RECOMMENDATIONS FOR SHAPE OFEARTHQUAKE RESPONSE …
Transcript of RECOMMENDATIONS FOR SHAPE OFEARTHQUAKE RESPONSE …
WASH-1254
RECOMMENDATIONS FOREARTHQUAKE RESPONSE SPECTRASHAPE OF
February 1973
Directorate of LicensingU.S. Atomic Energy Commission
Washington ,D.C.
LEGAL NOTICETM5 report %, prep.,rej js 4n 4%:%ount 3f '&otk %ponioted by tilt UnitedSt~ate% C~o~etniient. %:itlivr tile United States not the~ United States AtomicEnergy Cummimmion. nof An of their cmplo)cc%. not any of theirciontractfe.%, subt;ont rj,:tots. or their employees. makes, any wajrrnty.expre,.% or implied, or anumcs any troM~ hbiabyI3 or responsibility for thej~xurjcy. completenci% or uwefilness ot ny information. .ippiritus. pfodialor pro%:rs% disdowcd, or reprewenih that its uwe *ould nut inritnge privitclyowned riplit%
I
WASH-1254uc-1i
RECOMMENDATIONS FOR SHAPE OF EARTHQUAKE RESPONSE SPECTRA
prepared by
JOHN A. BLUME & ASSOCIATES, ENGINEERS
for the
DIRECTORATE OF LICENSINGUNITED STATES ATOMIC ENERGY COMMISSION
UNDER CONTRACT NUMBER AT(49-5)-3011
-w
February 1973
,it onit 1-y th. utu,rintnilont n o Dti) umonts. U.S. (u•'vrnmo.nt I nrsn ig: tlirt .
ACKNOWLEDGEMENTS
Assistance of Dr. A. G. Brady of California Institute
of Technology, Mr. D K. Jephcott of the State of Cali-
fornia Office of Artritecture and Construction, arid
Mr. V. Perez of the ieismological Field Survey, NOAA,
San Francisco, is ackno&wledged for supplying the punched
card decks of the digitized accelerograms. Mr. D. Lange
and Dr. K. Kapur of the Mechanical Engineering Branch,
Directorate of Licensing, AEC provided helpful comments
during review of a preliminary draft of this report.
- i I -
J
RECOMMENDAT IONS
FOR
SHAPE OF EARTHQUAKE
RESPONSE SPECTRA
LONTENTS
I. SUMMARY OF MAJOR FINDINGS AND RECOMMENDATIONS ----------------
A. Scope ----------------------------------------------------
B. Data Used ------------------------------------------------
C. Analytical Approach ------------------------------------
0. Major Findings and Observations --------------------------
E. Recommendations----------------------------------------
II. INTRODUCTION -----------------------------------------------
A. General ------------------------------------------------
B. Scope ---------------------------------------------------
III. RESPONSE SPECTRUM SHAPES -------------------------------------
A. Introduction -------------------------------------------
B. Selection of Historic Earthquakes-----------------------
C. Analytical Aspects of a Response Spectrum
D. Response Spectrum Shapes from Historic Earthquakes-------
IV. STATISTICAL ANALYSES OF SPECTRUM SHAPES---------------------
A. Introduction ---------------------------------------------
B. Statistical Analysis Techniques --------------------------
C. Statistics of Spectrum Shape Ensemble --------------------
D. Statistics of Grouped Spectrum Shapes-------------------
1. General -----------------------------------------------
2. Maximum Ground Acceleration -------------------------
Paqeý
2
4
7
78
9
9
91014
17
17
17
18
20
20
21
- iii -
CONTENTS (Cont'd)
Page
3. Soil Characteristics---------------------------------
4. Epicentral Distances of Recording Stations------------
Geographic Grouping of Spectrum Shapes--------------------
Summary of Findings---------------------------------------
E.
V. EFFECTS OC EARTHQUAKE DURATION-------------------------------
A. General-------------------------------------------------
B. Analytical Approach--------------------------------------
C. Pat Plot Studies and Results ------------------------------
D. Observations --------------------------------------------
E. Summary of Major Findings ---------------------------------
VI. RECOMMENDED SPECTRUM SHAPES----------------------------------
22
25
26
28
39
39
39
40
40
41
48
48
48
50
53
54
65
A.
B.
C.
D.
Introduction -------------------------------- ------
Recommended Spectrum Shapes-------------------------------
!. Probability Estimation for DAF -----------------------
Major Findings-------------------------------------------
Recommendations -----------------------------------------
VII. REFERENCES--------------------------------------------------
TABLES
I Salient Characteristics of Selected Accelerograms---------
2 Grouping of Accelerograms According to Maximum GroundAccelerations -------------------------------------------
3 Grouping of Accelerograms According to the Soil Impedanceat the Recording Stations--------------------------------
4 Grouping of Accelerograms According to Epicentral Distanceof the Recording Stations--------------------------------
5 Geographic Group 1: San Fernando Valley Earthquake -------
6 Geographic Group I[: Southern California -----------------
16
31
32
3334
35
- iv -
TABLES (cont'd)
Page
7 Geographic Group III: California ----------------------------- 36
8 Geographic Group IV: Western U.S.A.- -------------------------- 37
9 Strong Motion Durations 3 of the Accelerograms Selectedfor the Pat Plot Studies --------------------------------------- 43
10 Results from Pat Plot Studies of El Centro, 1940, NS ----------- 44
11 Results from Pat Plot Studies of El Centro, 1940, EW ----------- 44
12 Results from Pat Plot Studies of Taft, N21 0 E ------------------ 45
13 Results from Pat Plot Studies of Taft, S690E ------------------ 45
14 Results from Pat Plot Studies of Helena, NS ------------------- 46
15 Results from Pat Plot Studies of Helena, EW -------------------- 46
16 Results from Pat Plot Studies of Golden Gate Park, NIO°E -------- 47
17 Results from Pat Plot Studies of Golden Gate Park, NSO*W 47
18 Numerical Data for Recommended Spectrum Shapes ---------------- 56
FIGURES
I Single-Degree-of-Freedom System -------------------------------- 11
2 An Snsemble of n Response Spectrum Shapes for a Damping Ratio - 38
3 Basic Spectrum Shape ------------------------------------------ 52
4a Recommended Spectrum Shapes for Large (50%) Probability ofExceedance --------------------------------------------------- 57
4b Recommended Spectrum Shapes for Large (50%) probability ofExceedance ---------------------------------------------------- 58
5a Recommended Spectrum Shapes for Small (15.8%) Probab~ility ofExceedance ---------------------------------------------------- 5)
5b Recommended Spectrum Shapes for Small (15.8%) Probability ofExceedance ----------------------------------------------------- 60
6a Recommended Spectrum Shapes for Negligible (2.3%) Probabilityof Exceedance -------------------------------------------------- 61
6b Recommended Spectrum Shapes for Negligible (2.3%) Probability
of Exceedance ----------------------------- 62
7 Recommended Spectrum Shapes ----------------------------------- 63
8 Spectrum 16:ape Comparisons; Damping Ratio = 0.02 --------------- 64
" V "
APPENDICES
A Response Spectrum Shapes
B Ensemble Spectrum Statistics
C Group Spectrum Statistics
D Period-Amplitude-Time (PAT) Plots
F Spectrum Shapes Based on Log Normal Distribution
- vi -
I. SUMMARY OF MAJOR FIND!NGS AND
RECOMMENDATIONS
A. SCOPE
The major objectives of the studies described in this repor. were to analyse
and evaluate a numhbr of significant earthquake records and to utilize the re-
sults to develop "standardized" design snectrum shapes to be used in the seis-
mic design of nuclear power plant facilities. Because earthquak-s are complex
phenomena, and since it is not possible to exactly predict the nature of seis-
mic ground motions, statistical analyses of recorded ground motions must be
used. The major findings and recommendations from the studies based on such
statistical analyses, along with brief discussions of the data and the ana-
lytical approaches used, are discussed in this chapter.
B. DATA USED
Response spectrum shapes for thirty-three significant and different accelero-
grams generated by twelve major earthquakes were developed for damping ratios
of 0.005, 0.01, 0.02, 0.05, 0.07, and 0.10. The California Institute of Tech-
nology prepared most of the accelerogram digitizations, which are consistent
and reliable reproductions of the actual qround motions. The remaining digit-
izations were obtained from the State of California Office of Architecture and
Construction, Los Angeles, and the Seismological Field Survey, NOAA, San fran-
cisco. The thirty-three accelerograms for purposes of this study are termed
the ensemble.
The response spectrum shapes were studied as an ensemble and as groups cate-
gorized according to peak gro-nd accelerations, site soil characteristics,
epicentral distances, and geographical locale of the recording stations.
These studies were made to determine the influence of the various character-
istics on the shape of the response spectra. Pertinent literature was
searched to collect the most reliable available data concerning these earth-
quakes and their recording stations.
C. ANALYTICAL ADPROACH
The following approach was used for the statistical analyýes:
- I -
" Spectrum shape statistics, such as mean, median, and standard de-
viation were developed for the ensemble of all accelerograms and
for each of the groups to provide indications of the central ten-
dencies and uncertainties associated with the corresponding groups.
" A statistically acceptable probability model suitable for the com-
plete ensemble of spectrum shapes was determined. Smooth "stan-
dardized" design spectrum shapes were derived on the basis of this
probability model.
" Comparisons were made between the recommended spectrum shapes and
the current Atom!c Energy Commission (AEC) regulatory criteria,
Newmark, Housner, and Blume F-factor spectrum shapes.
" Period-time amplitude plot studies of eight accelerogranis generated
by four important earthquakes were made to investigate the effects
of earthquake duration.
The probabilistic approach was used because it was considered to be most
rational, realistic, and appropriate for seismic design, especially of crit-
ical installations such as nuclear power plants. Previous studies end pre-
dictions of the frequency characteristics of seismic ground motions have usu-
ally been based on the analysis of only a few records. The results of the
present studies are particularly significant for the following reasonis:
" The studies are based on the analyses of thirty-three different
accelerorrams, a larger number than used in other studies.
" The accelerogram digitizations are the most reliable ones available.
" The studies are comprehensive because they include the effects of a
number of parameters, detailed statistical analyses, and effects of
earthquake duration.
0. MAJOR FINDINGS AND OBSERVATIONS
The following major findings resulted from the studies:
- 2 -
o Statistical predictions of spectral characteristics of future-seis-
mic ground motions are plausible and desirable.
* Smooth "standardized" design spectrum shapes can be used to represent
probable severity of seismic motions. Recommended spectrum shapes de-
veloped from these studies are presented in Figures 4 through 6 for
large, small, and negligible probabilities of being exceeded.
* The approach of using spectrum shapes derived from analyses of the
ground motion with peak acceleration normalized to unity was vali-
dated because there was low correlation between the peak ground ac-
celerations and the spectrum shape values. Separate treatment of
these two as independent variables in these studies was therefore
appropriate.
The followitig significant observations can be made from the results of group
analyses:
* Various seismic parameters appear to influence response spectrum
shape.
o Larger spectral amplifications can be expected to occur at a softer
site. The predominance of long period motion for softer sites and
of short period motion for firmer sites was not confirmed or rejected.
* Distance from epicenter did not appear tc irnfluence spectrum shape.
Predominance of long period motion at longer epicentral distances did
nfrt seem conclusive. Neither, however, was there sufficier,t basis to
reject this possibility.
* The studies by geographic grouping revealed minor variations in the
central tendencies as indicated by the means and medians, although
there were significant variations in the standard deviations or un-
certainties. The ensemble represented a wide-range of frequency con-
tent much better than any other group and thus, it was adopted as the
basis for the recommended spectrum shapes.
-3-
The followring observationý can be made from the comparisons between the rec-
ommended spectrum shapes and the AEC. Newmark, Housner, and Blume F-factor
shapes for a 2% damping ratio:
" The current AEC design spectrum shape is below the small probability
of exceedance shape for periods shorter than 0.5 sec, and those
longer than 0.9 sec.
" Tla Newmark spectrum shape is consistently above the small probabil-
ity of exceed.ince shape, except for a short interval in the vicinity
of zero perioo.
" The Housner spectrum shape is below the large probability of exceed-
once shape for periods shorter than 0.4 sec, and it is above the
latter for longer periods.
" The Blume F-factor spectrum shape for a standardized normal variable
value of 1.0 is consistently higher than the small probability of ex-
ceedance shape.
From the period-ampiltude-time studies, the following observations can be
made:
e The earthquake duration effect on the response spectrum shape is
small for periods shorter than 0.5 sec, the period ringe significant
for the nuclear power plant structures. The dynamic amplification fac-
tor (OAF) at longer periods, however, would generally tend to bc iigher
for long duration motions than for those of short duration.
E. RECOMMENDATIONS
Because risk. minimization is Zhe basic rationale for the seismic design of
critical installations such as nuclear power plants, the seismic load criteria
should consider risk variability with such factors as regional seismicity,
geotectonics, etc.. Thus, design spectrum shapes representing variable -.evvr-
ity of seismic loadings should be specified to achieve appropriate minimiza-
tion of the cotal risk.
-4 -
'hfe follor..ing rLcommendations.pert nent tu the design spectrum sha,.cs are con-
skstent witO this seisr'ic design objective. Other ground motion character-
istics, surh is peak grournd acceleration .-:sd strong r.i-tion duratior (in the
case of time-history analyse! should oe thoroughly considered to fully satisfy
the seismic desion objective. The curves in Figures 4 through 6 are soectrum
shapes, not responst: speLtra theimselves. To derive pseudo absolute accelera-
tion response spectra, it is necessary to evaluate the joint probabilities for
peak ground accelerations and the sdectrum shape va!ues.
" The large probability of exceedance spectrum shapes shown in Fig-
ure 4 should be considered as lower bo'jnd spectrum :hapes.
" For sites associated withrelatively low risk!, e.g., located in
lowi seismic 'y areas, the design spectru',i shapes for different
damoing ratios should not he !o>.ier than the s,:iali probab;lity of
exceedance spectrum shapes shc.4n in Figure 5.
" For sites associated with relatively high risks, e.g., located in
high seismicity areas, the design spectrum ,hapes for different
damping ratios should not be lo.ver than the neg'iqible protability
of exceedance spectrum shapes shown in Figure 6.
It must be noted, howe-'er, that the negligible exceedance prob-
ability spectrum shapes represent extreme ground motion amplific.-
tion. An inappropriate use of these shapes could result in an ex-
tremely low probability seismic exposure and the corresponding de-
sign could be ultraconservative. For example, if a very high peak
ground acceleration estimated deterministically c,:, the basis of an
extreme earthquake expected to occur at a very short distance (say,
at a point on the nearby fault) from the site were to be combined
with the neS;igible probability shape, ar extremely low orobability
seismic exposure wcjld result. The corresponding design would be
ultraconservative because the risk considerations would then be un-
necessarily duplicated by first assuming a highly improbable earth-
.quake occurrence (because the probability of Pn extreme earthquake
occurring i a given Doint is extremely small) and then. determ nis-
tically combining the estimated peak ground acceleration with the
spectrum 5hape. ]h,1refore, it is rec ti:tl.,nded that for the sites
associ atvd with rc I at •,e 1 y hii h risk!., lý'c above rc•;ortmcndatiof
for tht neyl 9i b Ic probability shape b. applied in LIonjunct i cn
wlith an appropriate procedure for probabilistically estimating
the total seismicexposure.
* For sites judged to be significantly responsive to ground motion
compcmno:,ts with periods longer than 0.5 seconds, the above shapes
should riot be used without appropriate mudifications for the par-
ticular it:,. conridhiurns.
-6-
II. INTRODUCTION
A. GENERAL
Earth.quakes are complex phenomena that impose severe loadings on engineered
structures. The earthquake phenomena primarily consisL of large energy
releases in limited volumes of the earth's crust and propagation of signif-
icant parts of these released energies as seismic body and surface waves.
Engineered structures located in the propagation path of these waves re-
spond with vibratory motions. These vibrations generate forces in the struc-
tures which have to be safely resisted. Instances of inadequate structural
resi3tance to seismic loadings with disastrous consequences occur frequently.
Some of the major earthquakes that caused considerable damage are: San Fer-
nando (1971), Tokachi-Oki (1968), Lima (1966), Alaska (1964), Kern County
(1952), El Centro (1940), Long Beach (1933), San Francisco (1906), Charles-
ton (1885), and New Madrid (1811-12).
The severity of vibratory structural response to seismic motion largely de-
pends on the seismic ground motion characteristics and the structure's dy-
namic cuaracteristics. Some of the important grt)und motion characteristics
are the ptak motion parameters, such as acceleration, velocity, and displace-
ment, and the frequency content of the ground motion. The ground motion fre-
quency content can be generally described as a measure of relative predomi-
nance o` different fr..quencies present in the ground motion. Spectrum shapes
are one of the measures of the ground motion frequency content and are im-
portant in estimating seismic structural response.
Nuclear powJer plants are critically important structures, and as such must
be designed for appropriate seismic conditions. Because earthquakes are
complex phenomena and exact predictions of seismic ground motions are not
possible, it is appropriate to base the seismic design criteria on statis-
tical predictions of these motions. Various ground shaking intensity
levels and the probabilities of their being exceeded should be established.
and the seismic levels associated with appropriate probabilities of exceed-
ance should be used in the design.
-7-
One approach to developing statistical predcctiuns of seismic loadings is
to determine probabilities of occurrence of ground motion frequercy char-
acteristics.
The study described herein analysed the frequency ditribution of ground mo-
tions.generated by a number of major %arthquakes. The results were then used
to develop "standardized" design spectrum shape-.
