You are on page 1of 16

ANALYTICAL MODELING OF HYSTERETIC BEHAVIOR

OF R / C FRAMES

By Magdy S. L. Roufaiel 1 a n d Christian Meyer, 2 M e m b e r s , ASCE


Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

ABSTRACT: A theory for an enhanced mathematical model of R/C frame mem-


bers is presented and its accuracy is verified by simulating various laboratory
experiments for which data were available in the literature. New member and
global damage parameters are defined. These damage parameters are useful for
subsequent reliability analysis of damaged concrete frames.

INTRODUCTION

It has been the objective of the research, conducted by the writers, to


develop a rational analysis procedure that permits the safety assessment
of the existing building stock, including buildings that may or may n o t
have been damaged by previous exposure to earthquake ground mo-
tions.
The prerequisite for such an analysis procedure is a mathematical model
that can reproduce the hysteretic behavior of R / C members accurately.
It is also necessary to define parameters that are indicative of the dam-
age sustained by individual members in a structure as well as of the
overall damage in the structure.
It is the purpose of this paper to describe an analytical model that is
capable of reproducing the hysteretic behavior of R / C frame members
and to define damage parameters that correlate well with the m e m b e r s '
and the structure's residual strength a n d stiffness. The theory for the
analytical model is presented first, followed by the definition of the dam-
age parameters. The accuracy of the model is then verified by simulating
various experiments for which data were available in the literature. The
model is an extended and improved version of a model reported earlier
(Meyer, et al. 1983). It has been used in a reliability study of concrete
frames damaged by earthquakes, which is described in a companion pa-
per (Roufaiel and Meyer 1987).

R/C FRAME MEMBER MODEL

The mathematical model selected to simulate reinforced concrete frame


members is different from most others that have been reported in the
literature (e.g., Clough a n d Johnston 1966; Giberson 1969; Litton 1975;
Otani and Sozen 1972) in that it accounts for the finite size of the plastic
regions. The model requires the specification of only a few elementary
'Asst. Prof., Dept. of Civ. Engrg., Michigan Technological Univ., Houghton,
MI 49931.
2
Assoc. Prof., Dept. of Civ. Engrg. and Engrg. Mech., Columbia Univ., New
York, NY 10027.
Note.—Discussion open until August 1, 1987. Separate discussions should be
submitted for the individual papers in this symposium. To extend the closing
date one month, a written request must be filed with the ASCE Manager of Jour-
nals. The manuscript for this paper was submitted for review and possible pub-
lication on November 14, 1985. This paper is part of the Journal of Structural
Engineering, Vol. 113, No. 3, March, 1987. ©ASCE, ISSN 0733-9445/87/0003-0429/
$01.00. Paper No. 21305.

429

J. Struct. Eng. 1987.113:429-444.


material constants for steel and concrete and the geometric section prop-
erties. It is capable of simulating the complete behavior of R/C frame
members under strong cyclic loads up to advanced states of deteriora-
tion, even under the presence of high shear and axial forces.
To construct the analytical model, material constitutive laws for con-
crete and steel are assumed to obtain the primary moment-curvature
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

relationships for given R/C sections. Modified Takeda-type hysteresis


rules (Takeda, et al. 1970) are used to idealize the hysteretic moment-
curvature relationships. Effects of shear and strength degradation on the
flexural hysteresis are modeled empirically. During loading of the struc-
ture, members' end moments and curvatures are traced and the corre-
sponding stiffness matrices are calculated.
Stress-Strain Representation.—The stress-strain behavior of reinforc-
ing steel is idealized by a bilinear curve as shown in Fig. 1. For concrete,
a trilinear approximation of the stress-strain behavior is used (Fig. 2)
with the following characteristics:

1. The ultimate strength of concrete fcu and associated strain ec„ are
functions of the uniaxial cylinder strength f'c and the amount of con-
finement steel p":
/„ = (1 + 10 P ")/c; e» = (1 + 10 p")e0 (1)

