ML20094E096
| ML20094E096 | |
| Person / Time | |
|---|---|
| Site: | Millstone  |
| Issue date: | 06/15/1984 |
| From: | DAMES & MOORE |
| To: | |
| Shared Package | |
| ML20094E062 | List: |
| References | |
| NUDOCS 8408090092 | |
| Download: ML20094E096 (45) | |
Text
{{#Wiki_filter:- r t
,- ',1 I
I I e 4 .I , t l l L j SENSITIVITY OF SEISMIC HAZARD RESULTS l AT MILLSTONE TO LLNL STUDY ASSUMPTIONS ON ATTENUATION AND SEISMICITY l FINAL REPORT ! ! TO ! NORTHEAST UTILITIES l i u Dames & Moore l l i
~
1626 Cole Blvd. O Golden, CO .80401 w June 15, 1984 j e400090092 840731 - PDR ADOCK 05000423 A l > pop
~ ~
g( ..
-i:; , . -- +.
l7 t
+
.$ '. l 5'-
/
- : TABLE OF CONTENTS y
~
M
, INTRODUCTION...................................................... 1 BASIC COMPUTATIONS................................................ 2 EFFECTS OF ATTENUATION EQUATIONS.................................. 4 EFFECTS' 0F ACTIVITY RATE..........................................
5 EFFECTS OF LOWER-BOUND MAGNITUDE mb, min........................... 10
SUMMARY
........................................................... 11
.: REFERENCES
. TABLES FIGURES ei 4
m
's 6
)
e w* r- --- 1--r---, ,,. - ,.,ee .,-.- ,,..,4.-- , ----w . w#3%-----.9--m=-y--evv-- e-ww~ t
Er .l..o. ..
. ,- J .e l
j r l INTRODUCTION In Agil,1984 the U.S. Nuclear Regulatory Commission released a seismic hazard study for nuclear plants in the eastern U.S. which was performed by the i . Lawrence Livermore National Laboratory (LLNL) (1984). Millstone Nuclear Power Station was one of the sites studied in that reporc; site-specific results were presented in terms of ground acceleration hazard curves and uniform hazard spectra. 7 The purpose of this study is to examine the results produced in the LLNL report for Millstone, and to assess their dependence and sensitivity to various assumptions made in the LLNL study. The basic methodology used in seismic s - hazard assessments today (including that used by LLNL) is well-accepted. The specific inputs to any analysis require some judgment; these inputs, and their effects on the calculated seismic hazard, are the focus of this study.
, While. the results presented here are precise and reliable, they do not
. constitute a full examination of, or a final position on, the LLNL study. Such
'an examination or position would be impossible to attain in the brief period since the LLNL report was released. Indeed, the LLNL etudy itself is in a
- preliminary' form and has not had the benefit of outside review or of feedback
.'from the seismicity and attenuation experts who participuted in it. The final
.results issued.by LLNL might change substantially as a result of the review and feedback process. The sensitivity results obtained here d_o, o indicate the
. quantitative effects of alternative assumptions, and thus indicate how seisnic hazard estimates at Millstone might change as a result of the review and feedback process.
6 i e
(f:..qm+. l
-t- l
$[]_L .. i r M- - l P BASIC COMPUTATIONS ! E l I
. For purposes _of comparison and sensitivity analysis at Millstone, we have selected a subset of ' the seismicity and attenuation model input used in the LLNL study. This has resulted in a workable number cf seismic zones, para-meters, and attenuation models.
^
. ' The seismic zones used in this study are the "best estimate" zones in the northeastern U.S. indicated by each of the eleven seismicity experts. These
!ce a nr ' are: listed in Tabic 1; they verc =ciccted using Tabis 4.1 of the LLNL study which indicates the zones contri'.cing most to hazard at Millstone. Figures 1
- through 11 shows these zones plotted on a map of the northeastern U.S. The seismicity parameters used (upper-bound magnitude or intensity, "a" and "b" values) are the "best estimate" values indicated by each seismicity expert (Table A3 of the LLNL report).
The six attenuation functions most heavily weighted by the four LLNL E. L attenuation experts were selected to estimate peak horizontal ground accelera-tion and spectral velocity for 5% damping. These functions are listed in Table 2, along with the weights placed on these functions by the attenuation experts. (To compute overall weights we elected to give each of the four attenuation experta a credibility of 0.25.) These weights constitute some 85% (for acceleration) to 92% (for spectral velocity) of the credibility assigned to ground motion functions by the experts; the results obtained using these six functions can be considered representative of what would be obtained if all attenuation functions had been used. The uncertainty in ground motion was modeled here with a standard devia-tion of the natural logarithm of ground motion equal to 0.6. This is repre-L sentative of the values specified by the attenuation experts for ground acceleration and spectral velocity, with the exception of one expert who nreferred a best estimate value of 0.9 for spectral velocity. This difference will not have a substantial effect on the sensitivity results to be presented below.