B. SCOPE
The studies in this report were oriented ioward the delineation of appro-
priate shapes of earthquake response spectra to be recommended for seismic
design purposes, based on the rationale described above. The project was
authorized in the USAEC Division of Reactor Standards letter dpted August
18, 1971. The major steps in the studies were as follows:
* Develop response spectrum shapes for !he hr, izontal ground motion
components of a number of selected major earthquakes.
e Study the effect of earthquake duration on spectrum shape by de-
termining the relationship between the spectrum shape and nu.iber of
cycles at predominant periods.
* Perform statistical studies of the spectrum shapes, including group-
ing of the spectra according to various earthquake and site charac-
teristics.
" Recoinmend spectrum shapes appropriate for seismic design and evalua-
tion of nuclear paer plant facilities.
" Compare the results of the above analyses with the current AEC seis-
mic design review procedures.
Studies of each of the above items and the results are discussed in the follow-
ing text.
-8-
Ill. RESPONSE SPECTRUM SHAPES
A. IUT 7%CULIO1It
A number of important and reliable earthquake records of horizontal ground
motion components were selected, response spectrum shapes developed, and sta-
tistical studies made of the ensemble of response spectrum Fhapes. A fairly
large ensemble was used to derive significant results from the studies. The
development or the spectra shapes and the statistical studies are discussed
in this chapter. A number of important and reliable earthquake records of
horizontal ground motion components were selected and response spectrum shape
were developed. The ensemble of response spectrum shapes developed from the
selected earthquake accelerograms and the corresponding statistical studies
are presented in this chapter. To derive significant results from the sta-
tistical studies, a fairly large ensembie of important ground motions was
analysed.
8. SELECTION OF HISTORIC EARTHQUAKES
A total of thirty-three accelerograms generated by twelve different major
earthquakes with peak ground accelerations exceeding O.g were selected
for the analysis (Table 1). Oifferent magnitude ratings for an earthquake
are sometimes reported in the literature because the values reported are usu-
ally averages cf the values estimated at several recording stations. The mag-
nitudes given in Table I were selected because they have been quoted most fre-
quently. The rationale for selecting the earthquake records was as follows:
0 The accelerograms were considered to be reliable records of the
ground motions because most were recorded recently. Those not
recorded recently, such as El Centro and Helena. have been ex-
tensively and reliably documented.
* The accelerograms were associated with fairly intense ground shaking.
* A wide-range of response spectra characteristics was represented by
the selected accelerograms. The Linia and Hachinohe records were in-
cluded because the Lima records contain a predominant high-frequency
Q9
content and the Hachinohe records a predominant low-frequency con-
tent. Inclusion of these records Lhus increased the range of dif-
ferent spectrum shape characteristics.
" The accelerograms included eight different and important records
from the 1971 San Fernando earthquake.
" The accelerograms comprise a reasonably large ensemble, thus yield-
ing a higher degree of confidence and reliability to the studies.
" The effect3 of geographical variations were included because the
accelerograms were recorded at a number of different geographical
locales.
C. ANALYTICAL ASPECTS OF A RESPONSE SPECTRUM
A response spectrum is defined as a plot of the maximurn values of a response
parameter of a family of linearly elastic single-degree-of-freedom systems
with different frequency characteristics and with a given ratio of system
dai•ping to critical damping wher, subjectcd to a ground motion time-history
versus the frequency characteristics (such as natural periods or frequencies)
of the systems.
For the purposes of this report, the pseudo absolute acceleration response
spectrum shapes for damping ratios of 0.005, 0.01, 0.02, 0.05, 0.07, and 0.10
were developed for the selected accelerograms.
Figure I shows a model of a single-degree-of-freedom system. The spring is
linearly elastic with stiffness k and the dashpot indicates a viscous damper
with a damping coefficient c.
- 10 -
kg
FIGURE 1. SINGLE-DEGREE-OF-FREEDOM SYSTEM
The equation of motion of this system is:
m)n +c; + kx -i (1)
in which
m = mass
x = displacement relative to the ground [L]
d_ velocity relative to the ground [LT- 1 ]it
- d x• acceleration relative to the ground [Lf iId t
c = damping constant [;.IT- ]
k = spring constant [ MT- 2 ]
x = ground motion acceleration [LT" 2 ]g
. [!] indicates the dimension of the quantity. The basic dimensions
are mass, M; length, L; time, T.
Equation (I) can be rewritten as:
+ "+ - . (2)
in which
4-k = circular frequency Tm[t]C -
-2rr= damping ratio
The solution of Equation (2) is given by
tX(tI, - - f h(t-r) T () J
in which
h(T) = impulse rusponse function [T]
and
w damped circular frequency ( T-' ]
==
For a moderate amount of damping, w nearly equi-Is w. As example, for
a relatively high damping value of 20%, w is equal to 0.98;. Because
the damping values for most of the dynamic analyses of nuclear power
plant elements are considerably less then 20%, the difference between
w and w is negligible. Thus, w will be u3ed in place of w. The
system spring force, F, at an instant, t, is given by
F(t) = K.(t) _ mw2 (t) (5)
Let z (t) = W2X(t) (6)p
Then F(t) = m i (L) (7)
'"has the dimensions of acceleration, (LT-], and is termed pseudo
absolute acceleration. Comparison of Equation (6) with Equation (2)
indicates the reason for this quantity being termed "pseudo", namely,
- 12 -
the magnitude of P differs from that of the real absolute acceleration by
a quantity, 2cwý. The pseudo relative velocity, p, is defined by
x 2 [L11 (8)
ppThen, accepting •••
P ( t• = t (• .:t) (9 )
In accordance with the definition of a response spectrum, the pseudo absolute
acceleration response spectrum, S z is given by
0) I .0)
in which
n= atural period [:3
depends on the system period, , the system damping ratio, ,, and the in-
put ground motion, . Equation (1l) is derived from theoretical considera-
tions and Equation
Thus, for a rigid system =, 0), the pseudo absolute acceleration spectrum
value equals the maximum ground acceleration, a. This is one reason why nor-
malization of ground motion by peak ground acceleration is convenient. The
spectrum value generated from the ground motion normalized by the peak accel-
eration is actually the dimensionless ratio of the pseudo absolute acceleration
spectrum value to the peak ground acceleratior anJ is termed dynamic amplifica-
tion factor (DAF) for pseudo absolute acceleration.
- 13 -
Let D(T,,,) = dynamic amplification factor (OAF) for pseudo
absolute acceleration
(T •.- )
Then 1)(1 "") r_ (12)
and C)((;,.. ) = 1., (131
Thus, for zero period, equals 1.'].
0. RESPONSE SPECTRUM SHAPES FROM HISTORIC EARTHOUAKES
IThe obj~'ctive of the studies was to analyze the shapes of the respon.•e s.pec-
tra, not the response spectra themselves. To facilitate the c'amp'arison of
the spectrum shapes, the accelerograms w!re riormalized to a peak gruund a.:cel-
eration of 1.0g. i'hus, the spectral vaiues obtained are dimensiconless ratios
o'f spectral acceleration to peak 4"rouod acceleration, or dynamic'amplification
factors (DAF). The response spectra generated from such normalized ground
motions are term.td response spectrum shapes, or OAF. The response spectrom
shapes are useful for comparing the relative predominance of different fre-
quencies in an individual arcelerogram and for different accelerograr.s. Accel-
erograms can be normalIzed by several different methods. For t'ýe purposes of
his study, normalization by peak ground acceleration was co:widered most use-
ful and convenient, especially for the stotistical predictions of response
spectrum shapes.
A recursive algorithm' was used to compute the response spectrum shapes of
the selected historic earthquakes for damping ratios of 0.005, 0.01, 0.02,
0.05, 0.07, and 0.10. Because the method of digitizLiion of accelererjrams
significantly affects the response spectra,' the nwst receat and most reli-
able accelerogram digitizations were used in ihe in-jly-ses.
The spectrum shapes were computed for a period range: of 0.04 sec to 2.5 sec,
or a freqLency range of 0.4 cps to 25 cps. which was considere."- sufficient
to encompass the frequency characteristics of nuclear power plant components.
This range of frequency characteri,ýtics was discretized by us;g 108 points
to obtain good frequency resoiution of the spectra.
- 14 -
The Fourier content of digitized accelerogram data may be accurate up to
about 25 cps, 8 and thus the upper bound on the frequency range of the ýpc-
tra was set at 25 cps. Most of the accelerogram data were as described in
Reference I and hence the accelerogram corrections, such as smoothing of the
fixed trace, smoothing of the timing marks, and the root-mean-square minimi-
zation of acceleration were applied to the data. Generally, the accelero-
grams are also corrected by shifting and/or rotating the base line or using
a band-pass filter. Such corrections are necessary only for computing ground
motion velocities and displacements. Because they do not significantly af-
fect the computition of the pseudo absolute acceleration response,•,• these
corrections were not considered necessary.
Linear plots of the spectrum shapes are presented with period as abscissa and
alternatively with frequency as abscissa in Appendix A as Figures Al through
A66. These two way!. of plotting complement each other, representing the spec-
trum details more fully. Frequency-plots were used in determining the spectral
values for the high-frequency elements and period-plots were used for the long
period elements.
- 15 -
TABLE 1SALIENT CHARACTERISTICS OF
SELECTED ACCELEROGRAMSPeak GroundAcceleration,
Magnitude Component g UnitsEarthquake
El Centro
El Centro
Kern County
Olympia
Helena
San Francisco
Year Recording Station
1940 El Centro,Californi3
1934 El Centro,California
1952 Taft,CalifLrnia
1949 Olympia,Wasnington
1935 Helena,Montana
1957 Golden Gate Park,California
1966 Cholame-Shandon # 2,California
19%6 Chnlame-Shandon 0 5,California
Parkfield
Parkfield
Tokachi-Oki 1968 !!.chinohe,Japan
1966 Lima, Peru
7.0
6.5
7.7
7.1
6.0
5.3
5.6
5.6
7.8
7.5
6.6
6.6
6.6
6.6
6.6
6.5
5.6
N2 P ES69*E
S860W
UIS
1180 0W
UW5ES25OW
N5'1 W14115u 11
NSEW
EW
0.51Not Recorded
0.400.47
0.330.22
0.260.18
0.180.16
0.190.31
C. 130.16
0.110.13
IUSEW
Lima
San Fernando
San Fernando
San Fernando
San Fernando
Eureka
Olympia
Parkfield
1971 Castaic, ORR,California
1971 Bank of Calif.,California
1971 Universal-Sheraton, Calif.
1971 V.N. HolidayInn, California
1954 Eureka,California
196ý Olymoia,Washington
1966 Temblor,California
N8*°EN82 0W
N210 ES69°E
N11 0 EN79 0W
NSEW
NSEW
g790 E
S4 E
S860W
N650WN250 E
0.190.23
0.420.27
0.340.29
0.230.14
0.180.13
0.280.15
0.260.18
0.200.16
0.280.33
NOTE: Information presented in the above table is compiled from
References 1, 2, 3, 4, and 5.
- 16 -
IV. STATISTICAL ANALYSES OF SPECTRUM SHAPES
A. INTRODUCTION
The analyses in this chapter were oriented toward statistical predictiors of
the characteristics of future ground motions. Spectrum shape statistics such
as mean, median, and standard deviation were computed for the complete en-
semble of spectrum shapes discussed in Chapter Ill. These statistics were
also computed for the accelerograms grouped irn four differen: ,ýays:
" max;mum ground acceleration.
" so-l characteristics at the recording site,
" epicentral distance, and
* geographical locale of the recording stations.
The median and mean values indicate the central tendency of a sample. The
standard deviation measures uncertainty associated with the sample. In most
of the cases, the combination of either the mean or me-dian and the standard
deviation suffices for statisLical predict;ons.
B. STATISTICAL ANALYSIS TECHNIQUES
In an ensemble of r response spectrum shapes, shown in Figure 2, let i:f(!.,-)
denote the spectral ordinate of the ith spectrum shape for the period. , nd
the damping ratio, %. Then, the mean, -•.(T ,,)0 and standard deviation,
are computed by the following equations:
i( 1
r ,-" , II, ( Ss (T*~ -1- fT ) I!
- .17 -
The median of the ensemble, ry.)(T-,.), is determined as follows: First, the
spectral values, D.(Lt., i 1, 2 . ,are rearranged in a descending
order. Then, if ni is an even number,
7:) (16a)- J
If n is an odd number,
M, 0 •6b)
The mean spectrum shape, m,'(G'A, the median spectrum shape, , , and the stan-
dard deviation spectrum shape. , are constructed by computing means, me-
dians, and standard deviations for 108 periods in the period range under con-
sideration. The mean and median spectrum shape values For zero period equal
unity and the standard deviation spectrum value fer zero period equals zero be-
cause the DAF value for zero period is constant and equals unity. From Equa-
tions (14) and (16). the actual values of the mean and redian shapcs will not
be equal, if there is a skew in the data.
C. STATISTICS OF SPECTRUM SHAPE ENSEMBLE
The mean. median, and ýtandard dev;ation response spe,:trum shapes for the corn-
plete ensemble of the selected accelerograms and for all damping ratios were
computed as described above and are presented in Appendix B as Figures 81
through BIB. The following observatiuns regarding these spectrum shap o-.are
pertinent:
* Th,: mean and median spectrum shapes, indicators of the central ter;-
dency of the ensemble. are similar for each damping ratio. Th,.
small quantitative difference between these two s5atistics i'dicltes
that the DAF data his ,omv, -,kewnesý. See Chapt-r VI for a detailed
discussion of the DAF distributicon,.
* The mean arid median spectrum shapes are srmooth. As expected, the
smoothness increases with the damping ratios.
- 1 -
*A's the period approacnes zu-o the rmean ar'd rx!<;ar' jhpe ie
approach unity and thie s~andard deviat oni approaches zerc. as
expt2c ted . The s t.3toddrd Jvv i Is oar, dec rcta;e s a s t~ hep: n-r-j,*,
Tha 5 dec rea S tow * ,- r , ir s much smai I Ier t hav the cor rt~svo-e- .nq cc-
creije in the riwjr. arl:j lc~ ~ hu, *:" 'he it e~ne
in the OA-F increases w~itt t~e periodI. ese~al for ztcricds longer
mhan 0.25 -,cc For ;,v,-c,v,. i(~r'(er thr' 0.C9 sec. * " rvltjlt: .vJ-
certainty is st~s:ar't ialIlv larger t. i~ar for zhe shc~r te r per; wl
a ýa.i ri1- rap i ta~ trend Zk.nt ; nues tý 2 .5 ýt~c per; -;C.
The mo n reanor,~ fr.r '.!t: rc.c in xct.rt a nt,.. t per io
thle cl*,sLeml Ie Con' a ins sev..-,-I! I Lce I crc r.vr'. .,tr va !le~ Sricr t .4r
long s Y rong q -io iin a i&. . de~ .*i -~ r* f jte Ii
C e r tI i thv s ta)r u ura a o t: r r , is > ,ho- pr.* : ý i, ma r ir,C u
short per iods onl If *nrv- ~Ch~ t h Ifanq Jur at or~n 0-c thi. r~rv-
jo~r~i nance of both, short ar~d Ifloq 7a " WS . ts Mt' CV er-tr
dc ncy o tM hfe en bem b s a cc t!r. t u, i vs in th ) S~r IvE erat r arq t
an c unt i e r ab uItv mUe c~t 1 or J ;cfri k r lr.ge c Dus Ut. -I f-
t aI veIy l a r e d v v t i un~ 5f D£AF i i t rti v Inn tr~ r an t:. n j, %
ttv re iat vt uncur aart in tna. rant-e gr (r.'.t4'r *han (or
S h ()r tp t:riod r , jn ge
* The r~eafl and rmed ian spec ,rur, shapes dec reast: -. rL a~t ? ttr .tih
inc~reasing dam~pi ng rat ios . we;c ;nd;Cdte- reliti;vely qrvater dvc
crease fo. the ',hor t 'vtri od. t han for ihe Ir'nq per o(dS . The %tan-
dard deviation spectrum shav'ý,z also decreoae ane boeconeS flat*.?ýr ~..th
increasinq dam-inq ratio. This, dj~rre.se i, ia r hf-wt.~'er.t:K~
in the mea, and median and ;., e.spLuciaily t ruc for t~heŽ Ionq :)eriod!,
w.hich result,~ in great-.-r fIatnes.-. Thu0 5 * t he r-! 1 i vu wncer t a nt y
i n OAF va lues rumia ins PractLi ca I I i nd ifferent to t ho change', i T
d,,mping ratio. Increarer camping tend% to at.Cnuatv retpon-,L. more-
for sh.crt periods thanz fur long period',.
- '9
D. bkiAVSTICS OF GROUPED SPECTRUM SHAPES
I. General
The rationale of the grouping approach will be briefly discussed first
so that the group statistics may be appropriately interpreted.
A number of earthquake characteristics could influence response spectrum
shapesu Grouping of the records by a selected characteristic helps to
investigate and estimate the influence of such characteristics on re-
sponse spectrum shapes. Such an investigation has maximum sigificance
if performed under the ideal condition that only the characteristic under
consideration is varied wHile others are kept constant. It is apparent
fron. the sparsity of the available ground motion data, however, that this
is not fcs;ble. It is helpful, nevertheless, to understand what infor-
mation could be obtained from the grouping approach under ideal conditions.
