SIMPLE AND COMPLEX MODELS FOR NONLINEAR SEISMIC RESPONSE

OF REINFORCED CONCRETE STRUCTURES

by
t~hdiSaii di
and
Mete A. Sozen

A Report to the
NATIONAL SCIENCE FOUNDATION
Research Grant PFR-78-16318

Any opinions, findings, conclusions

or recommendations expressed in this
publication are those of the author(s)
and do not necessarily reflect the views
of the National Science Foundation.

University of Illinois
at Urbana-Champaign
Urbana, Illinois
August, 1979
 iii
ACKNOWLEDGMENT

The study presented in this report was part of a continuing experi-

mental and analytical study of the earthquake response of reinforced
concrete structures, conducted at the/Civil Engineering Department of the
University of Illinois, Urbana. The research was sponsored by the
National Science Foundation under grant PFR-78-16318.
The writers are indebted to the panel of consultants for their advice
and criticism. Members of the panel were M.H. Eligatorof Weiskopf and
Pickworth, A.E. Fiorato of the Portland Cement Association, W.D. Holmes
of Rutherford and Chekene, R.G. Johnston of Brandow and Johnston, J. Lefter
of Veterans Administration, W.P. Moore, Jr. of Walter P. Moore and Associates,
and A. Walser of Sargent and Lundy Engineers.
Special thanks are due to D.P. Abrams and H. Cecen, former research
assistants, and J.P. Moehle, research assistant at the University of Illinois,
for providing the writers with the results of their experimental studies and
for their criticism of the computed results.
Mrs. Sara Cheely, Mrs. Laura Goode, Mrs. Patricia Lane, and Miss Mary
Prus are thanked for typing this report.
The IBM-360/75 and CYBER 175 computing systems of the Digital Computer
Laboratories of the University of Illinois were used for the development
and testing of the analytical models.
This report was based on a doctoral dissertation by M. Saiidi directed
by M.A. Sozen.
 iv

TABLE OF CONTENTS

CHAPTER Page
.1 INTRODUCTION . . . . . . . . . . . 1

1.1 Object and Scope . . . . . . 1

1.2 Review of Previous Research. 2
1.3 Notation 5

2 ANALYTICAL MODEL . 9

2. 1 Introductory Remarks .. . 9
2.2 Assumptions about Structures and Base Motions . . 9
2.3 Force-Deformation Relationship 11
2.4 Element Stiffness Matrix . . 16
2.5 Structural Stiffness Matrix. 18
2.6 Mass Matri x . . . . . . . . . . . . . . · 19
2.7 Damping Matrix . • . . . . . 20
2.8 Unbalanced Forces . . . . . . · . . . 20
2.9 Gravi ty Effect . . . . . 21
2.10 Differential Equation of Motion . 22
2.11 Solution Technique 23
3 HYSTERESIS MODELS ..·· .... 25
3.1 Introductory Remarks 25
3.2 General Comments 25
3.3 Takeda Hysteresis Model· 26
3.4 Sina Hysteresis Model . 27
3.5 Otani Hysteresis Model 29
3.6 Simple Bilinear Model · ···· 30
3.7 Q-Hyst Model . . · · · . ··.· 30
4 TEST STRUCTURES AND ANALYTICAL STUDY USING MDOF MODEL 32
4.1 Introductory Remarks 32
4.2
4.3
Test Structures . · ·
Dynamic Tests . . · · · ··.· 32
33
4.4 Analytical Procedure 33
Analytical Study · · ···· 35
4.5 .
5 COMPARISON OF MEASURED RESPONSE WITH RESULTS CALCULATED
USING THE MDOF MODEL . · · . .. ..... ·· ·· 37
5.1 Introductory Remarks . . . . 37
5.2 Calculated Response of MF2 37
5.3 Calculated Response of MFl · . 38
5.4 Concluding Remarks . . . . 42
 v

Page
6 DEVELOPMENT OF THE Q-MODEL 44
6.1 Introductory Remarks 44
6.2 General Comments . . . 44
6.3 Q-Model . . . . . . . 45
7 ANALYTICAL STUDY USING THE Q-MODEL 50
7.1 Introductory Remarks . . . . 50
7.2 Structures and Motions . . . 51
7.3 Equivalent System. . . . . . . . . . . . . 52
7.4 Analytical Results for Different Structures. . 54
7.5 Analytical Results for Different Base Motions 56
7.6 Analytical Results for Repeated Motions. 59
8 SUMMARY AND CONCLUSIONS 62
8.1 Summary. . . 62
8.2 Observati.ons 64
8.3 Conclusions 65
LIST OF REFERENCES 68
APPENDIX
A HYSTERESIS MODELS 158
A.l General . . . 158
A.2 Definitions .. 158
A.3 Sina Model 158
A.4 Q-Hyst Model 161
B COMPUTER PROGRAr~S LARZ AND PLARZ . . 165
C MAXIMUM ELEMENT RESPONSE BASED ON DIFFERENT HYSTERESIS
~10DELS. . . . . . . . . . . . . . . 171

D COMPUTER PROGRAMS LARZAK AND PLARZK . . . . . . . . . 178

E MOMENTS AND DUCTILITIES FOR STRUCTURE MFl SUBJECTED TO

DIFFERENT EARTHQUAKES . . . . . . . . 180
F RESPONSE TO TAFT AND EL CENTRO RECORDS . . . . . . . . . 184
 vi

LIST OF TABLES

Table Page
4. 1 LONGITUDINAL REINFORCING SCHEDULES FOR ~1Fl AND MF2 . . 72
4.2 ASSUMED MATERIAL PROPERTIES FOR t1Fl AND NF2 73
4.3 COLUMN AXIAL FORCES DUE TO DEAD LOAD. . . . 74
4.4 CALCULATED STI FFNESS PROPERTIES OF CONSTITUENT ELEMENTS OF
STRUCTURES MFl AND MF2 . . . . . . . . . . . . . . . . . 75
4.5 CRACK-CLOSING MOMENTS USED FOR SINA HYSTERESIS MODEL. 77
4.6 MEASURED AND CALCULATED MAXIMUM RESPONSE OF MF2 RUN 1 78
4.7 MEASURED AND CALCULATED MAXIMUM RESPONSE OF MFl USING
DIFFERENT HYSTERESIS SYSTEMS . . . . . . . . . . . . 79
7.1 COLUMN AXIAL FORCES FOR STRUCTURES Hl, FWl, AND FW2 . 81
7.2 CALCULATED STIFFNESS PROPERTIES OF CONSTITUENT ELEr1ENTS OF
STRUCTURE H1 . . . . . . . . . . . . . . . . . . . . . . .. 82
7.3 CALCULATED STIFFNESS PROPERTIES OF CONSTITUENT ELEMENTS OF
STRUCTURE FWl . . . . . . . . . . . . . . . . . . . . . . . 83
7.4 CALCULATED STIFFNESS PROPERTIES OF CONSTITUENT ELEMENTS OF
STRUCTURE FW2 .. . . . . . . . . . . . . . . . . . . . 84
7.5 CALCULATED PARAMETERS FOR DIFFERENT STRUCTURES. 85
7.6 ASSUMED DEFORMED SHAPES FOR DIFFERENT STRUCTURES 86
7.7 MAXIMUM ABSOLUTE VALUES OF RESPONSE . . . . . . 87
7.8 MAXIMUM RESPONSE OF STRUCTURE MFl SUBJECTED TO DIFFERENT
EARTHQUAKES . . . . . . . . . . . . . . . . . . . . . . 89
7.9 MAXIMUM TOP-LEVEL DISPLACEMENTS FOR STRUCTURE MFl SUBJECTED
TO REPEATED MOTIONS . . . . . . . . . 90
7.10 WIRE GAGE CROSS-SECnONAL PROPERTIES . . . . . . . . . . 90
C.l MAXIMUM RESPONSE OF STRUCTURE MFl BASED ON TAKEDA MODEL 173
C.2 MAXIMUM RESPONSE OF STRUCTURE MFl BASED ON SINA MODEL. 174
C.3 MAXIMUM RESPONSE OF STRUCTURE MFl BASED ON OTANI MODEL. 175
C.4 MAXIMUM RESPONSE OF STRUCTURE MFl BASED ON BILINEAR MODEL . 176
 vii

Table Page
Deleted see
C.5 MAXIMUM RESPONSE OF STRUCTURE MFl BASED ON Q-HYST MODEL 177 page 172.a

E.1 MAXIMUM RESPONSE OF STRUCTURE MF1 SUBJECTED TO ORION .. 181 Deleted

page 1BO.a
E.2 MAXIMUM RESPONSE OF STRUCTURE MF1 SUBJECTED TO CASTAIC. 182
E.3 MAXIMUM RESPONSE OF STRUCTURE MF1 SUBJECTED TO BUCAREST 183
 viii

LIST OF FIGURES

Figure Page
2.1 Idealized Stress-Strain Curve for Concrete. 91
2.2 Idealized Stress-Strain Curve for Steel 91
2.3 Idealized Moment-Curvature Diagram for a Member 92
2.4 Moment and Rotation along a Member. 92
2.5 Moment-Rotation Diagram for a Member. 93
2.6 Rotation due to Bond Slip . . . 93
2.7 Deformed Shape of a Beam Member 94
2.8 Equilibrium of a Rigid-End Portion. 94
2.9 Deformed Shape of a Column Member 94
2.10 Biased Curve in Relation to the Specified Force-Deformation
Di agram . . . . . . . . . . . . . . . . . . . . 95
2.11 Treatment of Residual Forces in the Analysis. . . . . . 95
2.12 Equivalent Lateral Load to Account for Gravity Effect 96
3. 1 Takeda Hysteresis Model 97
3.2 Small Amplitude Loop in Takeda Model . 98
3.3 Comparison of Average Stiffness with and without Pinching
for Small Amplitudes. 98
3.4 Sina Hysteresis Model 99
3.5 Otani Hysteresis Model . 100
3.6 Simple Bilinear Hysteresis System 100
3. 7 Q-Hys t Mode 1 . . . . 101
4. 1 Reinforcement Detai 1 and Dimensi ons of Structures MFl and r,1F2 102
4.2 Test Setup for Structure MF1 103
4.3 Measured and Calculated Response for MF2 . 104
4.4 Measured and Calculated Response for MF1 Using Takeda Hys-
teresis Model . . . . . • . . . . . . . . . • . . . . . . . 107
 ix

Ifi gure Page

4.5 Measured and Calculated Response for MFl Using Sina Hys-
teresis Model. . . . . . . . . . . . . . . . . . . . 110
4.6 Measured and Calculated Response for MFl Using Otani
Hysteresis Model . . . . . . . . . . . . . . . . . . 113
4.7 Measured and Calculated Response for MFl Using Bilinear
Hysteresis Model . . . . . . . . . . . . . . . . 116
4.8 Measured and Calculated Response for MFl Using Q-Hyst Model. 119
5.1 Maximum Calculated and Measured Displacements (Single Ampli-
tude) for MF2 • . . . . . . . . . . . . . . . . . . . '.' . . 122
5.2 Maximum Calculated and Measured Relative Story Displacements
for MF2 . . . . . . . . . . . . . . . . . . . 122
5.3 Maximum Calculated and Measured Displacements (Single Ampli-
tude) for MFl . . . . . . . . . . . . . . . . . . . . . . . . 123
5.4 Maximum Displacements Normalized with Respect to Top Level
Di s placement . . . . . . . . . . • . . . . . . . . . . 124
5.5 Maximum Calculated (Using Takeda Model) and Measured Relative
Story Displacements . . . . . . . . . . . . . . . . . . . . . 125
5.6 Maximum Calculated (Using Sina Model) and Measured Relative
Story Displacements . . . . . . . . . . . . . . . . . . . . . 126
5.7 Maximum Calculated (Using Otani Model) and Measured Relative
Story Displacements . . . . . . . . . . . . . . . . . . . . . 126
5.8 Maximum Calculated (Using Bilinear Model) and Measured Rel-
ative Story Displacements. . . . . . . . . . . . . . . . . 127
5.9 Maximum Calculated (Using Q-Hyst Model) and Measured Relative
Story Displacements. 127
6. 1 The Q-Model . . . . . 128
6.2 Static Lateral Loads 129
6.3 Force-Displacement Relationships 130
7.1 Longitudinal Reinforcement Distribution for Structures Hl and
H2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

7.2 Longitudinal Reinforcement Distribution for Structures FWl

and FW2 . . . . . . . . . . . . . . . . . 132
7.3 Normalized Moment-Displacement Diagrams. 134
 x

Figure Page
7.4 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure Hl . . . . . . . . . . . . . . . . . . . . . . 135
7.5 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure H2 . • • . . . . • . . . . . . . . . .. 136
7.6 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure MFl . . • . . . . . . . . . . • . . .. 137
7.7 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure MF2 . . . . . . . . . . • . . . . . .. 138
7.8 Calculated (Solid Line) and Measured (Broken Line) Response
for Struct ure FW 1 . . . • . . . . . . . . . . . . . . . . . 139
7.9 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure Fl~2 . . . . . . . • . . . . . . • . .. 140
7.10 Calculated (Solid Line) and Measured (B.roken Line) Response
for Structure FW3 . . . • . . . . • . . • . • . .. 141
7.11 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure FW4 . . . . . . . . . . . . 142
7.12 Maximum Response of Structure Hl . 143
7.13 Maximum Response of Structure H2 . 143
7.14 Maximum Response of Structure MFl 144
7.15 Maximum Response of Structure MF2 144
7.16 Maximum Response of Structure FWl 145
7.17 Maximum Response of Structure FW2 145
7.18 Maximum Response of Structure FW3 ... 146
7.19 Maximum Response of Structure FW4 146
7.20 Q-Model (Solid Line) and MooF Model (Broken Line) Results
for Orion Earthquake. . . . . . . . . . . . . . . . . . . . 147
7.21 Q-Mode1 (Solid Line) and MooF Model (Broken Line) Results
for Castaic Earthquake. . . . . . . . . . . . . . . . . . . 148
7.22 Q-Model (Solid Line) and MooF Model (Broken Line) Results
for Bucarest Earthquake . . . . . . . . . . . . . . . . . . 149
7.23 Q-Model (with Increased Frequency; Solid Line) and MooF
('v1odel (Broken Line) Results for Bucarest Earthquake . . 150
 xi

Figure Page
7.24 Maximum Response for Orion Earthquake .. . . . . 151
7.25 Maximum Response for Castaic Earthquake 151
7.26 Maximum Response for Bucarest Earthquake 152
7.27 Repeated Earthquakes with 0.29 Maximum Acceleration . 153
7.28 Repeated Earthquakes with 0.4g Maximum Acceleration. 154
7.29 Repeated Earthquakes with 0.8g Maximum Acceleration 155
7.30 Repeated Earthquakes wi th l.2g Maximum Acceleration. 156
7.31 Repeated Earthquakes with 1.6g Maximum Acceleration 157
A. 1 Sina Hysteresis Rules 163
A.2 Q-Hyst Model 164
B.l Structure with Missing Elements 167
B.2 Block Diagram of Program LARZ . 168
B.3 Storage of Structural Stiffness Matrix 169
a&b
B.3c Storage of Submatrix K22 . . . . . . 170
C.l Element Numbering for Structure MFl . 172

D.l Block Diagram of Program LARZAK . . . 179

F.l Response for Structure MFl Subjected to El Centro NS 185

F.2 Response for Structure MFl Subjected to E1 Centro EW 186

F.3 Response for Structure MFl Subjected to Taft N21E . 187
F.4 Response for Structure MFl Subjected to Taft S69E . 188
 1

CHAPTER 1
INTRODUCTION

1.1 Object and Scope

The primary objective of the work reported was to study the possi-
bilities of simplifying the nonlinear analysis of reinforced concrete
structures subjected to severe earthquake motions. The study included
two distinct parts. One was a microscopic study of one of the particular
elements of the analysis, the hysteresis model, and development of simple
models leading to acceptable results. The other was a macroscopic
study which included the development of a simple model that, \'!ith
relatively small effort, resulted in a reasonably close estimate of
nonlinear response.
The first part was a continuation of the investigation initiated
by Otani (26). For this part, a multi-degree nonlinear model (LARZ) was
developed to analyze rectangular reinforced concrete frames for given
base acceleration records. A special feature of LARZ was that it was
capable of accepting a collection of hysteresis systems, some previously
used and others developed in the course of the present investigation. The
new systems were generally simpler. Chapters Two through Five describe
part one of the study.
In the second part, a given structure was viewed as a single-degree-
of-freedom system which recognized stiffness changes due to the nonlinearity
of material. The model is introduced and exa~;ned in Chapters Six and
Seven, respectively.
In both parts of this study. to evaluate the reliability of the
analytical models. the calculated responses were compared ".,ith the
 2

results of dynamic experiments on a group of small-scale ten-story rein-

forced concrete frames and frame-walls tested on the University of Illinois
Earthquake Simulator.

1.2 Review of Previous Research

Several investigators have studied the nonlinear modeling of
structures subjected to earthquake motions. The development of high
speed digital computers and the availability of numerical techniques
have had a substantial contribution to the ease of carrying such studies.
A comprehensive survey of earlier investigations in the area of
nonlinear analysis of plane frames is provided by Otani (26). Here, a
brief history of more recent studies will be cited in two sections:

a. Complex Models
In a complex model, there is a one-to-one correspondence between
the elements of an actual structure and the idealized system. The
choice of idealizing assumptions to represent structural members is a
crucial one in terms of computational effort and ease of formulating
stiffness variations. Giberson studied the possibility of using a
one-component element model with two concentrated flexural springs at the
ends, and compared the results with the calculated response using a
two-component element model (13). The inelastic deformation of a member
was assigned to member ends in the former model. It was found that the
one-component element was a more efficient model and it resulted in
better stiffness characteristics.
Due to relative simplicity, the one-component model attracted
considerable attention. Suko and Adams used this model to study a
mUltistory steel frame (33). To determine the location of the inflection
 3

point of each member, a preliminary analysis had to be done. Then the

points were assumed to remain stationary for the entire analysis.
Otani used the one-component model to analyze reinforced concrete
frames subjected to base accelerations (26). The point of contraf1exure
for each member was assumed to be fixed at the mid-length of that element.
The analytical results were compared with the results of tests on small-
-scale specimens. The one-component model was also used by Umemura et
a1 (38), Takayanagi and Schnobrich (35), and Emori and Schnobrich (11).
The force-deformation function assigned to a member can have a
significant influence on the calculated response. The more dominant the
inelastic deformations are, the more sensitive is the response to the
hysteresis model used. Therefore, as the research in nonlinear analysis
was continued, more attention was paid to the stiffness variation of
members. The trend was toward the establishment of more realistic hysteresis
functions.
Through several experimental works on reinforced concrete beam-to-
column connections, it was realized that the behavior of a reinforced
concrete member under cyclic loading is relatively complicated, and that
it is not accurate to represent such behavior by a simple bilinear
hysteresis function. Clough and Johnston introduced and applied a
degrading model which considered reduction of stiffness at load-reversals
stages (9).
Takeda examined the experimental results from cyclic loading of a
series of reinforced concrete connections, and proposed a hysteresis
model which was in agreement with the test results (36). This model,
known as the "Takeda ~'ode1 ," was capable of handling different possi-
bilities of unloading and loading at different stages. To accomplish
 4

this task, the model was expectedly complicated. Several investigators

have used the "Takeda t10del" in its original or modified form, and have
concluded that the model represents well the behavior of a reinforced
concrete connection in a frame subjected to ground motions (11,26,35).
The Takeda model did not include the 'pinching effect~' which are
observed in many experimental results (18). Takayanagi and Schnobrich
considered the pinching action in developing a modified version of
Takeda model (35). Later, Emori and Schnobrich used a cubic function
to include bar slip effects (11). In both models, the rules for the
first quarter of loading were the same with those of Takeda model.
Other more involved systems were constructed by superposing a set
of springs with different yield levels. In such systems, the hysteresis
function for an individual spring is a simple relationship, however,
because each spring yields at a different moment, the overall stiffness
of a member changes continuously. Pique examined the multispring model
to determine its influence on the calculated response (30).
Anderson and Townsend conducted a study on nonlinear analysis of
a ten-story frame using four different hysteresis systems. The models
included bilinear and trilinear hysteresis systems (3).

b. Simple Models
Despite the development of sophisticated and efficient digital
computers, complex nonlinear models for seismic analysis of structures
are involved and costly. Therefore, they impose a limit on the number
of alternative configurations and/or ground motions which may be desirable
to study, before the final design of a structure is made. As a result,
several studies have been aimed at finding less complicated nonlinear
models.
 5

Among the earlier work was shear beam representation of structures.

The stiffness of each story was assigned to a shear spring which included
nonlinear deformations. Aziz used a shear-beam model in the study of
ten-story frames, and compared the results with those obtained from
complex models (6). It was found that the maxima were in reasonable
agreement. A modified shear-beam model was introduced by Aoyama for
.reinforced concrete structures (4)~ Tansirikongkol and Pecknold used a
bilinear shear model for approximate modal analysis of structures (37).
Pique developed an equivalent single-degree-of-freedom model
assuming that structures deform according to their first mode shapes (30).
Three different structures with different number of stories were analyzed,
and the maxima were compared with the results of the shear-beam and
complex models. Reasonable agreement was observed between the maximum
response obtained from the single-degree system on one hand, and the
maxima obtained from shear-beam and complex models on the other hand.
1. 3 Nota ti on
The symbols used in this report are defined where they first appear.
A list of symbols are given below for convenient reference.