I \ ^~~- Idealized
/ ^--Actual

Strain

FIG. 1.—Stress-Strain Law for Reinforcing Steel

FIG. 2.—Stress-Strain Law for Concrete

430

J. Struct. Eng. 1987.113:429-444.


where p" = 2(b" + d")A"/(b"d"s), in which b" and d" = the width and depth
of the confined core; A" = the hoop cross-sectional area; and s = the
hoop spacing.
2. The loading branch is represented by a bilinear curve such that the
area under it is the same as the area under the parabola proposed by
Park, et al. (1972). The slope of the first straight line is equal to the initial
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

slope of the parabola:


_2 _ 1
e —
ley ~ 7 feu i n/ ~ en< • (2)

3. The coordinates of the cut-off point of the unloading branch are


chosen empirically as
fc = 0 . 2 / „ ; e„, = (2 + 600 p") ecu (3)
Experimental data suggest that concrete members, after having been
subjected to the "ultimate" strain e„, in the extreme compression fiber,
start to degrade during subsequent load cycles (Roufaiel and Meyer 1983).
This strain thus signals the onset of failure because the member will
sustain only a few load cycles of decreasing amplitudes after reaching
the critical value e„,.
Primary Moment-Curvature Relationship.—Based on the assumption
that plane sections remain plane even after cracking of concrete, and
assuming the stress-strain curves for steel and concrete as described pre-
viously, the primary moment-curvature relationship (i.e., for monoton-
ically increasing moment) of a beam section can be derived using con-
ventional reinforced concrete theory (Roufaiel and Meyer 1983). This
procedure allows the determination of the yield moment My and cur-
vature 4>i/ / as well as the moment M,„ associated with the curvature 4>„,
at which the maximum concrete strain reaches the failure value e„, (Fig.
3). The actual M-<$> relationship is nonlinear beyond the yield point, but
the approximation of this nonlinear branch by a straight line introduces
little error. The magnitude of the error depends on the value of the axial
load as will be discussed later. Further it should be noted that flanged
sections will introduce some errors because of neglect of the strain vari-
ation along flange widths.
Hysteretic Behavior of Reinforced Concrete.—Under load reversals
well into the inelastic range, the stiffness of a reinforced concrete mem-

E
o
2

$y •m
Curvature

FIG. 3.—Primary Moment-Curvature Relationship

431

J. Struct. Eng. 1987.113:429-444.


ber decreases due to concrete cracking and bond deterioration at the
steel-concrete interface. Takeda-type rules (Takeda, et al. 1970) are used
to describe the resulting hysteretic moment curvature relationships (Fig.
4). The model has basically five different branches. These are identified
in Fig. 4 by corresponding numbers in circles. A + or - superscript
indicates the direction of loading. These branches may be described as
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

follows:

1. An initial elastic branch with stiffness (E/)i = (E7)c characterizes elastic


loading and unloading as long as the yield moment My is not exceeded.
2. If the moment exceeds My, as well as the maximum moment reached
in any previous load cycle, the inelastic loading branch is represented
by the stiffness (EI)2 = p(EI)e, where p can be interpreted as a strain-
hardening parameter.
3. During unloading from an inelastic moment, the effective stiffness
(E7)3 is somewhat smaller than (E7)e. Laboratory experiments (Ma, et al.
1976) indicate that this softening is the more pronounced, the greater
the prior inelastic excursion.
4. Reloading in a direction in which the yield capacity has been pre-
viously exceeded consists of two branches. The first branch represents
loading during closing of any cracks.
5. The second branch describes the behavior after crack closure, when
the reloading stiffness can be observed to increase considerably and de-
crease again after renewed entry into the inelastic range.