Assume that a seismic characteristic, s, in the range, s; S < G, favors
a frequency range, f < f < f,,, over other frequencies. Then, the mean,
median, and standard deviation spectrum shapes for the accelerograms in
the group, s .: <, would be expected to display the following char-
acterisLics:
* In the frequency range, f < f : f the mean and median spectrum
values for this group will be generally higher than those for the
other groups, indicating the preference of this group. The same be-
havior will be expected when the ensemble mean and median are com-
pared with the group mean and median.
0 In the frequency range, f < f < f 2# the standard deviation spectrum
shapes will be generally lower than those for the other groups, in-
dicating a smaller degree.of uncertainty associated with the spectrum
values for the preferred frequencies. The same behavior will be ex-
oected when the standard deviation shapes of the ensemble are compared
with those of the group.
Higher spectrum shape values for the preferred frequency range will be
more probable than for the other frequencies. The reliability of such
- 20 -
an interpretaticn depends on two factors, nameiy, the number of accel-
erogram samples in the group, and the range of seismic parameter values
defining the group. The reliability will increase with an increase in
the number of samples, and decrease with an increase in the range of
seismic parameter values.
Statistics of the spectrum shapes grouped in four different ways were com-
puted to investigate the influence of different earthquake characteristics
on the response spectra. The accelerograms were grouped according to:
* the maximum recorded ground acceleration (Table 2),
* the soil characteristics at the recording station (Table 3),
* the epicentral distan.e of the recording station
(Tabie 4), and
* the geographical locale of the recording stations (Tables 5, 6, 7,
and 8).
The specific details of the groups and the analytical results are signif-
icant. Different values for a characteristic ef an earthquake are some-
times reported in the literature. In such cases, an apparently reason-
able value of the characteristic was adopted in the analyses without
further investigating the allthenticity of the value. It would be worth-
while, however, to c.)nduct such an investigation to improve the reliabil-
ity of the available information.
2. Maximum Ground Acceleration
Table 2 lists a grouping of accelerograms according to maximum ground ac-
celerations. This grouping of accelerograms described below achieves a
reasonable balance between group size and range of accelerations.
Maximum Number ofGroup Ground Acceler4 .. tion, a Accelerograms
Al a > O3g 8
A2 0.2g9 a < 0.g 10
A3 a < 0.29 15
21 -
The frean, median, and standard deviation spectrum shapes for the three
groups of accelerograms for a 0.02 damping ratio are presented in Appen-
dix C as Figures Cl through C9. The following are pertinent observations
regarding these spectrum shapes:
* The Group Al mean and median shapes are generally somewhat lower than
the mean and median shapes of the Groups A2 and A3 (Figures Cl, C2,
C4, C5, C7, and C8). The Group Al standard deviation spectrum shape
for periods longer than 1.0 second is considerably lower than the ones
of the other groups for the same period range (Fiqures C3, C6, and
CC). This means that higher OAF values for Group Al are less probable
than they would be for Groups A2 and A3. This is particularly true
for periods longer than 1.0 second; in other words, higher OAF values
occur for accelerograms with a 0.3,g.
* No cluar trends are apparent for the Groups A2 and A3.
In spite of the first observation above, the maximum ground accelerations
of the accelerograms generally have low correlation with the correspond-
ing OAF values. Thus, probabilities for the DAF values, as discussed in
Chapter VI, may be considered independent of those for the ma.6mum ground
accelerations. This conclusion leads to this important result:
From Equation (12),
Sa(T_~r,;j D U.(,•• (17)
Thus, the probabilities of exceedance for the spectral ordinate, 2 used
for computing response of structures, can be derived by a rather simple
combination of the probabilities for the peak ground acceieration, a, and
the DAF for pseudo absolute acceleration, D. This method of combination
of probabilities is discussed further in Chapter VI.
3. Soil Characteristics
The following site characteristics could influence the response spectrum
shape:
- 22 -
* Soil density
" Soil layering
" Layer thicknesses
• Depth to firm rock
* Water table elevation at the site
* Soil moisture content
" Shear and compressional wave velocities in the soil
* Nature of soil behavior -- linear or nonlinear
It is apparent from the above list that the influence of the soil char-
acteristics on the response spectrum shape is a complex phenomenon. In-
vestigation of this phenomenon is asid has been treated as an independent
field. 1 1 Thus, the subject can be treated only in an approximate way in
the present analysis.
It was shown 11,12 that the soil impedance, 1, is determined by Equa-
tion (18) is one of the factors of considerable importance in this kind
of analysis.
I (18)
in which
P = specific soil density
V = velocity of shear wave in soils
In the present analysis, soil impedances of the top layrrs at the record-
ing stations were used as the basis for grouping of the accelerograms.
In the case of a site with a number of shallow soil layers, an overall
average value of impedance was used. For the selected accelerograms, in-
formation on the soil characteristics at various recording sites is sparse
and contains a high degree of uncertainty. More detailed soil, geological
and geophysical investigations of various recording station sites are
needed to form a Firm basis for the kind of analysis described in Sub-
section I of this chapter.
- 23 -
The grouping of accelerograms according to their site impedance in
Table 3 can be condensed as follows:
Number ofLrm Impedance, I, (ft/sec)xlO Accelerograms
81 4.0 < I < 5.5 13
82 1 < 3.9 16
The accelerograms recorded at Helena and Temblor are not included in
these groups because the estimated impedance values for these sites,
based on the available data, are considerably higher than those of the
other sites and could result in Pn anomalous influence on the analysis.
The mean, median, and standard dev.iation spectrum shapes for the accel-
erograms in these two groups for 0.02 damping ratio value are presented
in Appendix C as Figures CIO through C15.
The following are pertinent observations regarding these spectrum shapes:
" The mean and median shapes for each group are generally similar
to each other.
* The mean shape of the Group BI is generally lov.ier than that of the
Group B2.
" No clear trend is seen in the standard deviation shapes.
It has been foundl 2 that the accelerograms recorded at firmer site5 gen-
erally show lower DAFs than those for the accelerograms recorded at
softer sites. This is indicated by the second observation above. It
would seem reasonable that ground motions recorded at soft sites would
show a predomitance of lower frequencies and those recorded at firm
sites would have higher predominant frequencies. Such a trend was not
strongly indicated by the results of this analysis because of the uncer-
tainty associated with the soil characteristics data, however, these re-
sults should not be considered as a negation of this belief.
- 24 -
4. Epicentral Distances of Recording Stations
The accelerograms were arranged according to the epicentral distances
of the recording stations as follows (Table 4):
Epicentral Distance, D Number of
Group Miks Arcel erog rams
Cl 15 < D < 45 14
C2 L < 15 15
The accelerograms recorded at Hachinohe and Lima were not included in
the analysis because the reported epicentral distances for these rec-
ording stations are much longer than those for the other stations and
hence could have an anomalous influence on the analysis.
The mean, median, and standard deviation spectrum shapes for the accel-
erograms in the Groups CI and C2 for a 0.02 damping ratio are in Appen-
dix C as Figures C16 through C21. The following are pertinent observa-
tions regarding these spectrum shapes:
* The mean and median shapes for each group are similar to each other.
* The mean shapes for both the groups are approximately equal.
* The standard deviation shape for the Group Cl is considerably lower
than for the Group C2 for periods longer than 0.5 second.
Although the comparisons of the median spectrum shapes do not clearly in-
dicate predominance of lower frequencies for accelerograms recorded 3t
longer epicentral distances or predominance of higher frequencies in
records with shorter epicentral distances, the third observation above
would tend to confirm these effects. As other parameters, such as soil
characteristics, would also influence the response spectrum shapes, the
absence of ciear trends in this grouping are not unexpected. In addition,
epicentral distance per se may not be a significant factor with respect
to the response spcctrum shapes. This may be rarticularly true for earth-
quakes associated with fairly Ior:ng aul breaks. In such cases, the
- 25 -
distance between the station and the nearest point of surface rupture
might be more appropriate. Also, the reported epicentral distances may
not be the actual ones because epicenters cannot be located exactly.
E. GEOGRAPHIC GROUPING OF SPECTRUM SHAPES
The selected accelerograms were recorded at a number of different geographical
locales in seismically active areas of the world. The accelerograms were gen-
erated by earthquakes that originated in different geological settings and per-
haps by different earthquake source mechanisms.
In all, spectrum shapes for five different geographic groups of accelerograms,
including the ensemble, were statistically analyzed as follows:
Number of
Group Geographical Locale Accelerograms
I San Fernando Valley 8earthquake
II Southern California 19
III California 23
IV Western U.S.A. 29
V Worldwide 33
The salient characteristics of the thirty-three accelerogram ensemble are
listed in Table I (Chapter III); those for the remainder of the above groups
are reproduced in Tables 5 through 8.
Mean, median, and standard deviation spectrum shapes for Groups I through IV
for 0.02 damping ratio are presented in Appendix C as Figures'C22 through
C33. Ensemble spectrum shape statistics for 0.02 damping ratio (Group V) are
in Appendix B as Figures B7 through B9.
A number of pertinent observations regarding comparisons of the spectrum shape
statistics for a geographical grouping can be made:
0 The mean and median shapes for each group are similar to each other.
The mean shapes are generally smoother than the median shapes.
- 26 -
" The Group I mean spectrum shape is generally higher than those for
the other groups, especially for the period ranges of 0.? - 0.4 sec
and 0.7 - 2.5 sec. In the latter period range, this shape is sub-
stantially higher than those for the other groups which indicates
a strong predominance of long period motion. In the short period
range, 0.04 - 0.02 sec, the Group I shape is slightly lower than
the Group V mean shape.
* The Group II mean spectrum shape is somewhat higher than those for
Groups III, IV,.and V. This tendency is more pronounced for the
longer period range (greater than 0.5 sec). This shape is slightly
lower than the Group IV and V shapes in the short period range, 0.04
- 0.2 sec.
* The mean spectrum shapes for Groups Ilt, IV, and V are close to each
other.
* The standard deviation spectrum shapes show more variations than the
mean shapes.
" The Group I standard deviation spectrum shape is higher than those
for the other groups in the period ranges of 0.2 - 0.4 sec and 1.0
- 2.5 sec, with the exception that the Group V shape is substantially.higher in the range of 1.0 - 1.25 sec but lower than the others in
the remaining period ranges.
* The Group II and IIl standard deviations are similar. Both are sub-
stantially lower than those for Group IV and V in the short period
range, 0.04 - 0.1 sec.
The Group IV standard deviation shape is similar to that for Group
V except it is lower than the latter in the period ranges of 0.04 -
0.1 sec and 0.8 - 1.2 sec.
* The ground motions recorded in Califor•iia are predominant in the
period ranges of 0.2 to 2.5 sec. This is particularly true for
the accelerograms generated by the San Fernane.dA earthquake. Incor-
poration of the spectrum shapes from the other parts of the western
U.S.A. and worldwide results in predominant dynamic amplification
in the short period range, 0.04 - 0.2 sec.
- 27 -
It appears appropriate to develop standardized design spectrum shapes on
the basis of the ensemble spectrum statistics for the following reasons:
0 The ensemble mean and standard deviation spectrum shapes appear to
encompass broader ranges of frequency content than those of any
other group. Thus, the predictions based on the ensemble statistics
would better represent the conditions not yet recorded in local areas
in addition to those recorded.
0 The reliability of the statistical measures and the predictions
based on them Increases with the sample size. The ensemble consists
of more earthquake records than any other group. Therefore, the
ensemble statistics and predictions based on them would be more re-
liable than those for any other group.
F. SUMMARY OF FINDINGS
Statistical analyses of the ensemble of response spectrum shapes and the
grouped spectrum shapes were presented in this chapter. The spectrum shapes
were grouped in four ways; by maximum ground acceleration, by epicentral
distance, by soil characteristics, and by geographical locales. The major
findings were:
& The mean and median spectrum shapes for the groups are smooth In com-
parison with the shapes for individual accelerograms, and they are
similar to each other. The median shape is less smooth than the
mean and shows small quantitative variations from the mean, indica-
ting some skewness in spectrum shape data.
• The mean and median shape values decrease for longer periods and
higher d'mping. The rate of decrease with longer periods, however,
decreases with higher damping, resulting in flatter shapes for
higher damping.
The standard deviation spectrum shapes show greater variations than
the mean and median, indicating greater variability of uncertainty.
- 28 -
" The standard deviation shapes also generally decrease with longer
period and higher damping. However, tI, rates of decrease are rela-
tively smaller than those for the means and medians, indicating in-
creased relative uncertainty for longer periods and an indifference
of relative uncertainty with respect to .1iping.
" The influences of various seismic parameters on response spectrum
shape were indicated from the grouping approach. The significance
of the results of this part of the study, however, would be enhanced
if the followiing conditions could be met:
I. The uncertainties associated with some of the available data
are investigated and minimized.
2. A large number of reliable accelerograms are available so as to
allow variation of only one parameter while the others are kept
constant.
Because reliable ground motion data are sparse, it is difficult to satisfy the
second condition. The following significant observations, however, can be made
from the results of group analyses:
* The approash of using spectrum shape derived from the ground motion
analysis by normalizing the peak ground motion to u.nity was confirmed
because the peak ground accelerations and the spectrum shape values
had low correlation. Separate treatment of these two as independent
variables was therefore appropriate.
" Spectral amplification at a soft site can be expected to be larger
than at a firm site. The expected long period predominance for soft
sites and short period preaominance for firm sites were not confirmed
or rejected.
N The influence of epicentral distance on spectrum shape diminishes
with the increasing epicentral distances. Tne expected long period
predominance at longer epicentral distances was not indicated; nei-
ther was the rejection of this belief indicated.
- 29 -
* The results of the geographic grouping revealed that the central ten•-
dencies of the groups, indicated by means and medians, do not vary
significantly with different geographical locales. Although the un-
certainties or standard deviations did show significant variations,
the ensemble represented a wide-range of frequency content much bet-
ter than any other group. Thus, the ensemble was adopted as the
basis for the recommended spectrum shapes.
- 30 -
TABLE 2
GROUPING OF ACCELEROGRAMS ACCORDING TO MAXIMUM GROUND
ACCELERATIONS
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Accelerogram
Parkfield #2
Parkfield #5
Lima
Parkfield #5
Castaic ORR
El Centro, 1940
Temblor
Olympia, 1949
Castaic ORR
V.N. Holiday Inn
Temblor
Lima
El Centro, 1934
Eureka
Hachinohe
Bank of California
El Centro, 1940
Olympia, 1965
Olympia, 1949
Hachinohe
El Centro, 1934
Taft
Universal-Sheraton
Eureka
Taft
Helena
Olympia, 1965
V.N. Holiday Inn
Bank of California
Helena
Golden Gate Park
Universal-Sheraton
Golden Gate Park
Component
N65°E
U85°E
118° E
N 50W
N210E
NS
N250 E
S860W
S690E
NS
tN65 0W
N820W
NS
N79 0 E
EW
N110E
EW
S4°E
N4°W
NS
EW
N210E
NS
N11IW
S69 0 E
EW
S86 0 W
EW
N790W
NS
N80OW
EW
NIO0 E
- 31 -
Maxi mum GroundAcceleration,
g Units
0.51
0.47
0.420.40
0.34
0.33
0.33
0.31
0.29
0.28
0.28
0.27
0.26
0.26
0.23
0.23
0.22
0.20
0.19
0.19
0.18
0.18
0.18
0.18
0.16
0.16
0.16
0.15
0.14
0.13
0.13
0.13
0.11
Group AI
Group A2
;roup A3
TABLE 3
GROUPINGOF ACCELEROGRAMS ACCORDING TO T11E
SOIL I1*EDAdCE AT THE RECORDING STATIONS
Numbe
12
3
4r
6
7
8
9
10
11
12
13
14
15
Recording Station Impift
Helena
Temblor
Golden Gate Park
Cholame-Shandon r2
Cholame-Shandon r5
Ilachinohe
Castaic, ORR
Eureka
Lima
Van Nuys Holiday Inn
Olympia
Bank of California
Universal-Sheraton
Taft
El Centro
Note: Information presented in the aboveand/or compiled from References 5,16, 17, 18 and 19.
edance, I
x IO•
19.1
10.3
5.4
4.34.3
4.3 Group B1
4.3
4.3
4.2
3.0
2.1
1.6 Group B2
1.6
1.3
1.1
table is deduced11, 13, 14, 15,
- 32 -
TABLE 4
GROUPING OF ACCELEROGRAMS ACCURDING TO
EPICENTRAL DISTANCE OF THE RECORDING STATIONS
Number
1
2
3
45
6
7
8
9
10
11
12
13
14
15
16
17.
Recording Station
Hachi nohe
Li ra
Taft
Olympia, 1965
Olympia, 1949
El Centro, 1934
Castaic ORR
Bank of California
Uni versdl-Sheraton
Eureka
El Centro, 1940
Van Nuys Holiday Inn
Golden Gate Park, S.F.
Helena
Temblor
Cholame-Shandon #5
Cholame-Shandon "2
Epi centralDistance,
M~iles
.100
10G
44
35
31
20 Group Cl
18
18
15
13
13
84 Group C2
43.3*
O.Ob*
* Shortest distance from the San Andreas fault
Note: Information presented'in the above tablefrom References 2, 3, 4, 5, 13, and 14.
is compilea
- 33 -
TABLE 5
GEOGRAPHIC GROUP I:
SAN FERNANDO VALLEY EARTHQUAKE
Peak GroundAcceleration,
q UnitsEarthquake
San Fernando
San Fernando
San Fernando
San Fernando
Year Recording Station
1971 Castaic, ORR,California
1971 Bank of California,California
1971 Universal-Sheraton, Calif.
1971 V.11. Holiday,Inn, California
Magnitude Component
6.6
6.6
6.6
6.6
N21°ES69"E
N179°W
NSEW
NSEW
0.340.29
0.230.14
0.180.13
0.280.15
- 34 -
TABLE 6
GEOGRAPHIC GROUP II:
SOUTHERN CALIFORNIA
Peak GroundAcceleration,
g UnitsEarthquake Year Recording Station Magnitude Component
El Centro
El Centro
Kern County
Parkfield
Parkfield
San Fernando
San Fernando
San Fernando
San Fernando
Parkfield
1940
1934
1952
1966
1966
1971
1971
1971
1971
1966
El Centro,California
El Centro,California
Taft,California
Cholame-Shandon #2California
Cholame-Shandon #5California
Castaic, ORRCalifornia
Bank of Califor'da,California
Universal-Sheraton, Calif.