As = area of steel
[C] = damping matrix
omax = maximum deformation attained in loading direction
D(y) = yield deformation
db = diameter of the tensile and compressive reinforcement
d~d' = distance between tensile and compressive bars
E = modulus of elasticity
Es = modulus of elasticity of steel
 6

e = steel elongation
Fr = external force at level r
F = total external force
t
f = flexibility of rotational spring
f = stress of concrete
c
f' = measured compressive strength of concrete
c
f = steel stress
s
f sy = yield stress for steel
g = gravity acceleration

hr = height at 1evel r
I = moment of inertia
j = number of levels in the original system
K = stiffness of the original system
[K] = instantaneous stiffness matrix
Leq = equivalent height
f = total length of a member
f' = length of elastic portion of a member
fa = anchorage length
[M] mass matrix
M = cracking moment
c
equivalent mass
e =
M

M ::: mass at nth degree of freedom

n
t1 = mass at level r
r
Mt = total mass of the original system
r1 u = ultimate moment
ty = yield moment
 7
fl M= moment increment at member end
flM' = moment increment at end of the elastic portion
P.1 = total vertical load at level i
Q = restoring force
Scy 1 = slope of the line connecting yield point to cracking point
in the opposite direction
Sl = slope of unloading for post-yielding segment
T = time
flt = time interval for numerical integration
u = average bond stress
Vi = shear force due to gravity load at level i
Xmax = maximum residual deformation previously attained
{X} = displacement vector
Xg = ground acceleration
{flX}, {fl50 t = incremental relative displacement, velocity, and acceleration
. {flX} vectors, respectively
x = distance from the point of contraf1exure
.
X,X,X = relative lateral displacement, velocity, and acceleration
of the equivalent mass with respect to the ground
{flY g } = incremental base acceleration vector
Z = slope of stress-strain curve at E:c?E:o
S = constant of the Newmark's S method

flo = incremental lateral displacement

EC = strain of concrete
E: o = strain at f c=f'C
E: u = ultimate strain of concrete
8 = rotation due to flexure
81 = rotation due to bond slip
M = incremental rotation at end of the elastic portion
 8
ec = rotation at cracking
eu = ultimate rotation
ey = rotation at yielding
A. = ratio of the length at rigid end to the length of elastic portion

E, •
1
= damping factor for ith mode
<P = curvature
<Pc = cracking curvature
</> = assumed displacement at level r, normalized with respect to the
r top level displacement
<Py = yield curvature
A</> = incremental rotation with respect to vertical axis
w = circular frequency of single-degree system
w. = circular frequency for ith mode
1
 9

CHAPTER 2
ANALYTICAL MODEL

2.1 Introductory Remarks

An analytical model was developed to study the dynamic
response of reinforced concrete frame structures subjected
to earthquake motions. Inelastic deformations were con-
s i dered in the mode 1 th'~ough hys teres is systems. The
mode 1 is capable of accepti ng different hys teres is func-
tions with different levels of complication.
This chapter describes basic principles used for treat-
ing the parameters involved in the analysis. It was not
the intention of this study to examine different alterna-
tive techniques for dealing with such parameters. Therefore,
methods were used which have proven to be. appropriate and
efficient. Similar to several other nonlinear models, this
model linearized the problem over a short time step. As a
result many assumptions used in an elastic analysis were
considered to be valid.

2.2 Assumptions about Structures and Base Motions

Several simplifications were necessary to avoid a compli-
cated and costly solution. Meanwhile, the simplifying assumptions
had to assure a relatively realistic representation of the problem.
The assumpti ons were the fo 11 owi ng:
1. A beam or a column is a massless line element consisting
of (a) infinitely rigid portions at ends, (b) a linearly elas-
tic portion in the middle, and (c) two flexural springs connect-
 10

ing the elastic portion to the end portions (Fig. 2.7).

The position of each member coincides with its centroidal
axis.
2. Axial deformation is neglected in all members.
Therefore, at each level, all the joints connected by beams
displace equally. Because of this assumption, vertical
displacements are not considered in the model.
3. The structure is a plane frame which displaces
horizontally in its plane, and rotates about an axis
perpendicular to the plane of the structure.
4. Deformations are considered to be sufficiently
small to allow the initial configuration of the structure
to prevail throughout the analysis.
5. Shear deformations of the members are neglected.
6. Joint cores at beam-to-column connections are
infinitely rigid.
7. Stiffness characteristics of the structure remain
unchanged over each short time increment.
8. Masses are lumped at locations where the horizontal
degrees of freedom are defined. There can be more than one
degree of freedom at the same level, if some beams are dis-
continued.
9. The foundation of the structure is considered in-
finitely rigid. Columns at the first floor are rigidly
connected to this foundation.
10. Gravity effects, usually referred to as "P_f! effects, II
 11

are taken into account.

11. Base motions occur in the plane of the structure in
the horizontal direction.

2.3 Force-Deformation Relationship

Flexural characteristics of structural elements for
monotonically increasing loads were calculated based on the
measured material properties. To simplify such evaluation
it was necessary to make a few idealizations which are ex-
plained in the following sections.

a. Stress-Strain Relationships for Concrete and Steel

A function consisting of a parabola and a linear seg-
ment, proposed by Hognestad (28), was adopted to idealize
stress-strain variation of concrete (fig. 2.1). The
mathematical formulation of the curve is as follows:

f (2.1)
c = f'c

and

fc = f c [1 - Z (e;c -
I
EO)] EO < EC (2.2)

where
fc = stress of concrete;
f'c = measured compressive strength of concrete;

EC = strain of concrete;

EO = s trai n at f C = fl.
C'
Z = slope of stress-strain curve at EC >E O ·
 12

The idealized stress-strain curve for steel is presented

in Fig. 2.2. The curve consists of three segments for linear,
plastic, and "strain-hardening" stages.

b. Moment-Curvature Relationship
The primary moment-curvature re 1ati onshi p for an element··
was idealized as a trilinear curve with two breakpoints at
cracking and yielding of the element (Fig. 2.3). Cracking oc-
curs when the tensile stress at the extreme fiber of the con-
crete under tension is exceeded. Yielding of the section is
associated with yielding of the tensile reinforcement.

c. Moment-Rotation Relationship due to Flexure

The idealized primary curve described in Section (~ is
used to determine the moment-rotation relationship. Moment
was assumed to vary linearly along the member as shown in
Fig. 2.4. With the point of contraflexure fixed at the mid-
dle of the member, it was possible to specify a relationship
between rotation and curvature. This relationship remained
invariable during the analysis. The end rotation in terms of
curvature is described as follows:

e = COlt; = t~f tl [~(x)] x dx (2.3)

o

in which
t l
= length of elastic portion of a member;
x = dis.tance from the point of contrafl exure;
~ = curvature
 13

CO is in effect the first moment of the area under curvature diagram

with respect to the point of contraflexure (Fig. 2.4). Be-
cause the variation of the moment along an element is linear
and because the skeleton curve was assumed to consist of linear
segments, the curvature varies linearly along the element (Fig. 2.4).
Hence, the computation is reduced to evaluation of the area
moment of a triangular part at the uncracked region of the
element and trapezoidal segments in other portions.
Based on the foregoing discussion, the end rotations at
cracking, yielding, and ultimate points are calculated as
follows:
1. Cracking stage:

i '
ec = 6EI Mc (2.4)
where
EI = elastic flexural stiffness;

Mc = cracking moment.
2. Yielding stage:

Q,I3
By =6- [(1_A ) <py_A2 <pc] (2.5)

where
<Pc = cracking curvature;
<Py = yield curvature;
A =
-Mc .
My
.
My = yield moment.
 14

3. Ultimate stage:

8 = ~'{[(2 + ).2)(1 - ).2)(u).2 + 1 - ).2)/u

U
(2.6)
+). (1+).) 2).3]!l + 2).2 4 ]/12
2 2 - 1).2 1 c
in which
M - M~ 4
u = (4 u _ <j> )
(2)
u y My
Mc
). 1=
Mu
). 2= ~
Mu

Mu= ultimate moment

With cracking, yielding, and ultimate breakpoints, the

moment rotation curve \';Ias idealized into the trilinear curve shown
in Fig. 2.5. Because the 8 values are proportional to the length
of the member, the curve was constructed for only unit length of
each member.

d. Rotation Due to Bond Slip

The rotation caused by relative movement between tensile steel
and concrete is calculated based on some simplifying assumptions
as follows:
1. The anchorage length of the reinforcement is sufficiently
long so that no pullout will occur.
2. Steel stress varies 1inearly from a maximum val ue at
the end of the flexible portion of the beam to zero
as shown in Fig. 2.6.
3. The rotation due to bond slip occurs with respect to
 15

the centroid of the compressive reinforcement.

4. The tensile stress in the reinforcement is proportional
to the moment.
When the tensile reinforcement is subjected to stress fs the
elongation e can be calculated from (Fig. 2.6)
2
db' fs
e = (2.7)
8 Es U

in which:
db = diameter of the tensile reinforcement;
Es = modulus of elasticity of steel;
u = average bond stress.

Then the rotation (e) can be expressed as

2
, db' f s x 1
e = 8E u d-d ' (2.8)
s
where d-d ' = distance between tensile and compressive bars.
Assumption (4) can be stated as follows:

f = f . ~ (2.9)
s .y My
where fy = yield stress of steel.
Substituting Eq. 2.9 in Eq. 2.8 will result in a parabolic
expression for the rotation in terms of moment:
d\ f2
e' = 1 2 x --Y (11..)2 (2.10)
8 Esu.. d-d My I

From this equation, the rotation due to bond slip is calculated

at the breakpoints of the curve in Fig. 2.5, and then, added to
ec , ey , 8 u values.
 16

2.4 Element Stiffness Matrix

Each element was assumed to consist of (a) an elastic
prismatic line member over a length equal to the clear span
of the members at the ends, (b) one concentrated rotational
spring at each end of the elastic part, and (c) two infinitely
rigid parts (Fig. 2.7). The springs were used to account for
elastic deformations, and their force-deformation function was
governed by a series of hysteresis rules. The rotational
spring and the elastic portion of the member behave as two
springs in series. The end rotation of the elastic segment
at one end is affected by the magnitude of the moment at the
other end. However, it was assumed that the rotation in the
spring at each end is not influenced by the moment at the
other end. As a result the relationship between the incremental
end moments and end rotations of a flexible portion of an element,
in combination with the flexural springs, can be stated as follows:

1
=

'~----------------------~v~--------------------~/
Stiffness Matrix [KI] (2.11)

where
f = flexibility of rotational spring;
~MI = moment increment at end of the elastic portion;
~e = incremental rotation at end of the elastic portion.
 17

The stiffness matrix for the entire element, including

the rigid end portions, is obtained by appropriate trans-
formation of the stiffness matrix in Eq. 2.11. The trans-
formation matrices were formed by considering the
equilibrium of rigid end segments (Fig. 2.8) as

llMA = llMA + AA (IlMA + llM')

B
I

or llM A = (1 + AA) llMA + AA llM'B

thus

llMA 1 + AA AA llM'A
= (2.12)
11MB AB 1 + AB llM'B
... v
,
[E]

where
llM = moment increment at member end;
A = ratio of the length at rigid end to the
length of elastic portion

Finally, the element stiffness matrix is formulated in the form

[K] = [E]T [K'J [E] (2.13)

which is a 2 by 2 matrix consistent with one rotational degree

of freedom at each end. Because no axial deformation is con-
sidered for members there are no lateral displacements at beam
ends. Therefore, the stiffness matrix in Eq. 2.13 is directly
applicable for beams. For columns, however, there are relative
 18

lateral displacements at the ends, consequently the number of

degrees of freedom is 2 at each end.
The stiffness matrix for a column can be formulated by
inclusion of the effects of lateral displacements. This is
accomplished by relating the total rotation and displacement
of a column enQ to the rotation with respect to the axis of
the column. The transformation matrix [T] serves the purpose
(Fig. 2.9),

/I.<'\A
A4>A
{:::} " [TJ (2.14)
MB
l\</>B

in which 1

~J
Q,
(2.15)
1
o t

where t = total length of member;

/1.8 = incremental rotation at member end with respect
to member axis;
/I.</> = incremental rotation with respect to vertical axis
/1.0 = incremental lateral displacement.

Finally, the column stiffness matrix is expressed in the fol-

1owi ng form:

[K] = [T] T [K] [T] (2.16)

2.5 Structural Stiffness Matrix

By accumulating the contributions of individual element
 19

stiffnesses, the structural stiffness matrix was constructed. First,

element indices were developed to relate local degrees of freedom to
global. Then, element stiffness matrices were added to the struc-
tural stiffness matrix at appropriate locations.
Since axial deformations were neglected, all the joints connected
by beams at a level displaced equally in horizontal direction, and in-
troduced one degree of freedom. In addition, each joint between
structural elements had one rotational degree of freedom.
The components of the structural stiffness matrix were divided
into three categories, and the matrix was partitioned accordingly
as described here:

(2.17)

Then the structural stiffness matrix was condensed to relate lateral

forces to horizontal displacements as

{L'lP} = [K] * {L'lU} (2.18)

where [K] * = [K ll ] - [K12 ] [K 22 ] -1 [K 21 ] (2.19)

2.6 Mass Matrix

The mass at any level of the structurewas considered to be
concentrated at that level. As a result, the mass matrix of the
system is a diagonal matrix.

o
[M] = (2.20)
o

No rotational inertia was considered for the masses.

 20

2.7 Damping Matrix

Damping forces were assumed to be proportional to the in-
stantaneous velocities of the points where the degrees of freedom
were defined. The damping matrix was considered at structural
level, and it was constructed by linear combination of the mass and
structural stiffness matrix.

[C] = a[M] + S[K] (2.21)

a and S can' be obtained from the following equations:

_ 1 2
~l -2w (a + SW1) (2.22)
l
1- (a + 2
['2 =2w2 SW2)

in which ~1 and ;2 = damping factors for the first two modes

It can be seen in Eq. 2.22 that when the damping matrix is

proportional only to the mass matrix (8=0), the damping factor is
small for higher frequencies of vibration. On the other hand, if
the damping is proportional only to the stiffness matrix (a_O), the
damping factor is large for higher frequencies. Therefore, the con-
tribution of the higher modes to the response will be less significant.

2.8 Unbalanced Forces

In a structure subjected to motions causing nonlinear deform-
ations the stiffness characteristics change continuously. However,
this cannot be reflected directly in a model which uses constant stiff-
ness during a time step. As a result, at the end of each time step
there may be residual forces at member ends. If these residual forces
are not eliminated, the analysis will converge to erroneous response
(Fig. 2.10).
 21

A force-deformation curve, consisting of linear segments, will

result in unbalanced forces only at break points; thus, the problem
of residual forces is less significant for this case. Nevertheless,
the accumulation of these forces will introduce errors in the calculated
results.
In the present study, at the end of each time interval the forces
were corrected (if necessary), and the new stiffnesses were used for
the next time interval (Fig. 2.11). Since the damping matrix is a
function of stiffness, there are also unbalanced forces due to change
in the damping. But these forces were considered negligible com-
pared with the inaccuracies which existed in the calculated damping
forces.

2.9 Gravity Effect

Generally, the inclusion of gravity effect (often known as
IIP_lI effectll) in the analysis results in softening of the structural
model. There have been many reports on the influcence of gravity
effect on the calculated seismic response. Goel found the effect
insignificant when he studied a multistory frame in nonlinear range
( 14). However, Jennings and Husid in their study of a single-
degree-of-freedom system concluded that the P-ll effect was substantial
(16). In the present study, the effect of gravity loads is taken
into account.
The resulting additional moment caused by P-ll effect can be re-
placed by a restoring force Q at the story level i. The shear force
due to the gravity load is (Fig. 2.12):

v.1 = P.1 (X.1 - X.1- l}/h.1 (2.23)

 22

and for the level i+l

(2.24)

in which Pi is the total vertical load on the column at level i;

and hi is the heig~of story i. The net restoring force Q is:

Q.1 = V,1 - V'+l (2.25)

1

It can be seen that, at each level i, Q.1 is a function of

displacements at level i and the two adjacent levels. Therefore,
a banded matrix [Kp] with the band width equal to 3, can be formulated
to relate the restoring forces to the story displacements {X}.

{Q} = [Kp] {X} (2.26)

[Kp] can be considered as a stiffness matrix which when sub-

tracted from the structural stiffness matrix, reproduces the soften-
ing effect caused by the gravity loads.

2.10 Differential Equation of Motion

By considering the equilibrium of all the forces, the equation of
motion can be formulated in an incremental form for a short time step,

[~1] {!:IX} + [C] {!:IX} + [K] {!:IX} = - [M] {!:I Y}

g
(2.27)

in which
[M] = mass matrix;
{!:IX} = incremental relative acceleration vector;
[C] = instantaneous damping matrix;
.
{!:IX} = incremental relative velocity vector;
 23

[K] = instantaneous stiffness matrix;

{~X} = incremental relative displacement vactor;
{~y }
g = incremental base acceleration vector.

2.11 Solution Technigu~

Several explicit and implicit methods are available for inte-

gration of the equation of motion. Newmark's S method (23) is
one of the most efficient algorithms, and has been widely used for
both linear and nonlinear problems. This method was adopted for
this analysis.
The value of S was taken equal to 0.25 which corresponds to
constant acceleration over the solution time interval. For linear
problems s= 0.25 results in unconditionally stable solutions. How-
ever, in a nonlinear problem, the method is unstable if large time
steps are used in the analysis ( 2 ).
The incremental velocities and displacements over a short time
step are calculated from:
... .. t,t
t,X = Xn ~t + t,X :2 (2.28)

(2.29)

From Eq. 2.27 the incremental relative accelerations can be formed

.. 6~X 6· ..
t,X = (~t)2 - t,t Xn - 3X n (2.30)
.
After substituting this equation in Eq. 2.25, t,X will be in the form:

. 2
t,X = t,t - 2X n (2.31)
 24

Substitution of Eqs. 2.28 and 2.29 in the differential equation of

motion will result in Eq. 2.32 for relative displacement vector:

{L~X} = [Ar 1 {B} (2.32)

in which [A] = [~i2 [M] + ~~ [C] + [K]]

and

{B} = [M]Cu4t + 2 {x"} - {~y}} + 2 [C] {X'}

{X.}
n n n
With the values of incremental displacements the incremental
velocities and acceleration were calculated from Eq. 2.30. Then,
the total values were obtained.
 25

CHAPTER 3
HYSTERESIS MODELS

3.1 Introductory Remarks

In this chapter, general considerations related to hysteresis
models are first stated followed by specific explanations about models
which had been used before or were developed in the course of this study.
Five hysteresis models were used. Two of them have been described in de-
tail elsewhere (25,36). The other three models, which are relatively sim-
pler, are documented in this chapter and in Appendix A.
In the sections on individual hysteresis systems, the degree of
complexity as well as the performance of the models for high- and low-
amplitude deformations are described. The primary curve in all cases is
assumed to be symmetric with respect to the origin.

3.2 General Comments

Since the nonlinear analysis of reinforced concrete structures under
cyclic (or dynamic) loading started, special attention was needed to be
given to the hysteretic behavior of the members. When experimental re-
sults on the cyclic loadings of the reinforced concrete members and joints
became available, it was evident that closed-form mathematical formulas
did not allow enough versatility to match the measured behavior. There-
fore, mu1tisegment hysteresis models consisting of linear portions were
developed which could reproduce the experimental results. In this group
was the Takeda model ( 36), which in several cases has proven to lead to
satisfactory results. This model is one of the alternative hysteresis
systems which can be used by the analytical model developed in this report.
 26

The Takeda model does not include the "pinching" effects (tendency
for very low incremental stiffness near the origin followed by a stif-
fening) which are often observed in the experimental results. And yet,
the model is complicated. Therefore, it seemed worthwhile to develop and
examine a simpler model which considers the pinching effect. This new
model was named IISina. 1I

Other less complicated hysteresis systems have been used by different

investigators. Otani ( 25) modified the Takeda model to use in conjunction
with the original model. Although this model was applied to account for
bond slip, it can be regarded as a complete hysteresis system.
Another model, which has been used widely, is a simple bilinear re-
lationship. The analytical model was equipped with the facility to use a
bilinear hysteresis.
For unloading and load reversal stages, the bilinear hysteresis
system results in incremental stiffness values which are considerably in
excess of the corresponding measured values. To obtain closer agreement
with test results without complicating the hysteresis system, a new bi-
linear hysteresis model was developed with softened unloading and load re-
versal branches. This system (Q-Hyst) along with Otani and the simple bi-
linear model are the three other alternative hysteresis systems considered
in the present study.

3.3 Takeda Hysteresis Model

Based on various experimental results, the Takeda model consists of
16 rules operating on a trilinear primary curve (Fig. 3.1). The primary
curve can include additional deformations caused by bond slip. However,
the rules do not cover the pinching effect which can also be caused by slip
of the reinforcement.
 27

The rules determine different stiffness characteristics at different

stages of cracking, yielding, unloading, and reloading in successive cycles.
The fact that the model considers cracking as a break-point, results in
some energy dissipation under cyclic loads even at pre-yielding stage.
This is realistic and desirable. Many of the rules used in Takeda's sys-
tem are concerned with developing realistic force-displacement relation-
ships during low-amplitude cycles which are within the bounds of large-
amplitude cycles previously reached. For example, the force-displacement
wave is specified to proceed from X3 to R3 (this action requires a srecial
rule) rather than, say, from X3 to R2 (which would not have required a
special rule, Fig. 3.2).
As it was previously noted, pinching is not included in the Takeda
system. As a result, the model ignores the softening that can occur for
beam-to-column connections at low amplitudes. This is illustrated in
Fig. 3.3. The Takeda rules are presented in full detail in Reference 25 .

3.4 Sina Hysteresis Model

This model was developed to account for the pinching effects while it
had a fewer number of rules than the T~keda model. The skeleton curve con-
sists of three parts similar to those used by Takeda. Nine rules define
this system. A complete description of the rules is presented inAppendixA.
The initial loading and unloading rules are similar to the Takeda rules
1 through 4. The slope of unloading for post-yielding regions (Sl) is
assumed to be:

(3.1)
 28

where
SCI Y = slope of a line connecting yield point to cracking
point in the opposite direction;
O(Y} = yield deformation;

°max = maximum deformation attained in loading direction;

a = constant (assumed to be 0.5).
When the load is reversed towards the direction previously yielded.
a low-slope branch followed by a stiffening part is considered (path
Xl BUm in Fig. 3.4). The portion X1B corresponds to the stage when the
crack (now in the compressive region) has not been closed. and the moment
is resisted only by the reinforcement. When the crack closes (at point B
in Fig. 3.4) the compression caused by the moment is resisted by both the
compressive steel and the concrete; hence the resistance increases.
The position of the crack-closing point has a significant effect on
the stiffness for small amplitudes. Based on the experimental results re-
ported in Reference 18. the following can be stated about the location of
this point:
1. For a given section. the value of moment at crack-closing pOint
remains almost constant throughout the loading.
2. If the anchorage condition does not allow any "push-in pUll-out"
to occur the value of the moment can be calculated from:

M= a1Asfsy (d-d
l
) (3.2)

in which
al = constant (a = 0.5 appears to give a reasonable
agreement w1th the experimental results);

As = area of steel;
fsy = yield stress for steel;
d-d = distance between the centroids of compressive and
l

tensile reinforcement.
 _ 29

Otherwise, the moment is resisted by bond stresses and can be

obtained from:

(3.3)

where
db = diameter of the compressive bar;
u = average bond stress;
la = anchorage length.
3. The rotation at which the crack closes depends on the maximum
rotation attained in the corresponding direction (Fig. 3.4).
It is assumed that

_ 3
D(B } - 4 (X max ) (3.4)

where
Xmax = maximum residual deformation previously attained.

