You are on page 1of 11

EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS, VOL.

2, 3 5 4 5 (1973)

EFFECTS OF STIFFNESS DEGRADATION ON DUCTILITY


REQUIREMENTS FOR MULTISTOREY BUILDINGS
ANIL K. CHOPRA* AND CHRISTOPHER K A N ~
Department of Civil Engineering, University of California, Berkeley, California, U.S.A.

SUMMARY
The effect of stiffness degradation in reinforced concrete structural members on the inelastic response of multistorey
buildings to earthquakes is investigated. In particular, the following question is examined. How do the ductility
requirements for multistorey systems with degrading stiffness behaviour compare with those for structures with
ordinary bilinear hysteretic property ? Inelastic dynamic responses of two idealized multistorey buildings, one
having a long and the other a relatively short fundamental period, to an ensemble of twenty simulated earthquakes
representative of moderately intense ground motions in California at moderate epicentral distances on firm ground,
are analysed for ordinary bilinear hysteretic behaviour and for bilinear hysteretic behaviour with stiffness de-
gradation property. The conclusions deduced from the results of this investigation include the following.
(1) It is, in general, not possible to predict the maximum response of a degrading stiffness system from results
for the corresponding ordinary bilinear system.
(2) The differences in ductility requirements due to stiffness degradation are generally smaller than those
associated with probabilistic variability from one ground motion to another.
(3) Stiffness degradation has little influence on the ductility requirements for flexible buildings, but it leads to
increased ductility requirements for stiff buildings.

INTRODUCTION
It is well known that buildings designed for lateral forces specified by the Uniform Building Codel can
withstand intense earthquake ground motions only if the structural system is capable of developing enough
~ , ~ on analysis of yielding systems having a bilinear hysteretic force-deflection relationship
d ~ c t i l i t y . Based
(Figure 1(a)) subjected to moderately intense earthquake ground motions, it has been concluded that multi-
storey buildings must be capable of undergoing storey drifts 4 to 6 times the elastic limit drifts, for adequate
per f ~r ma nc e . ~
The bilinear relationship of Figure l(a), in which the stiffness always returns to the original elastic stiffness
when yielding ceases and the strength remains unaffected by increase in inelastic deformation or in number
of load reversals, is a suitable idealization for behaviour of steel structural members. Experimental as well
as theoretical work”’ on reinforced concrete members subjected to repeated and reversed loadings has
demonstrated that their post-yield behaviour is considerably different from the bilinear relationship of
Figure I(a). In such members there is a loss of stiffness with cyclic increases in inelastic deformation, even
though there may be no apparent loss of strength. For analytical purposes, this degrading stiffness behaviour
was idealized as shown in Figure I(b) in an earlier investigation.8
As mentioned earlier, most analyses of ductility requirements for multistorey buildings subjected to earth-
quakes have assumed the bilinear hysteretic force-deflection relationship. The following question therefore
arises. How does stiffness degradation in reinforced concrete structural members affect the inelastic response
of multistorey buildings to earthquakes; i.e. all other factors being identical, how do the ductility require-
ments for multistorey systems with degrading stiffness behaviour compare with those for structures with
ordinary (without stiffness degradation) bilinear hysteretic property ?

* Associate Professor.
t Graduate Student.
Received 15 September 1972
Revised 30 January 1973
@ 1973 by John Wiley & Sons, Ltd.
35
36

/--
'6' I
ANIL K. CHOPRA AND CHRISTOPHER KAN

/-*
b
/A//
STORY SHEAR, F

I / / STORY DRIFT, X

-4-

-
o. ORDINARY BILINEAR HYSTERETIC SYSTEM b. BILINEAR HYSTERETIC SYSTEM WITH DEGRADING
STIFFNESS

Figure 1 . Hysteretic systems

Earlier investigations8? concerning this question have been limited to single degree of freedom systems
and their conclusions may not directly apply to systems with many degrees of freedom such as multistorey
buildings. This work therefore examines the above question for multistorey buildings. Inelastic dynamic
responses of two idealized multistorey buildings, one having a long and the other a relatively short funda-
mental period, to an ensemble of twenty simulated earthquakes representative of moderately intense ground
motions in California at moderate epicentral distances on firm ground, are analyzed for ordinary bilinear
hysteretic behaviour and for bilinear hysteretic behaviour with a stiffness degradation property. Results for
the two hysteretic systems are compared and the effects of stiffness degradation on earthquake ductility
requirements are discussed.