V.N. Holiday.Inn, California
Temblor,California
7.0
6.5
7.7
5.6
5.6
6.6
6.6
6.6
6.6
5.6
NSEW
NSEW
N21°E$69°0E
U65 0ENot Recorded
N5°WN850E
N21°E$69°E
NIl 0EN70°WUS
EW
NSEW
N650WN25=E
0.330.22
0.260.18
0.180.16
0.51
0.400.47
0.340 .29
0.230.14
0.180.14
0.280.15
0.280.33
- 35 -
TABLE 7
GEOGRAPHIC GROUP III:
CALIFORNIA
Earthquake
El Centro
El Centro
Kern County
San Francisco
Parkfield
Parkfield
San Fernando
San Fernando
San Fernando
San Fernando
Eureka
Parkfield
Year Recording Station
1940 El Centro,California
1934 El CentroCalifornia
1952 Taft,California
1957 Golden Gate Park,California
1966 Cholame-Shandon #2,California
1966 Cholame-Shandon #5,California
1971 Castaic, ORR,California
1971 Bank of California,California
1971 Universal-Sheraton, Calif.
1971 V.N. Holiday,Inn, California
1954 Eureka,California
1966 Temblor,California
Magnitude
7.0
6.5
7.7
5.3
5.6
5.6
6.6
6.6
6.6
6.6
6.6
5.6
Component
NSEW
NSEW
N210ES69°E
N1O0 EN90OW
N65°ENot Recorded
N50WN85°E
N210ES69°E
N11 0 EN79 0W
NSEW
NSEW
N79 0 EN11°W
N65 0WN25°E
Peak GroundAcceleration,
gi Units
0.330.22
0.260.18
0.180.160.110.13
0.51
0.400.47
0.340.29
0.230.14
0.180.14
0.280.15
0.260.18
0.280.33
- 36 -
TABLE 8
GEOGRAPHIC GROUP IV:
WESTERN U.S.A.
Earthquake
El Centro
El Ce',tro
Kern County
Olympia
ielena
San Francisco
Parkfield
Parkfieild
San Fernando
San Fernando
San Fernando
San.Fernando
Eureka
Olympia
Parkfield
Year
1940
1934
1952
1949
1935
1957
1966
1966
1971
1971
1971
1971
1954
1965
1966
Recording Station
El Centro,California
El Centro,California
Taft,California
Olympia,Washington
Helena,Montana
Golden Gate Park,California
Cholame-Shandon 62,California
Cholame-Shandon #5,California
Castaic, ORR,California
Bank of California,California
Universal-Sheraton., Calif.
V.N. Holiday,Inn, California
Eureka,California
Olympia,Washington
Temblor,California
Magni tude
7.0
6.5
?. 1
7.1
6.0
5.3
5.6
5.6
6.6
6.6
6.6
6.6
6.6
6.5
5.6
Component
NSEW
NSEW
U21°E569 0 E
N40 WS860 W
NSEW
N 10 0 EN801 W
N65°ENot Recorded
N50 W1485o E
N210ES690 E
N11 0EN790W
NSEW
NWEW
N79 EN11°W
S40ES860W
N650WN250 E
Peak GroundAcceleration,
Units
0.330.22
0.260.18
0.180.16
0.190.31
0.130.16
0.110.13
0.51
0.400.47
0.340.29
0.230.140.180.130.280.150.260.18
0.200.16
0.280.33
- 37 -
D1
-41.0
D2
1.0
Di
-1.
1. 0
DI (Tj, i)
T . Period, T (sec)
D 2 (Tj, r')
T .3 Period, T (sec)
D- (T.,
"j,
T.3 Period, T (sec)
(Tj, )
I 3T Period, T (sec)
FIGURE 2. ENSEMBLE OF n RESPONSE SPECTRUM SHAPES
- 38 -
V. EFFECTS OF EARTHQUAKE DURATION
A. GENERAL
It is generally thought that long duration shaking is more damaging to build-
ings than short duration shaking, even though the amplitues may be equal. Al-
though the duration of strong shaking may not directly affect the onset of
damage, damage could be increased by continued strong motion. Thus, it seems
reasonable to postulate that damage potential of an earthquake is directly
dependent on the response levels induced and their duration. For this reason
the effects of- duration of strong seismic motion were included in this study.
The response spectrum, which is normally used to define seismic input motion,
is a plot of the maximum values of a specified response paramLter. The re-
sponse spectrum permits ready evaluation of the frequency content of the
seismic input motion. However, the duration of the maximum responses and
their frequency of occurrence are not discernible from the spectrum curve.
It is therefore possible that response levels smaller than the maximums may
exist, which might induce greater structure damage. The study was structured
so as to investigateresponse levels, in addition to the maximums, that might
be significant from thp-viewpoint of potential damage, and how these levels
and their duration are affected by the earthquake duration.
Detailed information regarding the relationship between building damage and
structural response levels and durations is not available. Attempts, however,
are being made to estimate the influence of structural response level and its
duration on building damage by low-cycle fatigue. 2 0
B. ANALYTICAL APPROACH
A recently developed technique of period-amplitude-time (PAT) contour mapping21
of response time-histories was used in these studies.
The pseudo absolute acceleration response time-histories of a family of single-
degree-of-freedom systems with different natural periods and 2% damping
ratio subjected to a ground motion were generated and the response amplitudes
enveloped to obtain response envelope time-histories. The response envelopek
- 39 -
contours corresponding to a number of response levels were then plotted on a
period-elapsed time plane. The PAT plots thus generated were advantageous in
studying the duration of various response levels for each earthquake. The num-
ber of cycles at a given response level were approximately determined by first
obtaining the total time extent of the response level contour for a given per-
iod and then taking the ratio of the time extent to the period.
C. PAT PLOT STUDIES AND RESULTS
Eight different accelerograms generated by four important earthquakes were
selected for the PAT plot studies (Table 9). The PAT plots for these accel-
erograms, normalized to a 1.Og maximum ground acceleration, for a 0.02 damp-
ing ratio are in Appendix 0 as Figures nl through D8. It can be seen from
Table 9 that the El Centro and Taft acc(elerograms have a fairly long strong
motion duration and the Helena and Goldtn Gate Park have relatively short dura-
tions. The PAT plot studies of these accelerograms are indicative of the dura-
tion effects of a range of strong motion durations.
The results of the PAT plot studies of the selected accelerograms are sum-
marized in Tables 10 through 17. The predominant periods for these accel-
erograms are those with fairly large DAFs (see the response spectrum shapes
in Appendix A). The total durations of different response levels for each
period were scaled from the PAT plots. The DAF values corresponding to the
periods are also listed.
D. OBSERVATIONS
* Although Helena and Golden Gate Park records are dominant in the 0.10
to 0.20 sec period range, El Centro (NS) is also strong in this
range.
* All records participate in the 0.10 to 0.50 sec period range.
* El Centro and Taft are dominant for periods longer than 0.5 sec.
* The number of cycles decreases with period, as expected.
* Normalized responses as S-eat as 4.0 or more appear out to 0.5 sec
period.
- 40 -
The above observations lead to the following interpretation:
* Both the short and long duration motions show significant and compar-
able dynamic amplification for periods shorter than 0.4 sec. Thus,
the duration effect on the response spectrum shape for periods shorter
than 0.5 sec -- a significant period range for nuclear power F:.int
structures -- can be considered relatively small. For longer periods,
however, the short duration records show significantly smaller dyna-
mic amplification than the long duration records. The spectrum shapes
for longer periods would thus generally tend to be higher for long
duration motions than the. short duration motions. This trend is reason-
able because short duration shaking cannot contain significant long
period motion nor permit the longer response build-up times renuired
for major response of long period structures.
o The total durations of various normalized response levels for a given
period are generally longer for the long duration accelerograms than
those for short duration motion. This indicates that the almost-
free vibrations that ensue after the strong shaking are quickly
damped out and thus would not result in a prolonged large-amplitude
response.
The above observations confirm the higher overall damage potential of
long duration earthquakes because they would excite structures over a
much wider range of frequencies. Large-amplitude structural responses
would be more prolonged for long duration earthquakes, and damage once
started, would tend to progress.
SUMMARY OF MAJOR FINDINGS
* The e3rthquake duration effect on the response spectrum shape for
periods shorter than 0.4 sec -- a significant period range fnr nuclear
power plant structures -- is small. The shape for longer periods,
however, would generally tend to be higher for long duration motions
than for short duration motions.
- 41 -
* Long duration earthquakes could be potentially more damaging than
those of short duration because of the following reasons:
1. Long duration earthquakes would excite structures and structure
components over a much wider range of frequencies.
2. Large-amplitude structural responses would be more prolonged
for long duration earthquakes, and once damage is started, it
would tend to progress.
3. Short duration earthquakes appear to induce significantly fewer
cycles of damaging response for periods less than 0.4 sec, than
do longer duration earthquakes. For the purpose of this compari-
son, response accelerations greater than 0.5g are assumed to be
damaging.
4. Long duration earthquakes also induce a significant number of
cycles of damaging response in periods greater than 0.4 sec.
- 42 -
TABLE 9
STRONG MOTION DURATIONS 3 OF THE ACCELEROGRAMS
SELECTED FOR THE PAT PLOT STUDIES
Number Accelerogram ComponentStrong MotionDuration, sec
I
2
3
4
El Centro, 1940
Taft
Helena
Golden Gate Park, S.F.
NSEW
N21°ES69 0E
NSEW
NlO*EN90gW
24
17
4
3
- 43 -
TABLE 10
RESULTS FROM1 PAT PLOT STUDIES
OF EL CENTRO, 1940, NS
Number
I
PredominantPeriod. sec
0.136
2 0.248
3 0.451
4 0.551
S 0.900
NormalizedResponse in
Excess of
1.02.03.0
1.02.03.0
1.02.03.C
1.02.03.0
1.0
Total Duration.sec
2.900.900.15
4.100.900.20
10.005.001.10
6.804.400.70
8.5
No. ofCycles
217
1
1741
22
212
OAF
3.63
3.55
3.49
3.43
2.12
12q
10
TABLE 11RESULTS FROM PAT PLOT STUDIES
OF EL CENTRO, 1940, EW
Number
1
2
3
4
5
6
PredominantPeriod, sec
O.k:48
0.302
0.451
0.551
0.744
1.226
NormalizedResponse InExcess of
1.02.03.04.0
1.02.0
1.0
1.02.03.0
1.02.0
1.02.0
1.0
Total Duration,sec
17.47.02.70.8
21.13.3
12.4
7.76.72.5
16.75.6
18.13.7
8.5
No. ofCycles
7028113
7011
28
14125
22.8
15
3
4
OAF
4.36
2.90
3.84
3L79
2.27
2.25
1.477 2.10
- 44 -
TABLE 12
RESULTS FROM PAT PLOT STUDIES
OF TAFT, N210E
NormalizedPredominant Response in
Number Period, sec Excess of
1
2
0.244
0.369
0.499
0.673
0.822
1.02.03.0
1.02.03.04.0
1.02.0
1.02.0
1.02.0
Total Duration,sec
9.62.21.2
10.35.03.20.3
9.02.0
11.80.8
18.12.9
No. ofCycles
43105
281491
OAF
3.67
4.30
2.87
3.28
2.3
3
4
184
181
224
5
TABLE 13
RESULTS FROM PAT. PLOT STUDIES
OF TAFT, S69 0E
Number
1
2
3
4
S
PredominantPeriod, sec
0.203
0.334
0.451
0.609
0.822
NormalizedResponse inExcess of
1.02.03.0
1.02.03.04.0
1.02.03.04.0
1.02.0
1.02.0
Total Duration,sec
12.28.80.8
18.15.31.20.2
12.44.02.20.6
10.70.7
7.21.8
No. ofCycles OAF
6043
4
541641
28951
181
3.58
4.25
4.58
2.10
2.1192
- 45 -
TABLE 14
RESULTS FROM PAT PLOT STUDIES
OF HELENA, NS
PredominantNumber Period, sec
I 0.111
Normal izedResponse inExcess of
1.02.03.0
1.02.03.04.0
1.02.0
Total Duration,sec
3.61.60.1
3.81.91.10.6
5.41.7
No. ofCycles DAF
3.33
2 0.150
32141
251374
134
5.09
3 0.408 2.45
TABLE 15
RESULTS FROM PAT PLOT STUDIES
OF HELENA, EW
PredominantNumber Period, sec
1 0.183
Normal izedResponse inExcess of
1.02.03.0
i.02.03.0
1.02.03.0
Total Duration,sec
3.52.50.5
3.21.30.8
5.11.90.6
No. ofCycles DAF
2 0.274
19133
1253
14S2
3.86
3.64
3.283 0.369
- 46 -
TV.BLE 16
RESULTS FROM PAT PLOT STUDIES
OF GOLDEN GATE PARK, tIO 0 E
PredominantNumber Period, sec
1 0.111
0.150
Normal izedResponse inExcess of
1.J2.0
1.02.03.0
1.U2.03.0
2
Total Duration,sec
3.01.3
2.61.80.5
3.82.41.2
No. ofCycles
2711
17123
15105
2.97
3.74
3. 71
OAF
3 0.248
TABLZ 17
RESULTS FROM PAT PLOT STUDIES
OF GOLDEN1 GATE PARK, N80 0 W
PredominantNumber Period, sec
1 0.136
Normal izedResponse inExcezs of
1.02.03.04.0
1.02.03.04.0
1.0
Total Duration,ScC
1.61.31.00.6
4.12.01.10.5
2.3
No. ofCycl es OAF
1210
74
18952
4.78
2 0.244 4.32
3 0.408 6 1 .54
- 47 -
VI. RECOMMENDED SPECTRUM SHAPES
A. INTRODUCTION
Nuclear power plants are structures of ý._fmrnount :iportance because they are
sources of power and because enurmous'potential risks are involved in the
case of partial or total structural failure due to inadequate scismic resis-
tance. Ideally, statistical predictions of seisn.ic forces, structural be-
havior, and potential damage and loss should form a basis for design optimi-
zation to achieve mininum potential risk. 2 2 Designs of nuclear ptwer plant
structures, based on seismic design conditions having sufficiently low prob-
abilities of exceedance considerably reduce or eliminate the risks of struc-
tural failure.
Pseudo atsolutc acceleration respon•e spectra are quite important ir- reprvtent-
ing the severity and Frequency content ..f the seismic ground motions. As rioted
prevt.'usly. they are also useful in determining the seismic forces inc,.ced in
a structure. Statistica' ,'redictions of pseudo absolute acceleratin respont,e
spectra should therefore provide a rational basis for probabilistically esti-
mating the seismic input motion for structural designs.
Selection of spectrum shapes (i.e., pseudo absolute acceleration response'
spectra for ground motions normalized by the peak ground acceleration) is oat
important step in deriving the pseudo absolute acceleration response spectra.
Recommendations for the use of spectrum shapes for damping ratios 0.005, 0.01,
0.02, 0.05, 0.07, and 0.10 are presented at the end of this chapter. The
recommended spectrum shapes are based on the probabiiity distributions consi-
dered suitable for the spectrum shape ensemble data.
B. kuCOHMENDED SPECTRUM SHAPES
In view of the relationship between the pseudo absolute acceleration spectrum
and the DAF, probabilistic estimation of the peak ground accelerations and the
spectrum shape values is necessary for predicting the pseudo absolute accelera-
tion spectra. Such peak ground accelerations and spectrum shapes then shou!d
be combined probabilistically to derive the spectra. It must be noted that
deterministic estimates of very high peak ground accelerations based on a
- 48 -
postulated extreme earthquake occirrence when directly combined with spectrum
shapes representing fairly '.vei probab;lIties of exceedance could result in ex-
treme seismic conditions 3nd erence unduly conservative designs. Pseudo abso-
lute acceleration response sp-ctra, " for different probabilitriLs of exceed-
ance can be derived by using pribat-ility density functions (:cO) of peak
ground accelerations (0) and £Ais (..:) as discussed herein.
Let random variables ., ., and Y represent a, , and D, respectively. Then,.j
Equation (17) can ve rewritten as:
' = v .(19)
Because of the low correlation between and Y noted in Chapter IV. these
variables can be considered independent and their joint r,,:f can be derived
by Equation (20).
f V y) = ,C.) , (201
in which
f , y) = joint rdf' of X and Y
f x(A) = pOf of ;0 . :
.fy i) -- P f of Y, y 0 ,
tNow, the z.df of ., f,(;:) can be derived by Equation (21).I-
?2(z) t ' f (rfn) ,i•, d] , C (21)
in which
].',n =dummy variables of integration
The oCf of Z or 3a can be used to derive different probabilities for S..
- 49 -
Probabilistic estimation of tlhe peak ground accelerations is not within the
scope of the present study. Probability estimation for DAFs are presented
below.