Specific details about implementation of these various hysteresis rules


are given by Roufaiel and Meyer (1983).
Shear Effect on Hysteretic Behavior.—The reloading curve (branches
4 and 5 in Fig. 4) gives hysteresis loops the characteristic pinched shape
associated with diagonal shear cracks. By evaluating test data, e.g., Ma,
et al. (1976) and Popov, et al. (1972), a strong correlation can be found
between the degree of pinching and the relative magnitude of shear at

FIG. 4.—Hysteretic Moment Curvature Relationship

432

J. Struct. Eng. 1987.113:429-444.


the section under consideration. For small shear forces, no pinching is
observed. The reloading curve is straight and crosses the initial elastic
loading curve at a "point of no pinching" (M„ ,<$>„) (Fig. 5). At the other
extreme, i.e., very large shear forces, the stiffness during crack closure
can be ignored completely, so that the first reloading branch follows the
4>-axis up to the origin (0,0). Realistic reloading curves generally have
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

an amount of pinching somewhere between these two extremes. The


characteristic point (Mp, (J),,), where the reloading curve crosses the orig-
inal elastic loading curve (Fig. 5), can be determined as a function of the
shear span ratio a/d according to the following empirical relationship:

Mp = a,, M„; <]>p = otp <>


|„ (4)

where a„ = 0 for - < 1.5; • a„ = 0.4 - - 0.6 for 1.5 < - < 4.0;
d d d

a„ = 1 for - > 4.0 (5)


d
Member Stiffness Matrix.—To compute the tangent stiffness of a gen-
eral frame member, the member is subdivided into three regions (Fig.
6): (1) An inelastic region of length x, at node i, having the average sec-
tion stiffness (£Z),; (2) an inelastic region of length Xj at node ;', having
the average section stiffness (EI),; and (3) a central region of length (L
- Xj - Xj), having the initial elastic section stiffness (EI)e.
If the sectional stiffness is known at each point of the member, the
coefficients of the six-by-six tangent stiffness matrix can be computed
using standard methods of engineering mechanics (Roufaiel and Meyer
1983). Obviously, the stiffness coefficients depend on the lengths of the
plastic regions, which in turn are functions of the amount by which the
end moments exceed the member's yield moment (see Fig. 6).
Axial Load Effect.—An axial force has an important effect on the pri-
mary moment-curvature relationship of a reinforced concrete section (Fig.
7). Up to the balanced load Pb, the axial force increases the yield mo-
ment of a section but reduces the failure curvature considerably. Even
though for larger P-values the curves are very nonlinear beyond the yield
point, they can be approximated by a bilinear formulation as long as P

FIG. 5.—Pinched Reloading Curve


433

J. Struct. Eng. 1987.113:429-444.


Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

FIG. 6.—Frame Member Model

< 0.5P[,. For larger axial forces, the member's energy dissipation capac-
ity is reduced so much that the yield point can be considered as the
failure point.
When a building is subjected to an earthquake, the axial forces in the
columns vary in time as the result of the overturning moment and ver-
tical accelerations. The accurate computation of member properties tak-
ing into account this variation of axial forces is difficult and time con-
suming (Keshavarzian and Schnobrich 1984). Here the axial forces are
assumed to remain constant and equal to the gravity load effect present
at the beginning of the cyclic load history.
Another axial load effect is associated with the large lateral displace-
ments and is known as P-delta effect. Since concrete structures are gen-
erally stiff and do not deform very much, the P-delta effect can be ap-

\>-"~ P/P = 0.75

1200

^ — V ^ ^ v ^ - — P / Pb = 0.5

P/F
1000
r V °-25

800 \ " __JHJ£^T^~


I
i 600
As = 1.2 in 2
; As = 1.2 in2
1?"