3.5 Otani Hysteresis Model

This model, which is a modified version of Takeda system, was
originally used to represent the stiffness variation of a joint spring
in conjunction with a flexural spring. Because it was simpler than the
Takeda model, Otani system was applied here as an independent model.
The primary curve in this system is bilinear with the break-point
at yielding of the section (Fig. 3.5). Because the cracking point is not
recognized, the rules related to cracking pOints could be eliminated in
this model. There are eleven rules describing the Otani model which are
explained in Reference 25. The unloading slope from post-yielding branch
was
 30
The general trend for handling of the low amplitude loops is similar
to those in the Takeda system. As a result, this model is still complicated.

3.6 Simple Bilinear Model

Because of its simplicity, the bilinear hysteresis system has been
extensively used for both steel and reinforced concrete structures. The
model can be described by only three rules (Fig. 3.6). There are merely
two stiffnesses considered in the model: elastic and yielding stiffnesses.
Unloading and load reversal slopes are the same with the slope of the elas-
tic stage.
The general observation in this model is ,that: (l) large energy
dissipation is provided for high amplitude deformations, and (2) in low
amplitudes, no hysteretic energy dissipation is considered.
From a crude inspection of the model, it is evident that the stiffness
characteristics of the unloading and load reversal stages are substantially
different from what is observed in cyclic loading of a reinforced concrete
member. However, to obtain a better understanding of the influence of this
discrepency on the calculated response, the bilinear model is examined here.

3.7 Q-Hyst Model

This model was developed as part of the present study. It can be
considered as a modified bilinear hysteresis system. The basic purpose
of modification was to provide softened branches for unloading and load
reversal stages (Fig. 3.7). The model consists of four rules which are
described in Appendix A.
Unloading from a point beyond the yield point and reloading in the
other direction follows two different slopes:
 31

(1) The slope of the unloading portion (UmX o) is determined as

a function of the displacement Urn and the slope of the initial portion
OY in a manner similar to that for Takeda Model (See Appendix A).
(2) The reloading portion has a slope determined by the coordinates
of poi nts Xo and U~ where U~ represents a poi nt on the primary curve
symmetric to Urn with respect to the origin.
This assumption is desirable because: (1) it helps to simplify the
model and (2) for low amplitude deformations, provides some softening
comparable with pinching effects.
It should be emphasized that this model does not provide any
energy dissipation unless the system yields (this deficiency also
exists in Otani and simple bilinear models). Consequently, if the load
starts with small amplitude deformations below the yield point, the
model considers the section elastic. This is unrealistic in view of
the fact that nonlinear behavior in a reinforced concrete section
starts immediately after the section cracks.
 32

CHAPTER 4
TEST STRUCTURES AND ANALYTICAL STUDY USING MDOF MODEL

4.1 Introductory Remarks

Two test structures were studied using the multi-degree-of-freedom
model. First, this chapter briefly describes these structures and the base
motions used in the experimental studies. Then, considerations which
were given in computing different parameters involved in the analysis
are explained. Finally, the analytical program and the calculated
results are presented.

4.2 Test Structures

Dynamic behavior of two small-scale ten-story reinforced concrete
structures, called MFl and MF2, was studied. The structures were
identical except for (a) the discontinuity of beams at the first floor
of MF2, and (b) the first- and second-story column reinforcement which
was different for the two structures. Each structure consisted of two
i denti cal three-bay frames. In both structures, the fi rst and the
tenth stories were longer than each one of the other stories. The
overall configuration of the frames is presented in Fig. 4.1.
The mass at each level of each test structure was approximately
465 Kg, except for the first story mass in structure t~F2 which was
291 Kg. At each level, the mass was transferred directly to the
column centerlines such that each column carried 1/8 of the weight
(except for the first floor of MF2).
The structures were designed using the substitute-structure
method. The design maximum acceleration was 0.4g. Cross-sectional
 33

dimensions and reinforcement were chosen so that beams would develop

major yielding before columns yield. The distribution of longitudinal
reinforcement is depicted in Table 4.1 (See also References 15 and 21).

4.3 Dynamic Tests

The tests were conducted using the University of Illinois Earthquake
Simulator. In each structure, the frames were placed parallel to each
other on the test platform (Fig. 4.2). The direction of motion was
parallel to the plane of the frames and in horizontal direction. Each
structure was subjected to three simulated motions with increasing
intensities (normalized maximum accelerations) in successive runs.
In addition, before and after each earthquake simulation, free vibra-
tion and steady state tests were carried out to determine damping ratios
and changes in natural frequencies of the structures.
The base motion for the three earthquakes was modeled after the
measured north-south component of the earthquake at E1 Centro,
California, 1940. Because the structures were in small scale, the time
axis of the base motions had to be compressed by a factor of 2.5
to obtain realistic ratios between the earthquake frequency content
and natural frequencies of the test structures. For each earthquake
motion and steady state test, relative story displacements and total
story accelerations in the direction of motion, also total vertical
accelerations at the top of two corner columns (one in each frame)
were recorded.

4.4 Analytical Procedure

Based on the discussion in Chapter two, a computer program
C'LARZ") was developed to analyze rei nforced concrete structures
 34

subjected to base motions. Five different hysteresis models (described

in Chapter 3) can be assigned to calculate stiffness characteristics of
structural elements. A block diagram and some technical information
about the program is presented in Appendix B.

a. Flexural Properties of Members

Moment-curvature relationships for members were calculated based
on measured material properties and idealized relationships described
in Chapter Two. Nominal dimensions of cross section of each member was
used in this calculation. No ultimate limit was imposed on strength
(or deformation) of a member. The material properties used in the
analysis are depicted in Table 4.2.
Axial forces in beams were assumed to be zero. In columns,
although the axial forces vary during an earthquake, it was assumed
that axial forces remain constant. Consideration of changing axial
force would involve complicated hysteresis models. The axial forces
due to dead load and the assumed axial forces to calculate moment-
curvature relationships of columns are presented in Table 4.3. Moment-
rotation relationships were calculated by the computer program using
Eq. 2.4 through 2.6.
Rotations due to bond slip were calculated using Eq. 2.10. Values
of rotations corresponding to the cracking, yield, and ultimate moment
of each member were calculated. Then, they were added to the flexural
rotations. The rotations due to bond slip are listed in Table 4.4.
 35

b. Damping
In general, that part of seismic response caused by higher modes
of vibration is neither calculated nor measured accurately. To reduce
contribution of higher modes to calculated results, a stiffness dependent
damping was used in the analysis (a=O in Eq. 2.21). The damping factor
(~) was taken equal to 2%.

c. Time Step for Numerical Integration

In Newmark's 8 method (23), the limits on the time step to insure
convergence and stability of the solution are functions of natural
frequency of the structure. When a structure develops nonlinear
deformations, the natural frequencies change as the stiffness changes.
Hence, the limits are not directly applicable. Several authors have
cited different limits on time step of numerical integration (20,23).
In the present study, the time interval was taken approximately equal
to one-tenth of the shortest period of the structures. For the structure
MF2, ~t=0.0008 second was used; in analyzing structure MF1, it was
found that ~t=O.OOl second led to stable response as well.
Because a piecewise force-deformation relationship is used in the
analysis, it is not necessary to vary stiffness in each short time
interval. Some investigators have recommended to change the stiffness
once at every ten time steps (11). This was adopted in the present
study.

4.5 Analytical Study

To observe the performance of the model, structure t1F2 was first
analyzed subjected to the first six seconds of measured base acceleration
in the first earthquake run with the maximum value normalized to 0.38g
(design intensity). Takeda hysteresis model was used to govern stiffness
 36

variation. Because the time axis of the input base motion was compressed
by a factor of 2.5, the duration of the analysis corresponds to fifteen
seconds of the actual earthquake at El Centro. The base motion, used
for the analysis, included both large and small amplitudes. The cal-
culated and measured responses are presented in Fig. 4.3.
The measured base moment, base shear, and top story displacement
and acceleration are superimposed on the corresponding analytical
results to make possible a close comparison of the response. Because
the displacements were dominated by the first mode, comparison of the
measured and calculated top story displacement is a representative
measurement of the quality of the calculated story displacements. The
observed and calculated displacements and accelerations at every other
level are also depicted in the figure. The maximum response values
at all levels are presented in Table 4.6.
Sensitivity of the calculated response to the hystereSis models
was studied analyzing structure MF1. The measured base acceleration
during the first earthquake run was used as the base motion. This
earthquake was comparable with the design motion. The structure was
analyzed using the hysteresis models described in Chapter Three. In
each case, one hysteresis system was used for all structural elements.
The duration of the earthquake in each case was five seconds. This
duration was long enough to cover large- and small-amplitude ranges
of response. The calculated and measured response are presented in
Fig. 4.4 through 4.8. The maximum analytical and observed displacements
and accelerations are cited in Table 4.7. Maximum rotational ductilities
of member ends are pres~nted in Appendix C.
 37

CHAPTER 5
COMPARISON OF MEASURED RESPONSE WITH RESULTS
CALCULATED USING THE MDOF MODEL

5.1 Introductory Remarks

The results obtained from the multi-degree-of-freedom model are discussed
in this chapter. First, the calculated values are compared with the test re-
sults for structure MF2. In section 5.3, influence of the hysteresis models on
the calculated response are described and the performance of each system is
studied. Because the waveforms of top-level displacement, base shear, and base
moment are similar, only top-level displacement is discussed,in detail. To some
extent, the discussion is also applicable to top-level acceleration. Note that
the displacements are expressed relative to the platform of the earthquake sim-
ulator, while the acceleration response represents the total acceleration.
In the following sections, values of T refer to the abscissas in Fig. 4.3
through 4.8. In the discussions of maximum response, absolute values of response
are considered.

5.2 Calculated Response of MF2

Measured (broken curves) and calculated (continuous curves) response
histories of structure MF2 are presented in Fig. 4.3. The structure was ana-
lyzed using the Takeda hysteresis model.
Performance of an analytical model can be evaluated in various aspects.
One of the factors of importance is the maximum response. It can be seen in
Fig. 4.3 that the measured top level displacement includes two major peaks at
T ~ 1.4 and T ~ 2.4 seconds. Although the analytical model reproduces the first
peak very well, it fails to match the second one. The maximum acceleration at
Level 10 is calculated reasonably well. In the large-amplitude region, the
frequency content and the waveform of the calculated response is very close to
 38

what was measured. In the low-amplitude range, the calculated response has a
distinctly alternating character which was not observed in the test results.
The difference is even more visible in the acceleration response.
The overall deformed shape of the structure is presented in Fig. 5.1.
The numerical values of maxima are listed in Table 4.5. Very close correlation
was observed between the measured and calculated shapes. It appears that the
analytical model slightly overestimates the displacements at levels one through
six.
Relative story displacements are plotted in Fig. 5.2. Between levels five
and ten, the calculated values were smaller than those observed. The trend
is reversed in lower stories. It is worthwhile to notice that relative story
displacements are highly sensitive to slight changes in deformed shape of a
structure. Hence, the calculated results may be regarded as being satisfactory.

5.3 Calculated Response of MFl

a. Takeda Model
Figure 4.4 shows the observed and calculated response history of structure
MFl subjected to the base acceleration measured in Run 1. The analytical
results were obtained based on the Takeda hysteresis model. Excellent cor-
relation is observed up to T ~ 3.2 seconds. During this period, maxima, fre-
quency contents, and waveforms of experimental and analytical results are quite
close. This indicates that the overall hysteretic behavior and energy dissipation
of the structure were presented well by the Takeda model with a = 0.5 (in Eq.3.1)
The calculated response deviates from the measured curve at T ~ 3.2 seconds
when low-amplitude displacements are experienced. Differences can be seen in
amplitudes,waveforms, and frequency contents. In fact, the analytical model re-
sults in a response with some visible frequency content while the test results
have no clear low-mode frequency content. Such difference signifies that, in
 39

low-amplitude regions, the Takeda hysteresis model resulted in structural

stiffnesses larger than the actual stiffness of the structure.
The calculated story displacements along the height of the structure are
plotted against observed maxima in Fig. 5.3. The calculated deformed shape of
the structure is very close to the measured shape. Some minor differences are
observed in lower stories. At the tenth level, the calculated displacement was
only 2% smaller than the measured value.
To compare the shape of the structure at the time of maximum response,
the maximum story displacements are normalized with respect to the tenth level
displacement (Fig. 5.4). The calculated shape is reasonably close to what was
measured. It can be seen that the shape obtained from the analytical model is
smoother than the observed shape.
For relative story displacements, the difference between the observed and
analytical results seem to alternate and no uniform variation can be recognized
(Fig. 5.5). The discrepancy is attributed to the sensitivity of relative story
displacements to small changes in the deformed shape of the structure.

b. Sina Model
The analytical results based on the Sina hysteresis model are presented in
Fig. 4.5. The calcualted response seems to be in reasonably good agreement with
the measured response up to T ~ 2 seconds. Beyond this point and before T ~ 3.2
seconds (where low-amplitude response starts), the frequency content of the ana-
lytical results is almost the same as that of the test result. However, dis-
placement maxima are overestimated by the model. The fact that three of the
four peak points in this range overestimate the response indicate that dis-
sipated energy considered by the Sina model was less than what was ex-
perienced by the structure. Consequently, the model had to develop additional
displacements to compensate for the difference.
 40

Reasonably close correlation can be seen between low-amplitude response of

the measured and the calculated results. The agreement is more pronounced in
base shear, base moment, and top level acceleration. Inclusion of pinching ef-
fect in Sina model is believed to have resulted in a stiffness close to the actual
stiffness of the structure over the low-amplitude region.
The calculated maximum story displacements at all levels are shown in
Fig. 5.3. It can be seen that the calculated values consistently overestimate
the measured quantities over the height of the structure. The difference at the
tenth level is 19%. The deformed shape normalized with respect to the top level
displacement (Fig. 5.4) is found to be very close to the shape obtained from the
test results. The correlation is more satisfactory at upper levels.
As for relative story displacements, the calculated values are larger than
those measured in most stories (Fig. 5.6).

c. Otani Model
The calculated response using Otani model exhibits the same frequency
content as the measured response during the period when large-amplitude dis-
placements were obtained (Fig: 4.6). However, the maxima are overestimated at
the end of that range. At level ten, the calculated maximum displacement is 33%
larger than the observed maximum value. In low-amplitude range of response,
the calculated displacement deviates substantially from the observed dis-
placements.
The calculated displacements at other levels are larger than the measured
maxima (Fig. 5.3). The difference between the analytical and experimental re-
sults is even more pronounced at the fifth and sixth levels. There was con-
siderable difference between calculated and measured normalized shapes between
levels three and seven (Fig. 5.4).
 41

Relative story displacements corresponding to maximum displacements are

presented in Fig. 5.7. Again, at lower stories, the calculated values are in
excess of the measured quantities. The trend is reversed at upper levels.

d. Bilinear Model
Unsatisfactory results were obtained with the simple bilinear hysteresis
system. Except for the frequency before T ~ 1.7 seconds which is somewhat close
to the frequency of the measured response, the calculated results were consid-
erably different from the experimental values in all important aspects. The
fact that the response was generally underestimated indicates that the analytical
model had dissipated the input energy before it developed displacements comparable
to the measured values. Because the hysteresis model is the major source of
energy dissipation in a nonlinear structure, it can beconcluded thatthebilinear
hysteresis model has overestimated the energy dissipation. At the top level,
the calculated maximum displacement was 14% smaller than the measured value.
Along the height of the structure, at sixth level and below, the calculated
maximum displacements were close to the measurements (Fig. 5.3). However, a closer
inspection of the deformed shape of the structure reveals that this close cor-
relation is due to inconsistency of the model (Fig. 5.4). It has to be emphasized
that even at these levels, the calculated and observed maxima occur at different
times.
As it can be expected from almost straight deformed shape of the structure
above level six (Fig. 5.3), the calculated relative story displacements were
underestimated by the model at these stories (Fig. 5.8).

e. Q-Hyst Model
Reasonably close agreement is observed between the measured and calculated
response based on Q-hyst model (Fig. 4.8). The correlation is satisfactory in
both large- and small-amplitude ranges. The peak values were overestimated
 42

in most instances. At maximum point of the tenth-story response, the cal-

culated value was 17% larger than the meaured response (Fig. 5.3).
Figure 5.4 includes the normalized deformed shape of the structure for the
calculations with the Q-hyst model. It can be seen that the calculated shape
is reasonably close to the measured shape at all levels except for levels one
through three.
Relative story displacements corresponding to the maximum displacements
are plotted in Fig. 5.9. Except for the first, ninth, and the tenth stories,
the calculated quantities exceeded the test results.

5.4 Concluding Remarks

Different aspects of performance of the hysteresis models for structure
MFl were discussed in sections (a) through (e). In terms of computer memory
space and compilation times, the smaller hysteresis models were advantageous.
However, the execution time for all the cases was approximately the same, be-
cause at each time step only one rule of the hysteresis model is used and
whether there are few or many other rules is immaterial. Based on the study
reported in sections (a-e) the following conclusions have been reached.
The bilinear model resulted in a response considerably different from the
measured response.
Among the four other hysteresis models, the performance of the Otani
model was found to be less satisfactory than the others. Considering the
fact that two of the other systems (Sina and Q-hyst models) are simpler than
Otani system~ no advantage was realized in using the Otani model.
The performance of Sina and Q-hyst models appear to be similar. However,
between the two, the Q-hyst model is preferred because: (a) Q-hyst system is
presented by only four rules as compared with nine rules in Sina model, and
(b) in the Q-hyst model, no decision is needed to be made on the location of
 43

crack-closing point which is included in Sina model.

The final comparison is to be made between Q-hyst and Takeda models.
Maximum displacements were obtained considerably closer to the measured values
when Takeda model was used, although the difference between the results based
on Q-hyst model and observed maxima were within acceptable range (17% error).
In calculating low-amplitude response, Q-hyst model was more reliable than the
Takeda system. Perhaps one of the more important factors is that the Q-hyst
model is substantially simpler than the Takeda model. Therefore, this model
is easier to understand and apply.
Further study is needed to establish the reliability of the Q~hyst model
in representing the hysteretic behavior of connections in a reinforced concrete
structure subjected to earthquake motions. Based on this particular study,
however, Q-hyst model seems to be preferable to the other hysteresis systems
considered, because it is simpler and because it led to satisfactory simulations
of the displacement-time records at all levels of the particular test structure
analyzed.
 44

CHAPTER 6
DEVELOPMENT OF THE Q-MODEL

6.1 Introductory Remarks

This chapter introduces a single-degree-of-freedom model (Q-Model)
for calculating the displacement response of reinforced concrete multi-
story structures subjected to strong earthquake motions. Nonlinearity
of deformations is considered in the model. Using the Q-Mode1, displace-
ment-histories at all levels of the structure and base moment can be
calculated.
In this chapter, "origina1 system" refers to the multi-degree
structure to be analyzed.

6.2 General Comments

The key requirement for representing the earthquake response of a
multistory structure by a sing1e-degree-of-freedom model is that the
deflected shape of the structure remain reasonably constant during an
earthquake. Experimental observations of the behavior of multistory re-
inforced concrete structural systems (1,5,8,15,21) have shown that the
deflected shape will tend to remain the same during the large amplitudes
of response. For the test structures, this was possible because the col-
umns were proportioned to experience limited yielding during the design
earthquake and the displaced shape was not sensitive to the extent of
yielding in beams.
For most earthquake motions, the elastic lateral displacement response
of multistory structures is dominated by the first mode. The results of
experiments mentioned above caul d be interpreted in terms of moderately
damped linear models with some effective stiffnesses smaller than the
initial values. In other words, the overall behavior of the test struc-
 45

tures was linear, despite the presence of local nonlinear deformations.

Observing (1) that the nonlinear displacement response of reinforced
concrete structures may be interpreted in terms of linear models, (2) that
the displacement response is dominated by the lowest mode, and (3) that the
defl ected shape remains essenti ally constant during the /ldesi gn earthquake, /I

it is plausible to use a shape similar to the shape of the first mode in

order to develop the characteristics of an equivalent SDOF model for ana-
lyzing the nonlinear response of a MDOF system.
For design purposes, lateral displacements caused by an earthquake
are of primary importance. Because, for a structure consisting of elements
with no abrupt change of stiffness, by controlling the displacements at
different levels, the member end forces (and rotations) can be kept below
the critical limits.

6.3 Q-Model
The equivalent system is shown in Fig. 6.1. The model consists of
a concentrated mass supported by a massless rigid bar. The bar is connect-
ed to the ground by a hinge and a nonlinear rotational spring. Damping
forces are exerted on the mass by a viscous damper. To define the system,
it is necessary to determine the equivalent mass, equivalent height (L eq ),
stiffness characteristics of the spring, and damping. Damping will be ig-
nored in the discussion which follows immediately, but it will be included
after the other parameters are developed.

a. Equivalent Mass
To define the mass of the single-degree model, first the dynamic equi-
librium of the system is considered. The differential equation of motion for
an undamped equivalent SDOF model representing a MDOF system as derived by
 46

Biggs (7), is:

(6.1 )

where
Ft = total external force;
Mt = total mass of the original system;
K = stiffness of the original system;
x = relative lateral displacement of the equivalent mass with
respect to the ground;
J
a,Q, = (L Fr <pr)/F t ;
r=l
J
a
m= ( L Mr<P~}/Mt;
r=l
Fr = external force at level r;
j = number of levels in the original system;
Mr = mass at level r;
<Pr = assumed displacement at level r, normalized with respect to
the top level displacement (see Section c).
For a structure subjected to an earthquake, the external forces can be
expressed as:

F
t
= -Mt X
g
(6.2)

where Xg = ground acceleration.