0 Probability Estimation for DAF
As liscussed in Chaoter III, the spectrum shape ensemble statistics, such
as mean, median, and standard deviation were computed for 108 different
periods in the period range under consideration. The coefficient of skew-
ness, or third moment, and the coefficient of kurLosis, or fourth mment,
were also determined for these periods. It is not surprising that the
third and fourth moments varied considerably from period to period in
view of the nature of response spectrum shapes, which tend to have pro-
nounced peaks and valleys, especially at low damping values.
The skewness coefficients are positive, which indicates a distribution
with a pronounced tail to the right, in this case away from the zero
value. This is acceptable because response spectra are constructed with
absolute peak values. The values of s':ewness vary widely between the
general limits of 0 and 4. However, the middle period range of, say
0.02 to 0.75 seconds, has skewness values much less than either the shorter
or longer periods, indicating for this range that the distribution is
closer to the normal distribution symmetry. Over a range of damping val-
ues, the skewness coefficient averages about 0.3 to 0.4 for periods from
0.10 to 0.75 seconds and about 1.00 to 1.40 for other periods. The co-
efficient of Ljrtosis or flatness also is less in the middle period range
with average values of about 2.3. This indicates more peaked'distribu-
tion than normal. At the other periods, however, the kurtosis coefficient
averages greater than 3, indicating flatter than the normal peaks.
These data, plus comparisons of mean and median values, indicate that
there is more variation in the short and Ionq period ranges than between
0.20 and 0.75 seconds. In general, a skewed distribution, with a 'ero
lower limit, occurs over the whole period range of interest.
It is apparent from these observations that any one of several distri-
butions would be effective. Three of these distributions, namely normal
or Gaussian with appropriate truncation in view of the postulated abso-
- 50 -
lute value for OAF, log normal, and extreme type II, were tried on the
OAF ensemble for 0.02 damping ratio. All of these distributions are
statistically acceptable within the range of the data.
The log normal distribution, given by Equation (22), was considered the
most desirable because it is convenient to use and has been found effec-
tive for ground motions generated by underground explosions.23
D (T., C; y) = 'M (T.,i.) (T., Y (22)
in which
y = standardized normal variable
p(Ti., r,) = geometric deviation for period, Ti., and damp-
ing ratio, C
r'D(Tj, 4) = median OAF for period, T., and damping ratio,
4, computed by using the mean and standard
deviation of OAF (as shown below). It differs
only slightly from the median described in
Chapter III
D (T., C; y) -OAF for period, T., and damping ratio, ., asso-J J'
ciated with the standardized normal variable, y.
The parameters, i"'M zd B are computed by using the ensemble mean and stan-
dard deviation as shown in Equations (23) and (24).
5T0 , T mD (Tj,r,) exp [- ½sD(T.,)] (23)
= exp n ( -. ,L)) 2 + i (24)
in which
rr D(T ,) = mean OAF for period, TV, and damping ratio, r,
- 51 -
SD(TJi) standard deviation DAF for period, Ti., and dampingratio, C
n= natural logarithm
exp = exponentiation
The parameters, mffD and 6 were computed for all 108 different periods to
derive the spectrum shapes for different y values of 0.0, 1.0, 1.645,
and 2.0 corresponding to probabilities of exceedance of 50%, 15.8%, 5%
and 2.3%, and for all damping ratios. These spectrum shapes are pre-
sented In Appendix E as Figures El through E6. Equation (22) shows
that the y = 0 curve is the median shape.
These spectrum shapes are smooth when compared with spectrum shapes of
individual accelerograms and form the basis for the recommended spec-
trum shapes.
The spectrum shapes in Appendix E for the above y values consistently
follow a basic shape shown in Figure 3.
DAF, D b
1"0
Period, T sec
FIGURE 3. BASIC SPECTRUM SHAPE
- 52 -
A moderate amount of smoothing was required to derive smooth spectrum
shapes from those in Appendix E. The amount of smoothing decreased rap-
idly with smaller y values, longer periods, and higher damping ratios.
It is noted that these spectrum shapes were not obtained by enveloping
b-tt by appropriate visual fitting to the different ,-value curves.
Three spectrum shapes, defined as large (50%), small (15.81), and neg-
ligible (2.39) probabilities of being exceeded, were developed corre-
sponding to y-values of 0.0, 1.0, and 2.0 and are presented in Figures
4 through 6. Four-way log plots of the shapes are presented in Fig-
ure 7. In addition, numerical data for the shapes, including the con-
trol points, A, 8, and C, and parameters b and -3 are listed in Table
18 to facilitate the reconstruction of these curves. The curves in the
vicinity of point A are extrapolaticns of the ..- curves in Appendix E
because the original shapes, due to limitation of the input data, were
computed to 25 cps frequency or 0.04 sec period. These extrapolations,
however, do not result in any significant errors.
Comparisons of the recoinmended spectrum shapes for a 0.02 damping ratio
with the current AEC criteria and other shapes proposed by Newmark."',
Housner2 5 , and Blume are presented in Figure 8.
The usefulness of the recommended spectrum shapes is demonstrated"F by
the fact that a time-history generated to match a spectrum curve for one
damping-ratio shape matches quite well with the curves for other damping
ratios. Such matching of spectrum shapes for different damping ratios
by a single time-history has not been satisfactorily obtained so far
and indicates the appropriateness of the recommended shapes with respect
to damping ratio relationships.
C. MAJOR FINDINGS
The major findings from these analyses are:
0 The log normal, truncated normal, and extreme type II probability
distributions were found statistically acceptable for the spectrum
shape ensemble data. The log normal distribution was adopted for
the further analyses because it is the most convenient to use.
- 53 -
0 The current AEC design spectrum shape for a 2% damping ratio is
below the small probability of exceedance shape for periods shorter
than 0.4 sec and for periods longer than 0.9 sec.
* The Newmark spectrum shape is consistently above the small probabil-
ity of exceedance shape, except for a short interval in the vicinity
of zero period.
a The Housner spectrum shape is below the large probability of exceed-
ance shape for periods shorter than 0.4 sec, and it is above the latter
for longer periods.
* The Blume F-factor spectrum shape for a 2% damping ratio and stan-
dardized normal variable value of 1.0 is consistently higher than
the small probability of exceedance shape.
D. RECOMMENDATIONS
To minimize risk in the seismic design of important installatiuns, such as
nuclear power plants, the seismic load criteria should incorporate such fac-
tors as regional seismicity, geotectonics, etc. The following recommenda-
tions, pertinent to the spectrum shapes described herein, are intended to
partly achieve the seismic design objective. Other ground motion charac-
teristics, such as peak ground acceleration and strong motion duration, in
the case of time-history analyses, should be given thorough consideration
to fully satisfy the design objective. The following are recommended:
* The curves shown in Figure 4 should be considered as lower bound
spectrum shapes for any site.
* For sit-s associated with relatively low risks, e.g., sites lo-
cated in low seismicity areas, the design spectrum shapes for
different damping ratios should not be lower than the small
probability of exceedance spectrum shapes shown in Figure 5.
* For sites associated with relatively high risks, e.g., sites lo-
cated in high seismicity areas, the design spectrum shapes for
- 54 -
differcnt damping ratios should not be,lcr.-er than the negligible
probability of cxceedancc spectrum shoaps -.hown In Figure 6. Be-
cause these curves represent extreme ground motion amplification,
their use should be carefully coordinated with the selection of
the peak ground acceleration. Care must be exercised to ensure
that the total seismic exposure for a site is compatible with the
risk exposure involved and not unduly conservative.
* For sites judged to be significantly responsive to ground motion
components with periods longer than 0.5 seconds, the above shapes
should not be used without appropriate modifications for the par-
ticular site conditions.
- 55 -
TABLE 18
NUMERICAL DATA FOR RECOMMENDED SPECTPUM SHAPES
Probabilityof BeingExceeded
Large
(50•;")
Smal 1
(15. 8;)
DampingRatio
0.005
0.01
0.020.05
0.07
0.10
0.005
0.01
0.02
0.05
0.07
0.10
Point A
T OAF
Point B
T DAF
Poi nt C
T DAF
0.03
0.032
0.034
0.036
0.038
0.040
0.028
0.029
0.030
0.031
0.032
0.033
0.025
0.026
0.027
0.028
0.029
0.030
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
0.12
0.12
0.12
0.12
0.12
0.12
0.11
0.11
0.11
0.11
0.11
0.11
0.09
0.09
0.09
0.09
0.09
0.09
3.22.8
2.5
2.0
1.851.7
5.1
4.1
3.5
2.6
2.2
2.0
0.350.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
4.03.5
2.9
2.3
2.0
1.75
6.2
5.0
4.2
3.1
2.6
2.3
Parameters for,the curve, bT
b o
1.20 1.46
1.08 1.16
0.93 1.075
0.76 1.053
0.67 1.038
0.59 1.032
2.342.00
1.73
1.35
1.13
1.02
4.32
13.68
3.12
2.40
2.03
1.77
0.928
0.872
0.843
0.794
0.790
0.776
0.761
0.692
0.604
0.511
0.489
0.470
Negligible
(2.3-4)
0.0050.01
0.02
0.05
0.07
0.10
1.0
1.0
1.0
1.0
1.0
1.0
8.16.2
4.8
3.2
2.6
2.3
9.67.6
5.9
4.1
3.4
2.9
- 56 -
8 1 t.
7.1---61-i-~--
U,
I
5 dl, €-
4
3.
2-
6•
Damping Ratios: 0.005, 0.01, 0.02,0.05, 0.07, 0.10
1 -- .. . - -
IiiS -4-~----- I ~-~-~-- 4 ~ I P 4*-* 4
v ii - I N -- 4 - -- -.----- 4 ~ F ~ A~ ~ I 4 1
IJ gJ I
114 I t 4
0.0 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.0O 2.25 2.50
Period, T sec
FIGURE 4a. RECOMMENDED SPECTRUM SHAPES FOR LARGE (50%) PROBABILITY OF EXCEEDANCE
Damping Ratios:5.
4-
0.005, 0.01, 0.02,0.05, 0.07, 0.10
PCo
I
3
2
m I
SLL.
NY
I
0IF
l ~I. -I I-
I t I I I 4
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0
Frequency, cps
FIGURE 4b. RECOMMENDED SPECTRUM SHAPES FOR LARGE (50Z) PROBABILITY OF EXCEEDANCE
8
7
0 0
Damping Ratios: 0.005, 0.01, 0.02,0.05, 0.07, 0.10
~JI
5.--
LL 4.
0.0 0.25 0.50 0.75 1.00 1.25 1.50 .75 2.00 2.25 2.50
Period, T sec
FIGURE 5a. RECOMENDED SPECTRUM SHAPES FOR SMALL (15.8%) PROBABILITY OF EXCEEDANCE
6-Damping Ratios: 0.005, 0.01, 0.02
0.05, 0.07, 0.10
3
I _ _ _ _ _ _ ____ _ _ _ _ _ _
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0
Frequency, cps
FIGURE 5b. RECOMMENDED SPECTRUM SHAPES FOR SMALL (15.8%) PROBABILITY OF EXCEEDANCE
10
9
7 - Damping Raties: 0.005, 0.01, 0.02,0.05, 0.07, 0.10
- ., .
1 -
4- I.5- _
0.0 0.25 0.50 0.75 1.00 ±.25 1.50 1.75 2.00 2.25 2.50
Period, T sec
FIGURE 6a. RECOMMENDED SPECTRUM SHAPES FOR NEGLIGIBLE (2.3%) PROBABILITY OF EXCEEDANCE
10
8 i
__ -Damping Ratios: 0.005, 0.01, 0.02,0.05, 0.07, 0.10
9. .. i - .- ..-.4 . ......... _ ___ _LL.
- i .
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0
Frequency, cps
FIGURE 6b. RECOMMENDED SPECTRUM SHAPES FOR NEGLIGIBLE (2.3%) PROBABILITY OF EXCEEDANCE
0.00s002 /Dampi ng
0:01 Ratiosto. 0
i IityExceedance
I-,-..-
K -Small Probabilityof Exceedance
r...
V. Ju
Large Probabilityof Exceedance
t
0.01 0.10
FIGURE 7.
Period, T so.
RECOMIIMENI DE D- 63 -
SPECTRUM SHAPES
10.0
10
9
8
Probability of4 Exceedance
S..Negligible
- Recommended Shapes
------ tNewmark1 Blume
--.-- Current AEC
------ Housner
La..
5
4
3
2
1
0
Large[l
0.25 0.5 0.75 1.00 1.25 1.50 1.75 2.0 2.25
Period: T s ec
2.50
FIGURE 8. SPECTRUM SHAPE COMPARISONS; DAMPING RATIO = 0.02
VII. REFERENCES
I. "Strong Motion Earthquake Accelerograms, Digitized and Plotted Data,"
Volume I, Parts A (July 1969) and B (October 1970), Earthquake Engi-
neering Research Laboratory, Calirornia Institute of Technology, Pasa-
dena, California.
2. "Response of Earth Dams to Strong Earthquakes," N. N. Ambraseys and
S. K. Sarma, Geotechnique, Volume XVII, Number 3, September 1967.
3. "Intensity cf the Ground Shaking Near the Causative Fault," G. W.
Housner, Proceedings of the Third World Conference on Earthquake Engi-
neering, Volume I, New Zealand, 1965.
4. "The Peru Earthquake of October 17, 1966," C. Lomnitz and R. Cabre,
S. J., Bulletin of Seismological Society of America, Volume 59, No.
2, April 1968.
5. "The Tokachi-Oki Earthquake, Japan, May lb. 1962, A Preliminary Report
on Damage to Structures," Report NIo. 2, International Institute of
Seisriology and Earthqudke Engineering, Japan, June 1968.
6. "Evaluation of Response Routines." R. E. hunroc, John A. Blume , Asso-
ciates Research Division, San Francisco, A Letter Report to Nevada
Operat ions ()ffice. AEC. September 1371.
"nalysis-of Strong i.ctiun *,ccvltcr,9rrap.h Records," D. E. Hisdon. N. L.
Nigan:, and 11. D. Tril'unac, Proceeding% uf the Fourth World Cocr,'r.;.ct'
of L,,rthqtjaý.- Engineering, Volure' I, Santiago, Chile. 19 6'i.
6. "High Frequency Errors and Instrument Correction, of Strong Motion
Accelerooram.s,' If. D. Trnfunac, F. E. UL-!wadia and A. G. Brady, Earth-
quake Engineering Research Laboratory, California Institute of Tec!!-
nolnny, Pasadena, California, July 197!.
- 65 "
9. "Low Frequency Digitization Errors and a New Method for Zero Baseline
Correction of Strong Motion Accelerograms," M. D. Trifunac, Earthquake
Engineering Research Laboratory, California Institute of Technology,
Pasadena, California, September 1970.
10. "Elementary Seismology," C. F. Richter, W. '-4. Freeman and Company, 1958.
I1. "Site Characteristics of Southern California Strong Motion Earthquake
Stations," C. M. Duke and b. 0. Leeds, Report No. 62-65, Department of
Engineering, University of California, Los Angeles, November 1962.
12. "Earthquake Ground Motion and Engineering Procedures for Important In-
stallations near Active Faults," John A. Blume, Proceedings of the Third
World Conference on Earthquake Engineering, Volume III, New Zealand, 1965.
13. "Studies on the Parkfield, California Earthquakes of June 1966," Bulletin
of Seismological Society of America, Volume 57, No. 6, December 1967.
14. "Strong Motion Instrumental Data on the San Fernando Earthquake of
February 9, 1971,11 D. E. Hudson, Editor, EERL, CAL TECH and Seismological
Field Survey, NOAA, U.S. Department of Commerce, September 1971.
15. Files of Seismological Field Survey, NOAA, San Francisco.
16. Geophysical Investigation of Recording Station Site, made by John A.
Blume & Assorates, Engineers, San Francisco, California.
17. "Report of a Foundation Investigation, Sheraton-Universal Site, Universal
City, California," L. T. Evans, Inc., Los Angeles, California, September
1966.
18. "Report of Foundation Investigation, Proposed Office Building and Parking
Structure, Ventura Boulevard East of Sepulveda Boule'ard, Sherman Oaks
Gistrict, Los Angeles, California," Le Roy Crandall and Associates, Los
Angeles, California, February 1969.
- 66 -
19. "Report of Foundation Investigation, Proposed Seven Story Motel, Roscoe
Boulevard and Orion Avenue, Panorama City, Los Angeles, California,"
Dames and Moore, Los Angeles, California, February 1966.
20. "Low-Cycle Fatigue Failure of Seismic Structures," I. Kasiraj and J.T.P.
Yao, Technical Report CE-li(68) NJSF-065, University of 1le* e Mexico, Albu-
querque, flew Mexico, 1968.
21. "Production of Time Dependent Response and Bond Pass Filter Spectra,"
JAB-99-10, K. F. Schopp, R. E. Scholl, Prepared for Nevada Operatians
Office, USAEC.
22. "Damage aid Risk Analysis for the Greater San Francisco Bay Area due to
Earthquake Loadiiig," Hiaresh C. Shah arid Jaiat S. Dalal, Coisference on
Microzonation, Seattle, Washington, 1972.
23. "Respot'se of High-rise Buildings to Ground Motion from Underground
Nuclear Detonations," John A. Blum'e, Bulletin of Seimological Society
of America, Volume 59, December 1969.
24. "Seismic Design Criteria for lucll•,r Reactor Facilities," N. M. tNcwmark
and W. J. Ilal], Proceedings of tlhc Fourth World Conference of Earth-
quake Engineering, Santiago, Chile, 19b9.