2.06-t-
400
~ |2„ fsy = 55.2 ksi
E. = 28,500 ksi

200

Curvature ( I 0 ^ r o d / i n )

FIQ. 7.—Axial Force Effect on Primary Moment-Curvature Relationship

434

J. Struct. Eng. 1987.113:429-444.


Lonct Point of Actual max. Artificial mox.
Incipient Failure Load paint^Loarj Point

\ Strength
— Dropoff

/ ^

V^ A
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

t 1^ ^ " ^ *m »x ?x •

FIG. 8.—Modeling of Strength Degradation

proximated for most practical purposes using a simple consistent geometric


stiffness for each member (Bolotin 1964).
Strength Degradation during Cyclic Loading.—If a reinforced con-
crete member is strained beyond a certain critical level during cyclic
loading, its strength will deteriorate. The member will no longer be able
to carry the same load at a given deformation level. This load capacity
will decrease continuously during subsequent load cycles. From evalu-
ating experimental data (Ma, et al. 1976; Popov, et al. 1972), it can be
concluded that the critical point is associated with the curvature 4>m at
which the concrete reaches the strain value e„,.
The experimental data suggest also that the strength drop-off is pro-
portional to the amount by which the critical curvature has been ex-
ceeded. To simulate this behavior, the present model replaces the actual
point of maximum loading as focus for the reloading branch by an ar-
tificial point (Fig. 8). In laboratory experiments, members usually fail
after very few load cycles once the curvature $„, has been exceeded. The
actual number of cycles to failure depends on the amount of confine-
ment and on the magnitudes of axial and shear forces. For practical pur-
poses, an approximate simulation of this strength degradation is ade-
quate (Roufaiel and Meyer 1983), because a failing member can dissipate
only small amounts of energy before the steel buckles or ruptures or the
member experiences some other mode of failure (e.g., bond or torsional
failure).

DAMAGE INDICATORS AND FAILURE DEFINITION

For subsequent reliability analysis of damaged concrete frames, it is


desirable to define two damage parameters: (1) A member damage pa-
rameter to describe the degree of damage sustained by individual mem-
bers in a structure; and (2) a global damage parameter that is indicative
of the overall damage in the structure.
The reliability of a building to withstand earthquakes depends greatly
on the strength and stiffness degradation, which its major structural
components may have experienced in earlier earthquakes. For this rea-
son it is appropriate to adopt a damage definition related to residual
strength and stiffness. A modified flexural damage ratio, MFDR, is de-
fined here as a member damage indicator as follows (Fig. 9):

435

J. Struct. Eng. 1987.113:429-444.


Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

•m •*

FIG. 9.—Definition of Modified Flexural Damage Ratio

MFDR = max [MFDR+; MFDR~] (6)


& - ^1. 4>* 4>y
My
in which MFDR+ MFDR" (7)
4>m <t>y
M + M; M" M~
This damage ratio represents the ratio between the secant stiffness at
the onset of failure Mm/$„, and the minimum secant stiffness reached so
far Mj/cfix . The term My/<\>v = the initial elastic stiffness. Superscripts +
and - denote the loading direction. The value MFDR = 0 indicates that
the yield moment of a member has not been exceeded, therefore damage
is nonexistent. At the other extreme, if the failure curvature c|>„, has been
reached, then MFDR = 1, which signifies the onset of failure of the
member. It should be noted that an assumption is made that reinforce-
ment details follow recommended practice, thus eliminating the poten-
tial for premature local failure in the joints or elsewhere.
The following global damage parameter, GDP, is introduced to define
the overall degree of damage in a frame building;

GDP = (8)

where dR = the maximum roof displacement; dY = the roof displacement


at which the first member in the frame reaches its yield moment capac-
ity, assuming the frame displaces in its first mode only; and dp = the
roof displacement at which the frame is assumed to fail. There are var-
ious failure modes possible, but at this point the definition will be re-
stricted to excessive roof displacements. An evaluation of various tests
(Roufaiel and Meyer 1983) indicates that
dF = 0.06H (9)
is a reasonable definition of failure displacement, where H = the build-
ing height. In very few instances has a cantilever beam withstood a tip
displacement larger than 0.06L without failure. Likewise, observations
of sub-assemblies and scale frame models point to the same limit. (Ob-
viously, a building frame that has displaced less than 6% of its height
may already be beyond repair.)