- g. I
.. . .. =
v a 5 Nine of the seismicity experts elected to represent seismicity with d body-wave magnitude ab, and two (nue.bers 1 and 5) used Modified Mercalli (MM) intensity for the northeastern U.S. Experts number I and 10 suggested special _i m relationships between ab and MM intensity; the others deferred to the Nettli- - Herrmann relationship presented below as equation (3). All of these preferences were considered and used to develop the proper relationship between seismicity -@
, and ground motion estimates for each seismicity expert. ia m _
A basic set of results was produced to replicate the LLNL calculations and f _ ensure that no errors in input had been made. This set of results was obtained ; using the best estimate seismic zones and parameters, and the best estimate - attenuation r.odels for acceleration. Two LLNL attenuation experts chose the i Nuttli (1983) model, one chose the Campbell (1982) model and one chose the T Trifunse (1976) model. The resulting probabilities of exceedance at various 3 acceleration levels for these attenuation models and each seismicity expert [ were then averaged. The resulting probabilities of exceedance, which LLNL
]
3 labels the "best estimate" results, are shown in Figure 12 as reported by LLNL g (in its Figure M1-1). Our attempt to duplicate these results is also shown in - _== Figure 12 and is labeled "This study - average of best estimates." Our results y lie slightly below the LLNL "best estimate" curve because we have not modeled 5 all seismic zones and have elected to use approximate, more efficient calcu- y lational procedures, but the two top curves in Figure 12 are generally in j reasonable agreement. j E Also shown in Figure 12 is our median hazard curve for Millstone, produced h using the normalized weights shown in Table 2 for the attenuation functions and 5 weighting all seismicity experts equally. The median curve lies below the j average-of-best-estimate curve by a factor of about 0.5 in probability. This y indicates, in a loose way, the difference between median and mean results for - highly-skewed distributions of probabilities of exceedance. As most of the
- 9 results in the LLNL study are presented in terms of constant percentile (median and f ractile) hazard curves, we use the median curve in Figure 12 as our base 4
y - j3 s W= . }- _4_ r c.1 n. case, and compare later results to it. . This curve (and others compared to it) were Eproduced by sixty-six hazard analyses (eleven experts times six attenua-tion functions). j r
' EFFECTS OF ATTENUATION EQUATIONS
. .y i
\i The first effect examined is that of the weights assigned to the attenua- )
.x ; tion-functions. The six attenuation functions are plotted in Figures 13 '
t through 15 for peak ground acceleration and 5% damped spectral velocity at 9 $g' jand,5 ha,.respectively. The spectral velocity curves labeled RS4 and RSS were only available at 10 hz, which was considered a close enough approximation to 9
-hs to warrant.their use. The Trifunac (1976) and Trifunac-Anderson (1977) curves . (labeled G16 and RS6, respectively - see Table 2) lie substantially above' the others at distances greater than 10 km, for several reasons. Among
.them is that the original Gupta-Nutt11 (1976) intensity attenuation (based on isoseismals) was used, rather than the modified version of this function. A 4
'second reason which is particularly important for Millstone is that the form of
.[ these equations contains a term exp(cS), where S is a variable indicating soil 7 conditions (0 for soft soils ,1 for medium soils, 2 for rock). Most of the data ~ on which the calibration was based was for conditions 0 and 1, so the use of ' the equatio ' for rock sites (S = 2) constitutes an extrapolation based on the functional form of the equation, rather than on data. Moreover, the most
^
recent attenuation equations in California indicate little effect of soil conditions on peak acceleration. The Trifunac form might be more acceptable if the intensity attenuation equation with which it is used was based on intensity values at rock sites (not on isoseismals drawn for all sites including high
- -intensity, soft soil conditions as usually exist in towns located in river valleys) but this has not been done. A justifiable alternative is to drop the Trif unac (1976) and- Trifunac-Anderson (1977) curves from the analysis on the basis that they are overly conservative for a mean-centered analysis.