25. "Behavior of Structures During Eart!,iuakes," G. W. Houns.ner, Proceed-
ings of Air.erican Society of Civil En(.'ne,_,rs, Voi. 85, EM4, October 1959.
26. "Civil, Mechanical, and Electrical Aspectc, of.Se ismic De.,iyn of tNuclear
Power Plants," Roland L. Sharpe, Presented at Joint Po..jer Conference
(ASME, IEEE, ASCE), [Uo.ton, Mas-,.cht,-,wtt,-, 11112.
- 67 -
0
-j
CLJ
a
Lfd
12 .C
RESPO]NSEEr-L CENTRo>DAMhP. ICT IDS --*
SPECTRRI19e40
.010!.070o
.020
.100
GROUND MOTION NORMALIZED TO 1.0 G
- :~-"-t-.----.-
4L .~• .0 1 •2 .0
FIGURE Al p L R, I V) ci I SN L- c
I.-
crU.,-jI
Cf
-LJ
4r,
cr
LlU
I S . C0 RESPONSEEL CENTRGWDAMP. RPTIOS =
GROUND MOTION NORMALIZED
SPEC1940.O0s,.OO,07
TO I.OG
TRO9 EW.0101.0701
.020
.10012 . a
9.4
S.G
. • -.i
o ._ iI_______________- - -.- -I
.0 1.0 2.0 2.5
FIGURE A2 PERtODrY SEC
7-4
cr.
U'
C-)
03
:3CrJ
1 r . ( RESPONSEEL CENTRIT]DPi:P RflTI1JS _
SPECTRR11934.00,o.Osol
~NSr.)10,0 0?01
.020
.100
i C+GROUND MOTION NORMALIZED TO I.2G
9 .
6.rk
3
0
V.-........ t -
p
4..--~ + ... ........ .t, ... .......... . ... --t.
0.0 .E 1 .0 1 .S 2.0 2,5
FIGURE A3 PERIOD' SEC
ci
I.
CLi
t-J
cli
al
CL
(~) 4
ifýI
RE.-SPFiNGEEL CENTRF[19D~4r1p. RflTIOc z
SPEC1934.005?.OSO7
~EW~.020.100
GROUND MOTION NORP.ALIZED TO 1.0 G
1I
-...- -. - -.-----------1 ----- "----*-.-----------* --
A
t
* ~- - - -4-~ .*j *I* . -~
.: 1.0 1 .£ 2.0 2.S
FIGURE A4 FUERI 00i SEC
1 -
TFqFT 49~2)
C2 T; RJ PT
r* )C~JLJ
G-ROUD MOTION NORMALIZED TO 1.0 G
I---
P A
t
*1*1.*
- -F.
1.'~- I .cl 2 .C 2 .
FIGURE A5
IL°I •"*
7
c-j
-j
U
-- 4
ti
'a-,
-'a,
L.
VqF, E P)L2NT~1~ (fl*FwT
'DC.
rwn.r13
GROUNMD MOTIONJ NOP.iALIZED TO 1.0 G
2
-F
C. -I
6.
.4.
1.0 2.C2 .~
FIGURE A6 F-R .IEC
-z
.1. .
~ A1 [71
GROU14D MOTION NQR.4A1LIZED TO 1.0 G
TB414 n0*').~** I ~**
-j
1..
D
12
.2
IT
( .
1.4-~.-*-
....*... . ...... ,...........
-; ~) (**, r-, £:
FIGURE A7f r
it: ~
~1. R~ESP ENSIE SPEC49:•S
rj"_, I •
1 640-.(
u.£ ~..
U I
.020
GROUND MOTI ON NOW1ALIZED TO 1.0 C
D .C~
,- (I.
It'
1 m -') (":
FiCURE A8 PICE k I EA8 )ý PEFU ',-
tie
RES)PfNSE PEW
GROUND MOTIONI NOR?4ALIZED TO 1 .0 G
J
i . 132* 100
44.
"I
I II JI
ls
•__, II†- ~l
1.0 V. 2 .0
FIGURE A1O
n RESPOILSEGOLDEN C_-D P M R n T I E,06J'J
S PE C"'T RPF
Li
-J
-4
1.1 1
cL2
LA
0 c9
.070
N10E020'Co
•. ; -,
GPOU-ND W)TIOM NORMALIZED TO 1.0 G
2 -
I
'2 Ci.--0.0 f 21 2.S
FIGURE All PERIOD" SEC
1 -C-' rP [N, SE SPECTRPGOLLDEN I-F1hI5~iG
0,17 0 .0?Oý .1-00
GROUJD) M.OTION NORMALIZED TO 1 .0 G
FIUE A2P R LE E
? .0
Is .rJf RESPONSEPPRKFIELD-DAMP. RATIOS =
SPECTRP
.-J
CL
V1,co
2 5.19,00,5,.O0SO: 0
ISG.9.010* 070
NGSE.020.,100
1 2 .(*GROUND t4OT ION NORMALIZED TO I .0 G
3.O0
09 .( %.# 1 .0 1.5 2.0
FIGURE A13 PERIODSEC
I- III I I I I I _,.
<RESPO]NSE GPETF?PPRKF'iFLD- sI 6 3 5J
A M Fs . n, 020
r1 2 GROUNDU I4OT I ON NOR?.IAL IZED TO 1 .0 G
CL
~~~1~~ L..a. . r.0
FIGURE A14ER2, E
F E R 127 ýý- 3 Sl E C
iS .G RESPONSEPPRKFIELD-DAMP. RATIOS =
C3
I-
Li
a
cm
CL
SPECTRP51 IG~i N,00S3 .0103.0503 .0?0)
B5E.020.100
12 .01GROUND MOTION NORMALIZED TO 1.0 G
iII
I
II
I
2.0A t - - - -
irl ý ----- 4---.
2.0
FIGURE AIS FIGURE AlS ERIdD)G EC
is .0 R E S-1P GN SEHiPCHIINDHE 5C, f? Cý. P, Af IC,- =
SP E CT RP
U j
196. O0'n-Osc,
El5NS.- 020
.10012 P+
GROUND 14OTlO .i NOF%.¶ALIZED TO 1.0 G
t
I.C0
UDLxJ
r .4,,p.) •. - .r - --- r -r - +
.5 1.S 2.0
FIGURE A16 PERIF)]5 SEC
1 :.Cli. r~SL~ S PPECT R ý7:
Li
crncrEl
HlCHINLHE9 9•962wLi .IP. rIzS = .D -
GROUND- MOTIOON NORMALIZED TO I.0 G
.070).020.100
12 . C+
9 -4
1-
I
!
0I- .n I I
L)•I-• S1.0 2.0 2. S
FIGURE A17 PERIODSEC
C-
z
ci_LLI
LUJ
ci
RESPTINSE SPELLINfll 19669 NBEDAMP. RlrIos = .. 0053
.0s0,
GROUND MOTION NOPRALIZED TO 1.0 G
.0101 .020
.070, .10012.
9.
L;J
mcr
6.
CL 3.
i.0 2.0 2.S
FIGURE A)8 PERIOD3 SEC
u-.
ciLii-j
CL
E7.
'12
R E-S)P NIiNS -E SPEO TRiL I N ýqo n~ rl1 ~.
ý 19 G., N 132 W. 010 Iq .020
* 100
GROUND ?4OTION NORMALIZED TO 1.0 G
4L*
S
...
.2 2'. 0
FIGURE A19 R I L"i S) E.
Luj
LJ
-i]
I-
Cfa2.4
PESP~jNSE71,ACPS5T PTC 5 19
R qr 1 ci z-
TRA21E.010,
.070,, lor .020.100
GROUND 1.101 I,1 "N NO•MALIZED TO I .0 G
*I r\ f~ .---- _______ ______
I
.- '~--
1.0 I.S 2.0
FIGURE A20 PERIrJCJ.SEC
r~EF': N SShSET q T C , 1~)
S P E
*rJos.0 :5 0
CC'
TR qG9r'DE.010o. 13 7 0!( -I
C
.020
.100L 1 . -
GP t',. T.TIOIN N1ORMALIZED TO 1 .0 G
2 . 'it
Li
ti
I.,'-I
IV- ** * *; --
-I- lr1 4" 4-
r9~rj 1 {i.0 , .%rr -~ ~.* '7)
FIGURE A21 p 1 .' 1 ,1 _
u-
-7D
uj
is.
12.
3'.
C.
0.
RESPINSEEBP1N K [I F -CF PDAMPF'. R nrFI C S -m
GROUND MOTION NORMALIZED
Al
SPECTRflL IF -19 7 1 Ni11E7
.CO0, .('110 C'20
TO 1 .0 G
It
i I
0 .G 1.0 1 . S:
FIGURE A22 PERIC0 i SEC.
Li-
L;
w
-J
a_
is. RE SPOJN SE8BP NKDAMP~.
0 FR A' 11 1
SPECCALTF
,OO5,.OS0- .0-10)
N 7 N9'LJ
-i
7
GROUND MOTION NOW4ALIZED TO 1.0 G
r-. (
hi
I I
IJ0. --N
0.0 z.0 I...S 2 . 0
PE.RllJUD ECFIGURE A23
LA-
7-
r-3
ILi
:I~ Z , 0
:L 2
RE SP0NSE SPEUNIUL SHERPTODAMP .RATIOS-) .00S
0 G
GROUND MOTION NORMALIZED TO 1.0 G
.N0 10 ý070?
02NS.0201 00
'C"T R
ILlý
m
CL
r
K ~ (~ I
3*9 K -- -.-.-...
-~------ - I
I..) 1 ,0 1 -S 22.0
FIGURE A24 PERIOD) LEC
C3
cz'-j
adu
t Li
C3
C)
:DwI(1)a_
2 .cr
12.0
9.0
6.0
RE SPOiN SEUNIU .OF P. RP
S'HE RTIO'E =
SPECTRPA TEN5 197
00: , .010•.050) .0?0,
1L•,EH.020.100
GROUND MOTION NORMALIZED TO 1.0 G
3 Cl4 I.70.0 . C 2.0 2., e
PERIDD. S)ECFIGURE A25
u-
13
0
0
mi
1s .0& RESPDNSE SPECTRPU.N. HLILIDF1Y IDANri. RATIOlS .00s
o~ol
NN5 I.0107.0703
971.020.100
12 0ý---
9.%
GROUNI) MOTION NOVEALIZED TO 1.0 G
6 .0w
3.0
0 C0.0 . 1.0 1 .5 2.0
FIGURE A26 PERIODOSEC
U- is. RESPONSE
2:u-
ai
U .N~DAMP.
HU0LI DRAlTIOS =
SPECTRAPY INN919719EW.OOS, .0107 .020.•OSO .070, .100
12. GROUND MOTION NORMALIZED TO 1.0 G
9.
CIE
f--
-J0
O0
Dnw
U._
G
3.
0.
Jýv
1 <7if~ -. ~-.1
0.0 .S I .0 1.5 2.0 2. S
FIGURE A27 PERIODI'SEC
a
C-)
ci
C1)cr.
15.
12.
RESPONSEEUREODMP.
•KR 5 195RATIOS =
SPECTRP4,N79E. 006, .010, .020.050, .070, .100
GROUND MOTION NORMALIZED TO 1.0 G
9
G.
U')CL
3.
0. +
0.0 1.5 2.0 2.5
FIGURE A28 PERIODSEC
1S .O%u-
aw-J.wL)U.
Lii
C3
RESPONSE SPECEUREKPA 19S34 .NIDAMP. RATIIJS =.00
.050,
GROUND MOTION NORMALIZED TO 1.0 G
TRAI.w..010".070. .020
.10012.0
9.0
6.0
3.0
0 .0 I I -~-- -~I0.0 1.0 1.s
I
2.0I
2'.5
FIGURE A29 PERIODi SEC
U-
I.
IdC-)u-cr.ui
c-J
0
a-
IS.
12.
9.
6.
3.
0.
RESPONSE SPE(OLLYMPIP9 19;G5DAMP. RATIOS .00s,
.050:GROUND MOTION NORMALIZED TO 1 .0 G
CTRAS04E
.010, .020
.070) .100
0.0 1.0 1.5 2.0 2.5
FIGURE A30 PERI00~ SEC
10
0ý
CE)
'LI
0
LU
a.
Is Gcr RE SP UN S)ElULYM'PTP3,1-9DAMIP. RATIOS =
SPECTRFP65938GW.0051 .010).0s0, .0702
.020
.1001' c• CROUND MOTION NORMALIZED TO 1.0 G
64
3.C4
0. a OL,
0.0 1.0 2.0
FIGURE A31 PERIODSEC
U-
0-
LIi
C-Jw3
0
a-
iS .0C,
12.0
RESPONSETEMBL.F]R i19DAMP. RATIOS =
S PIECJRPPGG qN.00sl.0s0l
.0101 .020.100
GROUND MOTION NORMALIZED TO 1.0 G
.G .14-
7~~1
0.00.0 1.0 1.5 2.0
. FIGURE A32 PERIOD, SEC
IL
Lit
cr
-J
10.
is5 ,C RESPONSETEMBLUR5 19DAMP. RATIOS =
SPECTRPG65N25E.005, .010,.050, .070,
.020
.100
12 .%1 GROUND MOTION NORMALIZED TO 1.0 G
9.0
6,.0ý
3.i~c~
0 C4.. -
0 . 1 .0 2.0 2. S
FIGURE A33 PERIOODSEC
7iiI~TI
K:
'-' 4 24
I j Au*1 lift
-~ fv~+ ti1~4~(iII ~II~~Jl
-r 1 pj'
U- Ivci
CI I
RESP[I Ns E P F C T R qzL
D A INC L--~
1Ills1j cl r) *,Io
. C f:
CGRUND '1OTICT' NORMIALIZED TO t .0 G
it!
*1-
V.2V »A
f!
10 . 0 4. t
FgCURE A34 1'ý U y r
-77-- - 7 ' I
ý- i i~ ~ \' K... <' ,j ~..
a-
I f_ __ 11KL A.
L'
1 ) ~~
1T' R*
0 9.~cC '. "D
? ,2C GROUND) MOT I ON 1,10fý!.1AL 1 1-[!) TO I .0 G
S .44-
3 .6CJ .- '
1 ..BC
-1------ -
0.0 5.0 10.0 15.0 20.0 25.0
F RE 0U E NC Y, C PSFIGURE A35
U. I'I:
L1I
ICl
12 .Cn
91.6.
iEL CENTR~iD IMhP. RATFIOS
RE,-SPDNS-E
? .2
GROUND MOTION NORMALIZED
iJ
SPECTRA19345NS.0D5S 1 .0i0, .020.0 ,- .070, .100
TO 1.0 G
Lii.
D-J
LAcnci
00:2Li
(1.
I
4.F•.
t2.4 ........
Vit!|
1-C' I
25.00 .0 f'-" . 0 10.0 15.0 20.0
FIGURE A36 FREQUENCYiCPS
9.LL.
0i
2u0
F-3
ma
U3crwD
RESPONSE SPE[TAFT 1952 9.N211 2DIPN. RPTIOS- .00,
GROUND MOTION NORMALIZED TO 1.0 G
ZT!RP
.01'0,'20.0701 .100
5.4
3.6
1.8
-. 0
0.0 10.0 1, 20.0 25.0
FIGURE A38 FIGUIE A38FREOUENCY, CPS
z
ci
ILl
cr.
0
02
0-
7
6~.
RESPONSE SPEIKTqFT, 2952, S691
I. . OPMP. RprIos = .00S,
iGROUIND MOTION NORMALIZED TO 1.0 G
.0101 .02-0
.0?0) .100
3.0
0.0
0.0
I,iI
0.0 10.0 isl0 20.0 23 .0
FIGURE A39 F REQUENCY, CPS
0L
z03
a
w
LiiI-
-LJu
Uf)
cr
13
U)0(.
7 .5s
6.0
4.5
3.0
I.S
RESPONSE SPECOLYMPIAh I9499 £ODMP. Rerios = .0051
.050,
GROUND MOTION NORMALIZED TO 1.0 G
VVRRB86W.0iO! .020
.100
0.0 .5.0 10.0 15.0 20.0 25.0
FIGURE A41FREQUENCYCPS
0
V1-4
ciw
(-j
ci
Lii
-J
Cr;
a..
