic maps included in the codes);
2. the site factor, depending on the nature of soil layers overlying the bed-
rock;
3. the rigidity factor, depending on the rigidity of the structural elements
and the mass of the building;
4. the reduction factor, depending mainly on the ductility of the materials
and on the detailing of the structural elements;
5. the importance factor of the given building;
6. the total weight W of the building.

In the case of regular structures (i.e. structures that are not very slender and do
not display significant irregularities) the seismic forces may be determined
according to the static lateral force procedure, as detailed in the following.
 Evaluation of seismic forces: regular structures 161

The total seismic force takes the form

(6.1 )

where c = F T/W denotes the seismic coefficient. The seismic coefficient is ob-
tained by multiplying several factors:

c = (Seismic intensity factor) x (Site factor) x (Rigidity factor)

x (1/Reduction factor) x (Importance factor)

The total seismic force F T is then distributed to each storey by multiplying it

by specified distribution coefficients d;. The force F; acting at storey i results in

(6.2)

Structures may exhibit 'irregularities', which affect the distribution and magni-
tude of seismic forces acting on the various resisting elements. These can be
classified in two categories:

• irregularities in the horizontal plane of the structure, mainly asymmetry,

leading to significant eccentricities and subsequently to torsional forces;
• irregularities in the vertical plane, mainly discontinuities in the rigidities
of structural elements and in the masses.

When the structure is very slender or presents significant irregularities, we

shall refer to it as a special structure and shall determine the seismic forces
based on a dynamic analysis, usually a modal analysis (see section 6.2).
It is important to point out the different approaches in seismic design
embodied in US codes (SEAOC-88, UBC-91) and in the European codes. The
US codes, unlike most of the European codes, use factored loads as well as
strength-reduction factors in proportioning of structural members; consequent-
ly the reduction factors also differ (Luft, 1989).
Therefore, in order to compare the seismic forces obtained from most of the
European codes with the seismic forces yielded by the US codes, we have
either to multiply the forces resulting from European codes by 1-4-1' 5 or,
alternatively, to obtain the reduction factors used in most of the European
codes by dividing the corresponding US reduction factors by 1-4-1' 5.
In the following, we shall refer mainly to the Californian Code SEAOC-88,
as a typical modern code.

6.1.2 SEISMIC INTENSITY FACTOR

The size of an earthquake is generally related to the amount of energy released;

this is usually measured by the magnitude M defined by Richter (1935) as

M=logA (6.3)
162 Earthquake design

where A denotes the maximum amplitude in ~m registered on a standard

seismometer placed at a point 100 km from the epicentre (as the existing
seismometers are generally located at different distances from the epicentre,
equation (6.3) has to be corrected accordingly).
Different procedures are used in seismic design, in order to determine the
intensity of the seismic motions, indirectly related to the magnitudes of past
earthquakes in the given area. We shall relate to two procedures: the intensity
scale, and the predicted horizontal maximum peak acceleration.

(a) Intensity scales

Intensity scales, such as the Modified Mercalli (MM), are based on historical
information dealing with the effects of past earthquakes and human feelings.
An abridged form of the Modified Mercalli scale is given in the following:

I. Usually not felt, except under exceptional circumstances.

II. Felt by persons at rest, especially in upper floors of buildings.
Suspended light objects may swing.
III. Felt indoors. Vibrations sensed like those of a passing car.
IV. Windows, doors, dishes disturbed. Standing motor cars noticeably
rocked.
V. Felt by nearly everyone, many awakened. Dishes, windows broken.
Cracked plaster. Unstable objects overturned.
VI. Felt by all, many frightened and run outdoors. Some heavy furniture
moved. Fallen plaster and damaged chimneys.
VII. Difficult to stand. Everybody runs outdoors. Moderate damage in
well-built structures, major damage in poorly built structures.
VIII. Slight damage to specially designed structures; substantial damage to
ordinary designed structures, with possible partial collapse. Fall of
chimneys, factory stacks, columns, monuments and walls. Heavy fur-
niture overturned. Sand and mud ejected, changes in well water.
IX. General panic. Masonry damaged or destroyed. Considerable damage
to specially designed structures, partial or total collapse of ordinary
buildings. Buildings shifted off foundations. Conspicuous ground
cracking. Underground pipes broken.
X. Most masonry and framed structures destroyed. Ground severely
cracked. Rails slightly bent. Considerable landslides from riverbanks
and steep slopes.
XI. Few, if any, masonry structures remain standing. Bridges destroyed.
Broad fissures in ground. Underground pipelines put completely out
of service. Earth slumps and land slips in soft ground. Rails greatly
bent.
XII. Damage total. Waves seen on ground surfaces. Objects thrown into
the air.
 Evaluation of seismic forces: regular structures 163

Esteva and Rosenblueth (1964) proposed a relationship between the MM

intensity, magnitude M and the epicentral distance; a graphical representation
of this relationship is shown in Figure 6.1, following Wakabayashi (1986).

( b) Predicted horizontal maximum peak acceleration

In procedures based on the predicted horizontal ground peak accelerations
(amaJ, usually, the codes refer to a relative acceleration Z = ama.)g, assumed
with a prescribed probability of exceedance (e.g. 10%) within a certain time
span (e.g. 50 years); the assessment is based on instrumental and historical
regional earthquake data.
The concept of peak ground acceleration (PGA) overlooks several data,
important for a correct evaluation of the probable structural damage. The data
not taken into account relate especially to the amount of energy released by
earthquakes, such as the duration of the acceleration peaks (very short peaks
lead to local damage only, without producing extensive collapse) and the
number of peak accelerations during a given earthquake (a small number of
peaks produce only limited damage). Newmark and Hall (1982) proposed
replacing the concept of peak ground acceleration by the more suitable con-
cept of effective peak ground acceleration, which now constitutes the basic
parameter in the seismic map of the Californian code SEAOC-88, and it is
probable that it will be also used in the future in other countries.
We should emphasize that consideration of the energetic aspect in evaluat-
ing the probable structural damage involves the maximum velocities occurring
in earthquakes (see section 6.6); consequently, maps including the probable
maximum velocities accompany in the SEAOC-88 code the probable maxi-
mum accelerations maps.

Hypocentral distance
(km)

1000

500

200
100
50

20

10
I I ~MM
2 3 4 5 6 7 8 9 10 intensity

Figure 6.1
164 Earthquake design

Average peak velocities (vmaJ and average relative peak ground acceler-
ations (Z = amaJg) corresponding to the MM scale intensities, as proposed by
Belt (1978), are given in Table 6.1. The values usually prescribed in seismic
codes are Z = O·05~O·40.

6.1.3 SOIL OR SITE FACTOR

The nature and thickness of various soil layers that overlay the bedrock
strongly affect the ground motions at a given site. The phenomenon is quanti-
fied by the site factors prescribed in seismic codes.
From the site factor point of view, the best soils are the rock-like, as well as
the stiff or dense soils. The most dangerous are soft clays, saturated sediments,
saturated cohesion less soils and loose or recent sediments; sandy soils, where
liquefaction may occur, are particularly dangerous. The site factors usually
prescribed in seismic codes vary between 1 and 1· 5, and exceptionally 2.

6.1.4 RIGIDITY AND MASS FACTORS; FUNDAMENTAL PERIOD

The intensity of seismic forces acting on a structure depends on its rigidity. In
order to quantify this effect we have first to define three important dynamic
concepts: degree of freedom, mode of vibration and response spectra.

( a) Degree of freedom
Consider a vibrating structure with lumped masses (Figure 6.2). The degree of
freedom (DO F) is defined as the number of parameters required to determine
the position of the masses at any moment.
Referring to the plane structure shown in Figure 6.2, by assuming a flexible
but inextensible axis, DOF = 3 (e.g. the coordinates Xl' xz, X 3 ); by also con-
sidering the vertical displacements, DOF = 6 (Xl' Yl' XZ, Yz, x 3 , h)·

Table 6.1 Average peak velocities and average relative

peak ground accelerations
Intensity v(m/s) Z=ama)g
(MM scale)

IV 0·01~0·02 0·015-0·02
V 0·02-0·05 0·03-0·04
VI 0·05-0· 08 0·06-0·07
VII 0·08-0·12 0·10-0·15
VIII 0·20-0·30 0·20-0·30
IX 0-45-0·55 0·50-0·55
 Evaluation of seismic forces: regular structures 165

Referring to the frame shown in Figure 6.3, the masses are generally as-
sumed as lumped at the levels of tl:ie slabs (by neglecting the mass of columns
and walls); by assuming flexible but inextensible columns and by considering
the beams as rigid, DOF = 4.
Note: If a given deformed axis (e.g. a sine curve) is assumed, the structure has a single
degree of freedom (SDOF), as a single parameter suffices to determine the position of
all the masses at any time. Referring to Figure 6.2, if the coordinate Xl is known, then
X 2 , X3 may be deduced from the condition that the deformed shape is imposd (e.g. a

sine curve).

Figure 6.2

m3

m2

m,

."". ."". ." ".

Figure 6.3
166 Earthquake design

( b) Modes of vibration
Consider the vibrating structure shown in Figure 6.4 (OOF = n) due to a short
'perturbation'; let u/t) be the deflection of the mass i. The maximum value of
u;(t) is the amplitude ¢i' After removing the perturbation the structure vibrates
'freely'; the amplitude decreases with time, owing to damping. It may be shown
(see Appendix At) that the deformed shape of the structure ui(t) may be ob-
tained as a sum of the amplitudes of n modes of vibration ¢i" ¢i" ¢i", 00. ,

mUltiplied by the functions C 1 (t), C Z (t), Cn(t): 00. ,

ui(t) = Ui, (t) + Ui2 (t) + 00. + Ui,,(t)

= ¢i, C 1 (t) + ¢i CZ(t) +
1
00' + ¢i" Cn(t) (6.4)

Each mode has a given form, which can be determined by dynamic analysis.
As the shape of the mode j is known, a single parameter will suffice to define
the positions of all the masses; in other words, each mode is a SOOF system.
Each mode has a natural period Tl > T z > ... > Tn' resulting from the dy-
namic analysis (see Appendix A). The period is defined as the time that elapses
during a 'cycle' (after which the motion repeats itself); the period is usually
measured in seconds (s). The number of complete cycles in a unit of time is
defined as the frequency of the vibration:

1
f=- (6.5)
T

Ui (t)

---c
1_-

Mode 3
T,=T T3

Figure 6.4
 Evaluation of seismic forces: regular structures 167

The frequency is measured in Hz = lis.

The first mode of vibration is the fundamental mode, and its period, which is
the longest, is the fundamental period T j = T.

(c) Fundamental mode; fundamental period

In the present section, we deal with regular structures, i.e. structures that are
not very slender and not very irregular (most building structures belong to this
category).
F or such regular structures, the effect of the first (fundamental) mode is pre-
dominant, and the effect of the other (higher) modes can therefore be neglected.
The length of the fundamental period T j = T is a criterion adequate to
define the type of building from the point of view of its rigidity:
Rigid T<0·3s
Moderately rigid T=0·3-0·7s
Slender T=0·7-l·5s
Very slender T> 1·5s
Buildings with periods larger than 2-3 s are rather uncommon.
An approximate evaluation of the fundamental mode can be obtained by
using empirical formulae recommended in the various codes or manuals.
These formulae usually take into consideration the effect of both structural
and non-structural elements (facade and partition brickwalls) and therefore
result in lower periods (i.e. higher seismic forces) than the 'accurate' methods.
During the first earthquake shocks, the assumption of collaboration be-
tween structural and non-structural elements is justified, but towards the end
of the earthquake this collaboration is doubtful (if special measures are not
taken to ensure it for strong motions); consequently, the seismic forces de-
crease during the earthquake, tending to those provided by 'accurate' methods
(where the structural elements only are taken into account). On the other
hand, the damping ratio (see Appendix AI) decreases towards the end of the
earthquake, and this slows down the rate of decrease of the seismic forces.
Using empirical formulae for evaluating the fundamental period is usually the
safer course; moreover, these formulae provide more accurate values than the
'accurate methods', at least for the first stage of the earthquake.
This paradox is emphasized in the commentaries to the proposed seismic
French code (1990):
The use of empirical formulae is at the same time the simplest and most accurate
method of taking into account the stiffness of nonstructural elements.
In the case of moment-resisting frames (Figure 6.5), the fundamental period
TJ = T is given as a function of either the number of storeys (n) or the total
height of the structure (H):
T=O·ln (6.6)
168 Earthquake design

T
H=nh

.100 ... ~ ,. J
Figure 6.5

or (according to the SEAOC-88 code):

T = 0·0732 H 3 /4 (6.7)
for reinforced concrete moment-resisting frames, or
T = 0·0854 H 3 / 4 (6.8)
for steel moment-resisting frames (H in m).
An average value:
T = 0.0793 H 3 / 4 (6.9)
can be used without significant error.
Another type of empirical formula takes into account the width L of the
building, too e.g. those recommended by the French code (1976) (Figure 6.6):
0'09H
T=-- (6.10)
fi
for reinforced concrete frames, or
O'lOH
T=-- (6.11)
fi
for steel frames (H and L in m).
Table 6.2 points out the differences between periods computed acCording to
the above-mentioned empirical formulae (reinforced concrete frames with a
width of 20 m and storey heights of 3 m have been considered).
In the following we shall refer to a procedure for evaluating the fundamen-
tal period of multi-storey uniform frames by considering only the structural
elements. In a first stage we shall deal with a uniform one-bay frame by
assuming that the beams are rigid (Figure 6.7a). The period T can be expressed
 Evaluation of seismic forces: regular structures 169

Table 6.2 Fundamental periods T(s) for moment-resisting frames, ac-

cording to various formulae
H Equation (6.6) Equation (6.7) Equation (6.9) Equation (6.1 0 )

20 0·7 0·7 0·7 0-4

30 \·0 0·9 \·0 0·6
40 \·3 \·2 \·3 0·8
50 \·7 \-4 \·5 \·0
60 2·0 \·6 \·7 \·2

in the form
T~ 8n TO

1)~)
where

(6.12)

and W denotes the total weight of the structure.

The curve of Bn versus the number of storeys n is shown in Figure 6.8. When
the deformability of the beams is taken into account (Figure 6.7b), the period
T will increase accordingly. In this case we can use the equation

(6.13)
The coefficient f1 was defined in section 1.2.5 as f1 = urnax/ uR, where Urnax de-
notes the maximum deflection of the considered one-bay frame under a given
pattern of horizontal forces and UR the same deflection when the beams are
rigid; a curve of f1 versus v = kb/ kc is shown in Figure 1.21.
We have to point out that the uniform one-bay frame shown in Figure 6.7b
may represent the equivalent (substitute) frame of a uniform multi-bay frame
(see section 1.3). Consequently, equation (6.13) provides an approximate value
of the fundamental period of a uniform multi-storey, multi-bay frame.
Equation (6.13) leads to periods close to those computed by 'accurate
methods', as it considers only the effect of the structural elements; it usually
yields higher periods than the empirical equations (6.6)-(6.11).
In order to obtain an approximate shape of the fundamental mode we may
use the following procedure. We load the given structure with the weights W;
directed horizontally instead of the actual, vertical directions (Figure 6.9): the
corresponding elastic line (u w ) is close to the shape of the fundamental mode
(the Rayleigh method).
The accuracy of this technique has been checked on a lO-storey building
frame (Figure 6.10), where:
 '" '~l T
+-

:::+ I It 3"" " Ib-oo

H=nh

e:-~ +-/4-
1 +-I-J.
(a) (b)
Figure 6.6
Figure 6.7

r1.28
T:i:EnTO
1.0 TO=v'(~~n

0.5

0.086 0.056 0.042 0.023

0.033 __ n

2 3 5 10 15 20 (no. of
25 30 storeys)

Figure 6.8
 Evaluation of seismic forces: regular structures 171

~ r-------,-----~

Wi _ t-------+------I
~ t-------+------I
--. t-------+------I
--. t-------+------I
--.t-------~----_+.

Figure 6.9

. -------

(a) (b)

Figure 6.10

(1) = accurate shape for rigid beams;

(1 ') = approximate shape for the same case;
(2) = accurate shape for very flexible beams;
(2') = approximate shape for the same case.

The greatest differences between deflections occur in the case of rigid beams,
but they remain less than 10%, and hence the approximation is justified for
practical purposes.
By the same method we may assess the fundamental period:

(6.14a)
172 Earthquake design

As a particular case of equation (6.14a), we consider a single-degree-of-freedom

system; U w is the horizontal deflection obtained by considering a horizontal
force equal to W Then we obtain Geiger's formula:

(6.14b)
where Uw is in m and T results in s.
Most modern seismic codes allow consideration of the deformed shape of
the fundamental mode of usual structures, with uniform or nearly uniform
storeys, as a straight line. This assumption leads to an expression of the
distribution coefficient in the form

d=2J~ H;) (6.15)

I ~)~ HJ
where Hi is the height up to level i and ~ is the weight of storey i. When the
structure is uniform (the weights ~ are constant; the storey heights hi are
constant), the distribution of seismic forces is inverted triangular (Figure 6.11);
the maximum horizontal force (at the top) results in
2FT
F max = n + 1 (6.16)

The distribution coefficient then becomes

d.= 2i (6.17)
I n(n+l)
In the case of RC structural walls, SEAOC-88 code recommends (Fig-
ure 6.12)
T = 0·0488 H 3 / 4 (6.18)

n
1
j
H=nh
2

.,,.
1

"
.,,,. 1
L

Figure 6.11
 Evaluation of seismic forces: regular structures 173

whereas the French code (1976) recommends (Figure 6.13)

(6.19)

(H,L in m).

0
0
0 H

Figure 6.12
J
~ Direction of seismic forces

J....-- L ----4
Figure 6.13
174 Earthquake design

Dynamic analysis for uniform structural walls without openings, by neglect-

ing the effect of shear forces on deformations, yields (see Appendix AI)

T= 1·787 )(;:1 3

) (6.20)

where g = 9·81 m s - 2 is the acceleration of gravity, W is the total weight of the

building tributary to the given shear walls, and I is the total moment of inertia
of the structural walls. The above assumption, that we may neglect the effect of
shear forces on the deformations, is only valid for 'high walls', i.e. for ratios
H / L > 5. In the case of medium walls (H / L = 3-5) it is recommended to
consider this effect; in the case of short walls (H / L < 3), it is essential.
For this purpose, we have to use the corrected moment of inertia Iv instead
of I; for instance,
I
I ~..,----=--=-
V-I + 0·68

where
6fEi 2 .2 I
8 = - -2 · 1=- (6.21)
GH ' A
fdenotes the shape factor (see section 2.2);j = 1·2 for rectangular sections and
2.2- 2.5 for I sections.
In the case of rectangular sections:
1-41 L2
s=---
H2

(6.22)

The periods computed from the above-mentioned equations are compared in

Table 6.3 (assuming storey heights of 3 m and a constant width of 20 m).
The openings in the structural walls lead to a significant decrease in their
rigidity (see section 2.3) and accordingly to an increase in the fundamental
period T. For flexible lintels, an increase of 1·6-2·5 may be expected; for
average lintels 1·4-2, and for stiff lintels 1·2-1·6.
In order to obtain the shape of the fundamental mode, we may use the
procedure described for the frames (the Rayleigh method). The accuracy of this
procedure has been checked for a 10 x 30 m wall (Figure 6.10b) and it has been
found that the errors remain within the same limits (lO%)in the upper zone of
the wall, but may increase significantly in the lower zone (H /3). Note that
deviations in the lower zone do not adversely affect the final results, so that we
may consider the approximate procedure as acceptable. Obviously, for 'squat
(short) walls' we have to consider shear deformations.
 Evaluation of seismic forces: regular structures 175

Table 6.3 Fundamental periods T(s) for structural walls,

according to various formulae
H(m) Equation (6.18) Equation (6.19 ) Equation (6.20)*
20 0·5 0·25 0-4
30 0·6 0-4 0·75
40 0·8 0·6 1·2
50 0·9 0·75 1·55
60 1·0 0·95 2·2
*W= 42 000-120 000 kN and E I = 6·2 X 107 -17-4 X 107 kN m2 were
taken into account.

For dual structures (moment-resisting frames + structural walls), the SEA-

OC-88 code recommends equation (6.18). In cases where the frames are more
rigid than the structural walls, an average value between equations (6.7) or
(6.8) and (6.18) is reasonable.
The effect of soil deformability on the rigidity of structural walls is import-
ant; it may decrease their rigidity significantly (see section 2.2.2 for structural
walls without openings and section 2.3.7 for coupled structural walls). In sec-
tion 2.3.7 a procedure is given to compute approximate fundamental periods T
in the form T ~ J(n + T;), where To denotes the period computed for elastic
walls on rigid soil and T, denotes the same period computed for rigid walls on
deformable soil.

6.1.5 DUCTILITY FACTOR; REDUCTION FACTOR

The concept of ductility, developed during the past 40-50 years, has an im-
portant role in explaining structural behaviour during earthquakes.
Ductility is defined as a property of the material, element or structure
subjected to cyclic loads, to display large inelastic deformations before failure.
The most commonly used ductility ratios are:

• displacement ductility:

(6.23)

where bm is the maximum displacement expected to be attained, and by is

the displacement at yield point.
Sometimes bm is replaced by the displacement at failure bu , although
bm ~ ba ; then

(6.23a)
176 Earthquake design

• curvature ductility:

({Jm ~ ({Ju
J1=-=- (6.24)
({Jy ({Jy

where ({Jm is the maximum curvature expected to be attained, ({Ju is the

curvature at failure, and ({Jy is the curvature at yield point.