.1 l
_ - . _ ,__-_ _ ____._ ,. ,_.,_ _., ,,,, , _,,.,_..._~ _.,,_, m,-r, m.,%.,_, - -._____ _ _ ,
7: e o;*., ;
.I . . ,
s* To; accomplish' this, we used the " alternative normalized weights" shown in
. Table ~ 2 to weight the remaining five attenuation functions. This has the
,[;.
effect of removing the Trifunac equations from the analysis and redistributing its weight among the.other.five attenuation functions. The:resulting median hazard curve for peak acceleration is shown in Figure iI !16. ' The change in hazard varies with acceleration level: at the SSE level' (170 cm/sec2 ) the probability at a given acceleration level decreases by about
~
-25%-and the acceleration at a given probability level decreases by about 10%.
g;p V q ~ )For.a typicalislope of the hazard curve these two changes are related by:
.(ground motion reduction factor)
(probability reduction factor).375 (1) C . For example, (.90) = (.75).375 At high accelerations the alternate weights
- - imply a slightly higher hazard than the base case.
. Figure 17 'shows equivalent results for spectral velocities at 9 hz and 5 hs. Here the'effect of alternate attenuation weights is largers at 10 cm/sec.
7-g spectral' velocities decrease by about 40% in probability or 17% in spectral
.vel ocity. -We concldu e ht at the fe fects of attenuation functions on spectral velocities are not necessarily well-estimated by comparisons based on ground acceleration. 'The reason is related in part to the larger range in estimates
. for spectral . velocity than for peak acceleration (compare ab = 5.0 at 23 km in Figure 13 to Figures 14 and 15).
g a 4 EFFECTS OF ACTIVITY RATE One area which was lef t . entirely to the seismicity experts in the LLNL 1 . study was the estimation of seismicity parameters (activity rates and b-values) from the seismicity catalog. Each expert was supplied with lists of historical seismicity, one list for each of the expert's zones. The list apparently included all events of MM intensity 1 IV or ab 13.75. Aftershocks, where
&, q b.ga
= ---
p.- J jf.e
- 7. g g ra o; * .
? -..
P .N - identified by . original sources, were tagged. Given such a summary of histor- l le' al seismicity, producing meaningful estimates of activity rates and b-values j is not a trivial task. One must divide the data into magnitude ranges, determine fneervals of completeness for each range, select a method of mathe-
-matically' fitting an equation to the data, perform the fitting process, plot
~
L . the data both in the f requency sense (number of events within a magnitude range) and in the cumulative sense (number of events greater than a given c imagnitude) and assure oneself that the data, the functional form, and the calculated parameters are consistent. It is very difficult to allocate the n# .4 time and .. resources to go through this estimation proccas rigorously for every seismic sone in the eastern U.S. How this task was acconplished by each expert is not reported in the LLNT. study. Thero is some evidence that the methods used by the experts to assess 1 rates of activity have, in many cases, produced conservative estimates. These
+ conservatisms may have been the result of one or more factors:
y. 1- o inclusion of all 1:M intensity IV events as mb > 3.75,
, o inclusion of af.4rshocks in the data base,
^
o ' use of the Nuttli-Herrmann (1978) intensity-to-nagnitude conversion in New England, and imperfect methods used to fit equations to eeismicity data. o On the first point, the Nuttli-Herrmann (1978) intensity-to-magnitude conversion estimates ab = 3.75 for intensity IV, but (even assuming that this conversion is correct) this magnitude value is more properly represented by the range ab = 3.5 to 4.0. Thus the data presented to the experts are better characterized as all events with mb > 3.5 (when estimated from intensity) plus all instrumental magnitudes of ab > 3.75. I I _ ~
e,
"[r.f . .] . . . . ."- .
- c r . y. -
A =' As to the second point above, the experts were notified by LLNL that the data base contained aftershocks, but there is no indication of if, or how, ( '
'these events might have been removed or otherwise accounted for. If labeled aftershocks were not removed, this would account for about a 12% increase in j
l perceived activity rate.
'f 'On the third point, there is some evidence (e.g. Weston Geophysical Corp,
.. '1982).that the Nutt11-Herrmann (1973) conversion, derived for the central U.S.,
may not be appropriate for New England. While this point is controversial and
- g. ,u there certainly is no scientific consensus on it, the recent instrumental seismicity suggests .that use of the Nutt11-Herrmann conversion in New England
. over-estimates rates of activity (see figures described below). Thus if one asserts - that the Nuttli-Herrmann intensity conversion applies to New England,
.then one must also accept the assumption that either:
+ o the exponential magnitude law does not hold over the ab range 2.0 to 6.0, or-e o rates of sei'smicity in New England have declined in the last seven i years, relative to the previous several hundred years, or
< .; ;o- the instrumental record in central New England is incomplete for ab
. . >2.0 over the last seven years.