10.0
8.0•
6.0
4.0
S2 .0
0.;0
0.0 10.0 IS.0 20.0 2S.0
FIdURE A42 FREQUENCYsCPS
0L
2
.-1wuuu
-LJD
I?fn
0
a.=
6.0.1
4.E
3.6
2.4
1.2
-. 00.0 10.0 IS .0 20.0 25.0
FIGURE A43 FIGURE A43REOUENCYICPS.
ci
w
LUJuu
ci0-w
6.0
4 .Ei
3.61
2 .4i1
1.2'
-. 00.0 [0.0 1: .0 20.0 .2S.0
FIGURE A43 FIGUR A43FREQUENCY~ OPS
0L
z0
a
a
V;
Z)w
CL!
s.0
413
3.6
2.4
I
0.0 10.01 15.0 20.0 2S.0
FIGURE A44 FIGURE A41.FREQUENC~i~CPS
z0
a
0ý.
w
I
-J
w
'n
U
C-
0E
w
tIJ
6.0
4.8
3.6
2.4
1.2
RESPONSE SPECPPRKFIELD- "2 , 19DAMP . RPTIOS = .00O,
GROUND MOTION NORMALIZED TO 1.0 G
T R(S6 1 N tD^5E.0101 .020.0701 .100
0.0 5.0 10.0 15.0 20.0 25.0
FIGURE A46 I E .F QUENNY, C PS
RESP[ONSE ~SFECTVRP
PARKFILELD--Sl9GtSN~HE2 O~~~DPP . rI 10nS .rf~ Cv.0~20
S 4 H- GROUND) MOTION NORMAL IZEID TO 1 .0 G
-IJ
F IGURE 14. ý7 C.:--H
r ~
7 RE S P' NS ..
cr
tit
cr
~1:
CI
CL
HPCHINL]HE)onriD. Pnrin's -- .301"-, .010, .020
6.01
4 .S
0.0
- GROUND MOTION NORMALIZED TO 1.0 G
K ~
, .......... '.'t-
0.0 5.0 10.0 15.0 .20.0 25.0
FIGURE A49 FIGU~E 49 REOUENCYI CPS')
.9 - - I
U-J 9.0
z0
cr
wY
cr
cc)
00
Lu('0)a-
RESPONSE SPECHPCHINLHE9 196[DAMP. RATIOS .OOs,
GROUND MOTION NORMALIZED TO 1.0 G
:T RP15 EL
.0109 .020
.0?0) .100
3.6
1_.8
0.0 £5.0 10.0 15.0. 20.0 25.0
FIGURE -A50 FREQUENCYiCPS
2< -2. RESPONSE SPECTRALIMlqi 1%9G NBE
z DAMP. RPTIOS = .005, .OlO, .020n .050, .070, .100
.c" GROUND MOTION NORM4ALIZED TO 1.0 G
ww I
LI.-
J I i ___
o 4 E •
a_
r. C 5.C 00 16 c 20.
FREQUENCY. CPS
.. * 'x_..-)
.oo1-t- RESPONSE SPECTRAI CPSTPIC,19 iSG9E
_ DAMI RAnIn r .OO:, .010 .020
cI* 4 [.80 -](GROUND MOTION NORMALIZED TO 1.0 G
-J
U I
L 2 .
CIIN
ui 2 40--
2OL --- -.-. ....
o.0 .O .10.0 nS.0 20.0
FIGURE A54 FREQUENCY., CPS
u-
0
1-4
-JIQuar
Ld1I-
-JCl
co:
00.0Li)(U3a-
9.0
5.4
3,6
1.8
'-. 0
RESPONSE SPEBPNK [IF CPLIFOrMP . R rIioS = .005
.050
GROUND MOTION NORMALIZED TO 1.0 G
CTRP•,1971, NI-E
, .010, 020.0701 .100
Vtq N
0.0 5.0 10.0 15.0 20.0 25.0
FIGURE A55 FREQUENCYi CPS
La-
z
Ld-i
Ld
C3
CL
i 2 RESPONSE SPE(BPNK OF CPLIFCAMP. RPTICS = ,C0S,
GROUND MOTION NORMALIZED TO 1.0 G
.070, .ico
4.
2.
0.0 10.0 15.0 20.0 25.0
FIGURE A56 FREQUENCYCPS
cr..
0
cil
-juii
E3
(LLa--
C_
Iw(na.
0 RESPONSE SPECTRA
4.. e a --
A
UNIVDIMP .
SHERRATIOS =
P"l VEN.005N.0OOS,
, 1971, NS.0101 .020.070, .100
"'tROUND MOTION NORMALIZED TO 1.0 G
I3.60
2.40
? ?.q
- 4 J - ~~~I .44..L~it.4.~.J...II I
C'~AWIW
rýUAA'
1
/W - fmc-t td~~~t ~ -4-A~'4 ~ + 4 1 - -
11(\ ý Výjjj
Nv V
1IiKP,.-, .
04~y- '- t i +p - -A I I I
-r * I I I.
0.0 I0.0 20.0 25 .0
FIGURE A57 FREQUENCYCPS
<0 0.0(
z0
9- .01a
Lu
S6.01
w
o3 4.0'
cr_
C30
w 2.0CL
0.0
LINSE SPEC. SHERPTON
RATIOS = .00S.050
ION NORMALIZED TO 1 .0 G
0.0 5 .0 I0 .0 .15.0 20.0 25.0
FIGURE A~58 FREQUENCYCPS
La3 12.
z
crwL
-Jua
RESPONSE SPEU .N, HOLIDPYDAMP. RArIOs =.OOS,
S.050'GGROUND MOTION NORMALIZED TO 1.0 G
CTRAINN,• I 19 NS
.010, .020
.070, .100
6
.
7 .
El)mcr.4.
0. 2.
0. ,0 10.0 20.0 25.0
FIGURE A59 FIG~fE A59FREQUENCY) CPS
0
CLL
-JLuu
LUI
0Z)
a-
12. RESPONSE SPECU N. HOLIDOY IDAMP. RATIOS = .005,
.0501
GROUND MOTION NORMALIZED TO 1.0 G
TRFANN, 19?1? EW.010Y .020.070, .100
0
4.
)
0.0 10.0 15.0 20.0 25.0
FIGURE A60 FREQUENCYi CPS
?b .so •RESPONSE SPECTRPEUREKA, 1954) N79E
Z DAMP. RATIOS = .005, .010C .02C3 ".0s0. .0C,? .100
6.0cCE GROUND MOTION NORMALIZED TO 1.0 G
wL-
u
w_J
S3.00
cr.CN
0
r-,i
C . .0 20.0 25.0
FIGURE A61 rmmuur-Nt-Tlur;D
RE SPONSEFEUREKPiI I9jDFPhP. RPT0DS
S PE C: U41N-1h
*r00S, .010.-0 C) ,rj100
Lu
CL
C-1
-J
Lu
S10201 00
6.0GRO1111D ?,TIO I N NORMALIZED TO 1.0 G
3 .0
1..5
- +- - I- -+ . -
0 0- , 15. 0 2" .0
FIGURE. A62 FIGURE A62 '. IEGDEflC~r'iCO
Pill
LMi
LJ
CL
12.
RE S PFNSE SPECTRAL1LYMP1PI5 1965~ 5S04EorqMp. ~RPIoS .005) .0105
.3S0' .0701.020.100
0 . 10.0 15.0. 20.0 25.0
FIGURE A63 FREQUENC'Yi CPS
*c 12.
z
CY
LJ
DJLU
D.(I
C3
a-
9.
4.
2.
0.0 5.0 10.0 15.0 20.0 25.0
FIGURE A64 FiGUR A64FREQUENCY5 CPS
6.0
pDrMP. R Is IS . o005, .010, .0200 •. O0SO .070, .100' 4 .1 -0 GROUND MOTION NOFRALIZED TO 1.0 GCT-
IL.
LuI
Il
--C• 2.40
LU 1.20U)
-. 0.
0.0 5.0 10.0 15.0 20.0 25.0
FIGURE A65 1r'l:ncnKmrv roDC
6.0
z0
I-wuj
a
RE SPO[NSE SPETEMBLOR915i66i
DPrMP. RPTIOS C005.050
GROLU4D MOTION NORMALIZED TO 1.0 G
N25E .2
S.07C, .1004. .84
3 .6C
-J
0
0
2.4
1.2
-. C0.0 5.0 10.0 15.0 20.0 25.0
FIGURE A66 FREOUENCY, CPS
ci
i S . r M E'qNrDqmpI I
RE S PONSENG R PIT1
SPEC TRUM~01005
GROUND MOTION NORMALIZED TO 1.0 G
12 .
C3
0
0 31)
00.0 1.0 2.0 2.
FIGURE 81 PERIOD) SEC
z
C3
a:
-j
an
LlJcna..
IS .0Ot N1EDIPN RESPOflHP. RPTILI
ONS5E SPEC TRUM'= nOOS
GROUND MOTION NORMALIZED TO 1.0 G
12 .C
9.q
6.%
.q
0.C
0.0 1.0 'I.s 2.0
FIGURE 82 FIGUE 82PERIDOD£EC
MEMO Wý
0L
z
w-J
I
a
0
U)CL
STD., DEU. RESPONSE SPECTRU[1
2.0
1.s
DFhP1 R -PTID[
GROUND MOTION NORMALIZED TO 1.0 G
.006
IZ"/N
0.0 .0 1.0 I.S 2.0
FI GURE . 3 PERIOOSEC
C3
cr
-j
F-i
a03
0L
15.
12.
9'
6.
MEPN RESPONSEDAMPING RPTIU
SPECTRUM= 0.010
GROUND MOTION NORMALIZED TO 1.0 G
I!
I
~~~1_
0.0 I .5 2.0 2.5
PERIODiSECFIGURE 84
I-
-J
ci
0
mI
a.
15.
S.
rEDIqN RESPON.D1MP. RPTIO =
GROUND MOTION NORMALIZED TO 1.0 G
E ISPECTRUM.010
3.
0.0.0 .S 1.0 2.0
FIGURE B5 PERIC02 SEC
L&.
z
'-4
a
V-)
-LJ
STID. DEU RESPOINSE SPECTRUMDAMP, RATIOI 1C010
GROUND MOTION NORMALIZED TO 1.0 G2.00
i .Sa
fMLA iII Oa I
.4' - '-*- _______ 6
/illI
K "I14- I I I I 4 4 ~ 4- -. ~---- I .~-.----- 4-
I
r~Li .UI - -, I I ------- ~-------------~-
T
c.0 1.0 2.0*1-
FIGURE B6 PERI00d SEC
ILL
z
0ý
wIJ
wu
clw
C-
0
a-
15;.(
12 .1
9.
6.
3,
0.
MEAN RESPONSEDAMPING RFITID
GROUND MOTION NORMALIZED TO 1.0 G
SPECTRUM= 0.020
0.0 .5 1.0 1.5 2.0
FIGURE 87 PERICDi SEC
.-
w-Jl
Lu
L!I--
-J
U)
cc:CE
CL
MEDIRN RESPONSEDPMP, RPTIO =
GROUND MOTION NORMALIZED TO 1.0 G
SPECTRUM020
12 .0
9.0
- 1r
01.0.0 . 1.0 1.S 2.0
FIGURE B88 PERIOD0GEC
C . ] STD DLYE' L RESPONSE SPECTIRUM
DP[P. RPP1I.. . ,020ZC3
GROUND MOTION NORMALIZED TO 1.3 G
2LU Ij .
.Jm
LU
C7,
C3
rri i • _ _ ____ ___ ___
-- ---- -- --- --
. .0S2.0
FIGURE 13 PE-RI00YSEC
2.5
L.-
0-- 1 . C- M1E: PN R E SPflNE ý- c) SP E CT Rulm0o.osor !'RIN RPqT Lr-i
7-
(2
Lu
Li
C:J
C,
.Ci
GROUND MOTION NORMALIZED TO 1.0 G
12.0
6 . j
2 ,.0
0 ,.0"1-
"[ I 1
1.0 2,0 C-
PERIQDSECFIGURE BIO
UY-4:
-jwuscu(-3arwd
W40
maC3
0~
toa.)
I1EDIE N RESPONSE -SPECTRUMD P M P RPTI Dl - 'OoD
12.0GROUND MOTION NORMALIZED TO 1.0 G
_1 --- , --.-. ...
6 qQ
3 G
I ic - 1. - r -9 1 r
0.0 .5 1.0
PERIO~iJSEC
2.0 2.5
FIGURE 811
U-H
a-4
ULU
Il---
cr'
CJ30
m
CL.
a..
2 .S
2.0
1 .Si
1 ,O1
STD. DEU. RESFDAMP. RPTI =
GROUND MOTION NORMALIZED TO 1.0 G
>ONSE SPECTRUM.OSO
0.00.0 1.0 2.0 2.5
FIGURE 812 PERIBODSEC
<5.r MEPN RESPONSE SPECTRUMDPfMPING RPTIO - .070
U7-
12. GROUND MOTION NORALIZED TO 1.0 G
Lu .uF--Cu- 9 G
cl
S, .0 1 CE 2.0 2 S
...... ... PERIGDiSECIIUUIL U0I.
z
Y-4
LUI
cr
wL
U)m
0
Ld
I3)
MEDIAN RESPONSEDPMP, RPTII --
GROUND MOTION NORMALIZED TO 1.0 G
SPECTRUM070
12.
9.
6.
3.
0.0.0 1.0 1.5 2.0 2.5
FIGURE 814 PERIODOSEC
t.-4
a3
wr-I
wJuu
0
0
wUJ
2 .50 STD.DfrPMP
rDEt RESPDNSE SPECTRUMRP~TIO .070
2 .OCGGROUND MOTION NOWtALIZED TO 1.0 G
I .
.s
r. r
Q.C 1.0 1.S 2.0 2.5
FIGURE B15 FERIG0D SEC
16
FMEAN RESPONSE SPECTRUM,DPMPINC RPTIEJ = 0.100
GROLND MOTION NORMALIZED TO 1.0 G
clLu
LU
uj
c:LI
LLJLU)
C-J I.x
:IuJLU
C. C1.01.52.0 2.S
FIGURE B16 PERIOD, SEC
C3
z
w-JLua
ui
a
0
U)Cl
15 .01 MEDIAN RESPONSE SPECTRUMDPMP. RATIO = .100
GROUND MOTION NORMALIZED TO 1.0 G12 .a
9.C
e.01
3.0 - 4-.----~-..-.. . - __________ - I -
I I I
I_I i -
O. 1.4' 9 1 -v 1 I
0 .0 .S 1.0 1.5 2.0
PERIODSECFIGURE 817
u-I
<- 2.0
a
Ii-JLuU
ciLLuI-
c3
Q-
0 L
00
FIG3o81
STD. DEUL. RESfDPAMP. RPTI-
GROUND MOTION NORMALIZED TO 1.0 G
PDNSE SPECTRUM.100
.0 1.0 :. .5 2.0
PERIOCiSEC
I0
I
MEPN RESPLNSEDAMPING RATIIOGROUP Ali
GROUND MOTION NORMALIZED TO 1.0 G
SPECTRUM= 0O.020
Z
-JLU
Lii
I-
CImcl
(L.
12.
6.
3.
0.0.0 .5 1.0 ._1.5 2.0 2,5
FIGURE CI PERIOD, SEC
u- 16 .q GROUP P.. READPIP., RPTIO
MEDIAN
S PONSE= ,020
SPECTRUN
zC
LU-JLu
Ua
12 2c
GROUND tIOTION NORMALIZED TO 1.0 G
LU
a-
G . r*
3 ). cit I
t
0 . S.0 1 .5 2.0 2.5
FIGURE .C2 PERICDE SEC
LL
2.5 GROUP PI RESPONSE SPECTRUMDP' R:PTIlIO : .020STPNCPRi C EUI,, r~ircN
S2.00-CI GROUND MOTION NORMALIZED TO 1.0 GLi"
-J
0.J
LLi
1.0
Z)
•-"m i "ci
LI I "
o. I " I I
It:
0 .0 .. .- . ._._2 0.
0.0 .0 .0. 1. 2.0 2.5
FIGURE C3 PERIOD, SEC
.4
c3
LA
Lii
rl-]
L7i
.D
12.4
MEqN RESPOINSEDAMP1PING RPTIWGROUP A2
GROUND MOTION NORMALIZED TO 1.0 G
SPECTRUM= 0 0420
S.) -
'1rts
-.
4
[
I
II-,-,-
-4
..., -- --t- , , , :- * '- - .- --t - - - - - --
I n0 S -~ 2.0 r
PEriC SEFIGURE C4
U.
0 I5 ,0. GROUP P2 RESPONSE SPECTRUMI DFPMP. RPTID
M E D It AN
'-I
Z0
c1
-LJ
Lii
ci
U]
Cl
ml
ci:
0~
12 .CtGROUND MOTION NOF•.ALIZED TO 1.0 G
6.
3.
0.0 1.0 1.5 2.0 2.5
FIGURE C5
C73
z1-l
ci
-LJwu
1*-
0
0
0-
GROUP P2 RESPEUPHP. RPTIO =STPNDARD DEUIATION
GROUND MOTION NORWALIZED TO 1.0 G
]NSE SPECTRUMo020
2 .0
.i .,:
1.0
0 10
I I
0.0 1 .0 i.5 2.0 2.5
FIGURE c6 PERIDDSEC
I -
IL
z
ciCL
000Lii
co(I
1 S oi.
12 .•
9.C
6.C
3n5
M1I N RE-SPOINSE SPECTRUMDiMPING RPTIOGROUP A3
GROUND MOTION NORMALIZED TO 1.0 G
0 , 02 0
C" 1.0 2.0 2.5
FIGURE Q7 PERIODSEC
z0
cr
c-4
Lu
U-)
a.i
I£,
12.
9.
6.
3.
0.
GROUP P.3 RESPODPAP. RATIO =
GROUND MOTION NORMALIZED TO 1.0 G
NSE SPECTRUM.020
0.0 .s 1.0 1.5 2.0 2.5
FIGURE C8 PERIDDsSEC
hL
AC 2.
a
C3
co
C-)ci
LU
a.
FIUR
GROUPOAMP,,STPNOPqRD
q ý,ý p r-- cr r, i *
SPECTRUM
RATJIO .020
DEUIPTION
NORM1ALIZED TO 1.0
G
I• r-1 L , ,-• _
GROUND MOT I ON
I .
I .&j
.s1.0
C9
1.,, r
2.0
PERIOD, SEC
cIs FIEAN RESPONSE SP
DPAMPING RPYT] =z GROUP Bi
12 .0 GRJUND MOTION NORMALIZED TO 1.0 GCL
.-JLu.
ci
_L1
•'-r' 1 . . 1'-
C1
1.0 1.