IDEALIZED BUILDINGS
For the purposes of this investigation, the building is idealized as a shear type structure with mass con-
centrated at the floors, which are assumed to be rigid, and the deflections are due only to lateral deformations
of the columns, resulting in one degree of freedom per storey.
The specific idealized systems analyzed are 8-storey buildings having the following properties. The lumped
mass at the floor levels are all equal. The relative values of the elastic stiffness for the various storeys presented
in Table I were selected so that the first mode of vibration is triangular, a property implied in the recom-
mendations of the Uniform Building Code (UBC). Relative values for the yield strengths of the various
storeys, also presented in Table I, were selected on the basis that yielding in all storeys occurs simultaneously
when the static lateral loading prescribed by the UBC is monotonically increased. The relationship between
shear and drift in each storey is described by Figure l(a) for structures with bilinear hysteretic behaviour, and
by Figure 1(b) for structures exhibiting bilinear hysteretic behaviour with degrading stiffness. The post-yield
stiffness k , is selected as one-tenth of the initial elastic stiffness k, for both hysteretic systems.

Table 1. Relative values of elastic stiffness and


yield strength

Storey Elastic stiffness Yield strength


1 .ooo 1.oooo
0.9725 0,9753
0.9167 0.9250
0.8333 0.8500
0-7222 0.7500
0.5833 0.6249
0.4167 0.4750
0.2222 0.3000
EFFECTS OF STIFFNESS DEGRADATION 37

The elastic stiffness k,, and the yield strength Ful of the first storey is given by
k---
, -4.54
w g

where
W = total weight of the building
w1 = fundamental frequency of vibration
TI = fundamental period of vibration = 27r/w,
g = acceleration of gravity
K = numerical coefficient in UBC depending on the structural
system; selected as 0.67, the value recommended for ductile
moment resisting frames
C = 0*05/(Q* is the base shear coefficient in UBC
It should be noted that the yield strength for the first storey is taken as twice the UBC design value, to
account for the difference between yield and design stresses as well as for the strengthening influence of non-
structural components. Reference may be made to Ruiz and PenzienlO.ll for further details on properties of
these idealized systems.
Energy dissipation in the structure in the elastic range of vibration is represented by a viscous damping
matrix constructed from damping ratio ( chosen to be equal for all modes, consistent with results of forced
vibration tests on some buildings.l2 Such tests generate extremely small vibrations and result in rather low
values of damping. For the larger amplitude motions considered here, a damping ratio [ = 0.05 has been
selected.

EARTHQUAKE GROUND MOTIONS


Earlier work by Clough and Johnston* on single-degree-of-freedom systems indicated that effects of stiffness
degradation on earthquake ductility requirements are influenced by the ground motion itself. It is also known
that there is significant dispersion in the response of hysteretic systems deforming well beyond the elastic
limit, due to probabilistic variations in the earthquake ground motion. It therefore appears important to
represent ground motion as a random process in studying the effects of stiffness degradation on response of
the two buildings described in the preceding section.
The random process model and simulation procedure adopted herein is identical to that developed by
Ruiz and Penzien.lo,l1 The simulation procedure consisted of generating samples of stationary Gaussian
white noise, multiplying the white noise by an intensity function of time to represent a segment of strong
shaking at constant intensity followed by a gradual decay in intensity, passing the resulting function through
a second-order linear filter with selected parameters to impart the desired frequency content and finally
performing a base line correction on the filtered function.
The earthquake accelerograms simulated for the purpose of this investigation are representative of ground
motions to be expected in the west coast of the United States on firm ground at moderate epicentral distances
and the filter parameters were selected accordingly. This is demonstrated in Figure 2 where the shape of the
average of the pseudo-velocity spectra for twenty simulated earthquakes is compared with Housner’s average
response spectra,13obtained from ground motions recorded on firm ground at 30-45 mile epicentral distances
at El Centro (1934), El Centro (1940), Olympia (1949) and Taft (1952).
38 ANIL K. CHOPRA AND CHRISTOPHER KAN