In the following we shall refer to displacement ductility. In the case of RC

structures, the reinforcement detailing has a strong effect on ductility.
The range of ductility ratios usually varies between J1 = 1 (non-ductile) and
J1 = 6--8 (high ductility). In special cases, very high ductility ratios of 8-12 can
be achieved. 'Ductile structures' have ductility ratios of around 4-5.
The ductility ratios do not appear explicitly as a factor in the evaluation of
seismic forces prescribed by the codes, but they are included in the reduction
(or behaviour) factors R. The reduction factors result from comparisons be-
tween the results obtained from elasto-plastic or non-linear analysis and elas-
tic, linear analysis.
These factors depend, in addition to the ductility ratio, also on other par-
ameters (e.g. the existence of alternative 'lines of defence', such as lateral force
system redundancy and non-structural elements, changed damping and period
modification with deformation).
We shall refer first to the dependence of the reduction factor on ductility (by
using a primary reduction factor R a), and we shall then give corrected values of
the reduction factor R by taking into account other effects. Finally, the reduc-
tion factor R thus obtained represents the ability of the structure to sustain
strong seismic motions and absorb energy in excess of the allowable stress
limit, without collapse.
The primary reduction factor Ra is defined as

_Fm~Fu
Ru --=- (6.25)
Fy Fy

where F m denotes the force, corresponding to the maximum expected displace-

ment; F u is the failure load; and Fy is the yield force.
The primary reduction factor Ra can be determined according to two cri-
teria, the equal maximum displacement and the equal energy, as follows.

(a) The equal maximum displacements criterion ( Figure 0.14)

Analyses have been performed by Clough (1970) for 20-storey RC moment-
resisting frames, subjected to dynamic forces yielded by given accelerograms,
and by Derecho et al. (1978) for 20-storey cantilever walls, both types of struc-
tures having relatively long natural periods. The computations were performed
for two alternatives:
 Evaluation of seismic forces: regular structures 177

Elastic linear

C
:~~IYSiS
,'1
,,' l
,
" Dynamic
,/ nonlinear
7I---....~1 analysis

Figure 6.14

storey

20

16

12

4 Displacement
(mm)
100 200 300

Figure 6.15

• a dynamic inelastic non-linear analysis (leading to maximum horizontal

deflections u:nax);
• an elastic linear analysis (leading to maximum horizontal deflections u;;'aJ

The two analyses yielded nearly equal maximum deflections:

These results can be explained by the significant reduction in stiffness

evidenced by the non-linear analysis, accompanied by two opposite effects: on
the one hand, it leads to an increase in displacements due to lower rigidity; on
the other hand, it leads to a decrease in the magnitude of forces due to the
longer periods, and thus to smaller displacements. The results of the above-
mentioned analyses show that the two opposite effects are nearly equal. This
property is illustrated in Figure 6.15.
178 Earthquake design

We now can deduce the primary reduction factor:

As f.1 = bul by, we deduce

(6.26)

(b) The equal energy criterion

Veletsos, Newmark and Chelapati (1965) and Newmark (1970) have shown
that, for low and moderate natural periods, we can accept that the elastic
energy Eel and the elasto-plastic energy Eep (defined in Figure 6.16) are ap-
proximately equal. This property enables us to determine the primary reduc-
tion factor Ro as follows:

R o =J(2f.1- 1) (6.27)

Chopra and Newmark (1980) pointed out that analysis of several spectra (see
section 6.2.2) show that the primary reduction factor Ro = f.1 fits for periods
T> 0·5 s, while the primary reduction factor Ro = J(2f.1 - 1) fits for
T = 0,1-0,5 s; in the range T < 0·1 s the primary reduction factor Ro = 1.
In a few specific structures (e.g. nuclear reactor structures) we have to
ensure that the main structural elements remain in the elastic range for all
levels of earthquake ground motions. For most structures such a condition is
economically prohibitive, and we have to allow for inelastic yielding, on condi-
tion that it will not impair the vertical load capacity of the building (see
Commentaries to SEAOC-88). This is expressed by assuming conventional

F
Elastic
energy

Fy[~~llllt~E:la~sto-Plastic
energy

o
Oy Ro Oy ou=1l0 y

Figure 6.16
 Evaluation of seismic forces: regular structures 179

seismic forces (F) equal to those expected for elastic, linear design (Fe I)' divided
by the reduction factor R:
F
F =--=l (6.28)
R

The reduction factor R represents, as defined above, a corrected value of the

primary reduction factor Ro. Values of the factor R are given in all modern
codes, and they usually vary between 3 and 8 (when non-factored seismic loads
are considered, such as in most European codes) or between 5 and 12 (when
factored seismic loads are considered, such as in the US codes). The values of
the reduction factors R were determined on the assumption that the accep-
ted inelastic deformations are distributed quite uniformly in all structural
elements.
Note that lessons derived from recent strong earthquakes have shown clear-
ly the much better behaviour of buildings relying on RC structural walls than
that of buildings relying on moment-resisting RC frames (Aoyama, 1981;
Fintel, 1991, 1994). This can be explained by two main factors (Scarlat 1993):
• Structural walls are usually designed on the assumption of rigid soil (fixed
base): this leads to an overestimation of their stiffness and subsequently to
their overdesign (see section 3.3) .
• Designers neglect the ductility provided by the soil surrounding the founda-
tions of the rocking structural walls; this may significantly change the
evaluated ductility of the walls.
Paulay and Priestley (1992) have already noted this aspect, and have stated that
It is now recognized that with proper study, rocking may be an acceptable mode of
energy dissipation. In fact, the satisfactory response of some structures in earthquakes
can only be attributed to foundation rocking. For rocking mechanisms the wall super-
structure and its foundation should be considered as an entity. In this context rocking
implies soil-structure interaction.

We have only to add that it is rather surprising that this essential aspect has
awakened little interest among researchers and designers. The interesting pro-
posal of Luong (1993) dealing with the concept of energy-dissipating index of
soils (EDI) should be noted.
The existing reduction factors (e.g. those prescribed by the SEAOC-88 code)
in fact encourage the wide use of ductile moment-resisting frames as preferred
earthquake-resistant elements instead of structural walls, and this contradicts
lessons of past earthquakes (Fin tel, 1994).
We have to determine in design the 'real seismic deflection' (check of drift,
pounding, P ~ effects). Most codes recommend determining the real deflections
~ in the form

(6.29)
180 Earthquake design

where R is the reduction factor prescribed by codes and ~cl is the elastic, linear
deflection due to the prescribed seismic forces.
Equation (6.29) overlooks the effect of several reserves, usually neglected in
design (e.g. the existence of non-structural elements), thus leading to overes-
timation of the real deflection by several tens of percent.
The SEAOC-88 code prescribes a computation of the real deflections ~ by
dividing ~el by the factor (3 Rwl8) instead of the reduction factor Rw' By
taking into account that in the code factored seismic loads are considered, this
prescription provides, in fact, a reduction of displacements computed accord-
ing to equation (6.29) by several tens of percent, and this is close to the actual
conditions encountered.
In concluding, it should be emphasized that the important concept of duc-
tility has been sometimes overstated, and this has led to exaggerated deforma-
bility of the structure as a whole, and to dangerous P ~ effects, which may have
been responsible for the collapse of several multi-storey buildings (Eisenberg,
1994); Dowrick (1987) also noted that
until the 1980s, research and codes had rightly been preoccupied with overcoming the
excessive brittleness and unreliability of ill-reinforced concrete. However, there may
have been too much emphasis on creating ductility for ductility's sake.

6.1.6 IMPORTANCE FACTOR

Seismic codes prescribe importance factors I, depending on the occupancy cat-
egories of buildings; the I-factor increases the lateral forces by up to 25-40%.
In the SEAOC-88 code the I-factor is limited to 1.25, considering that
the details of design and construction often dominate the seismic performance. There-
fore, increasing these aspects for essential facilities can improve performance more
effectively than relying solely on increased design force levels ...

6.1.7 VERTICAL SEISMIC FORCES

The horizontal seismic forces are accompanied by vertical seismic forces. Their
intensity is usually assumed to be i of the horizontal forces.
The effect of vertical forces is important for several structural elements:
• Horizontal cantilevers-where the downwards vertical forces must be ad-
ded. The SEAOC-88 code prescribes checking cantilevers for upwards
forces equal to 20% of the vertical loads.
• Prestressed concrete beams-where supplementary vertical seismic forces
may disturb the balance between stresses due to vertical loads and pre-
stressing forces. The SEAOC-88 code prescribes checking prestressed
beams by considering 'not more than 50% of the dead load, alone or
in combination with lateral force effects'.
• Precast elements-where the upwards seismic forces decrease the horizon-
tal friction. On the safe side, it is requested to ignore friction forces when
designing connections of precast elements (SEAOC-88).
 Evaluation of seismic forces: special structures 181

6.2 Evaluation of seismic forces: special structures

~2.1 GENERAL APPROACH

In cases where the structure is very slender or has significant irregularities, in

either the horizontal or in the vertical plane, we define it as a special structure
and we have to carry out a dynamic analysis. This analysis can be performed by
using one of the following procedures:
• modal analysis (see section 6.2.3 and Appendix A);
• direct integration of the equations of motion by a step-by-step technique
(a time history analysis).
For usual design it suffices to perform a modal analysis: we determine the
forces and resultant stresses for each considered mode of vibration according
to response spectra given in seismic codes and we superimpose the results
'statistically'. In very special cases we haye to perform a time history analysis
(usually a non-linear analysis) based on a number of ground motion time
histories; the dynamic structural response results from a numerical integration
of the equations of motion. The limit of slenderness beyond which a modal
analysis is required depends on the magnitude of the fundamental period T
(usually a limit of 1·5- 2 s is prescribed) or on the total height of the building
(usually a limit of 70-80 m is required).
The proposal for the revision of the French code formulated by AFPS
(1990) prescribes as a limit the height of 75 m or the period.

(6.30)

which yields T=0·74s for H=20m, T=I·24s for H=40m, T=I·68s for
H = 60m, and T= 2s for H = 75m.

6.2.2 RESPONSE SPECTRA

During earthquakes, the foundations are subjected to random displacements

UG of the ground (velocity uG ' acceleration lid. The total displacement U TOT
comprises, in addition to the ground displacement U G , the relative displace-
ment U of the structure (see Appendix A):
UTOT = u + u;
G (6.31)
where an overdot (.) denotes the time derivative.
The basic data for a dynamic analysis are the recorded displacements,
velocities or accelerations of the ground during earthquakes (Figure 6.17).
From these recordings we may evaluate the response of the simple pendulum
shown in Figure 6.18 during a given earthquake.
It is of practical interest to draw diagrams showing the response of pendu-
lums of variable height (i.e. of various rigidities, expressed by their periods T)
to the registered ground motions. By considering a number of records in a
182 Earthquake design

d(mm)
20

-20 +---~-_~__- l -_ _-1.._ _---L._ _ --L_. t(5)

5 10 15 20 25 30

v(m 5-')
4
2

o
-2

-4-t---~--"""""---L.---L----L----L-_t(5)
5 10 15 20 25 30

alg

0.3
0.2
0.1
o
-0.1
-0.2
-0.3
-0.4
t---+---+---'---I--~~---+ .......- t(5)
5 10 15 20 25 30

Figure 6.17

given area and retaining the maximum values, we may draw curves exhibiting
the variation of the maximum relative displacements (Sd)' maximum relative
velocities (SJ and maximum total accelerations (Sa) as functions of the period
of the pendulum, as shown in Figure 6.19.
These are the spectral responses of the structure to the given earthquake. For
each damping ratio ~ there is a different spectrum. Owing to the presence of
damping, the peaks of the recordings are smoothed out and the response spectra
are significantly flattened. Most codes assume a damping ratio of ~ = 0·05.
The spectra Sd' Sv and Sa are interrelated as follows:
(6.32)

where w denotes the circular frequency:

2n
w=2nf
. =-
T (6.33)
 Evaluation of seismic forces: special structures 183

rH
' M=~ M

1 ~
J- ua +Urel=U~
+- UTOT ---+
Figure 6.18

The equations (6.32) suggest the possibility of representing a given spectrum

and of reading its ordinates ('spectral values') on three different scales, which
are usually logarithmic. Two different spectra are drawn in Figure 6.20, one for
~ = 0 and the second for ~ = 0'10; these are based on a representation of the
1940 E 1 Centro earthquake, in direction N-S as drawn by Newmark (1970).
The straight lines represent the maximum ground displacements (d G ),
velocities (vG ) and accelerations (a G ). mH

In the c~'se considered dG = 3 r~m, vG = 32 cm s -) and aG . = 0.33 g

(y = the acceleration of gravity)': Let us consider a point A on the"'spectrum
~=0·1O, corresponding to the period T=0·5s. We read:

• on the scale Sd' d max = 3·5 cm = 0·11 dG ;

-) "'"'
• on the scale Sv' L'max = 45 cm s = 1-40 vG ;
• on the scale Sa' {lmax = 0·64 g = 1·94 (lGm,; "'H
That is, an SDOF (pendulum-like) system, with a natural period of 0·5 s,
located in El Centro during the given earthquake, would display a maximum
relative displacement of 3·5 cm, a maximum relative velocity of 45 cm s -) and a
maximum total acceleration of 0·64 y.
Inspection of the response spectra shows an important property: the abso-
lute maximum accelerations of very rigid structures of the considered SDOF
system (T < 0·1 s) approach the ground accelerations (the structures. behave
like rigid bodies attached to the ground) and the corresponding maximum
forces are very high (they tend to F max = maG ' where m denotes the consider-
ed mass). The displacements are relatively small. In contrast, the maximum
relative displacements of very slender structures (T > 3 s) are high, and the
maximum forces are relatively low (much lower than m (lG",).
184 Earthquake design

T(s)
0 0.5 1.0 1.5 2.0 2.5 3.0

Sv(ms- 1)
15
;=0%
10
5°;'
5
I
T(s)
0 0.5 1.0 1.5 2.0 2.5 3.0
Sa(m S-2)

15

10

T(s)
3.0

Figure 6.19

We should point out that the type of soil strongly influences the response
spectra, and therefore different response spectra have to be drawn.
We distinguish between two types of acceleration spectra. In Figure 6.21a,
the peak value of Sa lies in the low periods range: in such cases amplification
may occur for rigid structures, and consequently they are dangerous. In Fig-
ure 6.21 b, the peak value of Sa lies in the high periods range; in such cases,
resonance may occur for slender structures and these, too, are consequently
dangerous.
In order to allow for both types of earthquake, most modern codes also
provide spectra with an enlarged 'plateau' covering rigid and moderately slen-
der structures as well (Figure 6.21c).
 Evaluation of seismic forces: special structures 185

v(cm 5- 1)

d(cm)

....f'L--+----4-+--+-I~-_+_-_+-+__~--+_l~T(5)
0.05 0.1 0.2 0.3 0.4 0.5 2 3 45 10

Figure 6.20

6.2.3 MODAL ANALYSIS

It may be shown (see Appendix A) that the vibration modes can be 'decoupled'
so that each one can be studied independently. The corresponding modal
analysis yields the maximum forces:

C 1 WA..
F = ,'1-'" (mode 1)
" T~
C WA..
F. = z ,'1-'" (mode 2)
'2 T~

(6.34)
F.
C.wcp
= J , " (mode 1)
" Tj

C WA..
F. = n ,'1-',,, (mode n)
'" T~
186 Earthquake design

Sa

T
2 3
(a)

(b)
2 3

Sa

(c) 2 3

Figure 6.21

where Cj is a constant depending on the seismic zone, the response spectra,

including the effect of the type of soil, and the importance of the building; and
~ is the weight acting on the floor i. C j and the exponent :J( vary with the
codes; SEAOC-88 prescribes :J( =~.

(6.35)

It follows from equation (6.35) that the forces Fi (acting at level i, correspond-
ing to mode j) are proportional to the product' ~ ¢i' When the weights are
uniform, the forces F i , follow the shape of the mode j '(Figure 6.22).
 Evaluation of seismic forces: special structures 187

Figure 6.22

The number of modes to be considered in the analysis is defined by various

codes, or is chosen by the designer according to the desired degree of accuracy
(see Appendix A). Usually, no more than three modes need to be considered;
analysis of very slender structures, three-dimensional analysis and special dy-
namic analysis (involving, for instance, interaction between soil and structure)
need more than three modes.
The forces corresponding to mode j lead to the structural response rj (be-
nding moments, shear or axial forces, deflections). Taking into account that
the maximum responses are not simultaneous, a statistical superposition is
usually admitted (see Appendix AI):

rTOT = J(ri + r~ + r~ + ... ) (6.36)

Two main features of modal analysis are noted as compared with the static
lateral force procedure:
• It takes into account the effect of higher modes of vibration, in addition to
the fundamental mode; this affects the magnitude of the total lateral force
(the base shear).
188 Earthquake design

• The distribution of the base shear for each mode of vibration .i is per-
formed according to the shape of the mode.i and the values of the masses
acting on each storey.

6.2.4 APPROXIMATE METHODS IN MODAL ANALYSIS

(a) Approximate evaluation ()l higher modes' periods

• Moment-resisting frames: in the case of uniform structures with more
than eight storeys and fixed base, the following ratios hold:

These results are based on dynamic analysis of one-bay frames (which

may be considered as equivalent frames of multi-bay structures) up to 20
storeys with ratios kb/kc = (Ib/l)/(Ijh) ~ 0·1.
• Structural walls: if we neglect the effect of shear forces (allowable for
H/L>5) and the effect of soil deformability, we obtain for uniform walls
without openings (pilkey and Chang, 1978):
1) 1·425
(6.37)
Tl (2j - 1)2

resulting in:
T
~ Ts ~ 0.02
T-~ 0·16', T-
1 1

When we consider the effect of shear forces on the deformations these

ratios may change significantly, and they depend on the coefficient s =
6fEi 2/GH2; for s=0'35 we obtain
T
~~0'08'
T- ,
1

Note that the effect of soil deformability on the rigidity of structural walls is
usually important, and overlooking this effect may lead to significant errors
(see section 2.2).
In order to determine the order of magnitude of the effect of soil deforma-
bility, we consider an RC structural wall with a height of 30 m and a horizon-
tal rectangular section 0·20 x 6·00 m.
Assuming a fixed base, the following ratios are obtained:
T T
~~0'18'
y- , ~~0'07'
T- ,
1 1

When soil deformability is considered:

 Evaluation of seismic forces: special structures 189

• for subgrade modulus k, = 100000 kN m - 3:

T T
~~0'15'
T- , ~~0'09'
T- ,
1 1

• for subgrade modulus k, = 20000 kN m - 3:

The effect of shear forces was taken into account in all computations.

(h) Approximate assessment ()f higher modes' shapes

As the effect of the higher modes is usually much smaller than the effect of the
first mode, we may use rough approximations regarding their shapes without
adversely affecting the final results.
The Spanish code (1976) proposes the schematic shapes displayed in Fig-
ure 6.23. Their use, together with approximate evaluations of the periods,
enables the effect of higher modes to be taken into account, with no need to
resort to modal analysis computations.

(c) Envelopes of 'glohal effects'

Proposals have been made to construct envelope diagrams for bending mo-
ments and shear forces, including the effects of higher modes, as functions of
the basic diagrams (including the effect of the fundamental mode only), multi-
plied by given factors.
CEB-85 uses the envelopes displayed in Figure 6.24. Numerical examples
that we have performed show that the resultant stresses yielded by the CEB-85
envelopes are much higher than those obtained by a regular modal analysis.
The Romanian code (1978) recommends using the envelope shown in Fig-
ure 6.25. Numerical examples we have performed show that the coordinate
0·55 M is underestimated.

r-
H
-+H/3
-t
H/3
+-
HIS
+
HIS
-to
HIS
-.t-
-t HIS

L 2--
H/3
-,jo-
3
-
f.
HIS
-+-
Figure 6.23
190 Earthquake design

The SEAOC-88 code proposes a simpler method to account for the effect of
higher modes: in order to increase the shear forces in the top floors (where the
effect of higher modes is more significant) a fraction (Ftop) of the total horizon-
tal force FT is assumed to be concentrated at the top of the building (Fig-
ure 6.26). Then the remaining resultant force (FT - F top ) is distributed to each
storey. The computations are performed as for a 'regular structure', without
referring explicitly to the higher modes.
Capra and Souloumiac (1991) propose to take into account the effect of
higher modes by multiplying the resultant stresses due to the fundamental

(iY /
mode by a factor:

p = 1 + 0·05 3 > 1·05 (6.38)

where T.: is defined in Figure 6.27; TJ denotes the fundamental period.

6.3 Structural irregularities

Structural irregularities are considered as one of the main reasons for the poor
seismic performance of buildings. Most codes require special treatment of
structures with significant irregularities.
We may classify the irregularities into two main groups: plan irregularities
(in the horizontal plane) and vertical irregularities (in the vertical plane).