436

J. Struct. Eng. 1987.113:429-444.


The global damage parameter is set equal to 0 if dR < dY . In this case,.
no member has yielded, therefore damage is nonexistent. At the other
extreme dR > dF; the frame has failed.
The global damage parameter as has been defined is a reasonably good
indicator of damage in low to medium-rise structures. For tall buildings,
it may be necessary to refine the definition in order to include effects of
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

second and higher modes of vibrations.

SUMMARY OF ANALYTICAL PROCEDURE

The model described above has been added as a new element to the
well-known DRAIN-2D computer program (Kanaan and Powell 1973).
Besides incorporating the frame element model, the program was en-
hanced by a static analysis capability and an eigenvalue solver for de-
termining mode shapes and frequencies.
A brief summary of the analytical procedure is presented as follows:

1. Input frame geometry data, material properties, failure roof dis-


placement dF, and loading history.
2. For each member, determine initial elastic and geometrical stiffness
matrices.
3. Form global stiffness and mass matrices.
4. Determine fundamental mode shape and frequency.
5. Calculate roof yield displacement dY .
6. For each incremental loading step:
a. Form global system of equations and determine current dis-
placements.
b. If roof displacement exceeds previous values, consider the new
value to be the maximum roof displacement dR .
c. For each member:
i. Determine end moments and curvatures.
ii. If the loading branch on either end of the member changes,
then determine the new element stiffness matrix and update
the global stiffness matrix.
iii. If the loading branch is changing from 2 to 3 (inelastic load-
ing to unloading), then compute the value of the member's
damage parameter, MFDR, as given by Eq. 7. Also, deter-
mine the new slopes for loading branches 3, 4, and 5.
7. At the end of loading steps, calculate the value of the global dam-
age parameter GDP as given by Eq. 8.

MODEL VERIFICATION

The model just described requires a few material and section prop-
erties as input and will reproduce entire load-deformation histories. In
order to do this accurately, it was necessary to preset some empirical
constants for optimum agreement with a few selected experimental data.
The validity of the model was then demonstrated by simulating re-
sponses of other tests not utilized in the calibration phase. For this pur-
pose an extensive analytical testing program was carried out involving
the analysis of beams, frame sub-assemblies, and entire structures for
437

J. Struct. Eng. 1987.113:429-444.


TABLE 1.—Numerical Testing Program
Number of
specimens
Type of test specimen analyzed Reference Loading type
0) (2) (3) (4)
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

Cantilever beam 9 Ma (1976) Cyclic, quasi-static


Cantilever beam 3 Popov (1972) Cyclic, quasi-static
Beam-column 8 Atalay (1975) Cyclic, quasi-static
Frame subassem-
blage 12 Scribner (1978) Cyclic, quasi-static
One-story portal 1 Gulkan (1971) Monotonic and
frame cyclic, quasi-
static; harmonic,
dynamic
Ten-story three-bay 1 Healey (1978) Dynamic, shaking
frame table three con-
secutive time
histories
Two-story one bay 1 Clough (1976) Dynamic, shaking
frame table two con-
secutive time
histories

FIG. 10.—Load-Deflection Curves for Cantilever Beam (Ma, et al. 1976)

438

J. Struct. Eng. 1987.113:429-444.


EXPERIMENT
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.