The expected maximum acceleration for this ensemble of ground motions was selected as 0.30 g and the
duration 30 sec, consisting of 12 sec of strong shaking following by an exponential decay in intensity over
an 18 sec period. Six of the twenty earthquakes simulated are shown in Figure 3.

--- AVERAGE OF 2 0 SIMULATED EARTHQUAKES


15
-STANDARD RESPONSE SPECTRA (HOUSNER.19591

0 05 10 15 20 2 5 30
UNDAMPED NATURAL PERIOD-SEC.

Figure 2. Average pseudo-velocity response spectra (from Ruiz and Penzien)

I,----Ll---.
Time-secs
d
20 30 0 . 401 . . .
I . . . . . . . .I
10
Time-secs
20
. . . . .
04 04

0 0

I
0 0 pp(&$;;
Iyl '
04 04 L.--.----- ,
EFFECTS OF STIFFNESS DEGRADATION 39

CHARACTERISTICS O F EARTHQUAKE RESPONSE


The response of the two idealized buildings described earlier to the ensemble of simulated earthquakes is
computed by a step-by-step linear acceleration method.lO*l4 In each case, the response was analyzed for
ordinary bilinear hysteretic behaviour and for bilinear hysteretic behaviour with degrading stiffness. Actually,
results for the ordinary bilinear system have been obtained earlier.lO* l1 They were re-computed in this
investigation so that results for the two hysteretic systems are obtained from identical computational pro-
cedures to eliminate any spurious numerical differences. In comparing the computer results for the ordinary
bilinear system from this investigation with those obtained earlier, discrepancies not exceeding 6 per cent
were observed in the displacements. Careful checking revealed that it arose because the transitions from
elastic to yielding branch or from yielding to elastic branch were handled differently in the two computer
programs. In a straightforward step-by-step method with a fixed time increment, overshoots and under-
shoots in the shear-drift relationship would occur at transition points. In the present investigation errors
due to such departures from the actual relationship were eliminated by an iterative procedure.
A visual comparison of the time variation of storey displacements for the two hysteretic systems during an
earthquake revealed differences generally similar to those reported earlier for single-degree-of-freedom
systems. Lengthening of the period of vibration due to degradation in stiffness was apparent for both the
stiff (5"' = 0.5 sec) and flexible (5". = 2.0 sec) buildings. In case of long period single-degree-of-freedom
systems, it had been pointed out that yielding takes place during the early phases of the earthquake, and
thereafter the degrading stiffness model loses almost all of its tendency towards vibratory behaviour and is
not very responsive to the subsequent ground motion.8 Response of the flexible multistorey building revealed
generally similar behaviour except that the inhibiting of vibratory responses was much less pronounced and
occurred only towards the end of the ground motion. For the stiff multistorey building the response of the
degrading stiffness system was as violent as shown by the ordinary bilinear system, similar to the behaviour
of short period single-degree-of-freedom systems.s
The maximum values of drift in each storey for the two buildings, stiff and flexible, and both hysteretic
systems during each of the twenty simulated earthquakes were determined. These values were then expressed
as storey drift ductility factors. The maximum ductility factor for a storey is simply the ratio of maximum
drift divided by the drift at yielding for that storey.
Maximum ductility factors for simulated earthquakes nos. I , 3, 4 and 5 are plotted in Figure 4. A
significant aspect of the results is the apparent lack of any consistent relationship between the ductility
factors for the ordinary bilinear hysteretic system and those for the degrading stiffness system. Degradation