6.3.1 PLAN IRREGULARITIES

Significant irregularities in the horizontal plane of a structure qualify it as a
special structure, and a dynamic analysis (in most cases a modal analysis) is
required. We shall distinguish between several kinds of horizontal irregularity:
asymmetric structures (the most important one), presence of re-entrant corners,
diaphragm discontinuity, and non-parallel structural elements.
(a) In the case of asymmetric structures subjected to horizontal forces, the
slabs (assumed to be rigid in their planes) describe a displacement that can be
resolved into two components, as follows (see section 4.1):
• a translation parallel to the given forces; their resultant passes through the
centre of mass of the slab (usually very close to the centre of gravity of the
slab), associated mostly with translational forces (stresses); a planar analy-
sis can be used;
• a rotation with respect to an axis passing through the centre of rigidity of
the slab, associated mostly with torsional forces (stresses); a three-dimen-
sional analysis is required for determining torsional stresses.
As the centre of rigidity does not coincide with the centre of mass, a storey
torsion moment will develop: MT = Ve, where V is the storey shear (the
 r
H
T
2HI3

-+-
M,

1
L-J0.67i+0.85
HI3
+ L
+ 1 -+
W M TOT =WMI
J.-L+
\. ----- V-
Frames J
'----v-- .J
Structural walls

Figure 6.24

r
H
-t-
0.3H

-t
O.4H

--t
0.35 M1m"

0.55M1m"
Fi~
\r--J
-- r-t
FT-Flop
..
~
~
FlOp

11
Sa

1 \ I
H
0.3H ~

11
Hi
~
-J.- \ I
~ i
\1 I ~T
M1max 1.1 Mmax • Tc 7i

Figure 6.26 Figure 6.27

Figure 6.25
192 Earthquake design

resultant of horizontal forces above the given storey), and e denotes the eccen-
tricity of this force (methods of computing the torsional forces are given in
Chapter 4). The important effect of soil deformability in such analyses is em-
phasized, as this deformability leads to significant decreases in the torsional
forces (Scarlat, 1993).
Modern codes prescribe one of the following two approaches in order to
define an irregular structure from the point of view of torsional stresses.
According to most European codes, we compute the eccentricity e accord-
ing to Lin's theory (see Chapter 4); in the case where
(6.39)

the structure is irregular (see, for instance, the recommendations of the AFPS
for the new French code, 1990).
RT denotes the torsional radius:

Torsional rigidity
(6.40)
Translational rigidity
The torsional rigidity denotes the moment of torsion corresponding to a
relative rotation of the slab ({J T = 1, when vertical elements with fixed ends are
assumed; by using the torsion constant Co defined in section 4.1, we obtain the
torsional rigidity: 12 EC o/h3.
The translational rigidity denotes the horizontal force corresponding to a
relative displacement of the same slab ~ = 1; this gives 12 ELI /h 3 .
For a force F)' parallel to the axis y:

R;=IIiJ
F or a force F x parallel to the axis x:
(6.41a)

(6.41b)

The SEAOC-88 code defines torsional irregularity by referring to the deflec-

tions due to translation (u) and due to rotation (~u); see Figure 6.28. This
results in

urn in = U- ~Ul;

When Urnax> 1·2 uavcr ' the structure is considered as irregular, and we have to
use an increased eccentricity in the computations:

Urnax )2
eT = e rnin ( 1.2 U (6.42)
aver
 Structural irregularities 193

r----------------v U
I

CR • CM
•

Umax
Uave~ __ -~
____ 1,./llU2
--- I
I

Figure 6.28

emin is the minimum (accidental) eccentricity prescribed in order to account for

uncertainties in the location of the loads. It is usually limited to 5% of the
building dimension at the considered level, perpendicular to the direction of
the given forces.
In section 4.4 it was shown that asymmetric structures can be classified
from the point of view of their plane irregularity by using torsional index TI,
defined in equation (4.15), into three categories: regular, moderately irregular
and significantly irregular.
The torsional index is also used in the 'first screening' of existing buildings
(see Chapter 7).

(b) Re-entrant corners (Figure 6.29) lead to local stress concentrations, which
may be important when the sizes of the niches are significant; and structure is
usually considered irregular when cx >0'15L x and c}.>0·15Lv '

(c) When important non-parallel and asymmetric vertical elements are present
(Figure 6.30), the structure is considered irregular.

6.3.2 VERTICAL IRREGULARITIES

(a) Weak storey concept
The 'weak storey' concept is related to a discontinuity in the strength of
buildings with more than two storeys or higher than 10m. By denoting by
194 Earthquake design

Figure 6.29

• • • •
• • •
• • • /.
• • • • •
Figure 6.30

Llil Vu the sum of the ultimate shear forces in the vertical elements of the
storey i and by LIi+ II VU the same sum for the storey above it, a weak storey is
defined as

IVu<O'SIV u (6.43)
iii (i+11

In the case of columns and structural walls without openings we may use the
following approximate relation (Figure. 6.31):

(6.44)
iii (i+ II

For profiled structural walls and cores C.L, I, C, ... ) (Figure 6.32), an approxi-
mate relation based on the first stadium of reinforced concrete may be used:

I -
(i)
( I) Ymax
<O,S I
(1+1)
(- I )
Ymax
(6.45)

It has been assumed that the reinforcement ratio of the walls is nearly uniform.
A condition similar to equation (6.43) must be checked for the lintels (also
on the assumption that their reinforcement ratio is uniform).
 Structural irregularities 195

1
1
~F
Figure 6.31

Ymax

._.L

Figure 6.32

(h) Soft storey concept

The 'soft storey' concept is related to a discontinuity in the stiffness of the building.
By denoting by LU) K", the sum of the translational stiffnesses of the vertical
elements of the the storey i and by LU+I)K", the same sum for the storey above
it, the soft storey will occur when

IK", <0·7 I K", (6.46)

U) U+I)

We may consider an approximate relation:

(6.47)
U) U+I)

Ie (I ,w) represent the moments of inertia of columns (structural walls).

196 Earthquake design

It is assumed that the effect of shear forces on the rigidity is nearly the same
for both storeys i and i + 1 and can be neglected in equation (6.47), although
the assumption is questionable. The author (1994b) has proposed an alterna-
tive way to identify soft storeys: the structure is subjected to a horizontal force
at top and we determine the "storey drifts" (\ = U j - U j _ 1 (storey i) and
b j + 1 = U j + 1 - U j (storey i + 1); the soft storey occurs when b j + 1/ b j < 0.7.
Soft storeys have drawbacks that make them particularly dangerous:

• The vertical discontinuity leads to important stress concentrations, ac-

companied by large plastic deformations .
• Most of the deformation energy is dissipated by the soft storey columns,
and this leads to major overstressing of these elements; onset of plastic
hinges may transform the soft storey into a mechanism resulting in col-
lapse.

(c) Mass discontinuity

A mass discontinuity is present, according to SEAOC-88, when 'the effective
mass of any storey is more than 150% of the effective mass of an adjacent
storey (roofs excepted)'.

( d) In-plane discontinuities
In-plane discontinuities are present in vertical lateral force resisting elements
(Figure 6.33).

Figure 6.33
 Structural irregularities 197

(e) Set-back
A discontinuity, as shown in Figure 6.34, leads to a concentration of stresses in
the adjacent storeys and to the need for special reinforcement of these zones.
The SEAOC-88 code defines the set-back by Bb> 1· 3 B t • The ATC 1978
Model Code proposed a criterion based on rigidities, namely:
(6.48)

6.3.3 IRREGULARITIES AND STRUCTURAL DESIGN

Structural irregularities derive from real, often unavoidable, programmatic and
aesthetic architectural reasons. It is practically impossible, and usually not
necessary, to avoid theses irregularities completely.
However, we have to take into account that lessons from past earthquakes
demonstrate beyond all doubt that severe irregularities were among the main
causes of building collapse during earthquakes. Therefore we have to take care
to identify structural irregularities, and to quantify their damage potential with
a reasonable degree of precision. We can then introduce the structural modifi-
cations and strengthenings required to ensure adequate seismic resistance.
Significant inelastic deformations develop close to the structural discontinu-
ities during major levels of ground motion, which may lead to local distress
and sometimes to general collapse. In this context it should be borne in mind
that the values of the reduction factor R (see section 6.1.5) have been deter-
mined by assuming that the cyclic inelastic deformations are quite uniform in
all structural elements. In the case of irregular structures, these values are not
valid any more, and we have to decrease them drastically when designing
'irregular elements', i.e. to increase the forces acting on these elements com-
pared with those based on code formulae.

r 1 H

1 J
Figure 6.34
198 Earthquake design

In fact, only inelastic dynamic analyses, tests of structures subjected to

dynamic cyclic loads, or pertinent observations of structural damage after
major earthquakes can provide reliable data for quantifying possible damage
caused by various types of structural irregularities. As these are beyond the
practical possibilities of usual seismic design we have to be content with pre-
scriptions derived from such analyses, tests or observations and sometimes,
unfortunately, also from rules of thumb.
In the following, we shall refer to the prescriptions and recommendations
included in the SEAOC-88 code and its commentary.
Two ways are prescribed in this code for dealing with structural irregulari-
ties:
• performing a dynamic analysis, usually a modal analysis;
• using amplification factors in the design of the 'irregular elements', in
order to take into account the stress concentrations specific to structural
irregularities.
As shown in section 6.2.3, the modal analysis has two main features as
compared with the static lateral forces procedure: it takes into account the
effect of higher modes of vibration; and the distribution of the base shear is in
agreement with the vibration mode shape. This distribution can provide in
several cases a more accurate picture than the linear distribution yielded by
the static lateral force procedure. As an example, we refer to the structure
shown in Figure 6.35.
Regarding the effect of higher modes, we have to point out that the need to
take these modes into account depends only on the rigidity of the structure:
these modes do not provide the information needed to quantify the effect of
stress concentrations, which accompany structural irregularities.
SEAOC-88 provides several 'penalties' for structural irregularities, as fol-
lows.

(a) (b)
(a) Distribution according to static lateral force procedure
(b) Distribution according to model analysis (mode 1)

Figure 6.35
 Bounds on the seismic coefficient 199

(a) Plan irregularities

For torsional irregularity, as shown in section 6.31, in the case of siginificant
eccentricities we have to increase the torsional forces by increasing the eccen-
tricity resulting from a static analysis (equation (6.42)).
Connections between vertical resisting elements and horizontal diaphragms
of structures in seismic zones with relative peak ground accelerations Z > 0·3
must be designed by considering allowable stresses without the usual increase
of 33%.

(h) Vertical irregularities

Weak storeys must be designed to resist forces obtained by modal analysis,
multiplied by a factor of (3 Rw/8). As the average reduction factor Rw pre-
scribed by SEAOC-88 is 8, we have to multiply the design forces by 3. Note
that the recommended multiplying factor is independent of the extent of weak-
ness of the considered storey: it is the same for a decrease in strength of 35 % or
65 %. We point out that amplification factors should be applied also for soft
storeys (see section 6.6).
Recommendations for the analysis of set-back buildings are based on the
proposals formulated by Blume, Knox and Lindskog (1958).
The buildings are divided into four categories, depending on the ratios H t / H
and Bt / B (Figure 6.34). For each category a specific kind of structural analysis
is recommended.
The proposals were formulated before computer programs made dynamic
inelastic analyses possible, and their supporting evidence is not available. It
would be preferable to propose amplification factors for the design of the
structural elements close to the discontinuity level, based on dynamic inelastic
analyses.
As an incentive to the design of regular structures, a reduction of the base
shear of 10% is permitted for such structures.

6.4 Bounds on the seismic coefficient

6.4.1 GENERAL APPROACH

The seismic coefficient has been defined in section 6.1.1:

FT
c=-
W

where F T is the total horizontal force, equal to the base shear V, and W is the
total weight taken into consideration; W usually includes the dead load and a
fraction of the live load.
200 Earthquake design

The seismic coefficient represents the most important parameter in seismic

design. It is obtained by multiplying four or five factors; these include the
seismic intensity, the soil type, the structural rigidity and its ductility, some-
times taking into account correcting factors due to structural irregularities.
None of these factors has a solid scientifically determined basis, and their
product is exposed to wide scattering. Therefore, we have to limit the seismic
coefficient at both ends: cmax in order to avoid overdesign and ('min in order to
avoid underdesign.

6.4.2 THE UPPER BOUND

There are two methods for determining the upper bound of the seismic coeffi-
cient:

• by limiting the values of one or more factors included in the seismic

coefficient; for instance, the SEAOC-88 code limits

(6.49)

where S denotes the site factor and Tthe fundamental mode of the struc-
ture;
• by limiting the seismic coefficient itself. The Romanian code (1981) limits ('
to
C ~ (0' 30~0'45) Z (6.50)

depending on the type of structure.

Scarlat (1989) proposed:
for Z = 0·05: (' ~0'10 Z
forZ=0·30: c~0'20Z

Between these limits we can interpolate:

('~0'08 + 0·40 Z (6.54)

6.4.3 THE LOWER BOUND

In order to determine a lower limit of the seismic coefficient we can also apply
two procedures:
• to limit one or more factors; for instance, SEAOC-88 limits:
C 1·25 S
-= 2/3 ~0'075 (6.52)
Rw T Rw
where Rw denotes the reduction factor;
 Bounds on the seismic coefficient 201

• to limit the seismic coefficient itself. Scarlat (1989) proposed the following
limits:

for Z = 0·05: C? 0·02

for Z = 0·30: c?0·05
Between these limits we can interpolate:

c ? 0·014 + 0·12 Z (6.53)

It may be useful to translate the seismic coefficients into equivalent wind

pressure coefficients. To this end we shall refer to the building shown in Figure
6.36. Assuming an average weight of 3·5 kN m - 3, the total weight results in

w~ 3·5 BLH
The total seismic force:
F,=cx3·5BLH
The total wind force:

Fw= Pw BH
where Pw denotes the wind pressure.
Equating Fs = F w yields

Pw = 3·5 c L (6.54)
(Pw in kN m - 2; L in m).
For a width L = 20 m and by admitting

Figure 6.36
202 Earthquake design

where v is the maximum wind velocity (Pw in kN m - 2; v in m s - 1), we obtain

c=O'01: pw=0'70kNm- 2 v=33ms- 1

0·02: 1·40 47
0·05: 3·50 75

Note that inspection of buildings following the 1906 San Francisco earth-
quake has shown that buildings designed for wind pressures of 1· 5 kN m - 2
performed well (Key, 1988). Obviously, several specific factors must be taken
into account here, such as the large width of the external walls then used, the
important reserves of strength taken by the designer in order to compensate
for the poor technical state of the art, and the possibility that the earthquake
had a relatively long period - far from the natural fundamental period of the
buildings (althouth such periods are not specific for Californian earthquakes).
Yet the order of magnitude of the corresponding seismic coefficients remains
relevant. We see that a relatively modest seismic coefficient of 0·05 corre-
sponds to very high wind pressures.
It is interesting to add that in the Bucharest 1977 earthquake (rated at
8-8'5 on the MM scale), eight-storey buildings designed for seismic coeffi-
cients of 0,03-0,04 suffered little damage, and only three buildings out of
several thousand designed for such seismic coefficients collapsed.
The seismic design must not involve excessive horizontal forces, but has to
concentrate on other aspects: adequate structural solutions, avoiding excessive
irregularities, choosing structures with a reasonably high ductility, good detail-
ing and good workmanship.

6.5 Problems of deformability

6.5.1 GENERAL APPROACH

The plastic deformations (~pl) occurring during an earthquake are much

greater than the elastic one (~el)' One assumes (section 6.1.5, equation (6.29)):

where R denotes the reduction factor (which increases with ductility). It follows
that the actual, plastic deflections increase with the ductility; larger deflections
are expected in the case of ductile structures.
Excessive deflections are undesirable, as:

• they may lead to severe damage to non-structural elements;

• they may increase the P ~ effect;
• they may cause pounding of adjacent buildings.
 Problems of deformability 203

6.5.2 STOREY DRIFT LIMITATIONS

SEAOC-88 limits the storey drift to:

L1 0·04
H<20m: -<-- (6.55)
h Rw
L1 0·03
H>20m: -<--
h Rw
We have to bear in mind that the deflections L1 develop as a result
of factored loads considered, and that the reduction factors Rw used in the
SEAOC-88 are greater than the reduction factors used in Europe.
The CEB 1985 model code distinguishes three situations:

• brittle partitkms: L1/h < O'OI/R;

• deformable partitions: L1/h < 0'015/R; (6.56)
• no connections between partitions and structure: L1/h < 0·025/R.
By considering an average reduction factor R = 5·5 we obtain
L1 1 1
-=- to-
h 250 350
The ratio 1/350 seems exaggerated and difficult to satisfy, especially for mo-
ment-resisting frames.
Note: ~ includes translational and torsional effects. Inelastic, non-linear behaviour is
taken into account by letting ~ = ~e1 R, where ~el denotes the storey drift obtained by
elastic, linear analysis.

6.5.3 P L1 EFFECT

Equations of equilibrium are usually formulated for the undeformed structure

(first-order analysis). When deflections are excessive and large axial forces
exist, we are obliged to take into consideration the supplementary moments
and shear forces evidenced by formulating of the equilibrium equations for the
deformed structure (second-order theory or P L1 effect).
Referring to the cantilever shown in Figure 6.37:
First-order theory: M = F h
Second-order theory: M = F h + PL1
As the deflections occurring during earthquakes may be very large, the PL1
effect can sometimes be important. It can explain the collapse of several multi-
storey buildings (Eisenberg, 1994).
According to the SEAOC-88 code we are obliged to consider the PL1 effect
when:
204 Earthquake design

• the storey drift ratio:

L\ 0·02
->-- (6.57)
h Rw
• the stability coefficient (Figure 6.38):

PL\
8=->0,10 (6.58)
Vh
In equations (6.57) and (6.58), L\ denotes the actual, inelastic storey drift.
SEAOC-88 recommends computing L\ by multiplying the elastic drift by a
factor equal to 3 Rj8. When the PL\ effect must be accounted for, we have to

r
h

L
Figure 6.37

-+-A--+
t
T~
h

1 Figure6.3S
 Problems of deformability 205

design the columns by multiplying the shear force V from the first-order
analysis by the stability coefficient 0, defined by equation (6.58).
In high seismicity zones (Z = 3, 4), the storey drift limitations usually render
the check of the P ~ effect useless.

6.5.4 POUNDING OF ADJACENT BUILDINGS

Adjacent blocks are usually separated by narrow gaps, which are intended:

• to avoid excessive stresses due to temperature changes and shrinkage of

concrete (temperature joints);
• to avoid effects of differential foundation settlement (settlement joints);
• to avoid excessive torsional forces due to symmetric blocks, especially in
seismic areas (seismic joints).

As the natural periods of adjacent blocks may differ, seismic motions with
different phases may occur, leading to pounding (hammering).
Damage to buildings and local distress can occur, and general collapse due
to pounding has been sometimes reported.
We have to distinguish between three different situations, as follows.

I. The slabs of adjacent blocks are located at different elevations; this is the
most dangerous situation and may lead to collapse.
2. The slabs of adjacent blocks are located at the same elevations; limited
damage can be expected in this case.
3. The same as 2, but strong structural walls/cores are positioned close to
the joint (Figure 6.39), perpendicular to the gap line. Only minor damage
can be expected in this case.

The standard solution for avoiding the pounding effect is to design suffi-
ciently wide gaps in order to enable each one of the neighbouring blocks to
vibrate without contact. It entails widths equal to d = ~1 + ~2 at each floor. ~1
and ~2 are the maximum seismic deflections of the adjacent blocks 1 and 2,

- 0 0 -
1
- - 2

f-
d

Figure 6.39
206 Earthquake design

where non-linear behaviour must be considered, i.e. a total width of the gap
equal to

(6.59)

where ~1 and ~2 are the maximum deflections yielded by an elastic analysis,

and R is the reduction factor. The SEAOC-88 code prescribes an amplification
factor equal to (3 Rw/8).
By comparing various prescriptions, we reached the conclusion that a glo-
bal value of the gap width can be admitted in the form

(6.60)

when ~1 and ~2 are not computed.

We have to recognize that large gaps solve the pounding problem but lead
to difficult architectural problems not yet solved satisfactorily; also, we have to
ensure that the gaps remain clear of any debris.
An alternative approach, which implies the use of gap elements consisting of
a spring and dash pot allowing vibration of adjoining blocks together as a
coupled system, is now in the stage of research (Anagnostopoulos, 1988).
Extensive research is being carried out, dealing with behaviour of adjoining
structures during pounding (Leibovich, Yankelevsky and Rutenberg, 1994).

6.5.5 EFFECT OF RESONANCE

The accelograms recorded during earthquakes display the predominant per-

iods of the ground motion (TG)' When this period is close to the natural
fundamental period of the structure (T), an amplification of displacements and
stresses develops (see Appendix AI); this phenomenon is more dangerous when
the damping ratio is low (~ = 0·01 ~0·02).
When we have sufficient earthquake recordings on a given site displaying
long period ground motions, we have to design rigid structures, i.e. buildings
relying on strong structural walls and cores; the foundations must be as rigid
as possible in order to avoid exaggerated rocking effects due to soil deforma-
bility. Analysis of seismic behaviour of multi-storey buildings during the 1977
Bucarest earthquake illustrates this phenomenon. Predominant long periods of
the ground motion were registered during the earthquake (TG = 1~ 1·5 s). Most
of the modern multi-storey buildings were built with RC structural walls and
cores, either cast in situ or with precast panels, having short natural periods
(T = 0'2~0'3 s). Their behaviour was excellent. In sharp contrast, the old multi-
storey buildings, relying on moment-resisting frames, were more flexible
(T= 0'8~ 1·5 s) and suffered heavy damage, including 30 buildings that collaps-
ed. In the case of short period ground motion, amplification of displacements
and stresses due to resonance may occur in rigid structures, but in most cases
such structures have enough resistance to avoid collapse. We have to check
 Energy approach: application to soft storeys 207

whether the foundations can avoid excessive rigid rotation of the structures
(Figure 6.40), especially in the case of soils exposed to liquefaction.

6.6 Energy approach: application to soft storeys

Most of the modern energy approaches derive from a paper published by

Housner in 1956. Further investigations are due to Akiyama (1985) and
Bertero (1989).
Housner proposed evaluating the maximum energy absorbed by a structure
in the form

E= M S;.~ (6.61 )
2
where M is the total mass of the building and Sv.~ is the velocity spectrum for a
given damping ratio ~.
Housner assumed that this energy is constant for a given type of earthquake
and a given damping ratio ~; he based this assumption on the form of the
elastic velocity spectra obtained for several earthquakes, in the usual range of
rigidities (fundamental periods). In the case of inelastic velocity spectra, usual
ductilities, damping ratios and rigidities, the assumption of constant absorbed
energy remains practically valid (Veletsos, Newmark and Chelapati, 1965;
Aribert and Brozzetti, 1985).
We shall divide the total energy E, absorbed by the structural elements, into
elastic energy (Ee) and plastic energy (Ep):
(6.62)
Housner points out that
the energy input is the same when parts of the structure are stressed beyond the elastic
limit as it would be if the structure behaved elastically ...