Substitution of Ft from this equation in Eq. 6.1, and then dividing
through by a,Q, results in Eq. 6.3.

(6.3)
.. t'

or Me x + Kx = -Mt Xg (6.4)

in which
(6.5)
 47

Hence, the equivalent mass is a function of the total mass and the
assumed deformed shape of the structure.

b. Stiffness
The stiffness of the single-degree structure is provided by a rotation-
al spring at base (Fig. 6.1). Because the bar connecting the mass to the
base is rigid, all elastic and inelastic internal work takes place in the
rotational spring. The governing skeleton curve for force-deformation re~

lationship of the spring is directly related to the stiffness characteristics

of the multistory structure.
To obtain a representative function of the stiffness of the original
system, the structure is analyzed subjected to a set of monotonically
increasing static lateral loads at floor levels. The load at each level
is proportional to its height from the base of the structure. This part
of the analysis results in relationships between base moment and displace-
ments at different levels.
It is also necessary to determine the displacement at the height equal
to Leq (Leq = equivalent height; see Section c). If Leq is equal to the
height of one of the floor levels, the displacement at this level is directly
used. However, if Leq is between the heights of two levels, a linear inter-
polation is made between the displacements at these levels. The loading is
continued until large displacements well beyond the apparent yielding of the
of the structure are developed.
The triangular distribution (Fig. 6.2) is chosen based on the results
of the study reported in Reference 30. In this report, it was shown that the
triangular, first mode, and RSS distribution led to similar results. Because
the triangular distribution is simpler, it is used in the Q-Model.
 48

A typical moment-displacement curve obtained from the static analysis

is shown in Fig. 6.3. The vertical axis is normalized with respect to M*
J
where M* = I (Mrg)h r , in which g = gravity acceleration, and hr = height at
r=l
level r. The horizontal axis in Fig. 6.3 is normalized with respect to the
height of the equivalent system (L eq ).
The calculated curve is idealized by two straight broken lines. To ob-
tain the break point, the following procedure is used:
(a) A tangent to the initial part of the calculated curve is drawn (OT)
(b) From the horizontal axis at abscissas of 0.002 and 0.003, two lines
are drawn parallel to OT
(c) The break point is assumed to be in between the intersections of
these lines with the calculated curve
The slope of the second portion is established by joining the break
point to a point on the calculated curve at an abscissa of five times the
abscissa of the break point.

The procedure described above is not necessarily a general method.

However, for the cases studied here, the procedure yielded reasonable
idealizations.
To represent the hysteretic behavior of the spring, Q-Hyst model
is used (Appendix A). The model operates on the idealized curve described
above. It is assumed that the curve ;s symmetric with respect to the ori-
gin.

c. Deformed Shape of the Structure

During the static analysis of the structure, corresponding to each load
increment, the displacements at different levels are obtained. The shape cor-
responding to the moment equal to My (Fig. 6.3) is normalized with respect
to the top level displacement and is used as the deformed shape of the struc-
 49
ture ($). This shape is assumed to remain unchanged during the earthquake.
Equivalent height is calculated from
j
EM $ h
L - r=l r r r (6.6)
-~
eq EM $
r=l r r
After the displacement-history is calculated at this height, the dis-
placements at all levels can be determined based on the assumed deformed
shape.

d. Damping
Damping is assumed to be proportional to the relative velocity of the
equivalent mass, with respect to the ground. The damping factor is arbi-
trarily taken equal to 2%. The frequency based on the stiffness of seg-
ment OY (Fig. 6.3) is used to determine the damping coefficient (C in
Eq. 6.7). Damping coefficient is assumed to remain unchanged during the
entire analysis.

e. Eguation of r1otion
The complete equation of motion is stated as

(6.7 )

where C = 2;wMe and w = circular frequency of the single-degree system

based on the slope of the line OY (Fig. 6.3). Newmark's S method (23)
with S = 0.25 is used to integrate the differential equation of motion.
This value of S allows the use of relatively large time steps for numeri-
cal integration. However, because the Q-Model is simple and small time
intervals may be used without a significant increase in computer cost,
other values (e.g., S = 1/6) can also be assigned to s.
 50
CHAPTER 7
ANALYTICAL STUDY USING THE Q-MODEL

7.1 Introductory Remarks

A computer program named "LARZAK" was developed to implement the
dynamic part of the analytical procedure described in Chapter Six. The
computer cost for each run of this program is only 3% of that of LARZ
(Chapter Four), for a ten-story three-bay frame analyzed subjected to stx
seconds of base acceleration record. A block diagram of the program is
presented in Appendix D.
To examine the reliability of the Q-Model, a three-part investiga-
tion was conducted. This chapter first describes the test structures
which were used in the study. Then the analytical study and the related
discussions follow. The first part of the investigation was the analysis
of eight different small-scale structures tested using the University of
Illinois Earthquake Simulator. The analytical results were compared with
the measured responses.
In part two, the performance of the model for different ground motions
was studied. For this part, one of the test structures was analyzed sub-
jected to different earthquakes. Because no test results were available
for this part, the multi-degree analytical model (Chapters Two and Four)
was used to evaluate the results from the Q-Model.
Part three was concerned with the effects of repeated earthquakes on
the same structure. Responses of a particular test structure to five dif-
ferent intensities of the same motion was analyzed. In each case the
motion was made up of two identical earthquake records separated by a
period of no base motion.
 51

In the following sections, comparisons between measured and calculated

maxima are made for the absolute values of responses.

7.2 Structures and Motions

a. Test Structures
Eight small-scale, ten-story, three-bay reinforced concrete structures
were used for one or more parts of the analytical study. The structures
comprised either two frames (Group One), or two frames and a shear wall
(Group Two). Structures Hl, H2, MF1, and MF2 had no \oJalls. Structures
FWl, FW2, FW3, and FW4 had walls. The story mass in all cases was approx-
imately 465 Kg., except for structure MF2 which had a 291-Kg. mass at its
fi rs t story.
Two of the structures in the first group (MFl and MF2) were described
in Chapter Four. The reinforcement, principal material properties, and
the nominal dimensions of the other two structures (Hl and H2) are shown
in Fig. 7.1. For these structures, the story height was the same at
all stories. The assumed axial forces in columns and the stiffness prop-
erties of structural members are tabulated in Tables 7.1 and 7.2. De-
tailed information about structures Hl and H2 is provided in Reference 8.
In each of the structures of Group Two, a shear wall was centrally
located in between the frames. The wall extended along the full height
of each structure. The strong axis of the wall was parallel to those of
the frames. At each level, the wall was connected to the mass by a hinged
link. As a result, the lateral displacements of the wall and the frames
were equal at each level. Because the links were hinged, they did not
impose any rotational constraint on the wall.
The reinforcement distribution, basic material properties and the
 52

dimensions of the structures of Group Two are shown in Fig. 7.2. Note that
each pai r of structures FW1 and FW4, also FW2 and FW3 were i denti caL The
assumed axial forces in columns and the stiffness properties for elements
are listed in Tables 7.1, 7.3, and 7.4. The complete information on
casting and testing of these structures are given in Reference 1.
A conservative amount of shear reinforcement was provided for elements
of all structures so that any possible shear failure was prevented with
confidence.

b. Base Motion
During the experimental study, each test structure was subjected to
three simulated earthquakes (except for H2 which was subjected to seven
earthquakes), in addition to free vibration and steady state tests before
and after each earthquake run. The base motion for all the structures,
except FW3 and FW4, was modeled after the north-south component of the
earthquake recorded at El Centro, California, in 1940. The base accel-
eration for structures FW3 and FW4 was a simulated Taft (N21E component)
earthquake.
In all case, the time axes of the earthquakes were compressed by a
factor of 2.5, to obtain realistic ratios between the frequencies of the
earthquakes and the frequencies of the structures. For example, six-
second test duration equals 15 seconds of the original earthquake.

7.3 Equivalent System

To define each structure, force-deformation relationship, deformed
shape, equivalent height, and equivalent mass were calculated. These
parameters were sufficient to describe the equivalent structural models.
To obtain the moment-displacement curves (described in Chapter Six),
 53

program "LARZ2" which is a static version of the program LARZ (Chapter Four)
was used. Assumptions and idealizations made in LARZ2 are similar to those
in LARZ. Incremental loads are assumed to be applied at the levels where
the degrees of freedom are specified. The stiffness of the structure is
a function of previous load history. During each load increment, the
stiffness is constant. Considering the fact that different elements yield
under a different load set, it is important to apply sufficiently small
load increments to allow for gradual yielding of structural elements.
In particular, in the vicinity of the apparent yield point of the struc-
ture, a large set of load increments may result in an overestimated
apparent yield force.
The results of the static analyses are presented in Fig. 7.3. The
calculated curves are idealized using the method described in Section 6.3(b).
Because the calculated curves for the structures MFl and MF2 were identical,
they were represented by one idealized curve. The ordinates of the break
points and the slopes of the idealized curves for different structures are
listed in Table 7.5.
It is worthwhile to note that the initial slope of the idealized curve
for structure Hl is less than the initial slope for structure MF1. The
cross-sectional dimensions for both structures were the same. However,
because structure Hl was shorter and had a larger value of reinforcement
with higher yield point, structure Hl would have been expected to have a
larger lateral stiffness. When the problem was examined more closely, it
was noticed that the beam reinforcement in Hl was less than that of MFl
and that, as the lateral load was increased, beams yielded first (beams
were designed to yield first). As a result, beam reinforcement played a
more important role in choosing the initial stiffness of each structure.
 54

Hence, the structure with lower yield point for beams was idealized to
have a lower initial stiffness.
For each structure, the floor displacements corresponding to the
break point were normalized with respect to the top-level displacement
(Table 7.6). Then, the resulting shape (~) was used to calculate the
equivalent mass and height of the structure. The equivalant mass of each
structure was obtained using Eq. 6.5. The equivalent height was the
geometric centroid of the deformed shape (~). With the stiffness cor-
responding to the first branch of the idealized curve, the initial
frequency of the equivalent system was calculated and used to deter-
mine the damping coefficient (C in Eq. 6.6). The values of the equivalent
mass, equivalent height, and the initial circular frequency are presented
in Table 7.5.

7.4 Analytical Results for Different Structures

The structures were analyzed for the first six seconds of the measured
base accelerations during runs corresponding to the "design earthquakes'"
The first simulated earthquake for all the structures, except H2, had a
maximum amplitude approximately equal to that anticipated by the design
calculations. For structue H2, the third earthquake run corresponded to
the design motion.
Base accelerations, top-level displacements, and base moments are
presented in Fig. 7.4 through 7.11. The maximum floor displacements and
maximum relative story displacements are depicted in Fig. 7.12 through
7.19. The calculated and measured maxima are listed in Table 7.7.
In all cases, the calculated top-level displacement and base moment
had similar waveforms. Because base moment is a less sensitive measure,
the difference between the calculated and measured values are distinguished
 55

better in the displacement response. Therefore, comparison will be made

between the calculated and measured displacement response histories.
To evaluate the performance of the Q-Model, first the response for
structures Hl, H2, MF1, and MF2 are considered. The frequency contents
of the calculated response, in large-amplitude periods, were quite close
to those of the measured response in all four cases. During the period with
small amplitude response (from T ~3.3 to 4.5 seconds), the calculated curve
deviated from the measured curve (except for structure MF1). The cal-
culated peak values were reasonably close to the measured values. The
absolute value of the maximum top-level displacement was overestimated by
8% for Hl, 23% for H2, 19% for MF1, and 27% for MF2.
Comparison of the measured and calculated maximum floor displacements
at different levels (Fig. 7.12 through 7.15) showed that the model led to
reasonable maxima at all levels. The maximum displacements were generally
overestimated except for levels one through three of the structure MF1.
Differences were observed between the calculated and the measured maximum
relative story displacements. It can be seen that the model overestimated
the relative story displacements at ~ower stories, while it underestimated
the response at upper stories.
Results for structures FWl and FW2 are given in Fig. 7.8 and 7.9.
During large-amplitude response, the calculated values were in good
agreement with the measured response for each of these two structures.
In both cases, the calculated and measured top-level displacement maxima
were close. The model underestimated the response of structure FWl by
8%, but it overestimated the response of structure FW2 by 10%.
The performance of the model was not quite satisfactory during low-
amplitude response. In these periods of response, the waveforms were similar
but the calculated and measur~d responses did not match well.
 56

The calculated maximum story displacements at all levels were close

to the measured values (Fig. 7.16 and 7.17). No consistent trend was
recognized in comparing the measured and calculated maximum relative
story displacements.

Calculated and measured responses of the frame-wall structures to a

base motion simulating one horizontal component of the Taft 1952 record
are shown in Fig. 7.10 and 7.11. For FW3, the test structure with the
weaker wall, comparison of the calculated and measured waveforms is seen
to be satisfactory throughout the six-second period shown (Fig. 7.10). The
same is not true for the results of FW4, the test structure with the
stronger wall. The response of the test structure in the first three
seconds was considerably less than that calculated. For both structures,
the maximum single-amplitude displacement was overestimated by approximately
50%. Despite these discrepancies, the overall success of the model in
simulating the nature of the response is acceptable.

7.5 Analytical Results for Different Base Motions

Structure MFl was analyzed for seven different earthquake records con-
sidered~n two groups. The first group consisted of three records: Orion
NS, San Fernando, 1971; Castaic N21E, California, 1971; and Bucarest NS,
1977. The second group included E1 Centro NS and EW, 1940; Taft N21E, and
S69E, 1952. To have reasonable proportions between the input frequency
and the frequency of the structure, the time axis of each record was com-
pressed by 60%.
The maximum accelerations of all the motions were normalized such
that nonlinear displacements would be developed. Maximum base acceleration
is not necessarily a representative measure of the intensity of an earth-
quake. Many authors consider Housner's spectum intensity as a better
 57

index (10). However, it was not the intention of this study to compare
the response caused by different motions; rather, the objective was to
assess the performance of the Q-Model for each individual earthquake.
For the first group, the analysis was conducted using both the Q-
Model and the MDOF system (Chapter Two). Takeda hysteresis rules were
used for the MDOF analysis. The response-histories for top-level dis-
placements and base moments are presented in Fig. 7.20 through 7.26.
The maximum absolute values of the response are listed in Table 7.B.
Maximum element ductilities, obtained from the MDOF analysis, are presented
in Appendix E.
Earthquake records in the second group were similar to the simulated
motions described in Section 7.3. Therefore, results based on these
records were studied only qualitatively. No MDOF analysis was performed
for this group. The results and relating discussions are cited in
Appendi x F.
For the Orion and Castaic records, the Q-Model resulted in responses
comparable to the results of MDOF model (Fig. 7.20 and 7.2l). The fre-
quency of the response from the two models were close, and most of the
peaks occured at the same time. In both cases, the maximum top-level
displacements from the Q-Model were larger than those of MDOF system.
Along the height of the structure, for the Orion record, the results
from both models were quite close at first to fourth level (Fig. 7.24).
At other floors, the Q-Model results in larger values. Similarly, for the
Castaic record {Fig. 7.25}, larger values were calculated using the Q-
Model. In Fig. 7.24 and 7.25 it can be seen that the Q-Model resulted in
maximum relative story displacements equal or larger than those calculated
using the MDOF model.
 58

The Q-Model led to a top-level maximum displacement considerably larger

than that of the MDOF system, when the Bucarest earthquake record was used
{Fig. 7.22}. Study of the response revealed that the period of the struc-
ture, as assumed by the Q-Model, was close to the period of the input
acceleration between T ~ 1.2 to T ~ 1.8 seconds. Therefore, the structure
was in a state of near resonance during this interval. As a result, large
displacement was developed.
The MDOF model regarded the structure with a shorter period than the
period considered by the Q-Model. Hence, considerably smaller maximum
displacement was calculated using the MDOF model. To examine the validity
of the above observation, the initial period of the single-degree structure
was reduced by 10%. It was seen that, under the new condition, the Q-Model
resulted in a response comparable to the response from the MDOF model {Fig.
7.23}. In addition, maximum floor displacements and relative story dis-
placements along the height of the structure were in good agreement for
the two models {Fig. 7.26}.
It should be noted that the period of the SDOF system between T ~ 1.2
and 1.8 seconds is considerably different from the initial period (associated
with the slope of OY in Fig. 6.4). However, the initial period has an
effect on the period of the structure, at least, immediately after yield
displacements are developed. Therefore, a 10% reduction tn the intttal period
nas reduced the apparent period between T ~ 1.2 and 1.8 seconds enough so
that resonance did not occur.
The difference between the results from MDOF model and the first
solution using the Q-Model can be explained as follows: In the HOOF model,
because the stiffness of the uncracked section was recognized in moment-
rotation relationships, hysteretic energy dissipation started with low
 59

amplitudes of response. The Q-Model, using a bilinear moment-rotation

curve, did not dissipate any energy through hysteresis during low-amplitude
responses. Therefore, the Q-Mode1 resulted in relatively larger displace-
ments {at T ~ 1.4 seconds in Fig. 7.22}. Because of the large displacement,
the stiffness of the structure was reduced causing an increase in the
apparent period of the structure so that, in this particular case, the new
apparent period was close to that of the input acceleration. Hence, the
single-degree structure was in a state of resonance.

The Q-Mode1 resulted in responses reasonably close to the responses

from the MDOF model, for different motions. The results for Bucarest
earthquake records, which was not a typical motion, showed that the results
from the Q-Model need careful interpretation if the period of the input
acceleration is close to the apparent period of the system. However,
for more probable earthquakes the performance of the Q-Model was quite
satisfactory.

7.6 Analytical Results for Repeated Motions

In Reference 8, it is reported that structure H2 experienced almost
the same displacement history, when it was subjected to two identical
motions strong enough to cause inelastic deformations. In'this study, after
the first motion, the structure was allowed to come to rest before the
second motion started. To determine if such behavior can be simulated
using the Q-Model, structure MFl was subjected to five motions, each com-
prising two identical earthquake records. The base acceleration used was
the north-south component of E1 Centro, 1940.
At each case, the record consisted of two motions with the same max-
imum acceleration. The maximum acceleration was normalized to values ranging
 60

from 0.2 g to 1.6 g. The input acceleration in each case consisted of two
six-second durations and a 0.4-second quiet period in between. The quiet
period was included to separate the records. During the quiet period, any
free vibration was eliminated by setting the displacement, velocity, and
acceleration of the equivalent mass equal to zero. As a result, when the
second motion started, the structure was at rest, but with stiffness
characteristics the same with those at the end of the first motion.
The base accelerations, top-level displacements, and the base moments
are presented in Fig. 7.27 through 7.31. In each case, the response for
the second motion (between T = 6.4 to 12.4 seconds) is shown by broken
line and superimposed on the response for the first motion. The dis-
placement maxima are listed in Table 7.9. Because in some cases there was
a permanent drift, one-half of double-amplitude displacements were cited.
For the case with 0.2 9 maximum acceleration (Fig. 7.27), the apparent
frequency of the response for the second motion was smaller than that of
the first one. This showed a reduction in the structural stiffness from
the first earthquake to the second. During the first motion, major non-
linear displacement was not developed until T ~ 2.4 seconds. Beyond this
point, the structure had a smaller stiffness and longer average period.
This was seen more clearly during the first 2.4 seconds of the response for
motion two. During this period, larger displacements were developed result-
ing in further period elongation. The maximum double-amplitude displacement
for the second motion was 14% larger than that of the first one.
In the run with 0.4 g maximum acceleration (Fig. 7.28), the response
for the two motions coincided most of the time. The nonlinear displacement
started before T ~ 1.0 second of the first motion. So the structure lost
part of its stiffness early during the motion, and had an increased period
for the rest of the time. When the second motion started, differences were
 61

seen between the two displacement responses before T = 1.0 second. The
difference is attributed to the change in the stiffness characteristics of
the structure. Beyond T ~ 1.0 second, the response for the two motions
coincided. At maxima, the double-amplitude displacement of the second
response was 4% larger than that of the first one.
The above observation also applies to the cases with 0.8 g and 1.2 g
maximum accelerations (Fig. 7.29 and 7.30). The maximum double-amplitude
displacement, for motion two with 0.8 g, was 13% larger than that of the
first motion. For the case with 1.2 g, the maximum displacement was
increased by 8% in the second earthquake. Here (Fig. 7.30), the second
response exhibited some shift with respect to the time axis.
The frequency contents of the two displacement responses, obtained
from the two records with 1.6 g maximum acceleration, were close (Fig. 7.31).
However, the peak values were increased in the second response. The
double-amplitude maximum displacement of the second curve was 20% larger
than that of the first curve.
The findings in Reference 8 and the above observations suggest that
if a reinforced concrete structure has developed nonlinear deformations
(associated with the cracking of concrete and the yielding of reinforce-
ment) as a result of an earthquake, structural repair is not a necessity
if there are no bond slip or shear failure, and if a stronger earthquake
is not expected to occur during the service life of the structure.
In each of the five cases studied in this section, the Q-Model resulted
in similar responses for two consecutive records. This behavior is in
agreement with the experimental results on structure H2 which was subjected
to two identical motions (third and fourth simulated earthquakes, Reference
8).
 62

CHAPTER 8
SUMMARY AND CONCLUSIONS

8.1 Surrmary
This study consisted of two parts.· The first part was aimed at deter-
mining the sensitivity of calculated seismic response of reinforced concrete
structures to the hysteresis models used in the analysis. This part included
the development of a multi-degree analytical model for nonlinear analysis of
rectangular plane frames subjected to base excitations. In addition, two
new hysteresis systems were introduced which compensated for some of the
shortcomings, with respect to realistic response, of previously proposed
systems. The analytical model was formed so that it was able to work in
conjunction with the new hysteresis systems as well as three of the systems
used in earlier studies.
The previously proposed models were Takeda system (36), Otani model (25),
and the bilinear system. Takeda model (Fig. 3.1), which is relatively
complicated, was proposed based on experimental results on reinforced concrete
joints. Otani model was a simplified version of Takeda system (Fig. 3.5).
The bilinear model is a simple system which has been used extensively, despite
its poor correlation with experimental results (Fig. 3.6).
The two systems developed in the course of this study were Sina and Q-Hyst
models. Sina model was a version of Takeda model modified by adding pinching
effect (tendency for small incremental stiffness upon load reversal), and
simplified by eliminating some of the rules (Fig. 3.4). Q-Hyst system was,
in effect, a modified bilinear model which took into account: (1) reduction
in stiffness during unloading from the post-yielding segment of primary
 63

moment-rotation curve, (2) dependence of such reduction on the maximum

rotation experienced, and (3) reduction of stiffness at load reversal
stage (Fig. 3.7).
To study the influence of the hysteresis systems on the calculated
response of structures and to determine the system which best represented
the hysteretic behavior of the test frames, the multi-degree model was
used to analyze the small-scale ten-story three-bay structure tested by
T. J. Healey (15) using the University of Illinois Earthquake Simulator.
The results from the analytical model were evaluated assuming that the
experimental results provided a standard.
The objective of the second part of the study was the development of a
simple economical model to be used as an efficient tool for estimating the
overall seismic behavior of reinforced concrete structures undergoing
inelastic deformations. The available results of tests on numerous physical
specimens (1,5,8,15,21) served to develop and test the model. A very simple
model was introduced which treated each structure as a nonlinear IIsingle-
degree system (Q-Model) consisting of a mass, a viscous damper, a massless
ll

rigid bar, and a rotational spring (Fig. 6.1). The properties of the single-
degree model were related to those of the structure by assuming for the
structure a deflected shape corresponding to a linear lateral force distri-
bution. The backbone curve for the nonlinear spring of the Q-Model was
based on the calculated static force-displacement response for the structure.
Using the Q-Model, response histori.es for displacements at all levels and
base moment response are obtained. Computer cost for Q-Model analysis of a
 64

ten-story three-bay structure was approximately 3% of the cost for the MOOF
analysis. The proposed model was tested for a collection of eight different
test specimens including frames and walls (Fig. 4.1, 7.1 and 7.2), seven
different ground motions (Fig. 7.20 through 7.26 and F.1 through F.4), and
five repeated earthquake records (Fig. 7.27 through 7.31). The results were
compared with experimental results where available. Otherwise, the complex
model developed for the first part of the study was used to evaluate
responses calculated using the Q-Model.