L
0
I I I ' I I I ' I
- 1 Y , D EARTHQUAKE 5 - -
- / -
-
4
-- II ORDINARY BILINEAR -
\ SYSTEM
3 - __- DEGRADING STIFFNESS -
\
2 - \\ \
\ SYSTEM -
I0
I,
4
*L
.
" ws
- T,=20sec
'\
T,=0.5rec
-
I l l l i l l
8 12 16 0 4 8 12 16
DUCT1L l l'Y FACTOR
Figure 4. Maximum ductility factors
40 ANIL K. CHOPRA AND CHRISTOPHER KAN

of stiffness may for some earthquakes cause an increase and for others cause a decrease in the maximum
ductility factor for a particular storey. For a particular ground motion, the maximum ductility factor in a
number of storeys may increase while in some others it may decrease due to stiffness degradation. The stiff
building may be affected differently compared to the flexible building by stiffness degradation, i.e. the
maximum ductility factors during a particular earthquake may increase in the stiff building but decrease
in the flexible building due to stiffness degradation.
Inspection of Figure 4 reveals that the differences in maximum ductility factors for the same hysteretic
model from one earthquake to another, due to probabilistic variations in the ground motion, may be larger
than the differences due to stiffness degradation during a particular earthquake.
The structural model employed in this investigation dissipates energy by two different mechanisms,
viscous damping and hysteretic loops. The time histories of the energy dissipated per unit of storey mass in
these two forms are plotted in Figure 5 for simulated earthquake motions 6, 7 and 10. It appears that the

T, = 0 5 rec.

ORDINARY BILINEAR
SYSTEM
,,f ---- DEGRADING STIFFNESS
SYSTEM
I I I I I
w
0

I I I ' I I
FLEXIBLE BUILDING
Tt - 2 09ec T, =201dc

>
u
a
z
Y

I I I
I
0 5 10 15 20 25 30
T I M E , seconds

SIMULATED EARTHOUAKE 6 SIMULATED EARTHQUAKE 7 SIMULATED EARTHQUAKE 10

Figure 5. Energy dissipated (per unit storey mass) in hysteresis and in damping in ordinary bilinear and degrading stiffness
systems

energy dissipated in viscous damping is quantitatively similar for buildings with ordinary bilinear hysteretic
behaviour and those with degrading stiffness behaviour. In comparing the quantity of energy dissipated in
hysteresis in the two models, it appears that there is no simple and consistent trend from one earthquake
motion to another, an observation already made with reference to maximum ductility factors. Hysteretic
energy is, of course, related to the time history of inelastic deformations and ductility factors for all storeys.
The number of times the relative storey drifts cross the zero line is an indication of the predominant
frequency in the response. It is of interest to compare the number of zero crossings for the two hysteretic
models. The average number of zero crossings for all the storeys are plotted in Figure 6, bounded by envelopes
of one standard deviation on each side. (For the flexible building, the envelope is shown only on the left to
avoid excessive confusion of the lines.) The vertical dash-dot lines correspond to the average number of
zero crossings for elastic response in the fundamental mode of vibration.
The deterioration in the period of vibration due to inelastic yielding is very pronounced in stiff buildings,
particularly in the lower storeys where large ductility factors are observed. Increase in the predominant
period of vibration in the response of stiff buildings due to stiffness degradation is significant and apparent
in Figure 6. This indicates a further deterioration in the period due to degradation of stiffness with yielding.
The standard deviation for the number of zero crossings in the response of the stiff building is substantially
EFFECTS OF STIFFNESS DEGRADATION 41

-ORDINARY BILINEAR SYSTEM


--- DEGRADING STIFFNESS SYSTEM
T, = 2.0 sec T, = 0.5 sec
e-

7 -

6 -

5 -
>
U
0 4 -
v)

3-

2-

I -

0
I I
20
I
40
i
60
I
eo 100
I I
120 140
NUMBER OF ZERO CROSSINGS

Figure 6. Number of zero crossings for ordinary bilinear and degrading stiffness systems

smaller for the degrading stiffness system, but the coefficient of variation (standard deviation + average
value) is roughly similar for the two hysteretic models. Degradation of stiffness has little effect on the average
number of zero crossings for flexible buildings.