Figure 6.40
208 Earthquake design

This assumption holds true if

the inelastic deformations do not have a major effect on the stiffness characteristics of
the structure ...

We emphasize that the results of elasto-plastic time history analyses performed

by Clough (1970) on multi-storey frames and by Derecho et at. (1978) for RC
structural walls have confirmed this basic assumption (section 6.1.5). Finally,
we conclude that we can refer to the energy (E - Ee) as a measure of the plastic
energy dissipated by the structure.
The author proposed using an energy approach based on Housner's as-
sumption, in order to evaluate the amplification factor to be used in the design
of soft storeys (Scariat, 1994). Two 'extreme models' are considered (Fig-
ure 6.41): the first is a perfectly uniform structure with rigid beams, and the
second is a structure with rigid storeys, except the soft storey. By comparing
the elastic energies stored by both models we deduced the maximum amplifica-
tion factor C required to ensure that equal plastic energies are dissipated. This
factor results in

J(D1) (6.63)
Co =
V

where ~ is the shear at storey i and V is the base shear. The amplification
factor Co represents an upper bound, deduced by considering rigid structures
above the ground floor. A reasonable evaluation of the amplification factor c

\ :
\--...!I
, I

~
\ I
\
~
/
\ I ..c:
c:
\ I
~
I II
:t:
\ I
\ I

1
\J
\1
,I Ie Ie Ie
i ,,, ,,.
"
Figure 6.41
 Non-structural elements and non-building structures 209

can be obtained by interpolation between c = 1 (for uniform structures) and

c = CO' according to the formula
(CO - I)KGF
C=C o - (6.64)

where KGF is the lateral stiffness at ground floor and KST is the lateral stiffness
of the structure above ground floor. In the evaluation of the rigidities of RC
structural walls, in the case when the aspect ratio //H> 1/5 (H is the total
height of the wall), we have to take into consideration the effect of shear forces
on the deformations (see section 2.2.1); this can be performed by dividing the
moment of inertia of the wall by (1 + 28) where 8 = 6fE I / (GH~); f is the
'shape factor' of the cross-section of the wall.

6.7 Non-structural elements and non-building structures

6.7.1 NON-STRUCTURAL ELEMENTS

Non-structural elements are defined in this context as architectural compo-

nents (non-bearing walls, ceilings, parapets, ornamentations) and mechanical-
electrical components (such as elevators, sanitary installations and boilers).
Extensive damage to non-structural elements, accompanied by human
losses, have been repeatedly reported after major earthquakes. In some cases,
the losses in equipment were more important than the building itself.
Dynamic analysis of the behaviour of non structural elements is based on
floor response spectra. The seismic motion of the slab supporting these el-
ements is obtained by a standard dynamic analysis from a given ground
motion record; then, from the vibrations determined at the considered slab, the
response spectra of the slab can be obtained and the dynamic behaviour of the
non-structural elements attached to it analysed accordingly. Such dynamic
analyses are directly used only for very special equipment; usually, they are a
research tool for determining equivalent static forces to be considered in the
design of the non-structural elements and their connections, prescribed in
seismic codes.
The SEAOC-88 code requires that non-structural elements be checked as
subjected to a seismic force:
(6.65)
where Wp denotes the weight of the equipment and C p is a coefficient given for
various types of non-structural element. For parapets, ornamentations and
appendages, chimneys, stacks, signs and billboards, C p = 2. For non-bearing
walls and partitions, connections for precast elements and suspended ceilings,
as well as for mechanical and electrical equipment, tanks and vessels and
bookstacks, Cp = 0·75.
210 Earthquake design

The importance factor I varies between 1 and 1·5.

6.7.2 NON-BUILDING STRUCTURES

Non-building structures are designed for horizontal seismic forces computed in
the same way as for usual buildings. The formulae used for the fundamental
period T as a function of the total height H (section 6.1.4) are not applicable. It
is recommended to determine the period T by Rayleigh's method (section
6.1.4). The reduction factor Rw recommended by the SEAOC-88 code varies
between 3 and 5. When unfactored seismic loads are used (as in the European
codes), reduction factors R between 2 and 3·5 may be considered.

Bibliography
Akiyama, H. (1980) Earthquake-Resistant Limit-State design for Buildings, University of
Tokyo Press.
Anagnostopoulos, S. (1988) Pounding of buildings in series during earthquakes. Earth-
quake Engineering and Structural Dynamics, 16,443--456.
Anon (1982) The Earthquake of 4th of March 1977 in Romania, Ed. Academiei (in
Romanian).
Aoyama, H. (1981) A method for the evaluation of the seismic capacity of existing RC
buildings in Japan. Bulletin of the NZ National Society for Earthquake Engineering,
13(3),105-130.
Aribert, J. and Brozzetti, J. (1985) Comportement et concepts de dimensionnement des
constructions metalliques en zone sismique, in Genie Parasismique (ed. V.
Davidovici), Presses ENPC, Paris, ch. VII. 5.
Arnold, Ch. (1989) Architectural considerations, in The Seismic Design Handbook (ed. F.
Naeim), Van Nostrand Reinhold, New York, Ch. 5.
Bertero, V. (1989) Lessons learned from recent catastrophic earthquakes and associated
research, in Proceedings Primera CO'lferencia Internacional Torroja, Madrid.
Blume, J. (1958) Structural dynamics in earthquake resistant design. Journal of the
Structural Division of ASC E, 84 (ST 4), 1-45.
Blume, J., Knox, M. and Lindskog, C. (1958) Proposed setback provisions, Setback
Committee.
Blume, J. Newmark, N. and Corning, L. (1961) Design of Multistory RC Buildinys.
Portland Cement Association, Skokie, IL. San Francisco.
Bolt, B. (1978) Earthquake: A Primer, W. H. Freeman and Co. San Francisco.
Capra, A. and Davidovici, V. (1980) Calcul dynamique des structures en zone sismique,
Eyrolles, Paris.
Capra, A. and Souloumiac, R. (1991) Simplified seismic analysis for regular buildings, in
Recent advances in Earthquake Engineering and Structural Dynamics (cd. V.
Davidovici), Presses Academiques, Paris, Ch. IV-7.
Chopra, A. and Newmark, N. (1980) Analysis, in Design of Earthquake Resistant Struc-
tures (ed. E. Rosenblueth), Pentech Press, London, Ch. 2.
 Bibliography 211

Clough, R. (1970) Earthquake response of structures, in Earthquake Engineering (ed. R.

Weigel), Prentice-Hall, Englewood Cliffs, pp. 307-334.
Clough, R. and Penzien, J. (1992) Dynamics of Structures, McGraw-Hill, New York.
Davidovici, V. (ed.) (1985) Gene parasismique, Presses de I'Ecole Nationale des Ponts et
Chaussees, Paris.
Derecho, A., Ghosh, S., Iqbal, M., Freskakis, G. and Fintel, M. (1978) Structural walls
in earthquake-resistant buildings, dynamic analysis of isolated structural walls-par-
ametric studies. Report to the National Science Foundation, RANN, Constr. Techn.
Lab., Portland Cement Association, Skokie, IL, March.
Dowrick, D. (1987) Earthquake Resistant Design, J. Wiley, Chichester.
Eisenberg, J. (1994) Lessons of recent large earthquakes and development of concepts
and norms for structural design, in Proceedings of the 17th European Regional Earth-
quake Engineering Seminar, Technion Haifa, 1993, Balkema, pp. 29-38.
Esteva, L. and Rosenblueth, E. (1964) Espectros de temblores a distancias moderadas y
gran des. Boletin de la Sociedad Mexicana de Ingineria Sismologica, 2(1),1-18.
Fintel, M. (1991) Shear walls-an answer for seismic resistance? Construction Interna-
tional, 13 (July), 48- 53.
Fintel, M. (1994) Lessons from past earthquakes, in Proceedings of the 17th European
Regional Earthquake Engineering Seminar, Technion, Haifa, 1993, Balkema, Rotter-
dam, pp. 3- 28.
Green, N. (1987) Earthquake Resistant Building Design and Construction, Elsevier, New
York.
Gupta, A. (\990) Response Spectrum Method in Seismic Analysis and Design of Struc-
tures, Blackwell Scientific Publications, Boston.
Housner, G. (1956) Limit design of structures to resist earthquakes, in Proceedings of
the First World Conference on Earthquake Engineering, New Zealand, pp. 5.1-5.13.
Housner, G. (1970) Design spectrum, in Earthquake El1gineering, Prentice-Hall, Engle-
wood Cliffs, pp. 93-106.
Hudson, D. (1970) Ground motion measurements, in Earthquake El1gil1eeril1g, Prentice-
Hall, Englewood Cliffs, pp. 107- 126.
Key, D. (1988) Earthquake Desigl1 Practice for Buildil1gs, Thomas Telford, London.
Leibovich, E., Yankelevsky, D. and Rutenberg, A. (1994) Pounding response of adjacent
concrete slabs: an experimental study, in Proceedings of the 17th European Regional
Earthquake El1yineeril1Y Seminar, Technion, Haifa, 1993, Balkema, Rotterdam,
pp.523-538.
Luft, R. (1989) Comparison among earthquake codes. Earthquake Spectra, 5(4), 767-789.
Luong, M. (1993) Energy dissipating properties of soil, in Proceedil1Ys of Secol1d N ation-
al Earthquake El1yil1eerinfl Conference, Istanbul, pp. 607-617.
Mazilu, P. (1989) Behaviour of buildings during the 1977 and 1986 earthquakes in
Romania, Lecture at Technion, Haifa, November.
Newmark, N. (\ 970) Current trends in the seismic analysis and design of high-rise
structures, in Earthquake Enyineeril1Y (ed. R. Weigel), Prentice-Hall, Englewood
Cliffs, pp. 403 424.
Newmark, N. and Hall, W. (1982) Earthquake Spectra al1d Desifln, Earthquake Engin-
eering Research Institute, Berkeley.
Newmark, N. and Rosenblueth, E. (1971) Fundamel1tals of Earthquake Enflil1eering
Prentice-Hall, Englewood Cliffs.
212 Earthquake design

Okamoto, S. (1973) Introduction to Earthquake Engineering, University of Tokyo Press.

Paulay, T. and Priestley, M. (1992) Seismic Design of RC and Masonry Buildings, J.
Wiley & Sons, New York.
Pilkey, W. and Chang Pin Yu (1978) Modern Formulas for Statics and Dynamics,
McGraw-Hili, New York.
Pironneau, G. (1986) Seismic Joints: France-USA Workshop, CNRS, Paris.
Ravin, Sh. (1990) Why buildings are failing during earthquakes? in Israeli Association of
Earthquake Engineering Annual Conference, pp. 178-184.
Reinhorn, A., Mander, J., Bracci, J. and Kunnath, S. (1989) Seismic damage evaluation
of buildings, in Proceedings of the International Conference on Structural Saftey and
Reliability San Francisco.
Richter, Ch. (1935) An instrumental earthquake magnitude scale. Bulletin of the Seis-
mological Society of America, 25, 1-32.
Rutenberg, A. and Dickman, Y. (1993) Lateral load response of setback shear wall
buildings. Engineering Structures, 15(1) 47-54.
Scarlat, A. (1986) Introduction to Dynamics of Structures, Institute for Industrial and
Building Research, Tel Aviv (in Hebrew).
Scarlat, A. (1989) Criteres sismiques dans Ie projet du batiment Shikmona Haifa, in
2eme Coloque National Association Franraise du Gene Parisismique, St Remy-Ies-
Chevreuse, pp. C 13/2 7-15.
Scarlat, A. (1993) Soil deformability effect on rigidity-related aspects of muItistorey
buildings analysis. Structural Journal of the American Concrete Institute, 90(2),
156-162.
Scarlat, A. (1994a) Evaluation of existing buildings for seismic hazard in Israel, in
Proceedings of the 17th European Regional Earthquake Engineering Seminar, Tech-
nion, Haifa, 1993, Balkema, Rotterdam, pp. 481-498.
Scarlat, A. (1994b) Design of soft stories: an energy approach. Lecture given at the
Seminarion of the Faculty of Civil Engineering, Technion.
Taranath, B. (1988) Structural Analysis and Design of Tall Buildings, McGraw-Hili,
New York.
Veletsos, A., Newmark, N. and Chelapati, C. (1965) Deformation spectra for elastic and
elastoplastic systems subjected to ground shock and earthquake motions, in Pro-
ceedings of the Third World Conference on Earthquake Engineering, New Zealand,
pp. II 663-674.
Wakabayashi, M. (1986) Design of Earthquake-Resistant Buildings, McGraw-Hili,
New York.
Wang, T., Bertero, V. and Popov, E. (1975) Hysteretic Behaviour of RC Framed Walls,
Report EERC, University of California, Berkeley.
Wolf, J. and Skrikerud, P. (1980) Mutual pounding of adjacent structures during earth-
quakes. Nuclear Engineering and Design, 57, 253-275.
Yarar, R., Eisenberg, J. and Karadogan, F. (1993) A preliminary report on the Erzincan
earthquake, March 13 1992, in Proceedings of Second National Earthquake Engineer-
ing Conference, Istanbul, pp. 452-463.
7 Evaluation of existing
buildings for seismic hazard

7.1 Introduction

The sesimic resistance of existing buildings may be evaluated at three levels,

according to the required degree of accuracy and to the possibilities offered to
the evaluating engineer, as follows.

( a) Level I: Outside inspection (or 'sidewalk survey')

This is designed to be performed from the street, without entry to the building.
Some additional information may be provided by insurance companies.
This inspection is aimed at providing only statistical information regarding
the seismic vulnerability of a large group of buildings.

(b) Level I I: First screen in?}

This is based on visual inspection of the building, accompanied by measure-
ments of certain essential structural members at ground floor level only. We
assume that no technical documentation is readily available; some general
information can be provided by municipal archives or by direct questioning.
The first screening yields a seismic index, which quantifies the approximate
seismic resistance of the building and enables a decision to be taken as to the
need for further, more accurate investigation.

(c) Level II I: Accurate analysis

This is required when:

• the building is of special importance;

• the first screening signals that the structure is probably unsafe for poten-
tial seismic hazards.
Such an analysis is usually based on the eXisting building code, including
several specific amendments. A complete, or nearly complete, technical docu-
mentation is then required, on the basis of which we can decide whether the
building should be strengthened, and determine the main features of the re-
quired strengthening work.
214 Evaluation of existing buildings

7.2 Level I: outside inspection of existing buildings

As noted above, outside inspection can provide only statistical information. It

makes possible a very quick - but very approximate - evaluation of buildings.
Its cost is obviously low. A description of a typical outside inspection is given
in FEMA 154 (1988).
The author has proposed a similar approach, based on the main provisions
given in FEMA 154, but adapted to conditions in Israel. In the following we
shall give its main features.
The classification of the buildings is based on a structural score S:
(7.1)

where So is the basic score and I1S are modifiers. The basic score So depends
on the type of structure and the seismic zone factor Z (defined in the seismic
map of the country).
Three seismic zones are considered:

• H, high risk, Z=0'25-0'30

• M, moderate risk, Z = 0'15-0' 20
• L, low risk, Z ~ 0·10
where

Peak ground acceleration

Z = Acceleration of gravity

The proposed values of the basic score So are given in Table 7.1. Where several
types of structure are present, the predominant type will be considered. When
in doubt, the minimum basic score will be chosen.
The proposed modifiers are identical for all types of structure and all seis-
mical zones (Table 7.2). The structural score S < 1 denotes insufficient seismic
resistance. S ~ 1 denotes satisfactory resistance.

D Example 7.1

A RC frame building, located in a seismic zone Z = 0·15 (M), six storeys; poor condi-
tion; soft storey; possible pounding (adjacent slabs at same levels); year of construction
1970; soil type S 1.

Basic score: So = 1·5

Modifiers: dS = 0 (medium rise), -0' 3 (poor condition), -1 (soft storey), -0·2
(pounding), + O' 5 (year of construction), 0 (soil).
Structural score: S = 1·5 + 0 - O' 3 - 1 - O' 2 + O' 5 + 0 = O' 5 < 1. Insufficient seismic re-
sistance.
 Level I: outside inspection 215

Table 7.1 Basic score So

Type of structure Risk category
H M L
Wood frames 2·2 3 4·2
Steel moment-resisting frames 2 2·5 4·2
Braced steel frames 1·5 2 2'5
Concrete shear walls 2 2'5 3
Precast concrete large panels 1·5 2 2'5
Concrete frames I 1·5 2
Precast concrete frames 0'5 1 1·5
Reinforced masonry 1·5 1·7 2
Infilled frames 0·7 1 1·5
Plain brick/stone masonry 0·3 0·5 0·7

Table 7.2 Modifiers I'1S

Types of structure Modifiers t-..S
High-rise buildings (8 storeys or more) -0,5
Medium-rise buildings (4··7 storeys) o
Low-rise buildings (3 storeys or less) +0·3
Poor condition -0,3
Poor condition of precast concrete
structures - 0·5
Soft storey - 1
Significant eccentricity - 0·5
Pounding possible (for medium and high
rise buildings only)
Adjacent slabs at same level - 0·2
Adjacent slabs at different levels - 0'5
Heavy cladding (precast concrete or
cut stone) - 0·5
Short concrete columns - 0·5
Year of contruction
Before 1960 - 0·5
1960-1975 0
After 1975 + 0·5
Type of soil
S 1 (Rock and stiff clay) 0
S2 (Sand, gravel) - 0·2
S3 (Soft and medium soil or unknown) - 0·3
S3 + high-rise building - 004
216 Evaluation of existing buildings

7.3 Level II: first screening of existing buildings

7.3.1 GENERAL DATA

This technique, which originated in Japan (Shiga, 1977; Aoyama, 1981), is

based on statistical processing of data dealing with the behaviour of RC
structures during strong earthquakes. Subsequently, it was adopted in the
USA (Bresler et at., 1977; Hawkins, 1986), China (Chinese Academy of Build-
ing Research, 1977) and New Zealand (Glogau, 1980). In 1992, the author
proposed a similar technique, extended and adapted to conditions in Israel; it
is now being used for checking the seismic vulnerability of buildings in earth-
quake-prone areas.
The basic assumptions are the following: no technical documentation is
available, the structure is checked at ground floor level only, low-quality ma-
terials are assumed (for concrete, C 17 M Pal, and minimum steel ratios are
assumed. In a first stage, buildings with four storeys or less were considered;
subsequently, data were completed for buildings of up to 12 storeys.

7.3.2 BUILDINGS EXEMPTED FROM FIRST SCREENING

The following buildings located in areas where Z < O· 2 are exempted from first
screening:

• one-storey buildings with a light roof, where no more than five people are
usually present and no materials of special hazard or importance are
stored;
• one- or two-storey buildings for offices or dwellings where the ground floor
does not exhibit obvious features of a soft storey or large eccentricity;
• buildings up to four storeys where the perimetral structural walls are in
reinforced concrete at least 120 mm thick, the length of the RC walls on
each side of the perimeter (L) covers at least L;/2, and the RC walls reach
down to the foundations.

7.3.3 DATA NEEDED FOR COMPUTING THE SEISMIC INDEX

The seismic resistance of a building is quantified by determining a seismic

index Is, computed according to the answers to 25 questions included in a
questionnaire.
The questions refer to the following aspects:

I. General information (seismic area, year of construction, importance of

the building).
2. General description of building, including a sketch (Figure 7.1), number
of storeys, data required for determining the total weight of the building
above the ground floor.
 Level II: first screening 217

I mI I
Height: H=
I

H=
I ~ ~ IH= H=
I I
H=
No. of
storeys: n= n= n= n= n=
Figure 7.1

3. Geometric data dealing with the structural elements (Figure 7.2):

(a) For RC shear walls and cores: data required for computing the total
horizontal area of the walls Asw; walls at least 120 mm thick and 1 m
long are considered. In the case of openings with lintels less than
700 mm high (slab included), the corresponding horizontal areas are
neglected. The areas Asw in the main directions Land B are computed
separately. The walls of cores are taken into consideration in both
directions (L and B). In the case of masonry walls, similar data are
required (the horizontal areas Am for brick masonry and Asm for stone
masonry are computed). A minimum thickness of 150 mm for brick
masonry and 200 mm for stone masonry is required, as well as a
minimum length of 1 m; masonry including windows is considered as
without openings, but zones with doors are not taken into consider-
ation (Figure 7.3).
(b) For RC frames, the data for computing the sums "L(beh~) in both
directions are required (he is the size in the considered direction).
(c) For steel frames the data needed for computing the sums of the elastic
section moduli (Wx = Ix/ y max)·

7.3.4 CLASSIFICATION OF STRUCTURES

(a) Plan irregularities

Structures are classified from the point of view of their plan irregularities
according to the value of a torsional index TI defined in section 4.4 (equation
(4.15)). The classification is summarized in Table 7.3.

(b) Vertical regularity

Structures are also classified into three categories from the point of view of
their vertical regularity.
In regular structures, each storey is identical or stronger than the storey
above it, or the total horizontal area of the RC structural walls of the consider-
ed storey (usually the ground floor) Asw> 1·2Z"LA/l00 in both main direc-
tions Land B;"LA denotes the total area of the slabs above the considered
storey. For details relating to the computation of Asw' see section 2.4.
B
Masonry walls
t=0.2
iB
t",W • 11 1
20.0 [RC ~4
~ ~6.0t=OYi
20.0
RCwall

•
RCwall
t=0.3 .--1
L
R~wa311
t -0. eo.o
I

18 .<2.4.
30.0
j

L -,

Figure 7.2

·
D. M · I
. ··H· .
::::::::
.
........
. .. . .
.....

........
·
. . . · · · · ·
.::....
::::::::
. . .
......

....
Fn
...
..
········0······0·····0·
..

..
..............
:::/::::::::::::.:~<::::~:::: .:~:>:::~:
. ............................................... .
• ............... 0

........................
............................

.... e .............. 0
..

..