8.2 Observations
a. Part One
(1) The experimental results and the response calculated based on Takeda
hysteresis model were in excellent agreement during the high-amplitude displace-
ment response. The correlation was not close during the low-amplitude response.
(2) The inclusion of "pinching" (Fig. 3.4) in the hysteresis model im-
proved the response during the small-amplitude period, while it resulted in
a larger maximum displacement.
(3) The Q-Hyst system, which is a simple hysteresis model comprising
only four rules (Fig. 3.7), resulted in an acceptable waveform for the entire
response. The calculated maximum top-level displacement was 17% larger than
the corresponding measured value (Fig. 4.8).
(4) The simple bilinear model (Fig. 3.6) resulted in a waveform different
from the measured response. The results from this model were considered to be
unsatisfactory (Fig. 4.7).
 65

b. Part Two
(1) The displacement and base moment responses of the eight different
test structures, calculated using the Q-Model, had waveforms and frequency
contents similar to those of the measured responses (Fig. 7.4 through 7.19).
For all but two structures, the calculated and measured displacement maxima
were reasonably close. This was also true for the maximum displacements at
different levels of each structure. The agreement between the measured and
calculated maximum base moments was even closer. Despite the overestimated
maxima for two cases, the overall performance of the Q-Model was satisfactory.
(2) The Q-Model resulted in reasonable responses for different earth-
quake records. It was found that, for exceptional earthquake records similar
to Bucarest 1977, the Q-Model may view the structure at a state of near
resonance and hence, result in excessive displacements (Fig. 7.20 through 7.26).
(3) When the same structure was analyzed for repeated earthquake records
with the same maximum accelerations, the response did not change significantly
from the first motion to the second. Differences were observed only in low-
amplitude-response range occurring at the beginning of the run (Fig. 7.27
through 7.31). This was in agreement with the observations reported in
Reference 8.

8.3 Conclusions
Based on the results of the study in this report the following conclusions
were reached:
a. Part One
{I} Several stiffness characteristics may be included in a hysteresis
model (e.g., reduction in stiffness upon unloading from post-yielding portion
 66

of the primary curve, pinching effect. etc.}. The extent to which the inclu-
sion of these factors affect the response may differ from large- to small-
amplitude responses. 'For example the inclusion of the pinching effect may alter
the low-amplitude response significantly. while it has relatively small effect
on high-amplitude response. Therefore. to evaluate the influence of hysteresis
models used for the analysis of a structure. the calculation has to be extended
over both the large- and small-amplitude periods of response history.
(2) The assumed hysteretic behavior can have a significant effect on
the calculated maxima, waveform, and the apparent frequency of the response
of a structure subjected to base motions. If large-amplitude displacements
are developed early during the motion, the first one or two cycles are insensi-
tive to the particular hysteresis rules used.
(3) Observed response can be simulated faithfully by using more realistic
(and correspondingly more complicated) hysteresis models. However, a reason-
able estimate of the response waveform can be obtained by using simpler
models which represent the overall energy dissipation in the joints of a
structure.
b. Part Two
(1) The displacement and base moment waveforms of a multistory rein-
forced concrete structure with columns proportioned to develop limited
yielding, subjected to earthquake motions causing inelastic deformations,
was evaluated with acceptable accuracy, using the simple model introduced
in Chapter Six.
 67

(2) Local inelastic rotation requirements in a reinforced concrete

structure may be controlled satisfactorily by controlling the lateral dis-
placement as a function of the height of the building, provided the individual
elements do not have abrupt changes in stiffness. Therefore, determination
of the lateral displacements may be adequate to check the overall performance
of a structure subjected to a given earthquake.
(3) For design, a simple and inexpensive model is highly desirable
because by using such a model
(a) several preliminary designs with varying parameters can be
examined before the final design is reached, and
(b) the performance of a given structure can be evaluated using
a wide range of ground motions.
 68

LIST OF REFERENCES

1. Abrams, D.P., and M.A. Sozen, "Experimental Study of Frame-Wall

Interaction in Reinforced Concrete Structures Subjected to Strong
Earthquake Motions," Civil Engineering Studies, Structural Re-
search Series No. 460, University of Illinois, Urbana,r1ay 1979.
2. Adeli, H., J.f'1. Gere, and W. Weaver, "Algorithms for Nonlinear
Structural Dynamics," Journal of Structural Division, ASCE,
Vol. 104, ST 2, February 1978, p. 274.
3. Anderson, J. C. and W. H. Townsend, "Mode 1s for RC Frames with De-
grading Stiffness," Journal of Structural Division, ASCE, Vol.
103, ST 12, December 1977, pp. 2361-2376.
4. Aoyama, H., "Simple Nonlinear Models for the Seismic Response of
Reinforced Concrete Buildings, U.S.-Japan Cooperative Research
II

Program in Earthquake Engineering with Emphasis on the Safety of

School Buildings, August 1975, pp. 291-309.
5. Ari sti zaba l-Ochoa, J. D., and M.A. Sozen, "Behavior of Ten-Story Rein-
forced Concrete Walls Subjected to Earthquake Motion," Civil Engineer-
ing Studies, Structural Research Series No. 431, University of
Illinois, Urbana, October 1976~

6. Aziz, T.S., "Inelastic Dynamic Analysis of BUilding Frames," Publica-

tion No. R76-37, Department of Civil Engineering, Massachusetts
Institute of Technology, Cambridge, August 1976.
7. Biggs, J.M., Introduction to Structural Dynamics, McGraw-Hill Book Co.,
1964, pp. 202-205.
8. Cecen, H.M., "Response of Ten-Story Reinforced Concrete Frames to Sim-
ulated Earthquakes," Doctoral Dissertation, Graduate College, Univer-
s ity of III i no; s, Urbana, r·1ay 1979.
9. Clough, R.W. and S.B. Johnston, "Effect of StHfness Degradation on
Earthquake Ductility Requirements," Proceedings, Japan Earthquake
Engineering Symposium, Tokyo, October 1966, pp. 195-198.
10. Clough, R.W., J. Penzien, Dynamics of Structures, McGraw-Hill Book
Company, 1975, pp. 227-232.
11. Emori, K., and W. C. Schnobri ch, "Ana lys i s of Reinforced Concrete
Frame-Wall Structures for Strong Motion Earthquakes," Civil Engi-
neering Studies, Structural Research Series No. 457, University of
Illinois, Urbana, December 1978.
12. Gavlin, N., "Bond Characteristics of Model Reinforcement," Civil
Engineering Studies, Structural Research Series No. 427, University
of Illinois, Urbana, April 1976.
 69

13. Gi berson, M. F., liThe Response of Non 1i near t~ulti story Structures
to Earthquake Excitation,1I Earthquake Engineering Research Laboratory,
California Institute of Technology, Pasadena, California, May 1967.
14. Goel, S.C., and G.V. Berg, IIIne1astic Earthquake Response of Tall
Steel Frames," Journal of Structural Division, ASCE, Vol. 94,
ST 8, August 1968, pp. 1693-1711.
15. Healey, T.J., and M.A. Sozen, IIExperimental Study of the Dynamic
Response of Ten-Story Reinforced Concrete Frame with a Tall First
Story," Civil Engineering Studies, Structural Research Series No. 450,
University of Illinois, Urbana, August 1978.
16. Jennings, P.C. and R. Husid, "Collapse of Yielding Structures During
Earthquakes,1I Journal of Engineering r-~echanics Division, ASCE, Vol.
94, EM 5, October 1968, pp. 1045-1065.
17. Kanaan, A.E., G.H. Powell, "General Purpose Computer Program for In-
elastic Dynamic Response of Plane Structures," EERC 73-6, Univer-
sity of California, Berkeley, April 1973.
18. Kreger, M.E., and D.P. Abrams, "Measured Hysteresis Relationships for
Small-Scale Beam-Column Joints," Civil Engineering Studies, Struc-
tural Research Series No. 453, University of Illinois, Urbana, August
1978.
19. Livesley, R.K., Matrix ~1ethod of Structural Analysis, Pergamon Press,
New York, 1964, pp. 106-116.
20. ~1cNamara,J.F., "Solution Schemes for Problems of Nonlinear Structural
Dynamics," Transactions of the ASME, Pressure Vessels and Piping
Division, May 1974, pp. 96-102.
21. Moehle, J.P., and M.A. Sozen, "Earthquake-Simulation Tests of a Ten-
Story Reinforced Concrete Frame with a Discontinued First-Level Beam,"
Civil Engineering Studies, Structural Research Series, No. 451, Univer-
sity of Illinois, Urbana, August 1978.
22. ~1ondkar,D.P., and G.H. Powell, "ANSR-1, General Purpose Program for
Analysis of Nonlinear Structural Response," EERC No. 75-73, University
of California, Berkeley, December 1975.
23. Newmark, N.M., "A Method of Computation for Structural Dynamics," Journal
of Engineering Mechanics Division, ASCE, Vol. 85, EM3, July 1959,pp.69-86.
24. Newmark, N.M., and E. Rosenbluth, Fundamentals of Earthguake Engineerinq,
Prentice-Hall, Inc., 1971, pp~ 321-364.
25. Otani, S., "SAKE-A Computer Program for Inelastic Response of RIC Frames
to Earthquake," Civil Engineering Studies, Structural Research Series
No. 413, University of Illinois, Urbana, November 1975.
 70
26. Otani, S., and t1. A. Sozen, IIBehavior of t1ul tistory Reinforced Concrete
Frames During Earthquakes,1I Civil Engineering Studies, Structural
Research Series No. 392, University of Illinois, Urbana, November
1972.
27. Otani, S., and M,A. Sozen, IISimulated Earthquake Tests of RIC Frames,1I
Journal of Structural Division, ASCE, Vol. 100, No. ST 3, r,1arch
1974, pp. 687-701.
28. Park, R., and T. Paulay, Reinforced Concrete Structures, John Wiley
and Sons, Inc., 1975, pp. 11-15 and 312-411.
29. Perry, E. S., and N. Jundi, IIpullout Bond Stress Distribution Under
Static and Dynamic Repeated Loadings,1I ACI Journal, ~1ay 1969,
pp. 377-380.
30. Pique, J. R., liOn the Use of Simple t10dels in Nonlinear Dynamic
Ana1ysis,1I Publication R76-43, Department of Civil Engineering,
r~assachusetts Institute of Technology, Cambri dge, September 1976.

31. Powell, G. H., and D. G. Row, II Infl uence of Analysis and Design
Assumptions on Computed Inelastic Response of t10derately Tall
Frames,1I EERC 76-11, University of California, Berkeley, April
1976.
32. Sozen, M. A., IIHysteresis in Structural Elements,1I Applied ~1echanics
in Earthquake Engineering, ASME, t1MD-Vol. 8, November 1974, pp.
63-73.
33. Suko, M. and P. F. Adams, IIDynamic Analysis of r4ultibay Multistory
Frames,t1 Journal of Structural Division, ASCE, Vol. 97, ST 10,
October 1971, pp. 2519-2533.
34. Tada, T., T. Takeda, Y. Takemoto, IIResearch on Reinforcement of Beam-
Column Joint Panel for Seismic Resistant Reinforced Concrete Frame
(Part 1),11 (in Japanese), Report of the Technical Research Institute,
OHBAYASHI-GUMI, LTD., No. 12, 1976, pp. 33-37.
35. Takayangi, T., and W. C. Schnobrich, IIComputed Behavior of Reinforced
Concrete Coupled Shear Wall s, n Civil Engineering Studi es, Structural
Research Series No. 434, University of Illinois, Urbana, December
1976.
36. Takeda, T., t4. A. Sozen; and N. N. Nielsen, IIReinforced Concrete
Response to Simulated Earthquake,1I Journal of Structural Divsion,
ASCE, Vol. 96, ST12, December 1970, pp. 2557-2573.
37. Tansirikongkol, V. and D. A. Pecknold, IIApproximate t10dal AnalYSis
of Bilinear ~1DF Systems Subjected to Earthquake r~otions," Civil
Engineering Studies, Structural Research Series, No. 449, University
of Illinois, Urbana, August 1978.
 71
38. Umemura, H., H. Aoyama, and H. Takizawa, IIAnalysis of the Behavior of
Reinforced Concrete Structures During Strong Earthquakes Based on
Empirical Estimation of Inelastic Restoring Force Characteristics of
Members,1I Proceedings, Fifth World Conference on Earthquake
Engineering, Rome, Italy, June 1973, pp. 2201-2210.
39. Walpole, W. R., and R. Shepher~ IIElasto-Plastic Seismic Response of
Reinforced Concrete Frame,1I Journal of Structural Division, Vol. 95,
ST 10, October 1969, pp. 2031-2055.
 TABLE 4.1 LONGITUDINAL REINFORCING SCHEDULES FOR r1Fl AND r·1F2

Number of No. 13 9 Wires Per Face

Test Structure MFl Test Structure MF2
Interior Exterior Interior Exterior
Level Beams Columns Col umns Beams Col umns Columns

10 2 2 2 2 2 2
9 2 2 t
8 2 2
7 3 3
-...J
6 N

5
4 +
3 2 2
2 3 2 2 4
1 3 3 3 3 4 4
 73

TABLE 4.2 ASSut1ED f-1ATERIAL PROPERTIES FOR

r-1Fl and MF2

Concrete
f~ = Compressive strength 38. d r1PA
ft = Tensile strength 3.4 MPA
£
o
= Strain at f'c 0.003
£u = Strain at ultimate point 0.004
Ec = Young IS r-1odul us 20,000 r~PA

Steel
fSY = Yield stress 353 r1PA
Es = Young's Modulus 200,000 t-1PA
£sh = Strain at strain hardening 0.0018
fsu = Ultimate strength 372 r·1PA
£su = Ultimate strain 0.03
 74

TABLE 4.3 COLUHN AXIAL FORCES DUE TO DEAD LOAD

Level Nominal Force Assumed Force

(kN) (kN)

10 0.57 1.0
II
9 1.14
8 1. 70 II

II
7 2.27
6 2.84 3.2
II
5 3.41
II
4 3.98
II
3 4.55
2 5.12 5.2
1 5.70 5.2
 TABLE 4.4 CALCULATED STIFFNESS PROPERTIES OF
CONSTITUENT ELEMENTS OF STRUCTURES MFl and MF2
1. BEAMS AND THIRD TO TENTH STORY COLUMNS

Member Me M S** S** elt

(EI)~ncraeked y 2 3 c eylT eult
( Leve 1) (kN-r·, 2) (kN-t~) (kN-M) (kN-M 2) (kN-r~2 ) Rad Rad Rad

Beams 3.48 0.027 0.119 1. 31 0.039 0.0002 0.0033 0.0045

(l~7)

Beams 3.48 0.027 0.082 0.96 0.033 0.0004 0.0033 0.0052

(~1O)

Columns 8.40 0.079 0.179 2.84 0.045 0.0004 0.0021 0.0026

(3~) ......
(J'1

Columns 8.40 0.061 0.136 2.35 0.040 0.0004 0.0021 0.0029

(7~10)

* Effect of reinforcement not included

** See Fig. 2.3
t Rotations due to bond slip
e~ = Rotation corresponding to moment at €c = 0.004
 TABLE 4.4 CALCULATED STIFFNESS PROPERTIES OF CONSTITUENT
ELEMENTS OF STRUCTURES HFl AND MF2 (Continued)

2. COLUMNS (Levels 1 and 2)

s** s** e'T .!.

(EI)*uncracked e' e'+
Structure Level Mc r1y 2- 3 c y u
2
(kN-t4 ) (kN-M) (kN-M) (kN-H2 ) (kN_t~2 ) (Rad) (Rad) (Rad)

2 8.40 0.079 0.179 2.84 0.045 0.0004 0.0021 0.0026

Ext. II II
1 0.088 0.268 3.06 0.066 0.0002 0.0024
MFl II II II II II II II II
2
Int. II II II II II II II II
1
II II II II
2 0.321 4.52 0.076
-II
0.0002
Ext. II II II II II II II
1 ""'-I
0\
MF2 II II
2 0.079 0.179 2.84 0.045 0.0004 0.0026
Int. II
1 " 0.088 0.321 4.52 0.076 0.0002 0.0024

* Effect of reinforcement not considered

** Slopes of cracked and yielded section (Fig. 2.3)
t Rotations due to bond slip
e~ = Rotation corresponding to EC = 0.004
 77

TABLE 4.5 CRACK-CLOSING MOMENTS

USED FOR SINA HYSTERESIS
~10DEL

Unit = kN-M

l-BEAMS
Moment
Levels Exterior End Interior End

1-7 0.050 0.016

8-10 0.033 0.010

2-COLU~1NS

~1oment

Columns with 3 bars/face 0.160

Columns with 2 bars/face (levels 2-10) 0.107
 78

TABLE 4. 6 ~4EASUREDAND CALCULATED ~1AXmUM

RESPONSE OF t1F2 RUN 1

Level Displacement (mm) Acceleration (9)

f-1easured Calculated r1easured Calculated

10 24.4 24.1 0.59 0.57

9 23.4 23.4 0.48 0.49
8 22.8 22.7 0.43 0.45
7 21. 6 21. 7 0.39 0.44
6 19.7 20.3 0.38 0.48
5 17.3 18.5 0.35 0.44
4 14.3 15.7 0.39 0.40
3 12. 1 12.4 0.43 0.49
2 7.4 8.5 0.40 0.53
1 3.8 4.2 0.34 0.39
Base 0.38

Measured = 22.0
Base Moment
(kN-M) Calculated = 21.5

Measured = 13.0
Base Shear
(kN) Calculated = 14.0
 79

TABLE 4.7 . r~EASURED AND CALCULATED MAXIMUM RESPONSE

OF MFl USING DIFFERENT HYSTERESIS SYSTEMS

1. DISPLACEMENTS (mm)

Level ~1easured
Takeda Sina* Otani* Bil inear Q-hyst
10 23.6 23.1 28.2 31.4 20.7 27.8
9 22.8 22.5 27.3 30.9 . 20.3 27.0
8 21. 3 21. 7 26.3 30.0 19.7 25.9
7 20.7 20.6 24.8 28.9 19.2 24.2
6 18.6 19.1 22.0 27.4 18.5 22.0
5 16.7 17.1 19.1 25.4 17 .0 19.3
4 14.4 14.3 16.1 21. 7 14.4 15.8
3 12.3 10.9 12.1 16.4 10.9 11.8
2 8.3 7. 1 8.0 10.5 7.1 7.6
1 4.8 3.5 4.0 5. 1 3.4 3.6
Base
Moment 20.8 21.6 22.1 21.4 20.0 22.0
(kN-M)

*Measured and calculated maxima occur at different times

 80

TABLE 4.7 .r~EASURED AND CALCULATED HAXH1UM RESPONSE OF r~F1

USING DIFFERENT HYSTERESIS SYSTD1S (Continued)

2. ACCELERATIONS (9)

Level f1easured Calculated

Takeda Sina Otani Bilinear Q-hyst

10 0.76 0.62 0.68 0.61 0.60 0.56

9 0.60 0.49 0.50 0.46 0.47 0.51
8 0.51 0.46 0.48 0.47 0.53 0.49
7 0.49 0.50 0.51 . 0.48 0.48 0.43
6 0.41 0.48 0.51 0.43 0.48 0.42
5 0.40 0.41 0.44 0.38 0.46 0.40
4 0.43 0.51 0.54 0.50 0.37 0.48
3 0.46 0.56 0.53 0.56 0.45 0.49
2 0.50 0.45 0.37 0.41 0.48 0.33
1 0.40 0.35 0.35 0.35 0.43 0.30
Base
Shear 15.6 14.2 14.3 13.6 12.8 13.0
(kN)
 81

TABLE 7.1 COLUMN AXIAL FORCES FOR

STRUCTURES Hl, FW1, AND FW2

Unit = kN

Level Nominal Dead Assumed Axial Force

Load Hl FWl &FW2

10 0.57 1.2 0.0

9 1.14 1.2 0,0
8 1. 70 1.2 2,2
7 2.27 1.2 2.2
6 2.84 1.2 2.2
5 3.41 1.2 2.2
4 3.98 4.5 4.5
3 4.55 4,5 4.5
2 5.12 4.5 4.5
1 5.70 4.5 4.5
 82

TABLE 7.2 CALCULATED STIFFNESS PROPERTIES OF

CONSTITUENT ELEMENTS OF STRUCTURE H1

t t
Member (EI*)uncracked Mc My S2 S3
(Leve 1)
(kN-M 2) (kN-M) ( kN-~1) (kN-M 2) (kN-M 2)

Beams 3.48 0.027 0.] 07 0.90 0.027

(1+4)
Beams 3.48 0.027 0.078 0.70 0.023
(5+10)
Ext. Columns 8.40 0.072 0.628 6.04 0.103
(1+4 )
Ext. Columns 8.40 0.054 0.357 3.92 0.052
(5+10)
Int. Columns 8.40 0.073 0.530 5.25 0.087
(1+4)
Int. Columns 8.40 0.055 0.190 2.32 0.033
(5+10)

* Effect of reinforcement not included

tSee Fig. 2.3
 83

TABLE 7.3 CALCULATED STIFFNESS PROPERTIES OF

CONSTITUENT ELEMENTS OF STRUCTURE FWl

st i-
Member (EI )~ncracked c M My 2 S3
( Leve 1) (kN-M 2) (kN-M) (kN-M) ( kN-M 2) (kN-M 2)

Beams 3.35 0.026 0.080 0.89 0.029

(1+4 )
Beams 3.35 0.026 0.116 1.29 0.032
(5+9)
Beams 3.35 0.026 O.OBO 0.89 0.029
(10)
Ext. &Int. Columns 8.11 0.085 0.199 2.B4 0.047
(1+4 )
Ext. &Int. Col umns 8.11 0.067 0.159 2.65 0.038
(5+8)
Ext. Columns 8.11 0.047 0.117 2.09 0.037
(9+10)
Int. Columns B.l1 0.047 0.171 2.90 0.042
(9+10)
Wall 520. 0.76 13.7 515. 10.4
(1 +4)
Wall 520. 0.76 7.93 460. 5.59
(5+6)
Wall 520. 0.76 4.24 350. 2.73
(7+10)

* Effect of reinforcement not included

t See Fig. 2.3
 84

TABLE 7.4 CALCULATED STIFFNESS PROPERTIES OF

CONSTITUENT ELEMENTS OF STRUCTURE FW2

Member (E1)*uncracked ~1 My
st st
c 2 3
(Leve 1) (kN-~12 ) (kN-M) 2
(kN-r~ ) (kN-M 2 )
(kN-M)

Beams 4.00 0.029 0.088 0.99 0.031

(1+2)
Beams 4.00 0.029 0.118 1. 31 0.040
(3+7)

Beams 4.00 0.029 0.088 0.99 0.031

(8+10)
Ext. Column 9.66 0.090 0.202 3.21 0.050
( 1+3)
In t . Co 1umns 9.66 0.090 0.255 3.92 0.061
( 1+3)
Ext.&Int. Col. 9.66 0.072 0.162 2.75 0.041
(4+8)
Ext.&Int. Col. 9.66 0.053 0.118 2.19 0.036
(9+10)
Wall 731.1 0.85 4.23 350. 2.67
(1+10)

* Effect of reinforcement not included

"I" See Fig. 2.3
 TABLE 7.5 CALCULATED PARAMETERS FOR DIFFERENT STRUCTURES

Equivalent Equivalent M 2
(M*) x 10
Structure Mass Height Sl S2 Frequency
at break (cycle/sec.)
(kN/g) (M) point

Hl & H2 3.69 1.58 25. 48. 9. 17.