THEORY
M^^-
i/Jj!^pf^f~
WlM
KIPS)

^ j ^ S ^ ^ E F L E C T I O N (IN)
LOAD

-83- /}&
i #gf?J
-a- ^
l.i -l.I -0.1 e.s t:t o.i i.i

FIG. 11.—Load-Deflection Curves for Cantilever Beam (Ma, et al. 1976)

EXPERIMENT 4%s%=,

/
KZ5
^~~y//f/M
~- T- Za^^Sffl''
DISPLACEMENT
SETS 1 AND 2

THEORY

It- ^^^^^^^_^

•lo

-«0-

DISPLACEMENT (IN)

FIG. 12.—Load-Deflection Curves for Beam-Column (Atalay and Penzien 1975)

439

J. Struct. Eng. 1987.113:429-444.


which experimental results were available. These are summarized in Ta-
ble 1. Agreement between theory and experiment was excellent in all
cases and is documented by Roufaiel and Meyer (1983). In the following,
only a few selected cases are reproduced to illustrate the degree of agree-
ment and to point out a few aspects of concrete behavior.
Fig. 10 shows the load-deflection behavior of a cantilever beam tested
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

by Ma, et al. (1976). As can be seen, the hysteresis loops exhibit no


pinching, pointing to the absence of significant shear forces. The critical
curvature is associated with a tip displacement of about 50 mm (2 in.).
A sharp drop-off in strength can be observed after this displacement has
been exceeded and is correctly reproduced by the analytical results.
Fig. 11 shows a different case from the same test series (Ma, et al.
1976), but this time pinching of hysteresis loops indicates the presence
of high shear stresses. In this case, the failure tip displacement is some-
where around 33 mm (1.3 in.) and again is accurately predicted by the
model.
The response of a specimen with axial force tested by Atalay and Pen-
zien (1975) is shown in Fig. 12. Agreement between theory and exper-
iment is good, even though the displacement at which the model pre-
dicts strength deterioration is somewhat less than the failure displacement
indicated by the test.
The scale model of a 10-story frame (Fig. 13) was tested by Healey and
Sozen (1978) on the shaking table of the University of Illinois. The in-
dividual member damage ratios correlated well with the observed dam-
age (Roufaiel and Meyer 1983). A typical comparison of responses is shown
in Fig. 14. Likewise, Clough and Gidwani (1976) tested a two-story, half-
scale structure on the shaking table of the University of California (Fig.
15), and a typical comparison between theory and experiment is shown
in Fig. 16.

E s = 2 0 0 , 0 0 0 MPa
Ps = 0.01
fsy = 350 MPa
3 3 3
fc = 40 MPa
3 3 3 f 0 = 0,0034
/>"(BEAMS) = 1.6 %
3 3
5
? , ?
/(COLUMNS) = 1.17%
3 3 3
AB| = 8.5 mm2
2 A S 2= 12.75 mm2
3 3 3 MASS PER FLOOR =
2
0.227 N:-SEcVmm
3 3 3
2 2 2 DEAD LOAD PER FLOOR •
3 3 3 2228 N
? 2 2
DAMPING = Z %
3 3 3
2 2
4 4
2
:ross-section type
4 4 4

t f^^t r^™t f

"iVtr\ [of-
„, AS? t-l 1 4-
•oss-section 3 C-oss-section 4

FIG. 13.—Idealization of Frame Model Tested by Healey (1978)

440

J. Struct. Eng. 1987.113:429-444.


Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

^AM^ifi-

TIME (SEC)
lb)

FIG. 14.—Measured and Computed Top-Level Response of 10-Story Frame: (a)


Ground Acceleration; (b) Response

t 11 i % ~l l^r

• j" j. ,| <5=l GHOUND MOTION DIRECTION!

FRONT ELEVATION . SIDE ELEVATION

FIG. 15.—Two-Story Frame Tested by Clough (1976)

RttURDEO
PRESENT ANALYSIS

- v iJilihfi
' rff'^'fl
TIME (SEC)
lb)

FIG. 16.—Measured and Computed Top-Level Response of Two-Story Frame: (a)


Ground Acceleration; (b) Response

441

J. Struct. Eng. 1987.113:429-444.


The verification process, of which t h e above examples are only a small
sample, has demonstrated that the model reproduces the large inelastic
response of frame members reasonably well u p to the point of failure.