DUCTILITY REQUIREMENTS
The maximum ductility factors computed for a building during an earthquake are a measure of the defor-
mations beyond initiation of yielding which the various storeys must be capable of undergoing without
failure. They are basic design requirements to be met by the structural system. It is therefore important to
investigate how stiffness degradation affects the maximum ductility factors. As discussed in the preceding
section, the effects of stiffness degradation on response vary with the earthquake excitation in no apparent
systematic way. The only meaningful approach then is to examine the statistics of these maximum ductility
factors obtained for the ensemble of twenty earthquakes described earlier.
The average value of maximum ductility factor for each storey and a band of one standard deviation on each
side (only on the lower side for the flexible building) is plotted in Figure 7 for the stiff and flexible buildings
and the two hysteretic models. As was mentioned earlier, a yield load factor of two had been applied to the
lateral loads prescribed by the Uniform Building Code in designing the two idealized buildings considered.
As expected, for the same yield load factor relatively large ductility is required in stiff structures. As pointed
out in earlier investigations, there is an increase in ductility required in top storeys of flexible buildings due
to the ‘whiplash’ effect caused by higher modes of vibration.ll In the case of stiff buildings, the ductility
required is largest in the bottom storeys decreasing gradually in the upper storeys.
From Figure 7, the following conclusions may be deduced regarding effects of stiffness degradation on
ductility factors : (1) in stiff buildings, the average ductility factors are increased in the middle storeys (3-7),
but remain essentially the same for other storeys; and (2) in flexible buildings, the average ductility factors
for all storeys are essentially unaffected.
The maximum values of the ductility factors obtained for each of the twenty simulated earthquakes are
presented in Figures 8 and 9 on extreme value probability paper.15 Each plot, corresponding to a particular
building, shows the estimated probability distribution of the ductility factors, i.e. probability that a given
ductility factor is not exceeded, for all storeys. This may be interpreted as shown in terms of the expected
number of earthquakes required to exceed a given ductility factor. The wavy lines represent the probability
42 ANIL K. CHOPRA AND CHRISTOPHER KAN

- ORDINARY BILINEAR SYSTEM


--- DEGRADING STIFFNESS SYSTEM
8-

7-

6-

5-
>
a
0 4 -
I-
v)

3-

2-

IF-
T, = 2.0sec T, = 0.5 aec

I I 1 I I 1
0 2 4 6 8 10 12 14
DUCTILITY FACTOR

Figure 7. Maximum ductility factors for ordinary bilinear and degrading stiffness systems

RETURN PERIOD-NUMBEROFEARTHQUAKES RETURN PERIOD-NUMBER OF EARTHWAKES


If 2 4 5 10 20 2 4 5 10 20
I I 1 1 I I I 1 I I
I

ORDINARY BILINEAR SYSTEM DEGRADING STIFFNESS SYSTEM

li

a
0
u
c
2
> f
c
=!
I-
3
0
n

C 1 I I I I I1 I I 1 1 I 1
.I0 .30 .50 .70 .90 .95 .I0 .30 .50 70 .90 .95
PROBABILITY DISTRIBUTION PROBABILITY DlSTRl8UTlON
L I 1 1 I 1 I I 1
-1.0 0.0 I .o 2.0 3.0 -1.0 0.0 1.0 2 .o 3.0
REDUCED EXTREME VALUE y REDUCED EXTREME VALUE y

Figure 8. Probability distributions for ductility factors-stiff buildings

distribution computed from the response results, and the straight lines (not all are shown to avoid confusion)
are the theoretical distributions assuming that the maximum ductility factors are random variables with a
probability distribution of the exponential Type I. These plots are such that the ductility factor at the origin
of the reduced extreme value scale, y = 0, represents the most probable value and the slope is proportional
to the standard deviation.
As mentioned in an earlier investigation,ll the probability distributions estimated from computed responses
show good agreement with the theoretical straight line distribution ;this agreement being better in the case
of stiff buildings. Stiffness degradation apparently has little effect on the quality of this agreement.
EFFECTS OF STIFFNESS DEGRADATlON 43