, While any of these assumptions might be defended, each is unlikely. For a
.=
description of instrumencal' seismicity coverage in New England, ref er to Ebel (1984).
'To 'etermine the effect of alternate methods of estimating activity rates, L we adopted the.following procedure:
(a)~ historical MM intensities for New England were converted to mb using the relation (Weston Geophysical Corp., 1982): ab " (Io + .44)/0.67 (2)
(b) periods of completencss were chosen for each magnitude interval from the present back to a time just prior to an apparent decrease in activity rate (under the assumption that intervals for larger magnitudes were longer than, or equal to, intervals for smaller - magnitudes), (c) activity rates for mb 1 3.75 and b-values were computed for each zone using the best available statistical procedure (Welchert, 1980), and .
. (d) tne 66 hazard analyses for peak acceleration were re-run using the [
alternative Tates and b-Values. *' The Dames & Moore seismicity data base was used for the analysis. For New England this consists of the Chiburis earthquake catalog through 1982 with the . J revised event locations and magnitudes determined by Dewey and Gordon (1983). The only exceptions to this procedure were for seismicity experts 1 and 10 who ; chose to characterize seismicity using MM intensity. For these experts the seismicity characterization was not changed. , Figures 18 through 20 show seismicity data for zones 31, 7, and 4 speci- ; fied by experts 2, 3, and 6, respectively. These zones were selected for .= presentation because they include only the New England area where data from the New England seismic net are relatively complete down to the magnitude 2.0 level at least, since 1976. Three sets of data are shown on these figures. The x's indicate cumulative rates of seismicity from historical data using the Nuttli-Herrmann conversion to estimate mb fren intensity: . mb = 0.5 Io + J.75 (3) For these cumulative rates, completeness intervals were established as de-scribed it. ster (b) above. The squares in Figure 18 through 20 indicate cumulative rates calculated in an equivalent manner using equation (2) instead of equation (3) to estimate ab from Io . The circles indicate cumulative rates calculated from instrumental seismicity f rom the New England seismic net ( Pulli , personal communication, t h d 3
l"' j l'
*{
, _9_
_ <$ a 1984) for the period October 1975 to March 1982. Both sets of data generally ;
-lie below the rates which are obtained if the Nutt11-Herrmann conversion is
^
used for New England. Note that the instrumental rates at ab = 4.5 and 4.0 f r^- should not;be given much weight; they are determined from occurrences of one
~and-two earthquakes, respectively. On the other hand, tha rate for ab = 2.0 is '
t.
;.. determined from 140 or 150 earthquakes. The bars attached to the circles show
; one-standard-deviation bounds, calculated as the square root of the number of
, observations divided by the interval of completeness. The bars graphically a reinforce the last two observations about small and large sample sizes.
1
;ff;&: -
- The dashed line on .the figures indicated the seismicity distribution implied by each -expert's choice of a and b-value. (The use in the LLN' study
- of.a non-linear range between ab = 3.75 and some other value of ab specified by each' expert where the semi-logarithmic plot becomes linear -- see LLNL report L page A40 -- has little effect and is not followed in this study.) The dashed 3 flines in' Figures 18 through 20 are consistent with each expert's specification
.,- of:s and b-value. These dashed lines are generally conservative estimates of
. ,, the data, for'the reasons discussed above. For example, the x's in Figures 18 through 20 at ab = 3.5 were determined using all earthquakes given to the seismicity experts by LLNL and which were represented by LLNL as all events with ab > 3.75. Thus there may be a quarter-magnitude-unit bias in the dashed ,
lines compared-to the x at mb = 3.5. Ie The . solid: lines indicate the fit to the squares obtained by the maximus
~
likelihood procedure. This representation of seismicity lies below that
-indicated by_.each expert, and is more consistent with instrumental seismicity rates., In particular, the solid lines represent relatively well the cumulative rate of activity at ab = 2.0 (which is well-constrained with instrumental data) !
and 'for ab p_ 4.5 (where differences from the conversion of historical data to U magnitude by different means are reistively small and hence not controversial). i Thus the -solid lines are reasonable alternatives with which to examine the 4
' sensitivity of hasard results to the experts' choices of activity rates (and ;
b-values), vis-a-vis others. t I '
- , .V
- y. - >
g,
~k l
? :e a b ;q - 4;-
.,^
F 1A _to_ [$'b:;p m -
; g M
- Figure 21 shows the median acceleration hazard curve over all eleven I,
Tseismicity experts and six attenuation functions for Millstone, obtained using [ "
'g '
Lthe alternative activity rates and b-values. Also shown is the base case, for
~
-? comparison. : In general there is a reduction of 25% in probability or 10% in T acceleration implied by the alternative rates. Spectral velocity results are
.f .similar,and are shown in Figure 22.