CONCLUSIONS
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

An enhanced mathematical m o d e l h a s b e e n presented that is capable


of simulating the response of reinforced concrete frame members to strong
cyclic loads. It is different from most models found in the literature be-
cause the finite size of plastic regions is explicitly accounted for, instead
of assuming fictitious plastic hinges of zero dimensions. The model can
accurately reproduce the complete behavior of general R / C frame m e m -
bers under strong cyclic loads u p to advanced states of deterioration a n d
up to the point of failure, even u n d e r the presence of high shear a n d
axial forces. The model accuracy w a s verified by reproducing various
experiments for which data were available in the literature.
N e w damage parameters were introduced that correlate well with the
residual strength a n d stiffness observed o n t h e laboratory specimens.
The usefulness of these parameters will become apparent w h e n incor-
porated in a reliability analysis of damaged concrete buildings. Such
studies are reported in a companion p a p e r (Roufaiel a n d Meyer 1987).

ACKNOWLEDGMENTS

This study is a part of a research project conducted at Columbia Uni-


versity and was supported by the National Science Foundation u n d e r
Grant N o . CEE81-21359. This support is gratefully acknowledged.

APPENDIX I.—REFERENCES

Atalay, M. B., and Penzien, J. (1975). "The seismic behavior of critical regions
of reinforced concrete components as influenced by moment, shear and axial
force." Earthquake Engrg. Research Center Report No. EERC 75-19, Univ. of Cal-
ifornia, Berkeley, Calif.
Bolotin, V. V. (1964). "The dynamic stability of elastic systems." Holden-Day,
San Francisco, Calif.
Clough, R. W., and Johnston, S. B. (1966). "Effect of stiffness degradation on
earthquake ductility requirements." Proceedings of Japan Earthquake Engineering
Symposium, Tokyo, Japan, 227-232.
Clough, R. W., and Gidwani, J. (1976). "Reinforced concrete frame 2: seismic
testing and analytical correlation." Earthquake Engrg. Research Center Report No.
EERC 76-15, Univ. of California, Berkeley, Calif.
Giberson, M. F. (1969). "Two nonlinear beams with definitions of ductility." /.
Struct. Div., ASCE, 95(2), 137-157.
Gulkan, P., and Sozen, M. A. (1971). "Response and energy dissipation of rein-
forced concrete frames subjected to strong base motions." Structural Research
Series No. 377, Univ. of Illinois, Urbana, 111.
Healey, T. J., and Sozen, M. A. (1978). "Experimental study of the dynamic re-
sponse of a ten-story reinforced concrete frame with a tall first story." Struc-
tural Research Series No. 450, Univ. of Illinois, Urbana, 111.
Kanaan, A. E., and Powell, G. H. (1973). "DRAIN-2D, a general purpose com-
puter program for dynamic analysis of inelastic plane structures." Report No.
EERC 73-6 and EERC 73-22, Univ. of California, Berkeley, Calif. Revised Sept.,
1973 and Aug., 1975.
Keshavarzian, M., and Schnobrich, W. C. (1984). "Computed nonlinear seismic

442

J. Struct. Eng. 1987.113:429-444.


response of R/C wall-frame structures." Structural Research Series No. 515, Univ.
of Illinois, Urbana, 111.
Litton, R. W. (1975). "A contribution to the analysis of concrete structures under
cyclic loading," thesis presented to the University of California, Berkeley, Calif.,
in partial fulfillment of the requirements for the degree of Doctor of Philoso-
phy.
Ma, S. M., Bertero, V. V., and Popov, E. P. (1976). "Experimental and analytical
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