I I ~
~
0.80 I . 10 I
-r---t1 -+---+
·105
.. 120 600
.
+ ••
1.10 0.70

Figure 7.3
 Level II: first screening 219

Table 7.3 Classification of plan irregularities

Regular structures Symmetric or nearly symmetric in both
main directions, or asymmetric with:
TI > 1 for structures with RC structural
parallel walls in both main directions.
TI > 2 for structures with RC structural
parallel walls in one direction only.
Moderately 0.5 ~ TI ~ 1 for structures with RC
irregular structural parallel walls in both main
structures directions.
1 < TI < 2 for structures with RC
structural parallel walls in one direction
only.
Significantly All other structures.
irregular
structures
Ti = 5~]A,w d)/ L 2:A,w) where A,w is the horizontal area of RC structural
wall belonging to ~ pair of parallel walls; d is the distance between the pair of
parallel walls; L is the average horizontal dimension of the slab above
0;'
ground floor. F details see section 4.4.

Figure 7.4

A moderately irregular structure occurs when one storey is weaker than the
storey above it and

0'6ZLA 1'2ZLA
A ;:; ----,-=--
----,-=-- ;:;
100 "" SW"" 100

All other structures are classified as significantly irregular. When structural

short columns (Figure 7.4) supporting vertical loads are present, the structure
is classified as significantly irregular.
220 Evaluation of existing buildings

( c) Pounding of adjacent huildings

Classification of structures from the point of view of pounding of adjacent
buildings:

• Regular structures: the adjacent buildings have the same height, seven storeys
or less, and the corresponding slabs are located at the same elevation.
• Moderately irregular structures: the corresponding slabs of adjacent
buildings are located at the same elevation; adjacent buildings with slabs
located at different elevations have three storeys or less.
All other structures are classified as significantly irregular. In cases where the
gap between adjacent buildings at each level exceeds 0·03 Z times the corre-
sponding height, the buildings are classified as regular.

( d) Present condition
Classification of structures from the point of view of their present condition:
good, satisfactory, unsatisfactory. Criteria are given for the classification; when
cracks due to differential settlements are present the condition of the structure
is considered as unsatisfactory.

( e) Classification of precast structures

• High resistance: large panels with cast in situ horizontal and vertical joints
of a known and accepted type.
• Satisfactory resistance: large panels with joints based on welded bars only,
of a known and accepted type; RC frames with cast in situ joints of a
known and accepted type.
All other precast structures are classified as having an unsatisfactory resistance.

(j) Miscellaneous considerations

A special coefficient is reserved for unspecified aspects considered as important
by the examiner.

7.3.5 COMPUTATION OF THE SEISMIC INDEX [,

The seismic index has the form of a safety factor:

(7.2)

where V is the probable seismic force and Va is the resisting (allowable) force of
the structure.
 Level II: first screening 221

( a) Seismic Iorce
The seismic force V is determined as

V=cW (7.3)

where W is the total weight of the building above the ground floor and c is the
seismic coefficient, given by

C=C'(1 + II) (7.4)

where c·' is the basic seismic coefficient and I are modifiers.

For moment-resisting frames (RC and steel):

1·5Z
c (7.5)
Jii
For other structures:

2·5Z
c (7.6)
Jii
H is the height of the building (in m).
In the case of dual structures, we choose an intermediate value of the basic
seismic coefficient, between those yielded by equations (7.5) and (7.6). When
the ground floor is a 'soft storey' the corresponding shear force V is multiplied
by a factor f~. The basic seismic coefficients c) were determined as average
values corresponding to the provisions of the SEAOC-88 code. The 'modifiers'
I depend on the importance of the building, the type of foundation soil and on
the degree of horizontal and vertical irregularities; they vary between - 0·2
and + O' 3; Iv varies between 1 and 3.
Upper and lower bounds of the seismic coefficients were established as
follows:

In seismic areas where Z = 0·05: c = 2-10%

In seismic areas where Z = 0·30: c = 5-20%

Data for a quick computation of the total weight Ware provided.

(h) Resisting (allowahle) seismic fim'e

The resisting (allowable) seismic force

(7.7)
222 Evaluation of existing buildings

The basic allowable force V; is determined as follows (Scarlat, 1994):

• RC structural walls and cores (see section 2.4):

(7.8)

Asw denotes the total horizontal area of structural walls/cores at ground

floor level in a given direction: Asw = L Lw t.
Several restrictions imposed on the considered walls are detailed in
section 4.4.
Ta depends on the number of stories (n):

(7.9)

By assuming a vertical load of p = 10 kN m - 2 = 0·01 M Pa:

1A4
Ta = In ~ OA (MPa) (7.10)

• RC frames (see section 1.4.2):

(7.11)

where fek is the characteristic (cube) strength of the concrete (f~k = 17

MPa is assumed); he' he are the dimensions of the columns at ground floor
level (he parallel to the given forces); [; = 0·7 for regular beams and [; = 1
for slab beams; and h is the storey height.
• Steel frames (see section 1.4.3):

(7.12)

where L Wx denotes the sum of moduli of resistance of the columns at

ground floor level, and (T~ = 130 M Pa = 130000 k N m - 2.
• Masonry (see section 2.7):

(7.13)

where Am is the total horizontal area of the masonry at ground floor level
in the given direction. A number of restrictions imposed on the considered
walls are detailed in section 7.3.3.
The recommended values for Ta are as follows: plain masonry, Ta = 0·05
M Pa for solid bricks and 0·03 M Pa for hollowed bricks; infilled frames
(masonry infill), Ta = 0·2 MPa for solid bricks and 0·1 MPa for hollowed
bricks.
 Level II: first screening 223

• Reinforced masonry:

(7.14)

(see section 2.7.3)

• Stone masonry:
(7.15)
where Tu = 0.05-0.10 MPa.
The 'modifiers' a depend on the year the building was constructed, its present
condition, the type of the structure and the foundations type (in cases where
no information is available we use the minimum value). These vary between
- 0·5 and +0·2.

( (') Limit -state design

In cases where the structure includes several types of substructures we resort to
a limit-state design. The total design force takes the form

Va.T = I:J. i v" (7.16)

where the participation factors :J. are determined by considering the different
ductilities of each substructure. The following factors were chosen, by referring
mainly to data provided by Aoyama (1981) and Hawkins (1986) and to the
results of our own computations. The proposals set forth by Aoyama (1981)
are based on Newmark's equal energy criterion (see section 6.1.5): the primary
reduction factor Ro results in:

(7.19)

where the maximum ductility

<5
II ~-; (7.20)
Oy
By admitting a maximum force reduction factor for moment-resisting frames
in steel we obtain the participation factors as follows:

• Moment-resisting frames - steel: :J. = 1

-RC: :J. = 0·9
• Coupled structural wallsjcores- RC: :J. = 0·85
• Structural RC walls without openings: :J. = 0·8
• Reinforced masonry: :J. = 0·6
• Infilled frames: :J. = 0-4
• Plain brick and stone masonry: :J. = O· 3
224 Evaluation of existing buildings

7.3.6 CLASSIFICATION OF BUILDINGS

A building can be classified based on its seismic index Is into one of the
following categories:

1. Is> 1·3: a more accurate check is not required;

2. 1·1 < Is ~ 1·3: a more accurate check is not urgently required;
3. 0·9 ~ Is ~ 1·1: a more accurate check (according to the seismic code-see
section 7.4) is required;
4. Is < 0·9: a more accurate check is urgently required; strengthening of the
building is probably needed.

7.4 Level III: accurate analysis of existing buildings

As stated in the introduction, an accurate analysis of an existing building is

required either when the building is of special importance, or when the result
of the first screening is unsatisfactory.

7.4.1 EXISTING DOCUMENTATION

The ideal situation implies the existence of complete technical documentation,

including.

• 'as built' plans of all possible structural elements (including masonry

walls) and foundations with all corresponding details;
• in cases where alterations of the original structure were made, the plans
and details of these alterations;
• information regarding the quality of construction materials;
• a geotechnical report comprising data on the foundation soil.

Often, such complete documentation is not available and we have to restore

it, at least partly, by preparing schematic as-built drawings, including sketches
of the main details for the main structural elements.
The minimum data required are as follows:

• Geometric data of all possible structural elements (including masonry

walls), their dimensions and positions.
• For RC elements, reinforcement details for the main members: typical
columns, structural walls, typical beams and slabs; their identification
may be performed either by uncovering the reinforcing bars or by non-
destructive techniques (e.g. ultrasound). Concrete quality may be checked
by rebound hammer test (a minimum of 10 points per member) and, in the
case of unreliable results, by testing drilled samples.
• For steel elements, their dimensions and positions, as well as joint details.
 Level III: accurate analysis 225

• For brick masonry, whether the masonry is plain or is a part of in filled

frames; in the latter case we must check the dimensions of the RC or steel
columns and ascertain whether the brick infill is properly connected to the
main structural elements. We have to establish the type of the bricks too,
and check the quality of mortar (specifically, if the mortar can or cannot
be snapped away from the joints by hand with a metal tool).
• For precast structural elements, information regarding the dimensions
and reinforcement of the main elements and details of their joints, too.
• Information as to the nature of the foundation soil, either from the orig-
inal geotechnical report or by executing additional drillings.
• Information as to the foundations (type, depth), by special excavations or
by non-destructive techniques.
• Data as to the type of flooring, the type of facade and partitions, in order
to evaluate the permanent load.
We define two categories of documentation:
(a) good;
(b) satisfactory.

7.4.2 PRESENT CONDITION OF THE STRUCTURE

The effective capacity of the existing structure is affected by its present condi-
tion. We classify the structures from this point of view into three categories:
a. good;
b. satisfactory;
c. unsatisfactory.
The evaluating engineer will choose the proper category by taking into ac-
count the following criteria:
• age of the building (benchmark years are determined mainly according to
the publication of code revisions);
• alterations to the structural elements;
• damage due to fire;
• cracks due to differential settlements (generally diagonal cracks, visible on
both sides of the walls);
• presence of visible deterioration of concrete elements due to corrosion of
reinforcing bars and following spalling of the covering concrete; spalling
of concrete as a result of aggressive environment; visible segregation of
concrete;
• visible rusting or corrosion of steel elements;
• poor quality of mortar in masonry walls.
When the present condition of the structure is considered by the evaluating
engineer as very dangerous he will propose either strengthening or demolition
of the building or a part of it, without further examination.
226 Evaluation of existing buildings

7.4.3 PRECAST STRUCTURAL ELEMENTS

Precast elements are classified into four categories:
a. large panels with wet joints;
b. large panels with dry joints;
c. linear elements (columns and beams) with wet joints;
d. linear elements with dry joints.
The dry joints are based mainly on welded reinforcing bars or steel plates. Wet
joints comprise spliced reinforcement bars (by welding or overlapping) in-
cluded in concrete zones, having a transverse section at least equal to the
transverse section of the joining elements and a minimum length of 200 mm.
The definition of the present condition of precast structures depends mainly
on the presence of cracks or spalling of concrete in the joint zones.

7.4.4 EFFECT OF POUNDING (HAMMERING)

When adjacent buildings are separated by expansion joints that are insuffi-
ciently wide, pounding can occur, leading to possible damage and sometimes
structural failure.
Pounding is particularly dangerous in situations where the adjacent build-
ings correspond to the following definitions.
1. The floors of the adjacent buildings are not at the same elevations.
2. They have different heights, rigidities or masses.
3. At least one of the adjacent structures relies only on moment-resisting
frames, or the RC walls are not able to absorb the effect of pounding
(they are not perpendicular to the expansion joint and close to it).
In order to quantify the effect of pounding we shall refer to the following
categories.
(a) The adjacent buildings correspond to definitions 1, 2 and 3.
(b) The adjacent buildings correspond to definitions 1 and 2 or 1 and 3.
(c) The adjacent buildings do not correspond to any of the definitions 1,2,
or 3.
In cases where the gap between adjacent buildings at each floor is more than
0·03 Z times the corresponding height, the building will be considered as
belonging to category (c).

7.4.5 STRUCTURAL ANALYSIS

In principle, the structural analysis will be performed according to the
methods and specifications included in the existing seismic codes for earth-
quake design, subject to the recommendations and modifications detailed in
the present section.
 Level III: accurate analysis 227

The structural layout will include, according to the decision of the evaluat-
ing engineer, all possible structural elements (including masonry walls) or only
a part of them. Plain masonry walls or masonry infill not connected to the
main structural elements will not be taken into account for buildings with
three stories or more and for buildings in areas with moderate or high seismic
activity (the relative ground peak acceleration Z > 0,15).
Restrictions detailed in section 7.3.3 for masonry walls are valid for the
present section, too.
The force distribution among various resisting elements will be based on
elastic analysis in which soil deformability is taken into account; realistic
elastic properties of the soil should be used.
In special cases, an inelastic dynamic analysis is recommended, based on
simulated earthquakes.
A lower safety factor is usually allowed when checking existing structures,
owing to two main factors: a minimum of resistance of the structure (to
vertical loads, wind pressure and past earthquakes, even mild ones) has al-
ready been demonstrated; and the cost of strengthening an existing building is
relatively high when compared with the cost of increasing the seismic strength
of a building in the design stage.
The simplest way to consider this reduction consists in multiplying the
code-specified forces by given factors. These vary widely according to the
sources.
FEMA (1988) recommends values of 0·67 for structures with long periods
and 0·85 for structures with short periods; in Romania values of 0,50-0,70 are
recommended (Agent, 1994). We recommend a reduction factor of 0·70.
The forces F acting upon the structure are

(7.21 )

where F' represents the forces given by the code for the design of new struc-
tures, and m are 'modifiers' as follows:

• md , modifier to include the effect of the type of documentation. According

to the classification defined in section 7.4.1: category (a), md = 0; category
(b),m d =O·1.
• me' modifier to include the effect of the present condition of the structural
elements. According to the classification defined in section 7.4.3: category (a),
me = 0; category (b), me = 0'1; category (c), me = 0·2.
• mp ' modifier for structures with precast elements, only. According to the
classification defined in section 7.4.3: category (a), mp = 0; category (b),
mp = 0'1; category (c), mp = 0·2; category (d), mp = 0·3.
• m h , modifier to include the effect of pounding. According to the classifica-
tion defined at point 4: category (a), m h = 0·2; category (b), m h = 0'1; cat-
egory (c), m h = O.
228 Evaluation of existing buildings

The base shear (total seismic force) V = L F is limited by the following values:
in seismic zones where Z = 0'05, V = (0'02-0'10) W
in seismic zones where Z = 0'30, V = (0,05-0,20) W
where W is the total weight of the building.
For intermediate values of Z, we shall interpolate between the aforemen-
tioned limits.
The allowable stresses (forces) acting on the existing foundations can be
increased with respect to the allowable stresses taken into account in the
design of new structures by the following percentages: spread footings subjec-
ted to normal forces only, 50%;when subjected to normal forces and moments,
75%; piles, 50%.

7.4.6 NON-STRUCTURAL ELEMENTS

Non-structural elements that by their failure may cause loss of life or injury or
damage to equipment considered very important by the owner must be in-
cluded as a part of the overall building evaluation.
Special attention must be given to masonry walls that carry vertical loads
(slabs) and are not properly connected to the main structural members.
The checking of non-structural elements, except masonry walls, will be
performed by considering seismic forces:
(7.22)
where W is the weight of the element, and Cp = 2.
In the case of masonry walls we shall use the equations shown in section
7.3.5.

Bibliography
Agent, R. (\ 994) Guiding principles in the strengthening design of earthquake damaged
buildings in Romania and case implementations, in Proceedings of the 17th European
Regional Earthquake Engineering Seminar, Haifa, 1993, Balkema, Rotterdam, 1994,
pp.417-430.
Anon (1982) The 1977 March 4 earthquake in Romania, Ed. Acad. RSR.
Aoyama, H. (\ 981) A method for the evaluation of the seismic capacity of existing RC
buildings in Japan. Bulletin of the New Zealand National Society for Earthquake
Engineering, 13 (3), 105-130.
Bresler, B., Okada, T. and Zisling, D. (1977) Evaluation of earthquake safety and of
hazard abatement. UCB/EERC, 77/06, Earthquake Engineering Research Center,
College of Engineering, University of California, Berkeley.
Council on Tall Buildings and Urban Habitat (1978) Structural design of tall concrete
and masonry buildings, in Design of Masonry Structures (ed. A. Hendry, R. Dikkers
and A. York dale), American Society of Civil Engineers, Ch. CB-I3.
 Bibliography 229

Dowrick, D. (1987) Earthquake Resistant Design, 2nd edn. McGraw-Hill, New York.
Englekirk, R. and Hart, G. (1982,1984) Earthquake Design of Masonry Buildings,
vol. 1,2, Prentice-Hall, Englewood Cliffs.
Ergunay, O. and Gulkan, P. (1990) Earthquake vulnerability, loss and risk assessment:
National report of Turkey, Proceedings Second Workshop on Earthquake Vulnerabil-
ity, Trieste, December, pp. 1~45.
Fintel, M. (1991) Shearwalls~an answer for seismic resistance? Construction Interna-
tional, 13(July), 48~53.
Glogau, O. (1980) Low rise RC buildings of limited ductility. Bulletin of the New
Zealand National Society for Earthquake Engineering, 13 (2), 182~ 193.
Hart, G. (1989) Seismic design of masonry structures, in The Seismic Design Handbook
(ed. F. Naeim), Van Nostrand, New York, Ch. 10.
Hawkins, M. (1986) Seismic Design for Existing Structures, Seminar course manual
ACI-SCM-14, Part II, pp. 1~27.
Henpry, A. (1990) Structural Masonry, Macmillan, London.
Paulay, T. and Priestley, M. (1992) Seismic Design of Reinforced Concrete and Masonry
Buildings, J. Wiley & Sons, New York.
Scariat, A. (1993a) Asymmetric multistory structures subject to seismic loads, Contribu-
tion to the evaluation of torsional forces, in Proceedings Second National Earthquake
Engineering Conference, Istanbul, pp. 30~39.
Scariat, A. (1993b) Diagonstique preliminaire de la vulnerabilite des batiments existants
en Israel, AFPS Troisieme Colloque National, St Remy, pp. TA 57~68.
Scariat, A. (1994) Evaluation of existing buildings for seismic hazard in Israel, in
Proceedings of the 17th European Regional Earthquake Engineering Seminar, Tech-
nion, Haifa, 1993, Balkema, Rotterdam, pp. 481 ~498.
Shiga, T. (1977) Earthquake damage and the amount of walls in RC buildings, in
Proceedings Sixth World Conference on Earthquake Engineering, New Delhi,
pp. 2467~2470.
Stafford Smith, B. and Carter, C. (1969) A method of analysis for in filled frames.
Proceedings of the Institution of Civil Engineers, London, Part 2, 44, 31 ~37.
Stafford Smith, B. and Coull, A. (1991) Tall Buildings Structures: Analysis and Design, J.
Wiley & Sons, New York.
Wakabayashi, M. (1986) Design ()f Earthquake-Resistant Buildings, McGraw-Hill, New
York.
Wood, S. (1991) Performance of RC buildings during the 1985 Chile earthquake: impli-
cations for the design of structural walls. Earthquake Spectra, 7 (4), 607~638.
Postscript

1 Choice of structural solution

The choice of structural solution must take into account the seismicity of the
area. It is a nonsense to prescribe the same solutions for Scandinavia and
Armenia.
Limiting the general deformability is usually beneficial for multi-storey
buildings. It solves ipso facto several important problems: the P !l effect,
pounding, and the integrity of non-structural elements.
Buildings with more than five storeys (in high-seismicity areas) or 10 storeys
(in low-seismicity areas) relying on structural walls/cores at all levels have the
best chances of survival in major earthquakes.
Symmetric buildings are desirable, but pairs of strong structural walls in both
main directions offer a better chance of resisting significant torsional stresses.
High redundancy and the entailing 'invisible reserves of strength' are benefi-
cal. We have to think twice before separating the non-structural masonry walls
from the main structural elements in order to avoid asymmetry, so much more
that the needed details to achieve this are at best questionable. It is better to
ensure good connections with the main structural elements and to design
adequate pairs of structural walls in both main directions, able to resist tor-
sional forces.
Local stress concentrations due to vertical irregularities lead usually to
limited distress. In contrast, soft storeys are very dangerous in that they have
to dissipate most of the energy released by the ground motions; the onset of
plastic hinges in the vertical elements of soft storeys may lead to total collapse.
Pounding of adjacent blocks of the building is undesirable, but it is usually
accompanied by local distress only. The exception is adjacent structures with
slabs at different elevations. The best solution to avoid damage due to pound-
ing is to avoid joints when they are not absolutely necessary to avoid
excessive lengths or a very asymmetric configuration of the building.
Lessons from past earthquakes show that in high-seismicity areas, precast
elements in the form of large panels with 'wet connections' can perform as well
as monolithic structures.