MFl 3.68 1.59 29. 64. 8. 20.
~1F2 3.60 1.59 29. 64. 8. 20.
FWl & FW4 3.36 1.64 33. 113. 29. 27.
FW2 & FVJ3 3.36 1.63 38. 93. 12. 25.

co
0'1
 TABLE 7.6 ASSUMED DEFORMED SHAPES FOR DIFFERENT STRUCTURES

Level H1 &H2 MF1 MF2 FW1 & FW4 FW2 & FW3

10 1.0 1.0 1.0 1.0 1.0

9 0.98 0.97 0.97 0.93 0.92
8 0.95 0.92 0.92 0.85 0.83
7 0.88 0.86 0.87 0.75 0.74
6 0.79 0.79 0.79 0.64 0.63
5 0.66 0.69 0.70 0.51 0.51 00
0'1
4 0.52 0.57 0.59 0.37 0.39
3 0.37 0.43 0.46 0.24 0.26
2 0.22 0.27 0.29 0.12 0.15
0.08 0.13 0.13 0.03 0.05
 TABLE 7.7 MAXIMUM ABSOLUTE VALUES OF RESPONSE
Displacement Unit = mm

Hl RUN 1 H2 RUN 3 ~1F1 RUN 1 MF2 RUN 1

Level
t4easured Calculated Measured Calculated r~easured Calculated Measured Calculated

10 29.2 31. 7 24.5 30.1 23.6 28.1 24.4 31.1

9 29.0 31.1 24.7 29.5 22.8 27.3 23.4 30.2
8 26.0 30.1 22.2 28.6 21. 3 25.8 22.8 28.6
7 24.3 27.9 20.8 26.5 20.7 24.2 21.6 27.1
6 21. 2 25.0 17.5 23.8 18.6 22.2 19.7 24.6
co
-....J
5 17.2 20.9 13.2 19.9 16.7 19.4 17 .3 21.8
4 13.7 16.5 10.1 15.7 14.4 16.0 14.3 18.3
3 9.0 11. 7 7.0 11 .1 12.3 12.1 12.1 14.3
2 5.3 7.0 4.2 6.6 8.3 7.6 7.4 9.0
1 2.0 2.5 1.7 2.4 4.8 3.7 3.8 3.8
 TABLE 7.7 (CONTO.) MAXIMUM ABSOLUTE VALUES OF RESPONSE
Displacement Unit = mm

FWl RUN 1 FW2 RUN 1 FW3 RUN 1 RW4 RUN 1

Level
Measured Calculated Measured Calculated r~easured Calculated Measured Calculated

10 28.2 26.0 28.4 31.2 18.7 27.0 21. 5 34.1

9 26.5 24.2 25.6 28.7 17.4 24.8 19.7 31. 7
8 23.8 22.1 23.6 25.9 15.0 22.4 17.2 29.0
7 20.5 19.5 20.6 23.1 13.0 20.0 15.0 25.6
00
00
6 17.0 16.6 17.3 19.7 10.8 17.0 12.3 21.8
5 13.5 13.3 14.2 15.9 8.8 13.8 9.8 17.4
4 9.5 9.6 10.7 12.2 6.8 10.5 7.1 12.6
3 7.1 6.2 8.3 8.1 4.8 7.0 4.9 8.2
2 4.1 3.1 5.1 4.7 3.0 4.0 2.8 4.1
2.0 0.8 2.3 1.6 1.4 1.3 1.2 1.0
 TABLE 7.8 MAXIMUM RESPONSE OF STRUCTURE t·1F1
SUBJECTED TO DIFFERENT EARTHQUAKES
Unit = mm

Orion Castaic N21E 71 Bucarest 77

Level Q-~1odel
MDOF Q-f1odel ~1DOF Q-Model ~iDOF Original Frequency Increased Frequency

10 13.5 17.3 10.8 14.2 16.9 30.7 18.7

9 13.2 16.8 10.5 13.8 16.6 29.8 18.1
8 12.8 15.9 10.1 13.1 16.2 28.2 17.2
7 12.3 15.0 9.6 12.2 15.7 26.7 16.3
6 11.7 13.6 8.9 11.2 14.9 24.3 14.8 00
\0

5 11.0 11.9 7.8 9.8 13.7 21.2 12.9

4 9.0 9.8 6.4 8.1 11.9 17.5 10.7
3 7.0 7.4 4.7 6.1 9.5 13.2 8.0
2 4.7 4.7 3.0 3.8 6.6 8.3 5.0
1 2.3 2.3 1.4 1.8 3.4 4.0 2.4
 90

TABLE 7.9 MAXIMUM TOP-LEVEL DISPLACEMENTS* FOR

STRUCTURE MFl SUBJECTED TO REPEATED MOTIONS
Unit = mm

Max. Base Displacement Disp./Height

Acceleration Motion 1 r~otion 2 Difference Motion 1 Motion 2

0.2 9 13.5 15.4 +14% 0.6% 0.6%

0.4 9 21.4 22.2 + 4% 0.9% 0.9%
0.8 9 37.2 42.0 +13% 1.6% 1.8%
1.2g 64.9 70.0 + 8% 2.7% 2.9%
1.6g 94.0 112.0 +20% 3.9% 4.7%

* (Double Amplitude)/2

TABLE 7.10 WIRE GAGE CROSS-SECTIONAL PROPERTIES

Gage No. Diameter Cross-Section Area

(mm) (mm 2)

2 6.67 34.92
7 4.50 15.87
8 4.11 13.30
10 3.43 9.23
13 2.32 4.24
16 1.59 1.98
 91

f'c ------------=..,;;;;;;;;.--,."""-

\ fc::f~[I-Z(Ec-Eol
I Z=IOO
II)
I
II)
I
-
CD
~
I
en
I
I
I
I
I
Eo

o 0.00 ( 0.002 0.003 0.004

Stra in
Fig. 2.1 Idealized Stress-Strain Curve for Concrete

-
fSY--'~--------------~I~--------------------~
I I
II)
II)
I I
I I
-
CD
~

en I I
I I
I I
I I
I I
I I
I I
I I
I I
Esy

Sf ro;n
Fig. 2.2 Idealized Stress-Strain Curve for Steel
 92
M - c/> Curve for Axial Load P

My S3

-
c:
Q)
e
o
~

Me

Curvature
Fig. 2.3 Idealized Moment-Curvature Diagram for a Member

Moment

Curvature

I I
• Uncracked.Crocked II
. . . . Y·lei d e d

Fig. 2.4 Moment and Rotation along a Member

 93

Primary Curve

------=--=-=~~--
Mu

My

-
~
CD
E
o
:E

By Bu
Rotation
Fig. 2.5 Moment-Rotation Dia~ram for a Member

_ _ _ _ _--+fs

--- ---
~

-"? Average Bond

"0 Stress u

-- -- ---
Fig. 2.6 Rotation due to Bond Slip
 94

Rotational Spring

Rigid Zone

A~~------~--~~~----------~--B

.1'

Fig. 2.7 Deformed Shape of a Beam Member

Rigid Portion

Fig. 2.8 Equilibrium of a Rigid-End Portion

88

Fig. 2.9 Deformed Shape of a Column Member

 95

Biased Curve Converge To Wrong Resu It

Primary Curve To Be Used

Deformation

Fig. 2.10 Biased Curve in Relation to the

Specified Force-Deformation Diagram

Primary Curve To Be Used

Deformation

Fig. 2.11 Treatment of Residual Forces in the Analysis

 96

Xi+1
Girder at
VI + 1
Level i + I

+
.s::. Column at Level i+1

Xi V i +1

Q;' Vi

.s::.

Xi-! Vi

Fig. 2.12 Equivalent Lateral Load to Account

for Gravity Effect
 97

Primary Curve

Deformation

u'm

Fig. 3. 1 Takeda Hysteresis Model

 98

.
Q)
o
If

Deformation

Fig. 3.2 Small Amplitude Loop in Takeda Model

.
CI)
o
If /

Deformation

Average Stiffness .4I'Itt---Averoge Stiffness

When Pinching When Pinching Is
Is Considered Neglected

Fig. 3.3 Comparison of Average Stiffness with and

without Pinching for Small Amplitudes
 99

Primary
Curve

Deformation

Fig. 3.4 Sina Hysteresis Model

 100
Pri mary

Deformation

u~

Fig. 3.5 Otani Hysteresis Model

Primary Curve

Deformation

Fi g. 3.6 Simple Bilinear Hysteresis System

 101

Primary Curve

Deformation

Load Reversal

u:n

Fig. 3.7 Q-Hyst Model

 102

305 "I' 305 305 .. ,

Typical JOint
Reinforcement:
No.160 Wire
ty I
D
~- =-===t::::t::::
Typ~.1 Sho.,
Reinforcement;
No. 16 0 Wire -
~
:

flJ
/

I
:D.
D )-Tub"
/ 13 O:~

-
-- r _____
Typical Flexural - o
Reinforcement: ~
No. 130 Wire
CutOff For
Interior Column
~
St'el --~~--------
~
~
--<:::"'t:~t::::==:H=:t:~ 0
D CD

G:-~=~=h:o
~ - yl--ThiS Beam
Cut Off For / In MFI Only
Exterior
Column -1,-~-""~-C::ti====~=t:=~O::~I=~
...... V I .0.-+_ _--.

~"'~~ i

,--~-----'\~-+~13~72-------~O--5-J~~
~ Column F lexurol '\.
Reinforcement Welded \
To 102 x 51 x 3 'l ----"
Fi g. 4.1 Reinforcement Detai 1 and Dimensions of Structures MF1 and MF2
 r Telt 5tructwe
(All Di_ioIt, In Met.,,)
~ .. ~, --~

5t..1 Alfe_

F'_~

....
N'

a
w

o o

I. 091 2.29 091 0.91 Q9I

Fig. 4.2 Test Setup for Structure MFl

 104
MF2. RUN 1.

BASE ACCELERATION
.2
[ G )
o.

-.2
;-____~~______~4-______~--------+_------_+------__4TIME. SEC.
a 2 6
1. 3 5
BASE SHEAR
1.0 [ KN )

-1.0 - - - - CALCULATED
MEASURED
o 2 6
1. 3 5

BASE OVERTURNING MOMENT

1.0 KN-M ]
a
-1.0
_20
a 2 e
1. 9 5
DISPLACEMENT AT LEVE~ 1.0

1.0 C MM )

a
-:1.0

-20
o 2 e
1 8 5

ACCELERATION AT LEVEL 1.0

.9 r G
o.
-.9

o 2 e
1 a 5

Fig. 4.3 Measured and Calculated Response for MF2

 1'tF2 RUN1 TAKEDA HYSTERESIS

AT LEVEL B
10
10 ~ [ "" I
A A A A/\ AA A A f\ A f\ 1\
0 0

-10 -10

-20 -20 r V I T1:I£. SEC.

0 I I
2 . II 0 2 .I II
1 5 1 a 5
•
1\ f\ LEVEL II
10 ..
,. t oA A/\ f\ /\ f\ ~ A Ar 0

:10
OV\r\r~ \{\]"V \J\JVl] -10

_20
II 2 . Ii 0 2 .. II
1 5 1 a 5
10 "t"
• LEVEL of
" --'
0
0 t "'.., II I I I I I I I I ,_ ;I
,< I I I \ U"l

-10 i- \I \I
:,.WI
0 2 of II
0
1
2
• of
5
e
~~i!=
1 a 5

LEVEL 2

o , d \I I I I , I I I , I ; 't ~ _ J - '" I , I \ I

-5T \I \I :,I~~~
0 e of II 0 2 of II
1 5 1 a 5
•
CALCULATED
I1EASURED

Fig. 4.3 (cont'd). Measured and Calculated Response for MF2

~I! RUN1

.1!5 t ~ ~ ~ ~ A 1\ ~ ~ M 1\ AA A~EL~AT%ON AT ~E:E~ 8

O.
:: 1-116 ~ ~ A AA .A~ J AA
::2 'VIIl/lrT\fV ~~~ V\{\ -.1!5
I WI'· u ~ I I TDtE. HC.
0 2 4 e 0 I! .
1 a 5 1 S •
•
l LEVEL •
.It' 1 .II *.
~~ I.. AA 1\ h.l\ Itdtu,N\ A "
n O.

11 , , _.I!S
-.I! t \I U " •

0 2 .. C!I 0 2 .
1 S 1 S •
• •
--'
0
LEVEL .. O"'i
.I!' T A ~ ~I A~ If\ ~. M JLlA fIh A h.
O•

~ ~1~~-r=
::2··l·~
0
~ .M!wAA~ b.W/j:
~ ~~rn!VQVf~ 'rtf
I! 4
..,.25

2. 5 • 0 I! 4
• 1. S •
•5 •
.2 .1e5
LEVEL I!
L. ~n I MU. ft II.~. A. I AAA I. ..1\ .t.. uA
O. O.

-.1t5 . -- - ---- -.---

I I
-.2 0 Ie 4 IS 0 2 4· e
1 8 5 1 8 5

CALCULATED
I1£A8UltED

Fig. 4.3 (cont'd). Measured and Calculated Response for MF2

 107
MF~. RUN~. TAKEDA HYSTERESIS
•• ACCELERATION
.2
( G )
o.
-.2
TIME. SEC.
o 2
s 5

BASE SHEAR
~o
r KN )

_ _ _ CALCULATED
-:10 _____ MEASURED

o 2
1. 3 5
eo BASE OVERTURNING MOMENT
10 [ KN-M )

o
-:10

-eo
o 2
1. s
20 AT LEVEL ~O

10 [ MM ]

-20
o 2
1 s 5

.6 ACCELERATION AT LEVEL ~O

.3 [ Q

o.
-.3

-.6
o 2
:1 3 5

Fig. 4.4 Measured and Calculated Response for MFl Using

Takeda Hysteresis Model
...1 MIN:I. TAlCEDA WY8T1!JtEU8

15 AT LEVEL 8
10
r ... I

0 o
-10
-15
-20 ~._C.
r
0 e 4 o 2 4
1. 5 1 5
• •
10 10
LEVEL 8
0 o
-10 -10

0
l 2 4 o
I.
2 4
s
1. s •
•
10
LEVEL 4 --'
0 o
co
-10
~ 0 2 4
I. • S
0
1. • •
4
s
:J~~~
LEVEL I!

·1.AAA~AA ~~
o e 4
:. ~:'v:Y
o
1.
e
•
4
s
:.t~~: I.
• s

11EA8UIII£D
CALCULATED

Fig. 4.4 (cont'd). Measured and Calculated Response for MFl

Using Takeda Hsyteresis Model
~1 RUN1 TAKEDA HV8TERE8Z8

.4

fI AT
It";. ~~R~T%,* ~:E~ •
D.
"'j "~O A 1\
-.4
~
::.. -"V~VQ~
0 2 4
TZHE. _c.
0 2 4 1. 5
1. 8 5 • •
.4

.2 ~VE~ •
D.
' '1 ~"A AAll 6 ~ ~ ~ A
-.2 ::•• ~'V~~.
-.4 0 2 ..
0 e 4 8 s
1. 8 5 •
.4

.2 --'
~EVIE~ 4 o
1.0
D.

-.2
~ e 4
0 2 .. 5
1. 5 • •
•
~EVIE~ 2

0: :
1
-:e ~ V~! ~ V "VQl ' ~
!.~d~~.~j~.,~
D e . .
1 8 5
. • I!
8,
. 5

CA~CULATED
/1£A8UftED

Fig. 4.4 (conti d). Measured and Calculated Response for MFl
Using Takeda Hsyteresis Model
 110

MF~ RUN1. 8ZNA HY8TERE8Z8

ACCELERATZON
.2 ( Q )

O.

-.2
TIME. SEC.
o 2
1 s 5

BASE SHEAR
10 l KN )

o
- - \ ' - CALCULATED
-10 ------- MEASURED

o e
1 a 5
BASE OVERTURNING MOMENT

10 [ KN-H )

o
-10

-eo
o 2
1 a 5
OISPLACEMENT Af LEVEL 10
eo
t HH )
o

-eo
o 2
1 a 5
ACCELERATION AT LEVEL 10
.IS
( Q
.8
O.
-.8

-.IS
o 2
5

Fig. 4.5 Measured and Calculated Response for MFl Using

Sina Hysteresis Model
~1 _1 8XNA HY8TERE8%8

OI8P~ACEMENT AT LEVEL B
15
10 [ I1f1 I

o o
-10
-15
-20 t v
o 2 4 2 4 TXI£. SEC.
1 5 o 5
• 1 •
15
10
LEVEL e
o o

-10
-15

o 2 4 D 2 4
1 5 1 5
• •
10
10
LEVEL .. --'
o o --'
--'

-10
-10
D 2 4
D 2 4 1 5
1 5 •
•
LEVEL 2

0 2 4
:·1 o
~m!:'v:!
1
2
s
4
5
:.I~~ 1 8 5

HEA8UR£O CALCULATED

Fig. 4.5 (cont'd). Measured and Calculated Response for MFl

Using Sina Hysteresis Model
",,1 RUM1 8%MA HY8TE1llhl%8

... .e5
I Q I
o. o.
-.25 T%I1£. BEC.
- ... I I o 2 ..
o 2 .. 1 5
1 5 •
... • .5

.2 .25
LEVEL •
o. o.
-.2 -.e5
-... + ~
o 2 . o It .
1 5 1 5
• .5
• --'
... N
.2 .25
LEVEL ..
o. o.
-.2 - •e5 r ... I . . , ,. I
o e .
o 2 . 1 5
1 5 •
•
.e5
LEVE~ 2

o.
-.2 vv'r vih
: !1dv~MO~"~
c 1
~
•
.. 5
-.25+
o
,.
1
eI
I
8
. 5

I1£A~ C"LCULATED

Fig. 4.5 (cont'd). Measured and Calculated Response for MFl

Using Sina Hysteresis Model
 113

MF1 ~UN1. OTANI HV'TE~E'% •

... ACCELERATION
.1 [ G J

O.