PERIODDSEC
IC T R U M),,020
2.0
FIGURE CIO
FIIT~ I I ml
L.Is .Oi GROUP
DPMPMEDIPN
Bi RESPONSE SPECTRUMR PT 10i = 020
cr
F-
0
in
0
U)(m
12 0 GROUND MOTION NORMALIZED TO 1.0 G
6.0
3.0
0.0
0.0 .5 1.0 1.5 2.0 2.5
PERIODSECFIGURE C1I
GROUP BI RESPONSE SPECiDAMP. RATIO = .020
z STfNDPRC CEUIPTICN
2 0 GROUND M)OTION NORMALIZED TO 1.0 G
w=-Jw
a
1 .0
ciI I
U) II'
0.0 1E 00J . 2.0
PERIODSEC
rRUN
2.5
FIGURE CI1
IL
.MEPN RESPONSE SPECTRUM
DFr[NPING RPTIL = E C020GRCUP B2
H1S12.6
GROUD 1OTION NORMALIZED TO 1.0 G
uj
-J6.,
Il-Cr_
w 3j,0-
0.. .5 1,0 1.5 2.0 2.E
PERICCSEC
0-Is . 0ý GROUP B2 RESPONSE SPECTRUN
z
H
I--
-J
L)uJ
uj
F-
C-L)
.LJU)
Q
0.
DAMP, RATIO =.020MEDIAN
GROUND MOTION NORMALIZED TO 1.0 G12
9
c
01
6.%
3.G
0 .c0.0 1.0 1.5 2.0
PERIOD, SECFIGURE C14
LLJ
I-i
L*J
I-
0
cnr
Q..
2 .SO GROUPDPMP.STANDPRD
B2 RESPONSE SPECTRUMRATIODEUIATION
= .020
GROUND MOTION NORMALIZED TO 1.0 G2 .0a
I. .50
1 .00
C.0G
0 oG0.0 I.0 1.S 2.0 2.5
FIGURE C15 PERIODSEC
u-
CL
w-LJ
Lii
C3
CL
NEPN RESPENSEDPMPING RPT-IfGROUP Cl
GkOUND MOTION NORM$ALIZED TO 1.0 G
SPECTRUM= 0 .020
12.
9.
6.
3.,
0.0 .S 1.0 1.5 2.0 2.5
PERIOD)SECFIGURE C16
Lj~
frI
cruIcl
Cf
cc:u
(AJ
Is . 0- GROUPDPM'P.MIEDIAN
C.I RESPONSE SPECTRUMRPTID = .020
12 OýGROUND MOTION NORMALIZED TO 1.0 G
9 r.
6 ,C-
.A I
o .c0
iI
.0 . .. 1.0
PERIOOSEC
2.0
FIGURE C17
,ý V,ý . ý . k. . . .I I . I I ýý ! , . , - . - - - $1 .7L .* - . . -)ILL - i, - I- *. I ý ý . 11
L-
CD)
~ .•C+ G REJ UP 1Z R PE P FN SPEC T RUMR PTIT 1-
£Tflnr-4RD 51:U I Pi F'NO20
-~U-
GRCUIJIU M.OT ION NINM.ALIZED TO 1 .0 0ý
I . ~Ci
- . U'
i
/*. ,
-1~
2 . 0 2 .S
FIGURE C18
I 4*1
MEPN RESPONSED P MP.ING RPT ILGRCUP C2
GF;OUND A. OTION NORM- ALIZED TO I .0 G
SP ECTRUM0 .020
I.
1.0 1.5 2.0 2.5
'::": " !9 PERIODSEC
i GROUPDPMPMEDIAN
C2RA
R'ESPOT 10 =
NSE SPECTRUM,020
GROUND MOTION NORMALIZED TO 1.0 G
'4
\1.
4'
1.0 2.0 2 .,
"-z ! ' ',; 1". , C 2 0 PERIOD) SEC
G.ROUPOTMP .STANDARRD
C2 RESPRLATIONOEUIPr IoN
LINSE.020
SPECTRUrM
'4-
:, .0 .5 1.0 2.0 2.5
PERICOD SECF ~(WUR C21
.MEPN RESPONSE SP[DE MPING RPTID = (GROUP I
,I 2 '• - . GROUND MOTION NORMALIZED TO 1.0 G
1 . i
_______iO__
". I I
I I I
iI I
0.0 ." 1.0 1.5
•-,,,o• ,.,-,PERIODS SEC
-CTRUMD .020.
2.0 2.5
' UVINL L.LL
I .1- I'MEDIPN RESPONSEIl DMPING RPTIDL
GROUP
_ GROUND MOTION JORI.IALIZED TO 1.0 G
C3
cr'
1 .Lu24
J .
1E60
0.00 1 + ÷
0.0 .5 .0 1,5J
FIGURE c23. PERIDDSEC
SPECTRUMS.020
2.0 2 .E
UL-
0
arw-LJ
I
-LJ
0..L
4 .01
3.2'
2.4
1.6
.8
0.0
RESPONSE SPECTRUNRATIlO = 0.020DAMPIN
STPNDARDGROUP I
OEUIATION'
GROUND MOTION NORMALIZED TO 1.0 G
\AI0.0 .5 1.0 2. n 2.5
FIGURE C24 PERIODiSEC
Fr
.4.0 "HMEPN RESPONSE SPECTRUMDAMPING RATII = 0.020
.z GROUP II
'- 3.20 OG
a " GROUND MOTION NORMALIZED TO .3 G
ujLli
J..U: i I--II.cx
I • . .
-J 1
0.0 .5 2.0 1." 2.0 2.5
PERIBDSEC:'IGURE C;25
z1-4
L±J-Jw
C-J
0
co:(I
4.0
3-,,
2.4
1.6
0.0
MrEDITN RESPFINSDPMPING RqTITGROUP II
GROUND MOTION NORMALIZED TO 1-.0 G
E SPECTRUM= 0.,020
0.0 1.0 2.0
FIGURE C26 PERIODD SEC
S4. 0 RESPONSE SPECTRUI
DPMPING RPTIO =z STANDARD DEUIPTION
&GROUP IIc 3.2 GROUND MOTION N•ORMALIZED TO 1.0 G
wu, 2 .40
w
L.i ____
? 1.'6 -_._
cr
Cl'
0 .001I.0.0 1.0 1.5
PERIOD)SEC
D . 020
2.0
FluvKh LZr
IL
7-)
cr
4.0c
3 . .
VILHN <Lz
I DP MP I NGI GROUP III
SP [N, S E.R PT 101
SPEC- 0
TRU0020
2 .4r
CLI . 6C
C0
LW
a-
0.n 1.0 I.S 2.0
FIGURE C28 PERIODD SEC
!I
u-
-r
L-2
Lli
4 .0O RESPONZE SPECTRUMDPMPING RPTID 0,,020STrNDRD OEUIPTIONGROUP III
Gr,3UND MOTIOIN NORMALIZED TO 1.0 G
2 .40a . ,- -r
1.60
0.0o..0 1.0 1.5 2.0 2.5
FIGURE C30 PERIOD, SEC
4 .O0
cr
LUj
I-01
ILf
n)
1=~HEPN RESPONSEDOrPING RPTIOGROUP Iu
SPECTRUM= 0.020
3 . 201. GROMND IAT IOt1 NORM4AL I ZED TO I .0 G
i
.1=.2 . C[
DL .6
ft
0.0
I
I II I
0 .0 1.0 -I . 2.0
FIGURE C31 PERIODJSEC
'C
z
y-IL.I
ci
C-
4 .01
3.2,
2.4
1.6
.8
0.0
MEDIPN RESPLNSEDPMPING RPTIO =GROUP IU
GROUND MOTION NORMALIZED TO 1.0 G
SPECTRUM0 .020
0.0 1 . 0 1.5 2.0 2.S
PERIOCiSECFIGURE C32
U-
0
z
Y-JLUJ
a
Iy-
c,-)m-
4.01
3 2
2.6
0.0
,MEDTPN RESPLNSDAMPING RATIDGROUP IU
GROUND MOTION NORI4ALIZED TO 1.0 G
E SPECTRUM= 0o020
0.0 1 .0 1.6 2.0
PERIODOSECFIGURE C32
LL.
z
'-4
Li
LJ(1-
cr-
00
Li(i:)
CL
4,0•
3.20
2.40
1.6G
RESPONSE SPECTRUMDAMPING RPTIOSTANDARD DEUIATIONGROUP IU
GROUND W-TION NORMALIZED TO 1.0 G
=O0 20
.0.0 I.S 2.0
PERIODi SECFIGURE C33
J DHN A. E31. UMEc + q''SOCI A FE:: ./ENt IN.E 1 -N E E R ýS
rD A m -ni m c R P.r 14- .G2 0
Pc-EuDf nFl
CO NT SJU
S,~; fCCEL.
LE U EL S
.6000E+01
.ýO00E+oi
" sGOnGE+0i
" 20C'0E +01
* I000E.101
.600.'DE+00
.2 0,C0 E + Cl
RES~p DNS E E I: Lý' EL 1 P ~ESV rlqp
E L 'C'ENW7 [],' 1 '9 -! C"I H s
GROUND MOTION NORMALIZED TO 1.0 G
IFIGURE Dl
i , , . i i i i I ! i i i i i i ! .- --4- I I I - - - i i
I- It -
- -'S
-'S
a* . ~ o*
J~-
> b A
___ 'V ~o *'
In~ a ~ ~
~ ~. -'-.-.-
p __
-~~r ~r-Z... LI
6K>* .c~nZtŽ~-~ *I~1
v'I ~ . J ~I~
o ~ r~ ~~ ( ~V- _ N~1V
- --
c ~ ciiQ ~ (0Ia,.
~ .z,. ~r-'L~" I -z~~ ~ - -~
r~-~ ~ C) •0-.------.------.. '--------- -S
-~
I -.- - ,z~ -~ Y
~
a
~-~-'N~ 3IS4 3'
Q ---.I.
I 4
/
6
JOHN A. BLUME + ASSOCIPAES / EN6INEERS
DIPIN1PINC' RFTI O= .020
ACCEL.PSEUDO ABS.
CONTOUR LEUELS
.0oc0E+O1
.SOOOE+O1
.4000E+Oi
.3000E+Oi
*2000F+OI
.IO00E+OI
.6000E+O0
.2000E+O0
RESPCNSE EN'JELOPE SPECTRA (RES) MAP
EL CENTRDi. 19409 EIWGROUND MOTION NORMALIZED TO 1.0 G
FIGURE D2
263
LIW
'44
,73
1,4
U4
ISO
C1.2
Css
c-Li
04 X.4 J,
.9~~M1 .1UP P~1 ~ ii ~4~I NhJ r, ~fillIAA.L~~k ~~o HI I~flv' I
C. .. .0 10~ .1 1.0 e.u ' 1. IL.C. it1.0 L2.C 13-C 14 .- IS -C ±l-C 1V.C 18.l0 -'D0 "IX 2. Z2.C .. I c ±4. -,.116 z1I.0 2.1. Z, -H,.C
'CIIE ESECC:413..Fl
JOHN A. BLUME + PSSOCIfrES / ENGINEERS
DNPING RPrfIC = .02C
PSEUDO ABS. qCCEL.
CONTOUR LEUELS
.60CCE+OI
.5CCOE+O1* d000E+OI
.3000E+Oi
.2000E+Oi
.1000E+01
.6O00E+O0
.2000E+O0
RESPONSE EN'JEL OPE SNPECIRP (RES') MOcP
TqFT 19525 N21ECROUND N4OTION NOR•ALIZED TO 1:.0 G
FIGURE D3
0.4ý
Al 0 0,zc
ACA)
.0 i.c Z. .C 4.C S.0 6 .0 ?.0 ea 1. 1 CA ! I X 12 c I a. CS Isc tC . c U .C e.0 2.. zr C ±c a. 2 S.C Jt' . 1. 2?c Mnc
C It-;
F T
PS ELDE r) C4 6 ý-,. , q'ý '- E L -
CnNTqiUR LEk.ELS
t2 EC CE f C.1
E* C COE -4-01
2 D Cir;-i~ E
*6 c .C,§F KKC217 C- E C' I-
j E N J E L D PE SPIN ON
Yý r Ir 4 I-. /~ ~ Li
GROU.,D MOT ION NORMAL IZED TO ,..0 G
FIGURE D4
W -A
-- --- - -- --
..... .. D .6. ...
< ~ (K K
'I* zl ýA
.6 1 ____
3,:c )., ý.z 5, c C.. a .c 10 2.0 11.0 12.C L3.G 1. .S .. C IC..C O~.f 19.0 &9.1 ZCG 21.C :2.C ,].Z .. .~S-C .0C 21.C 20.
OAPMING RATIO = .020
PSEUDO ABS. PCCEL.
CONTOUR LEVELS
.6000E+01
* £OOOE+OI
* 4000E+Oi.
.3000E+01
.2000E+01
.1000E+01
* 6000E+00
* 2000E+00
4
rcJi3K ()9
(~ 0 a6•> RESPONSE ENUELGPE SPECTRP (RES) mnP
HEILENP2 193S5 NS
GROUND MOTION tJ(N6.ALIZED TO 1.0 G
(IME I&6CM10f.1 F I
/
OPrIRING- RAFTIO -- .0J20
PSEUDO PBSZ. LJCCEL.
0
CCNTOUR LEVELS
.6000E+O1
.SOO E.*O1
.400OE+OI
3000E+o1
.2000E+OI
.IO00E÷01
.6000Ec-O+
.200UE+r0
RESPONSE ENVELOPE SPECTRA (RES
HELENMP 19353 Eý
GROUND MOTION INOIALIZED TO 1.0 G
60.
P)
Cr3
a u.. ý wf.. A - 1 1 -- t-- --h -.-- f-- -t- - --t-'1---I---+- t-- --- $- i i i -.0 .0.2.c V~C 4.&3 S.O 6!G ?.a0 .0 3.c 10,011. ±2.0L s 1.1.0 14-' 1s.C .t6.0 17.0 10.0 19.2 n0 c 21.0 22.c 13.0 Z4.z 25.0 26.0 27. 20.0trflE hSECCOWS1
DAMrPING, RATIOD .0-L0
PSEUDS PBS. PCCEL.
CONTOUR LEVJELS
k-EY9S0 E +401
* 5000E+01
.4000E+01
.3000E+01
.2000E+0!
.1000E+0!
.6000DE+00
.2000E+00
"•1>
6RESPONSE ENUELOPE SPECTRA (RES) M'
GOLDEN GP'TE9 195 7 N10GR~OUNlD PtJTIotl NORMALIZED TO 1.0 G
r IM E ScoliCs) F Irime
DPIIPING RATIO =.020
PSEUDO PBS. QCCEL.
CON~TOUR LEUELS
.6000E+01
.500!9E+01
* '400E+Oi
.3000E+01
* 2000E+01
.1000E+01
*G0O0E +00
.2000E+00
C,~vRESPONSE ENUELOPE SPECTRA (RES) MAP
GL!_DEN GWTE 19E? I N8O1ý
GROUND IAOTIC-11 !,nORMALIZED TO 1.0 G
I I II* -+------.---4-~Io ~ 3. Z. 30 4. ~. .0 0 90 10.0 It1.0 12.C 13.0 1,. 1. 1,;.0 17.0 18.0 19.C 7:.C JL.C J2.9 ?3-G --2 Z.0 Zt .0 ? !0.0 'q
ul"E IsECalIDSt F I GU
LA.
z
Ir-
a
U)
0-
Is.0
12.4
RESPONSE SPECTDPMPING RPTIOMEDIAN) Y=N I Y=T.04SG
GROUND MOTION NORMALI ZED TO 1.0 G
R P0 .00O5
Y=2
9 . 1ý
G.O%
3
n ri - 1-r
0.c
-r
1.c 2.0 2. G
FIGURE El PERICCSEC
U-
z'-4I-
ci:w~~2wU
1±1
-J
cr~
c:JI LIU)a-
i~Lor RESPUNSE SPECTRPDPMPING RPTIZ zMEDIA:N) Y=z, Y=z1.645, Y=
GROUND MOTIO ON NORMALIZED TO I .0 G
0 .010
12 oi.
Itr• V
- -- ~ - -
L. 1.0 I .C 2.0 2.S
FIGURE E2 FIGURE E2 F £. RJii [~S E C
- w -
CL
(-I
LU
'is£~ ~E~;[NE3'L
D P INP PiW 'P" iTL IJI2M U~r~ P. Y -j 0..eP. 4Eýý Y
1.?
t3.d!34
GROUJr~ .'orio~j Pj.jR~.¶AI.t7j~ To I .0 t3
N
~''~N
N -'- - ------.- *-.--- --------- -
~-.--~
S.-.--.-..... -. ---.- .~-..- -~1-------------+4---
t" - Al n= a - P%
;IGURE t3 K;m .. r., - - C
z
a
LuaJ
15 aG
12 .O
9.0
RESPONSE SPECYRPDAMPING RPTIO-- 0ME DIRNfl Y=1, Y=i.6463 Y=2
GROUND MOTION NORMALIZED TO 1.0 G
16.%
3.0,-
0. w, I
0.0 I.0 2.0T -r
FIGURE E4 PERIOD)SEC
LL
z3
c-JwýUL
ci-w
C12
02
ci0)aL
IS. I0 RESPONSE SPECTRADAPPING RPTIE0MEDIRN3 Y=1. Y=1.645)GROUND MOTION NORMALIZED TO 1.0 G
0.0~70Y=2
12 .0[
9.%
6;. +
3.0
0 #,,,-I--r
0.1-r
.5T
0 1 S 2.0
PERIfJO, SEC.FIGURE E5