2 Choice of structural design method

When deciding on the degree of accuracy of the structural analysis we have to

consider the poor state of the basic supporting data. It is useless, and even
 Choice of structural design method 231

ridiculous, to try to replace a lack of basic seismic data (maximum acceler-

ations, frequency content of ground motions, filtering effect of soil layers, etc.)
by sophisticated, cumbersome and time-consuming dynamic non-linear ana-
lyses. They are useful and often necessary in research but not in design (except
some very special or very important structures).
Each seismic design must be checked by approximate techniques. To this
end, first screening methods are very useful tools.
Seismic codes must be short, simple and clear. These conditions are at least
as important as being 'correct'. In this context we should remember that each
new major earthquake provides 'structural surprises', so that the term 'correct'
must be accepted with a pinch of salt. A long, exhaustive and complicated code
suffers the worst possible fate: it will be ignored.
Neglecting soil deformability in dual structures entails completely distorted
pictures of stresses when rigidity aspects are involved: it leads to exaggerated
rigidities of structural walls and cores and consequently to significant errors in
the distribution of seismic forces, in the evaluation of torsional forces and in
the assessment of stresses due to temperature changes. Analyses based on this
assumption are often an exercise in futility.
At present, we have no reliable methods to assess the 'hidden ductility' of
structural walls/cores due to soil-structure interaction during rocking vibra-
tions. The current analyses that neglect this effect are at best questionable.
The lack of sufficiently simple procedures for considering soil deformability,
as well as other factors governing the real rigidity of structural elements,
underscores the need for a new approach in seismic design, based on limit-
design criteria. In order to make such techniques available for routine design,
some aspects have to be clarified (e.g. limitation of plastic regions in structural
elements subject to seismic forces, consideration of torsional forces).
Last but not least we have to remember that earthquake engineering is yet
in its infancy. The repeated appeals to 'sound engineering judgement' are a
symptom of this fact.
Competent analyses of the effects of major earthquakes, continuous con-
frontation of existing theories with these results, as well as progress in geo-
physics, are needed to enlarge our basic data, on which we can build more
exact methods of seismic design.
The present state of the art in earthquake engineering is best described by a
Japanese saying: 'Every error is a treasure. In the discovery of imperfection lies
the chance for improvement'.
Appendix A
Structural dynamics:
main formulae

A.l Single-degree-of-freedom (SDOF) systems

A.1.1 UNDAMPED FREE VIBRATIONS

See Figure A.1. Assume initial conditions: U o = 0; Uo = O.

U= -Miii5

Solution: U = urnax sinwt = (u o/ w) sin wt.

(A.1)

(w = circular frequency in rad s - 1)

f=!!!... (A.2)
2n
if = natural frequency in Hz = cycles s - 1).
1 2n
T=~=- (A.3)
f w
(T = natural period in s cycle - 1 = Hz - 1); see Figure A.2.
Approximately (Figure A.3):

(A.4)

.fu ....
M=Wlg

Figure A.1
 Single-degree-of-freedom (SDOF) systems 233

(U w in m; T in s): Geiger's formula.

Forces of inertia:

(A.5)

A.1.2 DAMPED FREE VIBRATIONS

See Figure AA. Viscous damping:

F damp = -bu (A.6)

Figure A.2

~w+
~--- -+W
w I
I
II
I
I

FigureA.3

u bu Mu
--~

Figure A.4
234 Structural dynamics: main formulae

Damping ratio:

(A.7)

Usually: ~ = 0·02-0·10.
By assuming wJ'(l-_-~""""2) = w* ~ w:

T= 2n/w* ~ 2n/w. (A.8)

If ~ > 1: overdamping (no oscillation)-Figure A.5.

A.1.3 UNDAMPED FORCED VIBRATIONS

See Figure A.6. Exciting force:

F(t) = F o.f(t)

u = u(t) = [-Mil + F o.f(t)J6

~
~~t
Figure A.S

f u t Mil
I---~~~"'+-
I F(t)=Fof(t)
I
I
I
I
I
I
I
I
I

Figure A.6
 Single-degree-of-freedom (SDOF) systems 235

(a) Harmonic excitation

!(t)=cospt

p = circular frequency of the harmonic excitation in rad s - 1.

(A.9)

Up = I 2/ 2) cos pt;
-(p w

When p = w ... resonance (u -> (0).

(b) Arbitrary excitation: p(r) = F of(r)

U= ;:.I'(t); .1'(t) = f/(r)SinW(t-r)dr (A.IO)

in which .f (t) is Duhamel's integraL

Dynamic amplification factor:

U
J1(t) = ~ = w.1' (t) (A.II)
p

A.1.4 DAMPED FORCED VIBRATIONS

See Figure A.7.

u bu F(t)=F6f(t)
t;o+-~

FigureA.7
236 Structural dynamics: main formulae

(a) Harmonic excitation

ii + 2wu + w 2 u = ,1p cos(pt)
U = UU) + up
UU) = e-U)~'(C 1 cosw*t + C 2 sinw*t); Up = Bcos(pt - Ie)

,1
B= p
J[(1 - p2/W 2)2 + 4 p2 ~2 / w 2]
, 2pw·~
tan A = 2 2
W -p

After a short lapse of time: e - "" -->0; UU) --> O.

U:::;: up = Bcos(pt - a)

('steady state vibrations': Figure A.8).

Dynamic amplification factor:

l1(t, c) =~:::;: up = cos(pt - Ie)

- ,1p ,1p J[(1_p2/W2f+4p2~2/W2]

For cos(pt - Ie) = I (maximum):

(A.I2)
I1
m x
• = J[(1 _ p2 / W2)2 + 4p2 e/w 2]
Resonance: p = w; 11 = I /2~ (e.g. for ~ = 0·2, 11 = 2·5). (A.I3)

(b) Arbitrary excitation

F(r) = F J(r)
Assuming w* :::;: w:

u
/TranSient state

FigureA.8
 Multi-degree-of-freedom (MDOF) systems 237

Dynamic amplification factor:

~ u up
p(t, s) = ~ ~~ = w.~(t, 0 (A.14)
p p

A.2 Multi-degree-of-freedom (MDOF) systems

A2.I UNDAMPED FREE VIBRATIONS: FLEXIBILITY FORMULATION

See Figure A.9.

u = u(x, t) = ¢(x) y(t) (AIS)
Computation of circular frequencies/periods:

(Mlbll-;H
M l b21
= 0 (AI6)

Solution: WI < W 2 < ... < W n ; T = 2n/w: TI > T2 > ... > 7;..
Computation of coordinates 4\ corresponding to the mode of vibration j:

(M Ibttw; - I)¢t,+ M 2bI2W;¢2, + M 3bI3W;¢3, + ... + Mnbtnw;¢n, = 0

M Ib2IW;¢I, + (M 2b22W; - 1)¢2, + M 3b23WJ¢3, + .. , + M nb 2n wJ ¢n, = 0

M t bntW;¢t, + M 2bn2 W; ¢2, + M 3bn3 WJ ¢3, + ... + (Mnbnn w; - l)¢n, = 0

(A.I7)
We choose an arbitrary value for one of the coordinates, e.g. ¢n = 1, and
deduce accordingly ¢t ' ¢2 ... from the set of equations; thus we ~btain the
modal shapes 1,2, .. .,j, : .. , n '(Figure A 10).

Figure A.9
238 Structural dynamics: main formulae

Mode 1 Mode 2

Figure A.10

The final shape of the vibrating elastic line will be obtained by superimpos-
ing the modal shapes multiplied by the corresponding functions Y(t):

u·l = u· + u· + ... + u· + ... + u·

l\ lz l) lu

Forces:

In matrix formulation:

bIn]
b2n
(flexibility matrix)

bnn

M -- [MI M2 Mn] (mass matrix)

I ~ [' J (un;' mu',;,)

 Multi-degree-of-freedom (MDOF) systems 239

2
Figure A.11

Computation of circular frequencies Wl' W 2 , ... , Wn:

det[D- (l/W 2). /] = 0; D=F-M (A.18)

Computation of coordinates 1> j corresponding to the mode of vibration j:

(A.19)

A.2.2 UNDAMPED FREE VIBRATIONS: STIFFNESS FORMULATION

See Figure A.II.

Computation of circular frequencies w:

[(k" - M,w') (k22 -

kl2

M 2W2)
k l3

k2l k23

knl kn2 kn3

240 Structural dynamics: main formulae

Computation of shape j:

(kll - M 1 WJ)¢I, + k 12 ¢2, + k I3 ¢3, + ... + k 1n ¢n, =0

k21 + ¢1,+(k 22 - M 2w;)¢2, + k 23 ¢3, + ... + k 2n¢n,=0
(A.21)

We choose one of the coordinates, for instance ¢n = 1 and then compute ¢1'
¢2/' .... I /

(c) Matrix formulation

kll k12 kin

k21 k22 k2n
K= =0 (stiffness matrix); K· F= I

knl kn2 k nn

det(K _w 2 . M) = 0 (A.22)

Computation of coordinates ¢ii

(K - wJ M).«\>j= 0 (A.23)

A.2.3 ORTHOGONALITY OF MODES hAND j (h #j)

Displacements:

(A.24)

In matrix form:
(A.25)

Rigidities:

In matrix form:
(A.27)
 Multi-degree-of-freedom (MDOF) systems 241

Figure A.12

A.2.4 UNDAMPED FREE VIBRATIONS: UNIFORM CANTILEVERS

See Figure A.12.

Uniform mass: m = W/(g H), where Wis the total weight of the tributary area
and H is the total height.
Mode j (by neglecting shear deformations):

. .
cPj=A j [ smCjx-smhCjx+
sinCjH+sinhCjH
h
cos CjH + cos CjH
x (coshCjx+coSCjx)
l
(A.28)
2
C4 = mWj
J El

A j = an arbitrary constant.

Periods:

1·425
T=--T····
J (2j - 1) 0'

where

A.2.5 UNDAMPED FORCED VIBRATIONS

See Figure A.13. The set of equations of vibration can be 'decoupled' so that
each mode will be described by a single equation.
242 Structural dynamics: main formulae

II/W"+
~ (t)
I~
If (t)

Figure A.13

(a) Flexibility formulation

We define, by referring to the mode of vibration j:
• Generalized mass:

(A.29)

• Generalized exciting force:

Fj(t) = 4>J- f(t) = cPi,F l(t) + cP2,F 2(t) + ... }
+ cP i,FJt) + ... + cPn,F n(t) (A.30)

where f(t) is the vector of exciting forces:

Fl (t)]
F 2(t)
f(t) = : (A.3J)
[
Fn(t)
The set of equations of vibration becomes
MjYj(t) + wy Mj}j(t) = Fj(t) (A.32)

(b) StifFness f()rmulation

We define the generalized rigidity
K*j = .,j,.T.
'I'j
K·..h..
'I'j
(A.33)

The generalized rigidity and the generalized mass are related by the equation
2
K*=w
j j
M*j (A.34)

The equations of motion become

M7 ¥;(t) + K7 }j(t) = Fj(t) (A.35)
 Multi-degree-of-freedom (MDOF) systems 243

A.2.6 DAMPED FORCED VIBRATIONS

Damping forces:

F D, = b II UI + b 12 U2 + ... + b I nUn
F D, = b21 uI + b22 U2 + ... + b2n un
(A.36)

Damping force vector:

Velocity vector:

(A.37)

Damping matrix:

b ll b l2 bin

B= b 21 b 22 b 2n

Equations of motion:

(A.38)
Equationj:

Mj lj(t) + Bj l](t) + Kj Y)t) = Fj (t) (A.39)

where Bj = 2w/ j Mj is the generalized damping coefficient. The property of

decoupling of modes of vibration also remains valid for forced vibrations.
244 Structural dynamics: main formulae

A.2.7 ARBITRARY EXCITATION

Solution of the equation of motion j for an arbitrary exciting force:

1
Y(t)=-- ft F *(r)e-W}~j(t-r)sinw.(t-r)dr (AAO)
J M*w. 0 J J
J J

A.3 Seismic forces

A.3.1 SINGLE-DEGREE-OF-FREEDOM STRUCTURES

The SDOF system shown in Figure 6.18 is acted upon by a random ground
motion; the displacement U G, the velocity uG and the acceleration iiG are given,
as records of the ground motion during earthquake.

The equation of motion

(AA1)
has the solution

du 1 d§(t,~) 1
u=-=- =-
ft udr)[e-WW-r)wcosw(t-r).
dt w dt w 0

- w~sinw(t - r)e-WW-r)Jdr (AA2)

The term w~sinw(t - r)e- wW - r) is usually negligible. Hudson (1962) has shown
that we may replace in the expression for velocity U the term cosw(t - r) by
sinw(t - r). Consequently the expression for the velocity becomes

U~ tUGe-wW-r)sinW(t-r)dr (AA3)

and u~u/w
According to Newmark and Rosenblueth (1971) this approximation affects
significantly the maximum velocities only for very long periods and the maxi-
mum accelerations only for very short periods.
The response spectra result as:
• Sd' the maximum relative displacement recorded for a given earthquake
in a specific direction:
1
Sd = urnax = -5 (t, ~) (AA4)
w
 Seismic forces 245

• Sv, the maximum relative velocity:

Sv = umax = §(t, ~)max = wS d (A.45)

• Sa' the maximum total acceleration:

Sa = UTOTm" = w§(t, ~)max = w 2 Sd = wS v (A.46)

The maximum forces:

(A.47)

The velocity spectrum Sv and the acceleration spectrum Sa obtained by using

the aforementioned approximation are sometimes called the pseudo-velocity
spectrum and pseudo-acceleration spectrum.

A.3.2 MUL TJ-DEGREE-OF-FREEDOM STRUCTURES

The effective mass M'J and effective weight ~e are given by

(A.48)

The sum of the effective masses (weights) is equal to the total mass M (weight
W).
The maximum modal force Fi at storey i, corresponding to the mode j,
results in 'm"

M<jJ.
I I
I,M<jJ.
j I I J
W<jJ.
l IJ
I, W<jJ.
I lj

Fi = __--'-i_ _ Sa, = _--i--Sa, (A.49)

J md , J J

The maximum base shear (sum of forces Fi ):

lma.

(A. 50)
246 Structural dynamics: main formulae

Soilty~

~iltypell
n-----~-.~-- . . / ....... r Soil type I

f3aj l-----.----.......
~~----------------~----------~T

Figure A.14

The force Fi acting at level i can be expressed in the form:

Jma~

Fi'm" = di, l-Jm", (A.51)

where the distribution coefficient dij is

d. = MicPi, = W;cPi, (A.52)

"
"LMicPi, "LW;cPi,
i i

Usually, the codes prescribe spectra Pa versus T (Figure A.14), where Pa =

Sal aG . = SalgZ; a G denotes either the peak ground acceleration or the effec-
tive ground accelera"iion (see section 6.1.2); usually several spectra are drawn,
depending on the type of soil. In order to take into consideration the import-
ance of the building and the structural behaviour (mainly the effect of ductil-
ity), additional factors are introduced (I, lIR). The base shear becomes

v. = Wj Sal = Hje Pal I = C. we (A. 53)

} gR R }}

The seismic coefficient cj for the mode j results in

c
P ZI
- -a R' - (A. 54)
j-

Pa is taken from the given code spectrum, for the period Tj and the considered
type of soil.
As the maximum modal forces Fi ,Fi , ... , Fi do not occur simulta-
neously, their effects r are superimpo~e"a 'st~milsticalll'

r= J(ri + r~ + ... + rD (A.55)

 Bibliography 247

The number of modes h to be taken into consideration depends on the re-

quired accuracy; usually
w~ + W~ + ... W~ ~ 0·9 W (A. 56)

where W is the total weight of the structure and W~, W~, ... , W~ denote the
effective weights of the modes 1,2, ... , h.
The total seismic force (base shear) yielded by modal analysis (Vmod) is
usually less than the total seismic force yielded by the static lateral force
procedure (VsJ Modern codes (e.g. SEAOC-88) require in such cases to multi-
ply the results obtained by modal analysis by a factor intended to reduce this
gap: (0.8 ... 1.0) x (V,t/Vmod)-the so called "scaling of results".

Bibliography
Anderson, 1. (1989) Dynamic response of buildings, in The Seismic Handbook (ed. F.
Naeim), Van Nostrand Reinhold, New York, Ch.3.
Blume, 1., Newmark, N. and Corning, L. (1961) Design of Multistorey RC Buildings,
Portland Cement Association, Skokie, IL.
Capra, A. and Davidovici, V. (1980) Calcul dynamique des structures en zone sismique,
Eyrolles, Paris.
Clough, R. and Penzien, 1. (1993) Dynamics of Structures, JY1cGraw-Hill, New York.
Davidovici, V. (ed.) (1985), Genie parasismique, Presses de ('Ecole Nationale des Ponts et
Chaussees, Paris.
Dowrick, D. (1987) Earthquake Resistant Design, 1. Wiley, Chichester.
Hudson, D. (1962) Some problems in the application of spectrum techniques to strong-
motion earthquake analysis. Bulletin of the Seismological Society of America, 52 (2),
417-430.
Mazilu, P. (1968) Curs de dinamica structurilor. Fac. Constructii (in Romanian).
Newmark, N. and Rosenblueth, E. (1971) Fundamentals of Earthquake Engineering,
Prentice-Hall, Englewood Cliffs.
Scariat, A. (1986) Introduction to Dynamics oj'Structures, Institute for Industrial and
Building Research, Tel Aviv (in Hebrew).
SEAOC (1988) Seismic Committee - Structural Engineering Association of California.
Appendix B
Techniques for finite element
computations

B.l Approximate analysis of structural walls by

finite elements
The finite element technique allows the most accurate analysis of coupled
structural walls, provided a sufficiently dense mesh is used. In order to simplify
the input work the mesh has to be regular - preferably with rectangular el-
ements. It is recommended that the elements be nearly quadratic; ratios of
length/width should be in the range 1/2 to 2/1.
The accuracy is usually proportional to the number of elements, but so are
the volume of the input and the computer time. In order to assess the order of
magnitude of the number of elements to be used in the analysis of coupled
structural walls with a reasonable accuracy, we refer to the structure shown in
Figure B.l a.
We choose a network that gives results that are not significantly im-
proved by a further increase in the number of elements; we define the
corresponding analysis as 'accurate'. In the specific example displayed in
Figure B.la a total of 760 elements has been used; the lintels (2'00 x 0·80)
have been divided into eight elements, each: (4 x 0'60) x (2 x 0·40)
(Figure B.l b.)
A simplification of the analysis may be achieved in the following wa~s.

• Use of a coarser mesh, as shown in Figure B.lc, with a total of only 210
elements. The maximum differences with respect to the 'accurate' results
are 10-20%.
• Use of a coarse mesh for the entire structure except for a given zone L,
where we need more accurate results, and where we have used a denser
mesh (Figure B.ld); within this zone the results are very close to the
'accurate' results.
• Replacing the lintels by horizontal beams as defined in Figure B.le
(the replacing beams have the same geometrical characteristics as the
given lintels). Obviously, greatest accuracy is obtained when the lin-
tels are shallow; then the results are nearly identical to the 'accurate'
ones.
 Use of symmetry properties 249

B.2 Use of symmetry properties

In the case of symmetric structures subjected to symmetric or anti-symmetric
loads we can perform the computations on a segment of the real structure
only, contained between the planes of symmetry (the 'structural segment' s*).
It obviously leads to a significant decrease in the computation volume.
We must ensure that the elastic line of the considered segment S* along
the plane of symmetry preserves the shape of the elastic line of the original
structure.
To this end we have to use the following means .

• Proper symmetry supports must be introduced in the nodes lying on the

planes of symmetry. They are meant to ensure that their displacements
preserve the type of displacements imposed by the symmetry conditions
(the symmetry supports must restr~in the directions of known zero dis-
placemen ts) .
• The geometrical characteristics of the elements lying on the symmetry
plane must be adjusted in agreement with the conditions of symmetry.

By referring to the symmetric structure shown in Figure B.2a, when symmetri-

cal loads are considered, the symmetry support displayed in Figure B.2b must
be introduced (the conditions of symmetry impose zero displacements (rota-
tions) in directions X 1 and X6). When anti-symmetric loads are considered,
the symmetry support displayed in Figure B.2c fits the conditions of symmetry
(zero displacement in direction X2).
By referring to the symmetric structure shown in Figure B.3a,where the
central column lies on the plane of symmetry, we have to adjust its area A and
moment of inertia 1 in order to ensure that the elastic line of the considered
segment is in agreement with the elastic line of the original structure; this
means that the corresponding column of the considered fragment has an area
A/2 and a moment of inertia 1/2 (Figure B.3b,c).
The proper symmetry supports and the needed adjustments of the geometri-
cal characteristics for any symmetric structure are detailed in Table B.l.
An element lying on the plane of symmetry will have the following adjusted
geometrical characteristics: area A * = A /2; moment of inertia (bending)
1* = 1/2; moment of inertia (torsion) 1~ = 1t /2. We have to bear in mind that
the symmetry supports have to be added to the existing supports of the orig-
inal structure.
Consequently, the fixed ends of columns and the pinned supports of finite
elements of the structural walls remain unchanged when defining the structural
segment S*.
In the case of springs lying on the symmetry plane and symmetrical loads
we have to adjust the corresponding spring constant k: k* = k/2.
 ~

C
I'
1-
'-
"..
i-
f"
,.
'-

'"
,.

+-"- - - - O ' O £ - - - - - - ' I r

t t t t t t t t
+-~r- - - 0'0£ IX! C'! IX! C'!
I tOtNftNt

~ E
DDDDDDDDDC -t 00
N
;::

'-----------l~ Jj
t t t t' t t t t t
 I I J ' .....
,.:.
I
rt:
:if,
I
...
~

~
~

t t t , ( ""
t ~
I
~
~ -
,;
c
<3

t t t t t t t ~
252 Techniques for finite element computations

Displacement
/ /Rotation
....;'... ~----- -+X1(X4)

fs(X6)
(a)

FI2

~
F ~~
restraints:
_ X1,X6
- S·

(b)

(c)

Figure B.2

After computing the stresses by using the segment S* we must double the
resulting stresses of the elements lying on the symmetry plane, in order to
obtain the actual resulting stresses.

o Numerical example B

The space structure shown in Figure B.4a is symmetrical with respect to two planes:
XI X3 and X2 X3. The given loads are symmetrical with respect to the plane XI X3
and anti-symmetrical with respect to the plane X2- X3.
 Use of symmetry properties 253

~
...... . ~X1(X4)

X3(X6)
(a)

FI2
F ~

r1
~-~~~~~nts:
A" =A12
- ((=112
irrelevant)
s"
(b)

Restraint: X2

r1
~~2
F F
... ... (=112
.... (A" =A12
s" irrelevant)
(c)

Figure B.3

The computations can be performed on the symmetrical segment S* shown in

Figure B.4b. According to the data displayed in Table B.I the restraints are as follows:
• XI, X2, X3, X4, X5, X6: nodes 11-23 (fixed ends in the original structure: no
adjustment is needed for the considered segment S*);
• X2, X4, X6: nodes 11-411 by 100, 12-412 by 100 (symmetry with respect to XI,
X3);
• X2, X3, X 4: nodes 18-418 by 100, 23-423 by 100 (anti-symmetry with respect to
X2, X3);
• X2, X3, X4, X6: nodes 13-413 by 100 (symmetry with respect to XI, X3 and
anti-symmetry with respect to X2, X3). 0
 X3 (x6)

F ___ 421 t
F--..!26.