RETURN PERIOD- NUMBER OF EARTHQUAKES


2 4 5 to 20,
I I. I I
8-

DEGRADING STIFFNESS SYSTEM /

I I I 1
O' .Ib 30 50 70 .90 .95
PROBABILITY DISTRIBUTION PROBABILITY DISTRIBUTION
L I I J
- 1.0 0.0 1.0 2.0 3.O - 1.0 0.0 1.0 2.O 3.0
REDUCE0 EXTREME VALUE y REDUCED EXTREME VALUE y

Figure 9. Probability distributions for ductility factors-flexible buildings

-
RETURN PERIOD NUMBER OF EARTHQUAKES R E N R N PERIOD -NUMBER OF EARTHQUAKES
2 4 5 10 20 2 4 5 10 20
I
1 I I I I
8
I I
, I

-
STIFF BUILDINGS TI = 0.5 SEC.
-ORDINARY BILINEAR SYSTEM

I
I I I I I 1 I I
.I0 .30 .50 .70 .90 .95 OL .lb 30 50 70 .90 .95
PROBABILITY DISTRIBUTION PROBABILITY DISTRIBUTION.
1 1 I I I II L I I I I [
- 1.0 0.0 1.0 2.0 3.0 -1.0 0.0 1.0 2.0 3.0
REDUCED EXTREME VALUE y REDUCED EXTREME VALUE y

Figure 10. Theoretical probability distributions for ductility factors

The theoretical distributions have been replotted in Figure 10 to facilitate examination of stiffness
degradation effects. The significant features to be noted for stiff buildings are the following: (1) the most
probable ductility factors for storeys 1, 2, 3 and 8 are little affected but those for the middle storeys are
increased by stiffness degradation; (2) the increase in standard deviation of the ductility factors due to
44 ANIL K. CHOPRA AND CHRISTOPHER KAN

stiffness degradation is small in storeys 1-6 but large in storeys 7 and 8. In the case of flexible buildings the
following conclusions may be deduced: (1) the most probable ductility factors for the lower four storeys are
unaffected but those for the upper four storeys are decreased by stiffness degradation; (2) the standard
deviations appear to be affected in a rather complicated way by stiffness degradation. They increase (compared
to an ordinary bilinear hysteretic system) slightly for storeys 1-4, increase significantly for storeys 5 and 6,
remain equal for storey 7 and decrease for the top storey.
Results of an earlier investigations of stiffness degradation effects on response of single degree of freedom
systems led to the following conclusion. The most probable values as well as the standard deviation of
maximum ductility factor is larger for the ordinary bilinear hysteretic system compared to the degrading
stiffness system. These differences are greater for stiff structures. These conclusions are similar in part to
those deduced above for flexible multi-degree-of-freedom systems but contradict those obtained for stiff
multi-degree-of-freedom systems. Consequently, effects of stiffness degradation on ductility requirements
for single-degree-of-freedom systems cannot, in general, be extrapolated to systems with many degrees of
freedom such as multistorey buildings.

CONCLUSIONS
The following comments may be made concerning the results of this investigation; obviously they pertain
to the specific model of stiffness degradation shown in Figure l(b), and may not apply to other softening
mechanisms.
1. There appears to be no simple or consistent relationship between the maximum ductility factors
computed for the ordinary bilinear and stiffnessdegrading hysteretic systems. In other words, it is not possible
to predict the maximum response of a degrading stiffness system from results for the corresponding ordinary
bilinear system.
2. The differences in ductility requirements due to stiffness degradation are generally smaller than those
associated with probabilistic variability from one ground motion to another. In designing a building for an
unknown future earthquake, it would therefore be inappropriate to treat the ground motion as deterministic
but expend effort in evaluating the influence of stiffness degradation on ductility requirements.
3. The apparent period of vibration in the response of stiff buildings is increased significantly due to stiff-
ness degradation, but remains essentially unaffected in the case of flexible buildings. Stiffness degradation has
little influence on the ductility requirements for flexible buildings, but it leads to increased ductility require-
ments for stiff buildings.
4. The conclusions drawn in earlier studies concerning effects of stiffness degradation on ductility factors
for single-degree-of-freedom systems cannot, in general, be extrapolated to multistorey buildings.
5. It must be recognised that inelastic deformations in a multistorey building during an earthquake are
concentrated at certain critical regions of the structural members ; and the ductility requirements implied by
these local deformations considerably exceed those representing the lateral storey drifts. Structural ideal-
izations more refined than the simple shear building considered in this investigation will be necessary to
investigate how stiffness degradation influences these local ductility requirements.
It also should be noted that stiffness degradation may have a significant influence on the load carrying
capacity and stability of columns, and thus might drastically reduce the earthquake resistance of reinforced
concrete buildings. This aspect of the problem has not been included in the present work but deserves to be
investigated.