-.1
TIME. SEC.
e
1 I 5

10 BASE SHEAR
[ I<N J
o

-10 - - - CALCULATED
------MEASUREO
o
1 IS
BAlE OVERTURNING MOMENT
10 t I<N-M J

o
-10
-10
o~------~~~--~------~------4-------4--
10
• 5
AT LEVEL 1.0
r MM ,
o

-10

.1 ACCELE~ATION AT LEVEL 10
.1 [ G )

O.
-.1
-.1 ~ ______ ~ ______ ~ ______ ~ ______ ~ ______ ~

o a
•
Fig. 4.6 Measured and Calculated Response for MFl Using
Otani Hysteresis Model
",,1 RUN1 DTANJ: ~

DISPI..ACEI1£NT AT LEVEL 8

( "" )
10 10

D 0

-10 -10

-20 -eo
r
0 2 ..
1 5 -8CI u T%fC • .e.
• 0

10 10
1
LEVEL •
0 0

-2.0 -10

l -eo
0 2 4
1 5 ......
• o a 4 ......
2.0 ..j::o

LEVEL 4
0
A'; A A: I\~ flat.
-10
'~:~"I ~_~IF
~
0 a 4 o a 4
1 5 1 5
• •
5
LEVEL 2
o
'I.AAA~AA ~~ -5
:.~':: -2.0+ v
0 2 4 o 2 4
1 8 5 1 5
•
I1£AauRED CALCULATED

Fig. 4.6 (cont/d). Measured and Calculated Response for MFl

Using Otani Hysteresis Model
~:I. RIJN:I. DTANI HyaTERESIa

.4

~ A" .-~"'~ .n~ ,

o.
. . t ". A ~ i\ ~ "'~ ' "
_.:~V V\{V LV}! vv,,: TZHE • •ec.
0 2 "
:I. • 5
2 4
:I. a 5

LEVEL a
~ ~~~ AM
_.:~ w~ ~ VJVVy~ V ~
··'f=.AA~AA
0 2 ..
2 .. :I. • 5
:I. a 5
-'
--'
U1

LEVEL ..
M . .A" I
::25· "VV~ W ytrV~ ~'~
2 ..
. ,i~., ~ r~ ~ ~I!~":w
o
1.
2
a
..
5
:I. 5
•
LEVEL 2
~. b!IM h~ ~ :r~ni. .~.6.
IjI:VJ"'~
.::... t ~~!fvvvr=
2 ..
:I. a 5 0 2 ..
:I. a 5
I1EA8URED
CALCULATED

Fig. 4.6 (cont'd). Measured and Calculated Response for MFl

Using Otani Hysteresis Model
 116
MF1 RUN1. BXLXNEAR HY8TERE8X8

ACCELERATXON
.2
( G )
o.
-.2
TIME. SEC.
o 2
a 5

BASE SHEAR
10
( KN )

o
-CALCULATED
-10 ----- MEASURED

o 2 4
a 5
MOMENT
1.0
( KN-M )

o
-10
-20~ ______~__L -__~__~__-+______-4______~~_
o 2 4
a 5
20 DXSPLACEHENT AT LEVEL 10

1.0
t MM )

o
-10
-20
o 2 4
1. a 5
ACCELEIltATXON AT LEVEL 10
.8
( G
••
o.
-.3

-.8 ~------+-------~------~-------+-------4--­
o 2 4
s 5

Fig. 4.7 Measured and Calculated Response for MFl Using

Bilinear Hysteresis Model
...1 MIN1 ~

~ACENENT. AT LKVEL •
10 UI r ... ,
0 0

-10 -UI
-eo t v J • I ttMR.8EC.
r II!
0 II! 4 :&. s
1 " •
• •
UI UI LnEL •

0 0

-10 -UI
0
• II! 4
l I I :&. I
0 II! 4 •
1

10
• • --'
--'
LrIIn. 4 '-J
0

-10 o II! 4
:&. • S
0
~
1
e
J~~
I
•
LEVEL II!

~ ~ I\A
:, V~~ o II! 4
'j .AAA~\J\L'::"V:V
o II! 4
~~.t ~~:vv: 1 • I
1 • 5

1tE~ CALCULATED

Fig. 4.7 (cont'd). Measured and Calculated Response for r1Fl

Using Bilinear Hysteresis Model
....:1. RUN:I. 8l:Ll:NEAR HV8TIERE8l:8
ACCELERAT%ON AT LIEYEL •

•4
•• I IJ J
.H

o. o.

NI ' '0 '4' • I • '~ T%tC. _C.

-.4 a 4
-.a. oI
:L
o a 4 • •
:L 8
.4
•
LavR\. •
.a •H

o. o.
-.a -.a.
-.4+ ,
o t 4 o ..
:L 8 :L
• • • ......
......
•4 00
•a
.a•
LEVEL ..

o. o.
-.a -.It.
o a 4 o ..
:L :L 5
• • •
.R' LEVEL R

0 o.
::
-~a .•~rv! V'frV . -.a5
!~~ !~.~A1~
0:L
2 4
• 5 o
2.
It
•
. •
HEA~ CALCULATIED

Fig. 4.7 (cont'd). Measured and Calculated Response for MFl

Using Bilinear Hysteresis Model
 119
MF1 RUN1. Q-HYST HYSTERESIS

BASE ACCELERATION
.e Ca )
o.
-.e
TIME. SEC.
o e
1

10 BASE SHEAR
[ KN )

- - - CALCULATED
-10 ----.HEASURED

o e
1 3 5

MOMENT
10 [ I<N-M )
o
-10

-eo
o 2
1. a 5

eo DISPLACEMENT AT LEVEL 10
[ MM )
o

-eo
o
1. 5

.e ACCELERATION AT LEVEL 10
.8 [ a
O.
-.3

-.e
o 2
1. 8 5

Fig. 4.8 Measured and Calculated Response for MFl Using

Q-Hyst Model
....~ IIUN1 O-Wt8T HY~

O%ItPLACEIt£NT AT LEVEL •
15
10 t "" ,

0 o
-10
-15
-eo r -+- v y TXM • .-c.
0 a 4 o e 4
1 5 1 5
• •
15
10
LEVEL.

0 o·

-10
-11

0
l a 4 o e 5
1 2.
• • • --I
10
N
LEVEL 4 o
0 .I .A
fI A~ AfI All . ~
-10
.~~ ~v~~r-v
~ o 1
a •
4 5
0 a 4
1
• •
LEVEL e

AA ~ ~ .AA A~ AfI ~ ~. ~
V>~V~\L? V wr-v
':<! .AAA~
o
vV~VYf'V:V
e 4
·:.I o 1
e •
4
5
~ 5
•
MASURCD CALCULATED

Fig. 4.8 (cont'd). Measured and Calculated Response for MFl

Using Q-Hyst r10del
11F3. RUH1 Q-HY8T HY8TER£8:r8

.4
.25
r a I
o. o.
-.25
T%f£. HC.
-.4
o 2 4
3. 8 5
0• 2 4
1 5
•
.4

.2 .25 LEVEL •

O. o.
-.2 -.25
;".4
. o 2 4
0 2 4 3. !I --'
1 5 • N
• --'
.4

.2
LEVEL 4
O. " ~ ,.,II"AI.A
::25~ , lAJ~fV ~VH~VVVVl i~
-.2

0
~
1
2 4
5
. . !. ~ h~
0 2 4
1
• 5
•
LEVEL 2
O ~"! ~~I~,.1
o~ ~ : .. WIIl!~
Wi
::! 'J\d~~.~MO~h~
-:2 V!
o
'V
e
VV'f VVl
4
_:.. =~v!~rvTC:V1'~Vvr~
3. • 5 D 2 4
3. 8 5
f£A8Uft£D
CALCULATED

Fig. 4.8 (cont'd). Measured and Calculated Response for MFl

Using Q-Hyst Model
Level Level
10 10
r I I I
I
I
9 r , 9- .... ..1

:[ J 7r -'--1
I
I
6t- 1/ 6t- ....,L_,
1
I
l
5t- II 51- .-,I
I
--'
4t- ,/ 4r- N
MF2 Run I
r-u.., N

Calculated
3t- 17 - - - - Measured 3r ------.--,I
I
2t- // 21- _,--...J

o' " I I I

25 (mm) o 2 3 4

Fig. 5.1 Maximum Calculated and Measured Fig. 5.2 Maximum Calculated and Measured
Displacements (Single Amplitude) Relative Story Displacements
furM~ for MF2
 123

0 10 20 30
I I I I
Level
10 ,
I
, I
9 f

MFI Run I
2 - - - - -Measured
----Takeda
A .Sina
o 0 Otani
o 0 Bi linear
- - -Q-Hyst

o 10 20 30 (mm)

Fig. 5.3 Maximum Calculated and Measured

Displacements (Single Amplitude)
for MFl
 124

o
I
0.2
I
0.4
I
0.6
I
0.8
I ,
1.0

Level
10

MFI Run I
2
- - - - - Measured
lJ A Takeda
" • Q- Hyst and Sina
o 0 Bilinear and Otani

o 0.2 0.4 0.6 0.8 1.0

Fig. 5.4 Maximum Displacements Normalized with

Respect to Top Level Displacement
 125

Level MFI Run I

10
I Calculated
I - - - -- Measured
9
I -,I
I
8 --l

7 --,
I
I
6

I
I
4
~
3 ,
I
2 -
I

----.,
I
I
o
o 2 3 4 5 (mm)

Fig. 5.5 Maximum Calculated (Using Takeda Model) and

Measured Relative Story Displacements
 Level Level
10 10

9r w,-, 9
I
I MFI Run I
8t- Calculated 8 r
1-'J
I
- - - - - Measured
I
71-
[L 7 L
I
I
I
&t- rJ ~ &

I
5t- L, 5
I I
I I
4t- r-' 4 I
J
.......
I I N
0\
I1_____ I
1______ ....,
3t- 3
I
I
2t- 2 r..J
I
I
L __ .,
'L-1 _, I I
I I
oI I I I I 0
0 2 3 4 5 (111111) 0 2 3 4 5 (111111)

Fig. 5.6 Maximum Calculated (Using Sina Model) and Fig. 5.7 Maximum Calculated (Using Otani Model) and
r·leasured Relative Story Displacements ~leasured Relative Story Displacements
 Level Level
10
10 rn
I
9 r"" L_l 9

I
I 8
8 r ~---- MFI Run I
I
Calculated
- - - - - Measured 7
7 r- -r-----,
I I
I
6 f- ~ rJ 6 I-
I I
I ~ L-,
5 r- ., --I
1 I
4 r 4
, ......
N
.....,
r I
I
I
3 I ----~ 3r-
I
f
2 r- r .... 21- r
I
I
---, If r--l
I
o I I I I II 0
0 2 3 4 5 (mm) 0 2 3 4

Fig. 5.8 Maximum Calculated (Using Bilinear Model) Fig. 5.9 Maximum Calculated (Using Q-hyst Model)
and Measured Relative Story Displacements and Measured Relative Story Displacements
 128

~x
Equivalent Mass (M e)

- cr
CD
..J
Massless
Rigid Bar

-
.c
.2'
Q)

-
J:
c
.!?

Fig. 6.1 The Q-Mode1

 129

Level r

i
....J

".~
"..,,.. ,.,'w
~
'"

Fig. 6.2 Static Lateral Loads

 130

o
o

6.
- x 100
Leq

Fig. 6.3 Force-Displacement Relationships

 131

en
c (/j
c (/j
E E E
:l 0
::s
u
0
0 ~
...:
)(
U
...:
c
1.305'1.305'1.3051
W

f
~
f DOD
,.. 51_

~ID
I
D
1
• • r::J
• ••
[y]DD 51

CJDD
0
=I:
f't)
-
~
-
U)

• DOD 0
Q)
N
N
v v U)

ODD II
Q)
N
- I- - r- - r- ODD -
N
0

D •
•
•
•
•
•
• •
• •
DDD Q

DOD
co
ODD
f't)
I'a
=II:
v •v =4t:
'lOt

&0
fsy =480 MPa o
rt)

f~ = 30 MPa
1372 ~I
# indicate gage wire number (see Table 7.10)

Fig. 7.1 Longitudinal Reinforcement Distribution for

Structures Hl and H2
 132

CI)
c: CI)
E E
::l
o
U
i
CD

, , + ,
-E n
~-

D - D- Oe r::l :tI=T
• • • • • •
~v
DOD
~ f+ei
~ [YJDD 51

D
- -
e

l.d
CJDD
D-- DOD o

DOD
(1)

-- - (\J
(\J
II

DOD (1)
(\J

-
(\J

D DDD
e
o
e _
o

cOrD-: Oe
DDD
~L · · · · DDD
LO
fsy = 352 MPa oI'f)
f~ = 33. MPa (FWI)
f~ = 42.1 MPa (FW2) I.. 1372 ~I
f~ = 34.5MPa (FW3)
FW3 = FW2
# indicate gage wire number (see Table 7.10) FW4 = FWl
Fig. 7.2 Longitudinal Reinforcement Distribution
for Structures FWl and FW2
 133

203
FW2 11
~

1 •
• I'l)
CD
• •

o
en
C\I
C\I

••••
••••

fay = 338 MPo

# indicate gage wire number (see Table 7.10)

Fig. 7.2 (cont'd) Longitudinal Distribution for

Structures FWl and FW2
 Structures HI 8 H2 Structures - - - MF , .
60 -MF2

Idealized

8
)(

-::t Structures FWI 8 FW4 Structures FW.2 a FW3

-.I

60 W
~

40

0.4 0.8

Ll:leq xlOO
Fig. 7.3 Normalized Moment-Displacement Diagrams
 135

H1 RUN 1
BASE ACCELERATZON [ G )
.8

o.

-.8
o 2 6
1 a 5

TIME. SEC.

CZSPLA [ 11M )

10

o
-10

-20

-80
o 2 6
1. a 5

BASE OVERTURNING MOMENTt ~N-M )

10

-10

-20~ ______-+________ ~~~ __ ~ ______-4________+-______ ~

o 2 6
1. 8 5

Fig. 7.4 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure Hl
 136

H2 RUN a
BASE ACCELERATXON ( Q )

.8

o.

-.a
o e e
1 a 5

TIME. SEC.

10

______ __ ______+-______ ______ ____

.
-ao~ +-~~ ~ ~ ~ ~

o e
1 a 5

BASE OVERTURNZNG MOMENT t KN-M ,

10

o
-10

o e
1 a

Fig. 7.5 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure H2
 137

SOOF MODEL MF:l RUN :l

SIMULATED ELCENTRO 1940 NS RUN 1

BASE ACCELERATION [ G )

.8

O.

-.S -f
a 2 a
1 a 5

TIME. SEC.

DISPLACEMENT [ MM )

" I
I
10 I
I
I
0 I
I
I
-10 I
I
I
I
I
-20

0 2 4 6
1 a 5

BASE OVERTURNING MOMENT ( KN-11 )

:1.0

o
-10
-20 ~ ______-+__ ~~ __ ________+-______ ________ ______
~ ~ ~ ~

o 2 6
5

Fig. 7.6 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure MFl
 138

SDOF MODEL MF2 RUN 1

SIMULATED ELCENTRO 1.g~O NS RUN 1.

BASE ACCELERATION [ Q )

.S

o.

-.a~----~-4------~~------~--------4-------~~-------
a
1. s 5
TIME. SEC.

...
,
)
II
,,\ I \

,
1\
I
I
a
-1.0

-20
-80
-F-------~---&----~--~--~--------~------~--------,
o e e
1. 8 5

BASE OVERTURNING MOMENT r KN-H )

3.0

a
-1.0

-20
o 2 e
5

Fig. 7.7 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure MF2
 139

SDOF MODEL FW1. RUN 1.

SIMULATED ELCENTRO 1940 NS RUN 1.

BASE ACCELERATION ( G )

.3

o.
-.8

1. 8 5
TIME. SEC.

DISPLACEMENT ( MM )

20

-20
o 2 e
1. 5

BASE OVERTURNING MOMENT ( KN-M )

eo
o

-eo

Fig. 7.8 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure FWl
 140
SOOF MODEL FW2 RUN 1

SIMULATED ELCENTRO 1940 NS RUN 1

BASE ACCELERATION [G)

.9

o.

-.8
o 2 6
1. a 5

TIME. SEC.

[ HM )

20

10

o
,
-10 ,,,
\I
-20
o 2 e
1. a 5

BASE OVERTURNING MOMENT £ KN-M )

20

o ,
i
-·20
.
o 2 e
a 5

Fig. 7.9 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure FW2
 141

SDOF MODEL FWa RUN 1

SIMULATED TAFT N21E RUN 1

BASE ACCELERATION ( G )

.3

o.

-.8

a 5
T:tME. SEC.

ME NT ( MM 1

-10

o 2 e
'- 5

BASE OVERTURNINQ MOMENT ( KN-M )

20

-20

Fig. 7.10 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure FW3
 142

SDDF MODEL FW .. RUN 1

SZMULATED TAFT N21E RUN 1

BASE ACCELERATZON [ G )

.3

o.

-.8

1 a 5
T:IHE. 8£C.

DZSPLACEMENT [ MM )

10
0
-10
-20
-80
0 2 6
1 3 5

BASE OV RTURNZNQ MOMENT I<N-M

eo

-eo

o e
1 a !5

Fig. 7.11 Calculated (Solid Line) and Measured (Broken Line) Response
for Structure FW4
 143
Level
10 SDOF Model
9

6
HI RUN I I
5 •.J
IL __
4
I
I
:3 I--~
I
2 Calculated ,J
Measured I

o 10 20 :30 (mm) o 2:3 4 ~ (mm)

Single Amplitude Malt. Calculated Malt. Calculated and Measured

and Measured Displacement. Relative Story Displacement.

Fig. 7.12 Maximum Response of Structure H1

Level
10 ,
,,
I
9
I
8 I
I

,,,
7 I

,,
6

5 ,
,,
I H2 RUN 3
4
,
3
I
I ,,I
Calculated
2
I
I
-- - -- Measured r.J
I

o 10 20 30(mm) o 2 3 4 5 (mm)

Single Amplitude MOl. Calculated Malt. Calculated and Measured

and Measured Displacements Relative Story Displacements

Fig. 7.13 Maximum Response of Structure H2

 144
Leve'
10
SDOF Model
9

6 MF I RUN'

3 Calculated
Measured
2

o 10 20 30 (mm) o 2 3 4 5 (mm)

Sino Ie Amplitude Max. CQlculated Max. Calcula ted and Measured

and Measured Di splacements Relative Story Displacements
Fig. 7.14 Maximum Response
of Structure MFl
Level
10 ,
I
I
9
, I

8
,,,
I

7 ,
I
6
I
I
I
5
MF 2 RUN I
t
4 ' - - - -.L.--.

o 2 3 4 ~ (mm)

Sin"e Amplitude Max. CalculGted Max. Calculated and Measured

and Measured Displacements Relative Story Displacements

Fig. 7.15 Maximum Response

of Structure MF2
 145
Level
10 SDOF Model

9 ,
8
._,
1

FWI RUN I 1I__ -,

7
•
6 r-J
I
5

3 Calculated
- - - - - Measured
2

o
o 10 20 3Q(mm) o 2 3 4 !5 (mm)

Single - Amplitude Max. Calculated Max. Calculated and Measured

and Measured Displacements Relative Story Displacements
Fig. 7.16 Maximum Response
Level of Structure FW1
10

9
fW2 RUN I
8

o o
o 10 20 30(mm) 234

Single -Amplitude Mox. Calculated Max. Calculated and Measured

and Measured Ojsplocemen's Relative Story Djsplacements

Fig. 7.17 Maximum Response

of Structure FW2
 146
Level
10 ,
I
I SOOF Model
9
I
8 I
I
7

6
FW3 Run' I
I
5 I

4
,
I
I
I
- Ca Iculated I
J
3
- - - - Measured I
2 ,J
I

o o 10 20 30(mm) 0 2 3 4 5 (mm)

SinQle-Amplitude Max. Calculated Max. Calculated and Measured

and Measured Displacements Relative Story Displacements
Fig. 7.18 Maximum Response
Level of Structure FW3
10

9
I
8 r'
7
,_,
I

I
FW4 Run'
6 ,J
I
I,
,
4 •.J
I
3 ,J

2 ,-,
I
I

o 0 10 20 30 (mm) o I 2 ! 4 5 (lftm)

SinQle-AmpUtude Max. Calculated Max. Calculated and Measured

and Measuted Displacements Relative Story Displacement

Fig. 7.19 Maximum Response

of Structure FW4
 '47

SDDF MODEL MF1

HaL! DAY INN

BASE ACCELERATION ( G )
.25

o.

o 2 4
1 a 5

TIME. SEC.
DISPLACEMENT ( MM )

1.0

-10
o 2 8
1. s 5

BASE OVERTURNING MOMENT ( KN-M )

1.0

-1.0

o
1 a 5

Fig. 7.20 Q-Model (Solid Line) and MDOF Model (Broken Line)
Results for Orion Earthquake
 148

SOOF MODEL MF:1

CASTAIC N2:1E 1971.

BASE ACCELERATION (G 1
.5

.25

o.

-.25
o 2
:1
TIME. SEC.

DISPLACEMENT (MM 1

1.0

-1.0

o 2
3 TIME. SEC.

BASE OVERTURNING MOMENT ( KN-M 1

1.0

-:10

o 2
:1 ~ TIME. SEC.

Fig. 7.21 Q-Model (Solid Line) and MDOF Model (Broken Line)
Results for Castaic Earthquake
 149

SO OF MODEL MF1

BUCAREST NS 1977

BASE ACCELERATION (G 1

.1

o.
-.1

o 2
1
TIME. SEC.

DISPLACEMENT ( MM )

20

-20

0 2
1 a TIME. SEC.

BASE OVERTURNING MOMENT (KN-H)

10

o
-10

-20
o 2
1 a TIME. SEC.

Fig. 7.22 Q-~1odel(Solid Line) and ~1DOF Model (Broken Line)

Results for Bucarest Earthquake
 150

SDDF MODEL MF1.

BUCAREST NS 1.977

BASE ACCELERAT70N (Q)

.1.

o.
-.1.

o 2
1.
T7ME. SEC.

D7SPLACEMENT [ MM )

1.0

-1.0

0 2
1. S TIME. SEC.

BASE OVERTURNING MOMENT (KN-H)

1.0

o
-1.0

o
1. 3 T7ME. SEC.

Fig. 7.23 Q-Model (with Increased Frequency; Solid Line) and MDOF Model
(Broken Line) Results for Bucarest Earthquake
 151

,,,
Level
10

MFI Orion
9
,,
,,
8

5
Q- Model
4
- - - - MDOF
3

o
o 10 20 30(mm) 0 2 3 4 5 (mm)

Single - Amplitude Maximum Maximum Relative Story

Displacements Displacements
Level Fig. 7.24 Maximum Response
for Orion Earthquake
10
I MFI Castaic
I
9

8
,
I

I
7

o o 10 20 30 (mm) o 2 3 4 5 (mm)
Single-Amplitude Maximum Maximum Relative Story
Displacements Displacements
Fig. 7.25 Maximum Response for
Castaic Earthquake
 152

Level
10
,,, I M F I Bucorest
I
I
9
I I ~
,
I
.I
8
I
I I
7
,
I -'J
•
1_'1
.I
6

/ 1.-
5
1
4 l-l
3 1_'1
2
- ._Q- Model (Orig. Freq.) 1'_1
- Q-Model (Increased
Freq. )
---MDOF
ri
o o 10 20 30 (mm) o 2 3 4 5 (mm)

Single -Ampl itude Maximum Maximum Relative Story

Displacements Displacements

Fig. 7.26 Maximum Response for Bucarest Earthquake

 153

SO OF MODEL MF1.