F--ro;
3011 1 I ~

201 1 _~I

1011 ~

21 3 311, .. ~

0:
211
!X2
21 22 (23 24 25 112
111
"J;18
16, 18

13 14
19 ,,20

1§-_X1 (b)
if 12 {3
.----
/ --~X1 (X4)

11
12 X3

6 7 8 9 10

1 2 3 4 5
(a)
Figure B.4 •. __ ~_~ .......................-o '* _
~.-~~~~
 Use of symmetry properties 255

Table B.l Symmetry supports

Plane of symmetry Symmetrical loads Antisymmetricalloads
Q)~
.... N
Q)§><:
~2'-:- XI, X3 X2, X6 XI
--><:
o...U)~
X2, X3 XI, X6 X2
N
><:
"0 •
·c :;;; XI, X3 X4 X3, X5
O~
Q)
X2, X3 X5 X3, X4
....
::l
Q)
u -u
ro ::l XI,X2 X3, X4, X5 XI, X2, X6
0,.':: XI, X3 X2, X4, X6 XI, X3, X5
U) U)

X2, X3 XI, X5, X6 X2, X3, X4

AppendixC
Methods of quantifying
the structural rigidity

Two main factors must be taken into account when we look for the order of
magnitude of the effective rigidities of structures subject to lateral loads: the
effect of soil deformability and the effect of cracking of reinforced concrete
structures. These factors strongly affect the magnitude of seismic forces and
their distribution between structural walls and frames, as well as the thermal
stresses.

C.l Quantifying the effect of soil deformability

The effect of soil deformability on the deflections/rigidities of structural walls
is very important, and overlooking it may lead to significant errors.

c.l.l SPREAD AND MAT FOUNDATIONS

(a) From a theoretical point of view, the most accurate method for determin-
ing the effect of soil deformability is the so-called interaction analysis, in which
the structure, the foundations and the surrounding soil are dealt with as a
single system (Figure C.l). The extent of the soil (the sizes L" B" H,) is chosen
so that the stresses on its periphery are negligible; consequently, the supports
along this periphery have no practical effect on the analysis. Obviously, such
an analysis has to be performed by tri-dimensional ('solid') finite elements only.
We must choose the moduli of elasticity E, (Young modulus) for normal
stresses and G, for shear stresses (or the Poisson's ratio) of the soil.
This method has several drawbacks:

• The computation involves a very high number of unknowns and, as such,

needs special computer capability.
• The validity of the usual assumption of elastic/uncracked soil elements is
rather questionable; we may partly correct the effect of this assumption by
a non-linear analysis that also includes gaps, but additional computa-
tional complications are involved.
• The soil is very non-homogeneous, and we cannot overcome this intricacy
by computational means.
 Quantifying the effect of soil deformability 257

Foundation ,
Hs
t

Bs

Ls

Figure C.1

• The recommended moduli of elasticity of the soil are uncertain, and this
strongly affects the results .
• No experimental confirmation of interaction analyses is available, es-
pecially of the frail assumptions made in a non-linear analysis; computa-
tions have shown that the corresponding rigidity of the soil may decrease
by 50% with respect to the rigidity resulting from an elastic analysis.

Consequently, when considering an elastic soil, it is advisable to consider two

values of the modulus of elasticity of the soil-a high value and a low
value - and to perform two corresponding computations; the design will be
based on the highest stresses.

(b) A simpler method to quantify the effect of soil deformability, albeit less
accurate from a theoretical point of view, is based on substituting a set of
discrete elastic springs (Figure C.2) for the continuous, deformable soil.
The spring constants are proportional to the subgrade modulus of the soil,
k, (kN m - 3) and the tributary area Ai (m 2):

(C.l)

This procedure is based on Winkler's assumption, and as such it neglects the

interaction of adjacent springs; the errors increase in the case of soft soils.

(c) If we assume that the foundation base remains plane after the soil's
elastic deformation, we also implicitly admit a linear variation of stresses.
258 Methods of quantifying the structural rigidity

V:: :. ::/£,

Figure C.2

(c) (d)

Figure C.3

Consequently, we can replace the set of elastic springs by three 'global' springs
in the centre of the foundation (Figure C3) with the constants
KV (kN m - 1) for vertical displacements;
KH (kN m - 1) for horizontal displacements;
K'r (kN m rad - 1) for rotations;

defined as follows:

L' = PI K';
u = FHI KH; (C2)
IfJ = MIK'r;
Usually, we assume k~ = k~1 = k~ = ks' although several tests have shown differ-
ent values (Barkan, 1962; SNiP 2.02.03, 1985). The magnitude of the subgrade
moduli vary with the sizes and the form of the foundation (Bowles, 1993).
 Quantifying the effect of soil deformability 259

It is worth noting that the method based on the assumption of linear elastic
stresses, which is the least accurate from a theoretical point of view, is the only
one based on tests and therefore, in fact, the most reliable. It is also consistent
with the usually accepted assumption that the design computations are per-
formed on the basis of linear elastic stresses. In this context we would like to
quote from the Seismic Design Handbook (Naeim 1989)

... arriving to a mathematical model to describe the inelastic behaviour of structures

during earthquake is a difficult task ... Procedures for utilizing inelastic spectra in the
analysis and design of multi-degree-of-freedom systems have not yet been developed to
the extent that they can be implemented in design (Mohraz and Elghadamsi, p. 77).
Great care must be exercised in soil-structure interaction analyses ... The current state
of art falls far short of modelling reality (Lew and Nissen, p. 375).

Paulay and Priestley (1992) recommend considering the columns of moment-

resisting frames modelled for seismic design as rotational springs (except for
columns supported on raft foundations or individual pads supported by short
stiff piles or by foundation walls in basements); when gravity loads are con-
sidered, fixed ends are recommended.
We consider it advisable to use sets of discrete springs for large foundations
and 'global' central springs for small foundations. In the numerical examples of
the present book two types of soil have been considered: soft soils (subgrade
modulus ks = 20000 kN m - 3) and hard soils (subgrade modulus k, = 100000
kN m - 3). Several relationships have been proposed between subgrade moduli
ks and moduli of elasticity Es. Calibration computations have shown that, by
considering elastic soil elements, subgrade moduli of 20000-30000 kN m- 3
correspond roughly to moduli of elasticity of 40000-60000 kN m- 2 , while
subgrade moduli of 80000-100000 kN m - 3 correspond roughly to moduli of
elasticity of about 200000 kN m- 2 •

C1.2 FOUNDATIONS ON PILES

Contrary to the prevailing opinion, foundations on piles do not ensure a high

degree of fixity. Significant horizontal as well as vertical deformations develop,
and we have to take them into account.

((/) Horizontal deflections

In order to compare the horizontal deformability of piles and spread footings,
computations have been performed for piles of 10 m length and footings with a
similar vertical bearing capacity, subjected to identical horizontal forces, and
the corresponding horizontal deflections were computed. We have considered,
in agreement with the Russian code SNiP 2.02.03 (1985), an equivalent soil
strip of 1·10 m for piles with 0-40 m diameter and 2'50 m for piles with 1·50 m
260 Methods of quantifying the structural rigidity

diameter. Denoting by Kp (Kr) the horizontal rigidities of the piles (spread

footings) the following results have been obtained:

Diameter of piles
(m)

0-40 1/5 to 1/2

1·50 1/2·5 to 1/1·5

This shows that, from the point of view of horizontal deflections, the piles are
much more deformable than the equivalent spread footings.

( h) Vertical settlements and rotations

In order to assess the order of magnitude of the vertical rigidity of piles, we
shall assume that, at service load, a settlement of 0·3-0·5% of the pile diameter
Dp is to be expected for reinforced concrete piles (Meyerhof, 1976; Poulos,
1980). Assuming that the piles are usually loaded at 80% of the maximum
service load only, we obtain

(C.3)

where rt. = 0·25-0-40%.

Let us consider an equivalent spread footing where the normal stress is (J.
Usually, the allowable stress is (Ja = 100-400 kN m- 2 , so that we may admit a
usual service stress (J = 80-320 kN m - 2. The corresponding settlement results
In

(C.4)

Equating (C.3) and (CA) yields the equivalent subgrade modulus:

(J
k eq = - (C.5)
s rt.D

In the case of small pile diameters (Dp = 0-40 m), we obtain k~q = 50000-
320000 kN m - 1; somewhat greater than the usual subgrade moduli obtained
for spread footings.
In the case of large pile diameters (Dp = 1·50m), we obtain
ks = 13000-85000 kN m- 3 : i.e. the same order of magnitude as for spread
footings.
Hence the order of magnitude of the piles' rigidities is rather close to the
values found for spread footings, and consequently we can accept the con-
clusions obtained for spread footings as qualitatively valid for pile founda-
tions, too.
 Quantifying the effect of cracking of RC elements 261

C.2 Quantifying the effect of cracking of RC elements

The effect of cracking in RC structures depends on the type of the structural
element, the reinforcement ratio and the stresses level. Chopra and Newmark
(1980) propose to evaluate the rigidity of RC elements by considering
an average of the moments of inertia ... between cracked and completely un-
cracked sections or as between net and gross sections ... , unless they clearly are
stressed at such low levels or that shrinkage is so limited that cracking is not
likely.
The global effect of cracking on the rigidity can be described by the ratio I e/ I g'
where I g is the moment of inertia of the uncracked element, and leis an
equivalent moment of inertia, determined so that it will yield deflections close
to the deflections of the cracked element. This ratio can be determined either
by tests or by computations using finite element techniques, which take into
account the effect of cracking. In the following we shall use data provided by
Paulay and Priestley (1992).

C2.1 MOMENT-RESISTING FRAMES

The ratio I e/ I g varies between the values

Rectangular beams: O· 3-0' 5
T and L beams: 0'25-0-45
Columns subject to high compressive strength:

(eJ =: > 0.5I ck ): 0,7-0,9

Columns subject to low compressive strength:

(eJ = 0.2fck ): 0,5-0,7
where A is the cross-sectional area of the uncracked column, and Ick denotes
the characteristic (cube) strength of the concrete. Computations based on these
values show that for the whole structure, a rough preliminary value of the
ratio lei I g ~ 0·5 can be assumed.

C2.2 STRUCTURAL RC WALLS WITHOUT OPENINGS

The proposed ratio is

Ie 100 eJ
-=-+- (MPa) (C6)
Ig I Ick y

Computations based on this formula show that a rough preliminary value of

the ratio I c/ I g ~ 0·5 can be assumed.
262 Methods of quantifying the structural rigidity

C2.3 STRUCTURAL RC COUPLED WALLS

The proposed ratio I e/ I g depends on the type of shear reinforcement of the

coupling beams.
When regular reinforcement is used (stirrups):

Ie 0.2
(C7)
Ig 1 + 3 (h b /ln}2

When diagonal reinforcement is used:

Ie OA
(C8)
Ig 1 +3(h b /ln)2

where 1m is the clear span and hb is the height of the coupling beams.
Computations based on these formulae show that rough preliminary values
of the ratio I e/ I g ~ 0·2 (regular reinforcement) and I e/ I g ~ 0·4 (diagonal rein-
forcement) can be assumed. Interesting information dealing with structural
behaviour during earthquakes can be obtained by comparing the natural per-
iods of buildings before and after major earthquakes. Ogawa and Abe (1980)
compared the periods of more than 200 buildings in Sendai, Japan, before and
after an earthquake of magnitude 6.8 and maximum ground accelerations of
0.25-0.40 g; the damage due to the earthquake led to increases in periods of
about 31 %. Similar comparisons were made in Leninakan, Armenia, before
and after the 1988 earthquake. Results cited by Eisenberg (1994) displayed
period increases of 50- 70% for three buildings and an exceptional increase of
230% for a building relying on moment-resisting frames.
By assuming an increase of about 40% of the natural period and accepting
the evaluation of seismic forces prescribed by the SEAOC-88 code we obtain a
decrease of seismic forces of about 20% during earthquakes.

C.3 Final remarks

We have to distinguish between two types of deformation analyses of struc-
tures.

(a) We aim at determining the effective behaviour of a structure or of a specific

structural element loaded by seismic forces corresponding to a given accelero-
gram (i.e. a time history analysis, where the structural response results from a
numerical integration of the equations of motion).
As was shown in section 6.1.4, comparative analyses have been performed
for a 20-storey RC moment-resisting frame and for a 20-storey RC structural
wall. Both computations reached the same conclusion: the maximum deflec-
tions computed by an elastic, linear analysis and by an elasto-plastic,
 Bibliography 263

non-linear analysis are similar. The result can be explained by the decrease in
rigidity due to cracking, a decrease that has two opposite effects: on the one
hand it increases the displacements; on the other hand, it leads to larger
periods and subsequently to a decrease in seismic forces accompanied by a
corresponding decrease in displacements. The aforementioned analyses have
shown that these opposite effects are nearly equal, and therefore the deflections
yielded by a non-linear, elasto-plastic analysis and by an elastic, linear analysis
are nearly equal.
This important conclusion allows us to use in design the simple elastic,
linear analysis.
(b) We aim at evaluating the real rigidities of various components of an RC
structure in order to obtain a correct force distribution among structural
elements. In this case, according to the definition of the rigidity, we have to
choose a given pattern of horizontal forces with an arbitrary intensity (usually
we choose either a force F = 1 concentrated at the top of the building, or a set
of triangularly distributed loads with a top load F rnax = 1) and to compute the
maximum deflection urnax by an elastic analysis; then we define the rigidity as
K = 1/ u rnax ' When we compute the rigidities of the structural elements in order
to obtain the maximum deflection u rnax ' we can use the aforementioned recom-
mendations for decreasing them due to cracking.
We note that the effects of soil deformability and cracking are interdepen-
dent: the soil deformability leads to a decrease in rigidity and seismic forces
and therefore the elTect of cracking will decrease, and similarly the cracking
leads to a decrease in rigidity and a corresponding decrease of the effect of soil
deformability. It is not allowed to take into consideration a direct cumulative
effect of soil deformability and cracking.

Bibliography
Barkan, D. (1962) Dynamics of'Bases and Foundations, McGraw-Hill, New York (trans-
lated from Russian).
Bowles, G. (1993) Foundation Analysis and Design, McGraw-Hill, New York.
Burland, 1., Butler, F. and Duncan, P. (1966) The behaviour and design of large-
diameter bored piles in soft clay, in Proceedinys of'the Symposium on Larye Bored
Piles, Institution of Civil Engineers, London, pp. 5\-72.
Chopra, A. and Newmark, N. (1980) Analysis, in Desiyn of' Earthquake Resistant Struc-
tures (cd. E. Rosenblueth), Pentech Press, London, Ch. 2.
Darragh, R. and Bell, R. (1969) Load Tests on Lony Beariny Piles, ASTM, Pub!. 444,
American Society for Testing and Materials, Philadelphia.
Eisenberg, J. (1994) Lessons of recent large earthquakes and development of concepts
and norms for structural design, in Proceedings of'the 17th European Reyional Earth-
quake Engineeriny Seminar, Technion, Haifa, 1993, Balkema, Rotterdam, pp. 29-38.
Meyerhof, G. (1976) Bearing capacity and settlements of pile foundations. Inst. Geotech-
nical Enyineeriny Dil'ision ASCE, l02(GT3), 195-228.
264 Methods of quantifying the structural rigidity

Naeim, F. (ed.) (1989) Seismic Design Handbook, Van Nostrand, New York.
Ogawa, J. and Abe, Y. (1980) The stiffness degradation caused by a severe earthquake,
in Proceedings of International Conference on Earthquake Engineering for Protection
from Natural Disasters, Bangkok, pp. 39-50.
Paulay, T. and Priestley, M. (1992) Seismic Design of Reinforced Concrete and Masonry
Buildings, 1. Wiley & Sons, New York.
Poulos, H. and Davies, A. (1980) Pile Foundation Analysis and Design, 1. Wiley, New
York.
Scarlat, A. (1985) Design for temperature changes, prejudices and reality, in Third
Symposium of the Israeli Association of Civil Engineers, Jerusalem, pp. 33. 1-20 (in
Hebrew).
Scarlat, A. (1993) Effect of soil deformability on rigidity-related aspects of multistory
buildings analysis. ACI Structural Journal, 90 (March/April), 156-162.
SEAOC (1988) Recommended lateral forces requirements and tentative commentary, Seis-
mology Committee, Structural Association of California, San Francisco.
SNiP (1985) 2.02.03-85 -Svaennie fundamenti, Moscow (in Russian), 1985.
Appendix D

Glossary of earthquake
engineering terms

accelerogram Diagram of acceleration versus time recorded during an earth-

quake (real or simulated). By integration we can obtain the corresponding
velocity and displacement diagrams.
amplification characteristics of surface layers Amplification of the amplitude
of a seismic wave when it propagates from the base rock towards the surface
layers.
aspect ratio of a wall Height/length ratio.
a!tenuation Characteristic decrease in amplitude of the seismic waves with
distance from source, due to geometric spreading of propagating wave, energy
absorption and scattering of waves.
base The level at which earthquake motions are considered to be imparted
to the structure.
base shear The shear force at the base of a structure (equal to the total
lateral force).
bearing wall An interior or exterior wall providing support for vertical loads.
bedrock A rock that is not underlain by unconsolidated materials.
braced frame A vertical truss system provided to resist lateral forces (con-
centric or excentric).
building Low rise: one to three storeys
Medium rise: four to seven storeys
High rise: eight storeys and taller
- according to classification proposed by the Federal Emergency
Management Agency (FEMA) 154/1988.
building frame system A structural system with an essentially complete space
frame providing support for vertical loads. Seismic resistance is provided by
structural walls or braced frames.
centre of mass (of a given slab) The centre of gravity of masses (structural
and non-structural) on the slab.
centre of rigidity (of a given slab) A point having the following property:
when a horizontal force passes through it, the slab displays only translational
motion (we assume that the slabs are rigid in both horizontal and vertical
planes and that the vertical structural elements are inextensible). For the gen-
eral case (deformable structural slabs and vertical elements) several definitions
have been proposed (see Chapter 4).
circular frequency The frequency multiplied by 2n (rad s - 1).
266 Glossary of earthquake engineering terms

collector An element provided to transfer lateral forces from a portion of a

structure to the vertical elements of the lateral resisting system.
concentric braced frame A braced frame in which the members are subjected
primarily to axial forces.
confined region The portion of an RC component in which the concrete is
confined by closely spaced special lateral reinforcement restraining the con-
crete in directions perpendicular to the applied stresses.
cycle The motion completed during a period.
damping Internal viscous damping: damping associated with material viscos-
ity, proportional to velocity. Body friction damping (Coulomb damping): damp-
ing taking place at connections and supports. Hysteresis damping: damping
taking place when a structure is subjected to load reversal in the inelastic range.
degree of freedom (in structural dynamics, by considering a system with
lumped masses) The number of independent parameters needed to define the
position of the lumped masses. Frequently, only the flexural deformations are
taken into account (the axial and shear deformations being neglected). Two
main types of structure are usually referred to: single-degree-of-freedom
(SDOF) and multi-degree-of-freedom (MDOF).
diaphragm A horizontal system designed to transmit lateral forces to vertical
resisting elements of the system.
dual system A structural system with an essentially complete space system to
provide support for gravity loads. The resistance to lateral loads is provided
by structural walls and moment-resisting frames. The two systems are designed
to resist the total lateral load in proportion to their relative rigidities; the
moment-resisting frames are designed to resist a prescribed percentage of the
total lateral force.
ductility Property of material, element or structure subjected to cyclic loads,
to sustain large inelastic deformations before failure. The ductility is quantified
by the ductility ratio, usually the ratio of a displacement (or curvature) at
failure, to the same displacement (or curvature) at yield point. Ductility is
usually included in reduction factors, prescribed by various codes.
duration of preliminary tremors (TpJ The time interval between the arrival at
the observation station of P and S waves. Usually, I;,s;;: 0·12 D s, where D is
the distance from the hypocentre to the observation point, in km.
dynamic analysis (elastic or inelastic) Analysis based either on modal analy-
sis or on direct integration of the equations of motion by using a step-by-step
technique; in this latter case an appropriate seismic accelerogram is considered
as input data.
eccentrically braced frame A braced frame with excentric connections, where
significant bending moments develop, besides axial forces.
eccentricity The distance between the centre of mass and the centre of
rigidity projected on a direction normal to the considered horizontal force.
epicentre (epifocus) The projection of the hypocentre onto the surface of the
earth.
 Glossary of earthquake engineering terms 267