; i'-
e
- EFFECTS OF LOWER-BOUND MAGNITUDE as ,a ,
Wh:gy ? .yg,7,f cthni parameters specified by the LLNL methodology (not by the seismicity .or ; attenuation experts) is the lower-bound magnitude ab, min. A value .of; 3.75 was used in .that study (or intensity IV, for experts who pro-
~
f erred intensity). ; This is quite conservative, at least for the assessment of
- seis'aic hazard to nuclear plant structures and major equipment, in that damage g - to . engineered structures -and equipment is not known for earthquakes of magni-1 s O 'tude less than about 5, regardless of the ground motions these events generate.
Thus 'the ~ inclusion of ground motion hazard from smaller events increases the
. mathematical risk but provides a conservative estimate of hazard for ordinary plant structures and components..
The ef fect 'of choosing different lower bounds was examined by using the experts' a and b-values, computing an activity rate f or each zone assuming P ab,'minL= 5.0, and recalculating seismic hazard for that lower bound and
~
i activity rate. ~ The process was repeated for ab, min = 4.5. The results for each case are shown in Figure 23. The median curve for ab, min - 5.0 is about
- 55% lower in probsbility (25% lower in ' acceleration) than the base case. The changes .for ab, min = 4.5 are-35% in probability and 15% in acceleration. Thus the'choi.ce by LLNL of ab, min = 3.75 has a major effect; the use of a more
', . realistic lower bound;for structures and mechanical equipment would lead to
'significant changes in the. median hazard curve.
Results- for spectral velocities at 9, and 5 hz are shown in Figure 24. The
~
'9 ths results lead to the same conclusions as for acceleration, namely that the use of ab, max = 3.75 is extremely conservative over more realistic alterna-i
r-
- .;r . 1
. n. . ,
.3 h, ,
tives',_by-a. factor of 35% to 25% in spectral velocity. For 5 hz, spectral velocity is governed by the ground-velocity-amplified portion of the response spectrum for magnitudes below ab = 5, and these contribute little to the seismic hazard results, so the curves are not sensitive to the choice of lower bound for spectral velocities below about 10 cm/sec.
' I[ SU12fADY Several assumptions and analyses used in the LLNL report have a major
~m u. imp'ac: on the seismic hazard at hi11 stone. The preliminary nature of the LLNL study, and the importance of these assumptions, require that care be taken in interpreting the LLNL results. Alternative choices and assumptions made in the
' f uture, during the review and feedback tasks, likely will lead to lower hazard curves-in future reports.
4 A quantitative estimate of the possible size of these effects is summar-7
~ized in Table 3 for accelerations around 170 cm/sec2 and spectral velocities
[ around 10 cm/sec. The combined reduction in spectral velocities for the effects examined here is in the range of 30% to 34% (reduction factor of .66 to p.
.70). The corresponding reduction in estimated probability level for a given
-L
. spectral' ordinate is 61% to 67% (reduction factor of .39 to .33). This implies, for example, that the uniform hazard spectrum currently identified by LLNL as having a 1000-year return period for Millstone, would have a revised estimate of return period of about 3300 years.
Thus, removing the major conservatisms identified in a preliminary review of the LLNL work means that the spectral amplitudes for a given probability level would decrease significantly. A visual representation of the effect is ) shown on' Figure 25. The peak ground acceleration hazard curve developed by Northeast Utilities is shown as the solid line, and compared to it are several
' original and revised hazard curves from the LLNL study. The long-dashed curves (labeled "SHCP" for " Seismic Hazard Characterization Program") are LLNL median curves, as originally published and as revised by reducing the accelerations at a given probability level by the reductions shown in Figures 16, 21, and 23 to I
I remove conservatisms. The revised median curve lies much closer to the Northeast Utilities' result: at the SSE level the two curves agree within a factor of 2 in probability. Figure 26 shows original and revised hazard curves from the LLNL-SHCP study, for three fractiles (the median, 15%, and 85%). This shows that the [ original LLNL-SHCP median curve lies almost at the revised 85% fractile curve,
, illustrating the conservatism associated with the preliminary LLNL results.