ACKNOWLEDGEMENTS

This investigation was supported in part by National Science Foundation Grant GI-31883.

REFERENCES
1. ‘Uniform Building Code-1967 Edition’, International Conference of Building Officials, Pasadena, California.
2. R. W. Clough, ‘Dynamic effects of earthquakes’, J. Struct. Div., ASCE, 86, No. ST4, 49-65 (1960).
3. A. S. Veletsos and N. M. Newmark, ‘Effect of inelastic behavior on the response of simple systems to earthquake
motions’, Proc. 2nd Wld Con$ Earthquake Engng, Tokyo, 895-912 (1960).
EFFECTS OF STIFFNESS DEGRADATION 45

4 . J. A. Blume, N. M. Newmark and L. H. Corning, ‘Design of multistory reinforced concrete buildings for earthquake
motions’, Portland Cement Association, Chicago, Illinois, 1961.
5 . N. W. Hanson and H. W. Conner, ‘Seismic resistance of reinforced concrete beam-column joints’, J. Struct. Div., Proc.
ASCE, 93, NO. ST5, 533-560 (1967).
6 . V. V. Bertero and B. Bresler, ‘Seismic behavior of reinforced concrete framed structures’, Proc. 4th Wld Conf. Earthquake
Engng, Santiago, Chile, 1969.
7. V. V. Bertero, B. Bresler and H. Liao, ‘Stiffness degradation of reinforced concrete members subjected to cyclic flexural
moments’, Report No. EERC 69-12, Earthquake Engineering Research Center, University of California, Berkeley,
California, 1969.
8. R. W. Clough and S. B. Johnston, ‘Effect of stiffness degradation on earthquake ductility requirements’, Proc. Japan
Earthquake Engng Symp., Tokyo, 1966.
9 . J. Penzien and S. C. Liu, ‘Nondeterministic analysis of non-linear structures subjected to earthquake excitations’, Proc.
4th Wld Conf. Earthquake Engng, Santiago, Chile, 1969.
10. P. Ruiz and J. Penzien, ‘Probabilistic study of the behavior of structures during earthquakes’, Report No. EERC 69-3,
Earthquake Engineering Research Center, University of California, Berkeley, California, 1969.
11. P. Ruiz and J. Penzien, ‘Stochastic seismic response of structures’, J. Engng Mech. Div., Proc. ASCE, 97, No. EM2,
441-456 (1971).
12. D. Rea, J. G. Bouwkamp and R. W. Clough, ‘The dynamic behavior of steel frame and truss buildings’, Report No.
66-24, Structural Engineering Laboratory, University of California, Berkeley, California, 1966.
13. G. W. Housner, ‘Behavior of structures during earthquakes’, J. Engng Mech. Div., Proc. A X E , 85, No. EM4, 109-129
(1959).
14. E. L. Wilson and R. W. Clough, ‘Dynamic response by step-by-step matrix analysis’, Proc., Symp. Use of Computers in
Civil Engineering, Lisbon, Portugal, 1962.
15. E. J. Gumbel, ‘Statistical theory of extreme values and some practical applications’, U.S. Department of Commerce,
National Bureau of Standards, Applied Mathematics Series, No. 33, 1954.