ELCENTRD kQ40 NS .2G

SOLID LINE • MOTION 1.
BASE ACCELERATION ( G ) BROKEN LINE t MOTION 2

.1.

O.

-.1.

-.2 4---~----~--4----+--~~--4----+~~~--4----+--~r---;--
o 2 4 e e 1.0
1. 5 7 9 1.1.
TIME. SEC.

DISPLACEMENT [ MM )

1.0

-kO

o 2 6
3 5

BASE OVERTURNING MOMENT [KN-M)

1.0

o
-1.0

o 2 6
1. 3 5

Fig. 7.27 Repeated Earthquakes with 0.2g Maximum Acceleration

 154
SO OF MODEL MF:1.

ELCENTRO :1.940 NS .4G

SOLIO LINE • MOTION :1.

BASE ACCELERATION ( G ) BROKEN LINE • MOTION 2

.8

o.

-.8

0 e 4 e e 1.0 1.e
:L. s 5 7 9 :1.:1.
T:tME. SEC.

DISPLACEMENT ( MM )

20

:1.0

-1.0

-20
0 2 6
:1. S 5

BASE OVERTURNING MOMENT [KN-M )

20

:to

o ,,
-:LO
,
-20~ ______ ~~
I
\ , ______ ~ ______ ~ ________ ~ ______ ~ ________ ~

o 2 e
:L 8 5

Fig. 7.28 Repeated Earthquakes with 0.4g Maximum Acceleration

 155
SO OF MODEL MFl.

ELCENTRD l.940 NS .BS

SOLID LINE • MOTION l.
BASE ACCELERATION r G ) BROKEN LINE • MOTION 2

.5

O.

-.5

o 2 4 e 8 1.0 1.2
l. 5 7 e l.l.
TIME. SEC.

DISPLACEMENT [ MM 1

40

20

-20

o 2 4 6
l. 5

BASE OVERTURNING MOMENT (KN-H)

20

-20
o 2 6
l. 5

Fig. 7.29 Repeated Earthquakes with 0.8g Maximum Acceleration

 156
SDOF MODEL MFl.

ELCENTRD :1.940 NS :1..2G

SOLID LINE • MOTION :1.

BASE ACCELERATION ( Q ) BROKEN LINE • MOTION 2

:1. •
•5
o.
-.5

-:1..
o 4 e B 1.0 1.2
:1. a 5 7 9 1.1.

TIME. SEC.

DISPLACEMENT ( MM )

40

-40

-BO

BASE OVERTURNING MOMENT (KN-H)

eo

-20

1. a 5

Fig. 7.30 Repeated Earthquakes with 1.29 Maximum Acceleration

 157
SDOF MODEL MF:1.

ELCENTRO :1.940 NS :1..6G

SOLIO LINE • MOTION :1.

BASE ACCELERATION [ g 1 BROKEN LINE • MOTION 2

:L

-:1.

a 2 4 e e :1.0 :1.2
1. a 5 7 1.1.
TIME. SEC.

DISPLACEMENT MM 1
:LOa

50
0

-50

-:LOO

0 2 6
:1. a 5

BASE OVERTURNING MOMENT [KN-M)

20

o
-20
-40~ ______ ~ ________ ~~ ____ ~ ________ ~ ______ ~ ________ ~

o 2 e
:1. 5

Fig. 7.31 Repeated Earthquakes with 1.6g Maximum Acceleration

 158

APPENDIX A
HYSTERESIS MODELS

A.1 General
The rules of the following models apply to both positive and
negative ranges of forces. If the current force is negative, it
has to be compared with corresponding forces at break-points in
negative region. In this case, the absolute value of the current
force is compared with the absolute value of the force at break-
point. For example, in Section 1.1 of Sina model it is stated:

1.1 Loading:
F(P) ~ F(C)

In negative range this rule should be read as:

1.1 Loading:

/F(p)1 :: IF(C')I· •.

A.2 Defi nitions

Loading: Increasing the force in one direction
Unloading: Decreasing the force in one direction
Load Reversal: Changing the force and its sign at the same step

A.3 Sina Model

There are 9 rules in Sina hysteresis system as follows (Fig. A.l)
Rule 1: Elastic stage
1.1 Loading:
F(P) < F(C) K = stiffness = slope of OC;
go to rule 1
F(P) > F(C) K = slope of CY; go to rule 2
 159

Rule 2: Current point on CY

2.1 Loading:
F{P} ~ F{y) K= slope of CY; go to rule 2
F{P} > F{y) K= slope of YU; go to rule 3
2.2 Unloading:
K= slope of PC·; go to rule 5
Rule 3: Current pont on YU
3.1 Loading:

K= slope of YU; go to rule 3

3.2 Unloading:

K =S ,
where D = maximum deformation
max attained in loading
direction
Rule 4: Current point on unloading branch from YU
4.1 Loading:

F{P} <- F{U m) K = S,; go to rule 4

F{P) > F{U m) K = slope of YU; go to rule 3
4.2 Unloading:
K = Sl; go to rule 4
4.3 Load reversal:
1. If not yielded previously K = ~~o~~ ~:l~o;·;

2. If formerly yielded K = smaller of the slope

of XoB· and X2U··
m'
go to rule 6
 160

Ru"le 5: Current point on RoC' (R 0 = Unloading point from CY)

5. 1 Loading:
F{P) < F{R )
o
K = slope of RoC'; go to rule 5

F(P) > F(Ro) K = slope of CY; go to rule 2

5.2 Unloading:
K = slope of RoC'; go to rule 5

5.3 Load reversal:

the same as 4.3
Rule 6: Current point on branch reaching crack-closing point
6. 1 Loading:

F{P) < F{B) K = slope of X,B; go to rule 6

-
F{P) > F(B) K = slope of BUm; go to rule 7

6.2 Unloading: (name the unloading point R3)

K = 51; go to rule 9

Rule 7: Current point on branch pointing towards Urn

If the section has not yielded previously, Urn is assumed
to be at Y.
7.1 Loading:
F{P) < F(Urn ) K = slope of X1U rn {or BUm}; go to rule 7
F(P) > F(Urn) K = slope of YU; go to rule 3

7.2 Unloading:
K = 51; go to rule 8
Rule 8: Current point on unloading from branch of rule 7
8.1 Loading:
F(P) < F{U~) K = 51; go to rule 8
F{P} > F{U')
rn K = slope of XoY (or BU~); go to rule 7
 161

8.2 Unloading:
K = Sl; go to rule 8

8.3 Load reversal:

the same as 4.3.2
Rule 9: Current point on unloading branch X1B
9.1 Loading:
F(P) < F(R3 ) K = Sl: go to rule 9

F(P) > F(R 3 ) K = slope of X1B; go to rule 6

9.2 Unloading:
K = Sl; go to rule 9
9.3 Load reversal:
the same as 4.3.2

A.4 Q-Hyst Model

There are four rules in Q-Hyst model as follows (Fig. A.2):
Rule 1:
1.1 Loading: if F(P) < F(Y) K = slope of OY; go to rule 1
if F(P) > F(Y) K = slope of YU; go to rule 2
1.2 Unloading: K = slope of OU; go to rule 1
1.3 Load reversal: K = slope of OY; go to rule
Rul e 2:
2.1 Loading: K = slope of YU; go to rule 2
2.2 Unloading: K = Sl = (slope of OY) x (~(Y))a;
go to rule 3 max
a = 0.5 in MDOF model
0.4 in SDOF model
 162

Rul e 3:
3.1 Loading: 1. If last unloading point on YU, go to 3.1.2
if F(P) < F(R) K = Sl; go to rule 3
if F(P) > F(R) K = (Slope of XoU~);

go to rule 4
2. If F(P) < F(U m) K = Sl; go to rule 3
if F(P) > F(U m) K = slope of YU;
go to rule 2
3.2 Unloading: K = 51; go to rule 3
3.3 Load reversal: K = slope of XU'·
o m'
go to rule 4
Rule 4 :
4.1 Loading: If F(P) < F(U~) K = slope of XoU~; go to rule 4
if F(P) > F(U~) K = slope of Y'U'; go to rule 2
4.2 Unloading: K = 51; go to rule 3
(name the unloading pOint R)
 163

Primary Curve
Q)
() @
"-
If

Deformation

Numbers In Circles
Indicate The Rul e #

Fig. A.1 Sina Hysteresis Rules

 164

w
u
'"" Primary Curve
&f

Deformation

1 Rule I
Rule 2

1: Rule
Rule
3
4

Fig. A.2 Q-Hyst Model

 165

APPENDIX B
COMPUTER PROGRAMS LARZ AND PLARZ

A special purpose computer program was developed to study the

seismic response of reinforced concrete rectangular frames subjected
to earthquake motions (LARZ). To plot response histories, a small
program (PLARZ) was written to be used in conjunction with LARZ.
The computer language of the programs is FORTRAN IV. The Cyber 175
computer system at Digital Computer Laboratory of the University of
Illinois was used to develop the programs.
In LARZ, two subroutines from IMSL computer library for matrix
inversion (LIN1PB) and for solution of simultaneous equations of
equilibrium (LEQ1S) have been used. For plotting purposes, the
graphic routines from GCS library have been applied.
Irregular frames similar to the one shown in Fig. B.l can be
analyzed by program LARZ. There can be more than one horizontal
degree of freedom at the same level. A special feature of the
program is that it can accept different hysteresis systems (models
presently implemented are those described in Chapter Three). In
fact, LARZ can be used to study steel frames using the bilinear
hysteresis system if this system is considered appropriate.
A block diagram of the program LARZ is presented in Fig. B.2.

a. Storage of Stiffness Matrix

The structural stiffness matrix is divided into three sub-matices
as shown in Fig. B.3. All matrix operations are performed in main core.
 166

Because [K ll ] is a symmetric matrix, only its lower half is

stored. No particular disadvantage ;s realized in storing the matrix
row-wise, so it is stored as a row-wise array. [K 12 ] is stored
completely. However, the location of non-zero elements are stored
using pointer arrays {ITK} and {JTK} as shown in Fig. B.3. Only
non-zero elements of [K 12 ] enter in matrix operations.
[K 22 ] is a symmetric banded matrix, hence, only half of the
banded portion needs to be stored. By the requirement of subroutine
LIN1PB, the lower half of this matrix is stored in a two-dimensional
array as shown in Fig. B.3.

b. Response-History Data
Secondary memory is used to store calculated response history.
Upon the execution of LARZ, if response plots are desired, the generated
data are written on three sequential files. At later stage, these data
are read by the program PLARZ, and plotted according to the scale
specified by the user. The response plots can be obtained in different
scales without a need to re-execute LARZ.
 167

p)".
"'~ ""'~ "'". "'~ ~~ ?J.: ~

41 .. Yg

Fig. B.l Structure with Missing Elements

 168

~T = TIME INTERVAL
OF INTEGRATION ~T

CALCULATE

ELEMENT CODES
HYSTERESIS
MODELS

CALCULATE ELASTIC
ELEMENT STIFFNESSES

TAKEDA AND STRUCTURAL

STIFFNESS MATRICES

SINA

CALCULATE CONDENSE STRUCT.

INSTANT ANEOUS STIFFNESS MATRIX
OTANI ELEM. & STRUCT. AND CALCULATE
TI FFNESS MATRI CES DAMPING MATRIX

BILINEAR

SOLVE DIFFERENTIAL
Q-HYST EQUATION OF MOTION

TIME =
NO TIME + ~T
CALCULATE MEMBER
END FORCES

Fig. B.2 Block Diagram of Program LARZ

 169
I
KII I KI2 NDF = Number of Degrees
of Freedom
... ---,-----
I
Structural
St iffness Matrix I
K21 I K22
NJ = Number of Joints
I Excluding Supports
I

(0)
Symmet r ic

NDFX NDF
Row
Number
t ITK JTK
I _ _ _ r----.., }
Column Numbers of Non - Zero
2
~::::: Elements In Row I of [KI2]
'--.l~--I

NDF

NJ
~ Non-Zero Elements
o

o
NDF XNJ

Fig. B.3a &b Storage of Structural Stiffness Matrix

 170

r-
-
r--
Symmetric

>

"- .
NJXNJ WXNJ

Fig. B.3c Storage of Submatrix ~2

 171

APPENDIX C
MAXIMUM ELEMENT RESPONSE BASED ON DIFFERENT
HYSTERESIS MODELS

The maximum moments and ductilities at the ends of flexible portions

of members, calculated based on different hysteresis models, are presented
in Tables C.l through C.5. Element numbering is shown in Fig. C.l. The
rotations are for unit length of each member. Ductility at a member end
is defined as the ratio of maximum rotation to the yield rotation.
 172

41 42 43

1 11 21 31
44 45 46

2 12 22 32
47 48 49

3 13 23 33
50 51 52
4 14 24 34
53 54 55
5 15 25 35
56 57 58
6 16 26 36
59 60 61
7 17 27 37
62 63 64
8 18 28 38
65 66 67
9 19 29 39
68 69 70

10 20 30 40

17 lop 17" HI" ".~

Fig. C.1 Element Numbering for Structure MF1

Pages 173-177 have been removed.

Due to legibility problems, the following computer printouts have been

omitted: "Table C.1 Maximum Response of Structure MF1 Based on Takeda
Model;" "Table C.2 Maximum Response of Structure MFl Based on Sina
Model;" "Table C.3 Maximum Response of Structure MFl Based on Otani
Model;" "Table C.4 Maximum Response of Structure MFl Based on Bilinear
Model;" and "Table C.5 Maximum Response of Structure MFl Based on Q-
Hyst Model."*

*For information concerning these pages contact Mete A. Sozen, University

of Illinois at Urbana-Champaign, Department of Civil Engineering, Urbana,
IL 61801.
 178

APPENDIX 0
COMPUTER PROGRAMS LARZAK AND PLARZK

A special purpose computer program was developed to calculate the

seismic response of a single-degree system consisting of a mass mounted
on a massless rigid bar connected to the ground by a hinge support and a
nonlinear rotational spring. The input base acceleration and the response
histories of displacement of the mass and the base moment are stored on
temporary tapes. A plotting program (PLARZK) was developed to read the
data and plot the response histories. The programs were written in
Fortran IV, using Cyber 175 computer at the University of Illinois.
Plotting routines from GCS library were used to plot the response histories.
To obtain hard copies of the plots, the Calcomp plotter at Digital Computer
Laboratories of the University of Illinois was used.
A block diagram of program LARZAK is presented in Fig. 0.1.
 179

TIME = L1T

L1T = TIME STEP OF CALCULATE ELASTIC STIFFNESS

INTEGRATION

CALCULATE DA~1PING

SOLVE DIFFERENTIAL
CALCULATE NEW EQUATION OF MOTION
STIFFNESS AND OBTAIN
DISPLACH1ENT

DETERMINE BASE MOMEN

TH1E =
TIME + L1T

CALCULATE NONLINEAR
FLEXIBILITIES USING THE NO
Q-HYST MODEL

Fig. D.l Block Diagram of Program LARZAK

 180

APPENDIX E
MOMENTS AND DUCTILITIES FOR STRUCTURE MF1 SUBJECTED
TO DIFFERENT EARTHQUAKES

Tables E.1 through E.3 present the maximum moments and ductilities
at the ends of flexible portions of the elements of structure MF1 sub-
jected to Orion, Castaic, and Bucarest earthquakes. The results were
obtained using the program LARZ,·which treated the structure as multi-
degree model. Element numbering ;s shown in Fig. C.1. In the tables,
the rotations are for the unit length of each member. Ductility is de-
fined as the ratio of rotation to the yield rotation.
Pages 181-183 have been removed.

Due to legibility problems, the following computer printouts have been

removed: IITable E.1 Maximum Response of Structure MFl Subjected to
Orion;1I IITable E.2 Maximum Response of Structure MFl Subjected to
Castaic;1I and IITable E.3 Maximum Response of Structure MFl Subjected
to Bucarest .*
ll

*For information concerning these pages contact Mete A. Sozen, University

of Illinois at Urbana-Champaign, Department of Civil Engineering, Urbana,
IL 61801.
 184

APPENDIX F
RESPONSE TO TAFT AND EL CENTRO RECORDS

Structure MF1 was analyzed for the measured records of E1 Centro NS,
El Centro EW, Taft N21E, and Taft S69W, using the Q-Mode1. In each case,
the structure was subjected to 15 seconds of the original record. The
time axes of the records were compressed by a factor of 2.5. The maxi-
mum acceleration for each earthquake was normalized to 0.4g which was
the design intensity for structure MF1. The base acceleration, top-level
displacement and base moment responses are presented in Fig. F.l through
F.4.
Except for the north-south component of El Centro, which was simulated
in the laboratory, no test results were available for structure MFl subjected
to the above records. Considering the fact that the Q-Model was successful
in simulating the measured response for structure MFl subjected to a simulat-
ed north-south component of E1 Centro (Sec. 7.4), and noting that the other
three motions were similar to E1 Centro NS, the calculated responses were
judged based on their overall appearance in relation to the measured re-
sponse for the simulated E1 Centro, NS.
The waveform in all cases (Fig. F.l and F.4) seemed reasonable; i.e.,
no unusual response was seen. The maximum absolute value of the single-ampli-
tude top-level displacement varied from l6.7mm (for Taft N21E) to 33 mm (for
E1 Centro EW). These values were in the same order of magnitude of that
from the experimental results (23.6mm). It is therefore possible to con-
clude that the Q-Model yielded reasonable overall responses for the e~rth­

quake records considered.

 185

seOF HODEL HF~

ELCENTRO ~840 NS .4Q

BASE ACCELERAT70N [G)

.a

o.

-.s
0 e 8
1. a 5
TZHE. SEC.

D::tSPLACEMENT ( MM 1

eo
~o

-1.0
-20
0 e e
1. s 5

BASE OVERTURN::tNQ MOMENT ( KN-M )

eo
1.0

-~o

-eo~ ____ ~~~ ____ ~~ ____ ______-+______-+______

~ ~

Q e
• 5

Fig. F.l Response for Structure MFl Subjected to El Centro NS

 186

SDDF MODEL MF1

EL CENTRO 1840 £W

8A8E ACCELERAT%ON ( Q )

.8

o.

o e
• 5
TIME. SEC.

D%SPLACEMENT [MM)

-eo

o e
1
•
8A8e DVERTURN%N8 MOMENT ( KN-M )

10

o
-10

-eo r-----~~----~------~------~----~4_----~e
o
• 5

Fig. F.2 Response for Structure MFl Subjected to El Centro EW

 187

SDOF MODEL

TAFT 1852

."SE ACCELER"T%ON [ G )
.8

o.

-.8

• !5

T:tt1£. SEC.

D%8PLACEMENT [MM)

1o0

-1.0
o
3 5

BASE OVERTURNING MOMENT [KN-M'

1.0

-10

2. 5

Fig. F.3 Response for Structure MFl Subjected to Taft N21E

 188

SDOF HODEL HF1.

TAFT S.. .4Q

BASE ACCELERATXDN t G )

.8

o.

-.8

til 5
TIME. SEC.

DXSPLACEMENT r HH )

10

o
-10
-eo~ ______ ______ ______ ______-+______-+______
~ ~ ~ ~

o 2
5

BASE OVERTURNXNG MOMENT r KN-H )

10

-10

-eo
0
1 5

Fig. F.4 Response for Structure MFl Subjected to Taft S69E

50272 .' 01
REPOIH DOCUMENTATION Il,-REPORT NO. 1
2. 1 3 . Recipient,'.S.AcceSSion .•No,
PAGE NSF/RA-790388 t~iJG!J 19; 24' 1 :3
f-4-.-~-t~e-n-la;-~-:---U-:-~I-d- Complex Models for Nonlinear Seismic Response to --po ReA~~~a~~ 1979'
Rei nforced Concrete Structures ~-
I--------------------~-----~-------
7. Authur(s) 8. Performing Organization Rept, No.
--------------------
M. Saiidi_, M. A. Sozen SRS No. 465
9. Performing Organization Name and Address 10. Project/Task/Work Unit No.

University of Illinois at Urbana-Champaign UlLU- ENG~?-=9_-_20=-=1=-=-3

_ _ _ _~
Department of Civil Engineering 11. Contract(C) or Grant(G) No.

Urbana, Illinois 61801 (C)

(G) PFR7816318
12. Spo,soring Organization Name and Address 13. Type of Report & Period Covered
Engineering and Applied Science (EAS)
National Science Foundation --------- _ . - - - - - ---------
1800 G Street, N.W. 14.
Washington, D.C. 20550
~----~~~~-----------------------------------~~~-=.-=.-=-~-=--=.~.~.~.=-.--~-~~~~--~
15. Supplementary Notes

---.-------------------------------------.----------------i
'''_ Ah .... ~~t (limit: 200 words)
This study investigated the possibility of simplifying the nonlinear analysis of
reinforced concrete structures subjected to severe earthquakes. One thrust was a
microscopic study of the particular elements of the analysis, the hysteresis model,
and the development of simple models leading to acceptable results. For this phase,
a multi-degree nonlinear model (LARZ) was developed to analyze rectangular reinforced
concrete frames for given base acceleration records. LARZ is capable of accepting
a collection of hysteresis systems, some previously used and others developed in
the course of this project. The new systems generally were simpler. The second
phase comprised a macroscopic study which included development of a simple model
that resulted in a reasonably close estimate of nonlinear response. A given structure
was viewed as a single-degree-of-freedom system which recognized stiffness changes
due to the nonlinearity of material. For both phases, the reliability of the ana-
lytical models was evaluated by comparing the calculated responses with results of
dynamic experiments on a group of small-scale ten-story reinforced concrete frames
and frame-walls tested on the University of Illinois Earthquake Simulator.

------------------------------------------------------------------------------1
17. Docliment Analysis a, Descriptors

Earthquakes Hysteresis Mathematical models

Nonlinear systems Buildings Earthquake resistant structures
Reinforced concrete Dynamic structural analysis

b. Identifiers/Open·Ended Terms

LARZ

C. COSATI Field/Group
•
I
vaiJ,.bility Statement 19. Security Class (This Report) 21. No. of Pages

NTIS ----------------------~--------------
20. Security Class (This Page) 22. Price

.39.18) See InstructIons on Reverse OPTIONAL FORM 272 (4-77)

(Formerly NTlS-35)
Department of Commerce