equivalent static lateral force procedure A method that replaces seismic lat-
eral forces by static lateral forces.
exceedance probability The statistical probability that a specified level of
ground motion will be exceeded during a specified period of time.
frequency The number of complete cycles in a unit of time (Hz = cycles s - 1);
equal to 1/period.
hypocentre (focus) The point where the earthquake originates.
intensity scale The scale of ground-motion intensity as determined by
human feelings and by the effect of ground motion on structures. The most
used: M9dified Mercalli (MM), 12 grades; Medvedev~Sponheuer~Karnik
(MSK), 12 grades; Japanese Meterological Agency (JMA), 8 grades.
lateral force-resisting system A part of the structural system, assigned to
resist lateral forces.
liquefaction Phenomenon whereby a saturated sandy layer loses its shear
strength, owing to earthquake motion, and behaves like liquid mud.
limit state Serviceability limit state is reached when the building becomes
unfit for its intended use thrO\.~gh deformation, vibratory response, degradation
or other physical aspects. Ultimate limit state is reached when the building
fails, becomes unstable or loses equilibrium.
magnitude (M) A measure of the amount of energy released by an earth-
quake. According to a definition proposed by Ch. Richter (1934): M = log A,
where A (/lm) is the maximum amplitude registered by a specific seismograph,
at a point 100 km from the epicentre.
major damage Repairs would cost approximately 60% of the building's value
(land or site improvement not included)~according to proposal of Federal
Emergency Management Agency (FEMA) no.154/1988.
modal analysis Dynamic analysis of an n degree-of-freedom structure based
on the decoupling of n single-degree-of-freedom deformed shapes (modes of
vibration) and their superposition in order to reconstitute the deformed shape
of the vibrating structure.
moment-resisting frame A frame in which the members are subjected primar-
ly to bending moments.
moment ratio (of a structural wall) The ratio base moment of the wall / total
overturning moment of the structure.
non-bearing wall An interior or exterior wall that does not provide support
for vertical loads other than its own weight.
overturning moment (at a given storey) The moment of the resultant lateral
forces above the considered storey with respect to an axis in the considered
slab's plan, normal to the considered seismic forces. The corresponding axial
forces in columns and shear walls must be added to the axial forces due to
vertical forces.
P,1 effect Development of significant additional stresses due to large horizon-
tal deflections of columns; they may be obtained by second-order analysis (the
equilibrium is formulated by considering the deformed shape of the structure).
268 Glossary of earthquake engineering terms

peak ground acceleration (PGA) The horizontal maximum acceleration dur-

ing earthquakes on the free surface of a rock or a similar stiff soil, with a
prescribed probability of being exceeded within a certain time span. Recently,
the concept of effective ground acceleration (EGA) has been introduced, where
a mean of low-frequency spikes of acceleration (having a more significant
influence on the response and behaviour of structures) is considered. The
zoning map of SEAOC-88 code is based on EGA.
period The time elapsed while the motion repeats itself (s). In modal analysis
each mode has its own period; the first one (the longest) is the fundamental
period.
radius of rigidity of a storey Square root of the ratio torsional rigidity /
translational rigidity.
reduction factor A measure of the ability of a structural system to sustain
cyclic inelastic deformations without collapse. It depends, besides the ductility
ratio, on the existence of alternative lines of defence, such as lateral force
system redundancy, non-structural elements, changed damping and period
modification with deformation.
regular structure A structure not very slender and with no significant plan or
vertical discontinuities of its resisting system.
reliability of a structure (in seismic design) The probability that the structure
will survive all the actions exerted upon it by seismic forces during a given
time interval.
response spectra Curves plotting the maximum values of relative displace-
ments (Sd)' relative velocities (SJ or absolute accelerations (S.) of a single-
degree-of-freedom system versus its period, for a given damping ratio,
according to recordings of earthquakes. In fact, slight simplifying modifica-
tions are allowed in the analytical expression for the maximum relative veloc-
ity and absolute acceleration, and hence we sometimes refer to spectrum of
pseudo-velocity and spectrum of pseudo-acceleration (see Appendix A).
response spectrum analysis An elastic dynamic analysis of a structure, utiliz-
ing the peak dynamic response of all modes having a significant contribution
to the total structural response.
rotational vibration Vibration involving rotations with respect to a horizon-
tal axis (also: rocking vibration).
seismic coefficient The ratio total lateral seismic force/total vertical load.
seismic waves (main) (a) Longitudinal waves (compression waves P); (b) trans-
verse waves (shear waves S); (c) Love waves (R) and (d) Rayleigh waves (surface
waves).
seismic hazard The physical effects of an earthquake.
seismic resistant system The part of a structural system that is considered to
provide the required resistance to seismic forces in design analysis.
seismic zone factor (Z) A dimensionless number mapped or defined in build-
ing codes that is roughly proportional to the ratio of a design earthquake
acceleration to the acceleration of gravity.
 Glossary of earthquake engineering terms 269

shear ratio (of a structural wall) The ratio base shear taken by the wall/total
base shear of the structure.
soft storey A storey in which the vertical resisting elements possess a lateral
stiffness significantly lower than the corresponding stiffness of the storey
above.
soil-structure interaction Analysis that considers the given structure (founda-
tions included) and the surrounding soil as a 'complex structure' with different
elastic-mechanical properties, subject to static or dynamic loads. -
space frame A structural system composed of interconnected members other
than bearing walls, that is capable of supporting vertical loads and that will
also provide resistance to horizontal seismic forces.
storey drift The horizontal displacement of one level relative to the level
below.
storey drift ratiQ The storey drift divided by the storey height.
storey torsional moment Moment of the storey shear multiplied by the ec-
centricity.
structural wall Wall proportioned to resist combinations of shears, moments
and axial forces induced by lateral forces.
torsional rigidity of a storey The total moment of torsion due to torsional
forces developing in the vertical resistant elements, as a result of a relative unit
rotation of the slabs above and below the storey.
torsional vibration Vibration involving rotations with respect to a vertical
axis.
translational rigidity of a storey Sum of shear forces developing in the verti-
cal resisting elements in a specific direction when a relative unit displacement
is impressed in the same direction.
weak storey A storey in which the total strength of the vertical resisting
elements is significantly less than the corresponding stress of the storey above.
Appendix E
Design codes and standards

For more information about National Codes and Standards, the reader is
referred to International Handbook of Earthquake Engineering, Codes, Pro-
grams and Examples, Edited by M. Paz, Chapman & Hall, 1995, 578 pp.

China
Chinese Academy of Building Research (1977) Criteria for Evaluation of Indus-
trial and Civil Buildings, Beijing, TZ 23-77. English translation: Department of
Civil Engineering, University of Washington, September 1984.

Canada
Canadian Prestressed Concrete Institute (CPCI) (1982) Precast and Prestressed
Concrete Design. Metric Design Manual. Ottawa, Canada.

France
Association Fran~aise du Genie Parasismique (AFPS) (1990) Recommendations
AF PS-90 pour la redaction de regles relatives aux ouvrages et installations
a reali.~er dans les regions sujettes aux seismes, Presses EN PC, Paris.

Groupe de Coordination des Textes Techniques (DTU) (1982) Regles Parasis-

miques 1969, Annexes et addenda 1982. (Regles PS 69, modifiee 82) Editions
Eyrolles, Paris, France.

Greece
Ministry for Environment and Public Works (1992) Greek Code for Earth-
quake Resistant Structures, Athens.
 Design codes and standards 271

International
Comite Euro-International du Beton (CEB) (1985) Model Code for Seismic
Design of Concrete Structures, Lausanne, April, Bulletin d'Information No.
165.
Commission of the European Communities (CEC) Technical Committee 250,
SC8 (1994) Eurocode 8. Earthquake Resistant Design of Structures: Part '1,
General Rules and Rules for Buildings, ENV 1998-1-1, CEN, Berlin.

Japan
Earthquake Resistant Regulations j()r Building Structures, Part 2, 1987. English
translation in Earthquake Resistant Regulations: A world list, compiled by the
International Association for Earthquake Engineering, July 1992.

New Zealand
Ministry of Works and Development (1981) Pile Foundation Design Notes,
Civil Engineering Division, Wellington, New Zealand, CDP 812/8.
Standards Association of New Zealand (1976 and 1992) Code of Practice for
.General Structural Design and Design Loadings for Buildings, New Zealand
Standards NZS 4203, Wellington, New Zealand.

Peru
National Institute of Research and Housing Regulations (ININVI) (1985) Code
.li)r Aseismic Design, Norma E-020, Cargas, Normas Tecnicas de Edificaci6n,
Lima, Peru.

Romania
Institutul de Cercetare (1978, 1981, 1992) N ormativ Pentru Proiectarea Antiseis-
mica a Constructiilor de Locuinte, Socio-culturale, Agro-zootechnice si Indus-
triale (Code for Earthquake Resistant Design of Dwellings, Socio-Economical,
Agro-Zootechnical, and Industrial Buildings) PI00-78, PI00-81, PlOO-92,
Bucuresti (in Romanian).
272 Design codes and standards

Spain

Presidencia del Gobierno, Comision Interministerial (1974) Norma Sismorresis-

tente P.G.S.-J (1974) Decreto 3209, Madrid, Spain.

United Kingdom
British Standards Institution (BSI) (1985) Structural Vse of Concrete, London.

USA

American Association of State Highways and Transportation Officials

(AASHTO) (1983) Guide Specifications for Seismic Design, Washington, DC,
USA.
American Concrete Institute (ACI) (1988) Building Code Requirementsi()r Ma-
sonry Structures and Commentary, Detroit, Michigan, USA, ACI 530-88,
550.1-88, ASCE 5-88, 6-88.
Applied Technology Council (ATC) (1974) Model Code, Redwood City, Cali-
fornia.
Applied Technology Council (ATC) (1978) Tentative Provisionsj()r the Devel-
opment of Seismic Regulations j()r Buildings, Redwood City, California, ATC-
3-06. National Bureau of Standards, Special Publication 510.
Applied Technology Council (A TC) (1988) Rapid Visual Screening of Buildings
j()r Potential Seismic Hazards: A Handbook, Redwood City, California, ATC-
21.
Applied Technology Council (A TC) (1989) A Handbook i()r Seismic Evaluation
of Existing Buildings, Redwood City, California, ATC-n.
International Conference of Building Officials (ICBO) (1988 and 1991) Vnij(Jrm
Building Code (VBC), Whittier, CA, USA.
International Conference of Building Officials (1992) V n!j()rm Code if)r Build-
ing Construction (V BC) Appendix, Chapter J, Seismic ProvisionsfiJr Vnreini()r-
ced Masonry Bearing Walls and Buildings (Advance version), Whittier, CA,
USA.
National Center for Earthquake Engineering Research (NCEER) (1992) A Pro-
cedurefor the Seismic Evaluation of Buildings in the Central and Eastern Vnited
States, by C. Polazides and 1. Malley, State University of New York at Buffalo,
Technical Report 92-0008.
 Design codes and standards 273

National Earthquake Hazard Reduction Program (NEHRP) (1988) N EH RP

Recommended Provisions for the Development of Seismic Regulations for New
Buildings, Federal Emergency Management Agency (FEMA), Building Seismic
Safety Council, Washington, DC, FEMA-97. pp. 1, 64, 65.
National Earthquake Hazard Reduction Program (NEHRP) (1992) A Hand-
hook for Seismic Evaluation of Existing Buildings, Building Seismic Safety
Council, Washington, DC.
NAVFAC P-335 Seismic Designfor Buildings, TM 5-809-10.1982.
Structural Engineers Association of California (SEAOC) (1990) Recommended
Lateral Force Requirements and Commentary, Fifth Edition, Seismology Com-
mittee, Sacramento, California, USA.
Index

Accurate analysis (existing buildings) assessment 17-21

(Level III) 213 lintel 59
documentation 224~5 Deformability 202
effect of pounding 226 Deformed shape 17
non-structural elements 228 Degrees of freedom (DO F) 164- 5
precast structural elements 226 see also Multi-degree-of-freedom
present condition 225 system;
structural analysis 226~8 Single-degree-of-freedom system
Active zone 132 Displacement ductility 175
Allowable seismic force, see Resisting Distribution coefficients 161
force Dual systems (walls and frames)
Amplification factors 198 classical analysis 86-8, 98
Arbitrary excitation 244 lateral force distribution 88-9
Arching effect 70 Ductility concept 180
Asymmetric structures 190 Ductility ratios 175-6
ATC 1978 Model Code 197 Dynamic analysis 181
Axi-symmetric system J35~43
Eccentricity, resultant force toO
Bars system, axi-symmetric 135~43 Elastic models, for substructure
Base shear 160 interactions 94
Behaviour factors, see Reduction factors Energy absorption approaches 207-9
Bracing effects 127~9 Energy-dissipating index (EDI) 179
Buildings, classification (by seismic Equal energy criterion 178-80
index) 224 Equivalent dual structure concept 89
European codes 161, 192
Californian Code, see SEAOC-88 code Evaluation of existing buildings
CEB-85 code 123, 189,203 Level I, see Outside inspection
Centre of mass 100, 190 Level II, see First screening
Centre of rigidity 69, 86, toO, 190 Level III, see Accurate analysis
Circular frequency 182, 239
Cohesionless soils 147 Facade panels, connection forces 81
Cohesive soils 146 Factored loads 161
Compression failure mode, masonry FEMA 154/1988 code
walls 70 outside inspection 214
Continuum approach 9~ 12, 50~3 reduction factor 227
Coupled walls, see Structural walls, with FEMA 178/1989 code, allowable shear
openings stress 78
Cracking, reinforced concrete elements Finite element analysis, structural walls
261-3 248
Curvature ductility 176 Finite joint frame 50
First screening (existing buildings) (Level
Damped forced vibrations 235- 7, 243 II) 34,65-8,78-81,213,216
Damped free vibrations 233-4 dual structures subject to torsion 121-3
Damping ratio 167, 182 exempted buildings 216
Deflections limit design approach 98
 Index 275

reinforced concrete frame 34- 5 Modified Mercalli (MM) intensity scale

steel frames 36 162-3
structure classification 217-20, 224 Moment of inertia, corrected 174
Framed tube structures 129-30 Mononobe-Okabe formulae (earth
equivalent flange analysis 132- 3 pressure) 156
planar structure analysis 130- 2 Multi-degree-of-freedom (MDOF) system,
Frames with finite joints concept 48, 130 formulae 237-44, 245-7
French code 167, 168, 173, 181, 192 Multi-storey multi-bay frames, substitute
Fundamental mode 167-9 (equivalent) frame method 23-5
Fundamental period 167-71 Multi-storey one-bay frames 2
evaluation 62 continuum approach 9-12
pin-supported 14, 16
Geiger's formula (vibrations) 143, 172 zero moment procedure 2-9
Glossary of earthquake engineering
terms 265-8 Natural period 166
Non-building structures 210
Hammering, see Pounding effects Non-structural elements 209, 228
Hazard seismic coefficient 66
One-storey industrial buildings
Importance factor (occupation category) bracing 127-9
160,180 facade panels 81
In-plane discontinuities 196 Outside inspection (existing buildings)
Intensity scales 161- 3 (Level I) 213, 214
Interaction analysis 256- 7
Irregular structures, see Structural Participation factors 98, 223
irregulari ties PLl effect 203-5
Peak ground acceleration (PGA) 163
Japanese seismic code 65, 68, 98 Performance criteria 160-1
Juravski's formula 56 Pile foundations
equivalent pile length 146-9
Limit-state design 98, 223 horizontal deflections 259-60
Lin's theory (torsional forces) \04-7 lateral forces in group 152- 5
approximate analysis 110-14 torsional force distribution 121
eccentricity 192 vertical settlements 260
slab deformability 116-17 Pin-supported frames 14, 16
soil deformability 118-21 Plan irregularities 190-3, 199
Lintels Plastic deformation 202
deflections 59 Pounding effects 205-6, 220, 226
shear forces 55-9 Precast structures, classification 220, 226
Present condition categories 225
Masonry structural walls 68- 70 Proportioned frame concept 24- 5
basic allowable force 222
first screening 78-81 Rayleigh method 169, 174
infilled frames 70- I, 80 Reduction factors 160, 161, 176-80,223,
out-of-plane bending 77-8 227
plain 70, 78, 80 Reference vertical load 66
reinforced 75-6, 80 Regular structures 122, 167
Mass discontinuity 196 Reinforced concrete frame
Maximum displacement criterion 176- 7 basic allowable force 222
Modal analysis 185-8 cracking effects 261- 2
approximate methods 188-90 first screening methods 34- 5
Modes of vibration 166 rigidity 261-2
276 Index

Resisting (allowable) base shear 67, 70, 71 Seismic maps 160, 163
Resisting (allowable) bending Seismic zones 214
moment 35, 75 Set-back discontinuity 197
Resisting force 34, 221- 3 Shape factor 40, 91, 104, 174
Resonance effects 206 Shear centres 106
Response spectra 181-4, 209, 244- 5 Shear failure mode, masonry walls 70-1
Retaining walls 156-9 Shear forces, structural walls 39-42
Richter magnitude scale 161-2 Sidewalk survey, see Outside inspection
Rigidity (existing buildings) (Level J)
evaluation 263 Simplification hypotheses I
translational 192 Single-degree-of-freedom (SDOF) system
Rigidity factor 160 165, 183
Romanian code 189, 200, 227 formulae 232-7, 244-5
Rosman method (continuum Site factors 160, 164
approach) 50-3 Slab coupling 48, 59
Slabs, vertical deformability 116- 17
Soft storey concept 195-6
SEAOC-88 code energy approaches 207-9
amplification factor 206 Soil deformability
dual structures 175 coupled walls 60, 62
effect of higher modes 190 and cracking 263
fundamental period 168, 172 effect 33,42-5,94-7
guidelines 160 elastic supports analysis 43, 97
importance factor 180 fixed supports analysis 43, 95, 97
mass discontinuity 196 quantifying effects 256-60
modal analysis exponent 186 structural wall rigidity 60, 62, 175
non-structural elements 209 torsional force distribution 118 21
p~ effect 203-5
Soil factor, see Site factors
real seismic deflections 180 Spanish code 189
seismic coefficients 68, 200, 221 Spectral response, see Response spectra
seismic forces 77, 262 Spread footing, equivalent 260
seismic intensity factor 161- 3 Spring constants (soil deformability) 33,
seismic map 163 257-8
set-back 197 Stability coefficient 204
shear force 66 Static lateral force procedure 160-- 1
soil deformability 44 Steel frame, first screening methods 36
storey drift 203 Stiffness radius 146
structural irregularities 198 Storey drift 17, 21, 203, 204
torsional irregularity 192 Storey moment of tension 86
vertical seismic forces 180 Storey shear force 100
Seismic coefficient 161, 199- 202 Storey torsion moment 100, 190
Seismic deflection, real 179-80 Strength reduction factors, see Reduction
Seismic design codes 160- 1 factors
see also under individual titles Structural analysis
Seismic forces choice of design methods 230
evaluation 160-1 existing buildings 2268
horizontal 160 Structural dynamics formulae
MDOF formulae 245- 7 multi-degree-of-freedom (MDOF)
SDOF formulae 244-5 237-44
special structures 181-90 single-degree-of-freedom (SDOF)
vertical 180 232 7
Seismic index 213, 216- 7, 220- 3 Structural irregularities 122, 161
Seismic intensity factor 160, 161-3 and design 197-9
 Index 277

plan 190- 3, 199 UBC-88 code, allowable shear

vertical 193-7, 199,217,219 stress 78, 80
Structural score (building UBC-91 code, seismic design 161
classification) 214 Undamped forced vibrations 234-5,
Structural solutions, choice 230 241-2
Structural walls 38-9 Undamped free vibrations 232, 237-9,
masonry, see Masonry structural walls 239-40,241
with openings 47-65, 248, 262
without openings 39-47, 261 Vertical grids (walls and slabs)
Subgrade modulus 43 123-6
equivalent 260 Vertical irregularities 193-7, 199,
Subgrade reaction 146-7 217,219
Substitute (equivalent) frame method Vertical reactions, distribution 60
23- 5
application 25 -8
Symmetry properties, use of 249, 252 Weak storey concept 193-4
Winkler's assumption 257
Time history analysis 181, 262
Torsional index (TI) 121, 193,217
Torsional radius 192 Zero moment point (ZMP) 2
Torsional rigidity 192 non-uniform frames 5-6
Torsional shear forces 100-6 uniform frames 4- 5
approximate analysis 110-14 Zero moment procedure 2-9
 APPROXIMATE METHODS IN
STRUCTURAL SEISMIC DESIGN
Adrian S. Scarlat
In the structural deSign of buildings. approximate methods of Slructural analysis are used In the initial
selection of the structural solution, assessment of the main elements' dimensions and In chedting
the final results.

The most difficult problem in the design of multi· storey buildings Is the effect of horlzootal forces,
either seismic loads or wind pressure, and approximate methods can be very useful. Two types of
approximate method are dealt with In this book:
• methods for determining the stresses and displacements of a given structure by using
substitute strlKlUres;
• methods based on 'global parameters' (seismic coefficients, total area of structural walls. etc.).

The author has taken advantage of recent developments In computerized Structural analysiS and
finite element dna lysis to re-dppraise existing approximate techniques and to define their scope and
limits more accurately. A number of new t~hniques are proposed. Many numerical examples and
comparisons with 'accurate' methods are given.

The book is the result of many years spent by the author in structural design and teaching. While
aimed principally at design engineers. it w[IJ also be valuable for non-specialists and as a teaching
ald. especially In connection with 'first screening' of existing buildings.

Dr Adrl.tn Su.rl.tt has throughout his career taught structural analysis at faculties of civil engineering
and simultaneously worked as a design engineer and consultant to structural engineering firms and
government agencies. He Is Manager of Research and Techoology and of A. Scarlat and N. Satchl.
Consulting Engineers. Tel Aviv, and is currently Chairman of the Israeli Association of Eanhquake
Engineering.

AlSO AVAIlABLE flOM [&. fN SPaN

LttttIqUl.ke lulsl....1 Contrtte SlnItlurn

G.G. Ptnelb and A.J. ~
Hardbad. {O 41 9 181200).600 p.lge5

Theory.and DnIp of Sebmk laiSI....1 Sled ftMICS

f.M. Man<>lanl and V. Piluso
Hardbatk (O 419 18760 X). S44 p.!geS

TIwe front cover shows the ur park of Northridge. CalIfornia

Stale UniYefslty. after the Lo~ Angeles e.mtlquake 01
17 January 1994. (Photo: Courtesy 01 Profe$SOf A.
Rutenberg.) ISBN 0-419-18750- 2

E & FN SPON
Ivo ........ d ChoprwI " Hot 9 )
London · ~ . WeMeim New Yor1r. · Tokyo · MeIbcMn& . Madru