studies on the hysteretic behavior of reinforced concrete rectangular and T-


beams." Earthquake Engrg. Research Center Report No. EERC 76-2, Univ. of Cal-
ifornia, Berkeley, Calif.
Meyer, C , Roufaiel, M. S. L., and Arzoumanidis, S. G. (1983). "Analysis of dam-
aged concrete frames for cyclic loads." Earthquake Engrg. Struct. Dyn., 11(2),
207-228.
Otani, S., and Sozen, M. A. (1972). "Behavior of multistory reinforced concrete
frames during earthquakes." Structural Research Series No. 392, Univ. of Illinois,
Urbana, 111.
Park, R., Kent, D. C , and Sampson, R. A. (1972). "Reinforced concrete members
with cyclic loading." /. Struct. Div., ASCE, 98(7), 1341-1360.
Popov, E. P., Bertero, V. V., and Krawinkler, H. (1972). "Cyclic behavior of three
R.C. flexural members with high shear." Earthquake Engrg. Research Center Re-
port No. EERC 72-5, Univ. of California, Berkeley, Calif.
Roufaiel, M. S. L., and Meyer, C. (1983). "Analysis of damaged concrete frame
buildings." Report No. NSF-CEE81-21359-1, Dept. of Civ. Engrg. and Engrg.
Mech., Columbia Univ., New York, N.Y.
Roufaiel, M. S. L., and Meyer, C. (1987). "Reliability of concrete frames damaged
by earthquakes." /. Struct. Engrg., ASCE, 113(3), 445-457.
Scribner, C. F., and Wight, J. K. (1978). "Delaying shear strength decay in rein-
forced concrete flexural members under large load reversals." Report UMEE
78R2, Dept. of Civ. Engrg., Univ. of Mich., Ann Arbor, Mich.
Takeda, T., Sozen, M. A., and Nielsen, N. N. (1970). "Reinforced concrete re-
sponse to simulated earthquakes," /. Struct. Div., ASCE, 96(12), 2557-2573.

APPENDIX II.—NOTATION

The following symbols are used in this paper:

A" =
stirrup cross-sectional area;
a =
shear span;
b" =
width of the confined core of R / C rectangular section;
d =
depth of a rectangular concrete section;
d" =
depth of the confined core of R / C rectangular section;
dF =
roof displacement at which frame is assumed to fail;
dR =
maximum roof displacement;
rfy =
roof displacement at which the first member in the frame
reaches its yield m o m e n t capacity;
EI = stiffness of a R / C sections (subscripts e, 1, 2, 3, 4, and 5
indicate loading branches s h o w n on Fig. 4);
EI = average stiffness of the plastic region;
fc = concrete stress;
f'c = uniaxial cylinder strength of concrete;
fcu = ultimate strength of concrete;
fcy = idealized yield stress of concrete;
GDP = global damage parameter;
H = building height;
MFDR = modified flexural damage ratio;

443

J. Struct. Eng. 1987.113:429-444.


M = moment;
M„, = moment capacity of a section associated with the failure strain
em;
M„ = moment coordinate of the intersection point of the reload-
ing branch with negligible shear force and the elastic branch,
Downloaded from ascelibrary.org by Northeastern Univ Library on 04/22/13. Copyright ASCE. For personal use only; all rights reserved.

Fig- 5;
Mp = m o m e n t coordinate at t h e point of intersection of t h e re-
loading branch a n d the elastic branch, Fig. 5;
Mx = current maximum m o m e n t during cyclic loading;
My = yield moment capacity of a R / C section;
P = axial force;
Pb = balanced axial force;
p = strain hardening parameter;
s = stirrup spacing;
ap = pinching factor given b y Eq. 5;
ec„ = strain associated w i t h concrete s t r e n g t h , fcu;
e„, = concrete strain at t h e cut-off point of t h e unloading branch,
Fig. 2;
e0 = strain associated with cylinder strength, f'c;
p" = confinement ratio = 2(b" + d")A'U{b"d"s);
<)> = curvature;
<|>m = curvature associated with the moment M,„;
4>„ = curvature associated with the moment M„;
4>p = curvature associated with the moment Mp;
<$>x = curvature associated with the moment Mx; and
4> y = curvature associated with the moment My .

444

J. Struct. Eng. 1987.113:429-444.

You might also like