I-tJ With these preliminary revisions, it is clear that the LLNL study results are consistent with those of the Applicant, given the uncertainty in seismic hazard calculations. There are strong scientific justifications f or the revisions we describe; use of the original LLNL results to evaluate seismic J safety at _ Millstone is inappropriate, given the preliminary and conservative nature of the LLNL results. e a e l l
.-m~
V:Y_. _
+..-i @]f .
J Je% . , y e. , -
~
\up REFERENCES b Campbell', ' K.W. l (1981) . "A Ground Motion Model for the Central U.S. Based on
, -Near-source Acceleration Data," Proc. Conf. on Earthquakes and Earthquake
, Engineering:- The Eastern U.S., Vol. 1,'pp. 213-232.
Campbell, K.W., (1982). "A Preliminary Methodology for the Regional Zonation
'W ' of Peak Ground Acceleration," Proc. 3rd International Earthquake g =
,. Microsonation Conference, Seattle, Washington, Vol. I, pp. 365-376.
q l' LDewey, J.W. ' and Gordon, D.W. (1983). " Seismicity of the Eastern United States m ay a; maa - n , ,e _ ;and Adjacent Canada. 1925-1976." U.S. Geological Survey Professional p 1 Ebel, J .E. (1984) . "What can be learned about New England Seismicity from 7 years of. Widespread Monitoring?," Preprint accepted for publication in u Bull.'Seis. Soc. Am.', Aug. 1984.. '
. Gupta,1.N. , Nutt11, 0.W. (1976) . " Spatial Attenuation of Intensities for
"? Central U.S. , Bull. Seis. Soc. Am. , Vol. 66, pp. 743-751.
c . Lawrence Livermore National Laboratory (1984), " Seismic Hazard Characterization
. [L of the Eastern United Statest Methodology and Preliminary Results for Ten L, -l, ,
Sites," Preliminary Report dated rebruary 1984.
.Newmark, ;.b b. N.M.. and Hall, W.J. (1982) . " Earthquake Spectra and Design," Earth-
- gy
- quake Engineering Research Institute. Berkeley, California.
M . Nuttli, 0.W. (1979) . - "The Relation of Sustained Maximus Ground Acceleration 4 'and Velocity td Earthquake Intensity and Magnitude," State-of-the-Art for J Assessing Earthquake Hazards in the United States, Report 16, Misc. Paper o+ LS-73-1, U.S. Army Engineer Waterways Experiment Station, Vicksburg, jg; . Mississippi. i j~ , 1 -Nuttli, 0.W. (1983), Appendix C-A to LLNL (1984), pp. C-117 to C-123.
^7
+
Nutt11, O.W. and Herrmann,' R.B. (1978). '" Credible Earthquakes in the Central
= United States," State-of-the-Art for Assessing Earthquake Hazards in the United States, Report 12, Miscellaneous Paper S-73-1, U.S. Army Engineer l1 Waterways Experiment Station, Vicksburg, Mississippi.
J: l ,) Trifunac,iM.D. (1976). "A Note on the Range of Peak Amplitudes of Recorded L Accelerations, Velocities and Displacements with Respect to the Modified 1 Mercalli Intensity," Earthquake Notes, Vol. 47, No.'1, pp. 9-24. s Trifunac, Mihailo D. and John G. Anderson (1977), " Preliminary Empirical Models for Scaling Absolute Acceleration Spectra," University of Southern California, Dept. of Civil Engineering Report No. CE77-03.
1
.,7 REFERENCES (Concluded) ,
)I r. bl . Weiche rt , D.J. (1980). " Estimation of Earthquake Recurrence Parameters for b-l Unequal Observation Periods for Different Magnitudes," Bull. Seis. Soc. .U f Am. , Vol. 70, No. 4, pp.1337-1346. I Weston Geophysical Corp. (1982). " Estimation of Seismicity Parameters for New [ England," Report prepared for Yankee Atomic Electric Co. , YAEC-1331, ,. Appendix D1. h t t .. Y i e 4 1: t P O L F e
' . oo.*-*.
1 s TABLE 1 SEISMIC ZONES USED
/ LLNL SBTSMICITY EXPERf ZONES S
1 20,22
- .2 31,32 3 6, 7
{ 4 16,18,19,23 5 1 6 4 r. 7 15,24 l' [ 10 2,4,22 11 1,5
^
!? 3,16,17,18 13 10,11,12 s
3 4
- J e
5
h) TABLE 2 ATTENUATION EQUATIONS USED Alternative LLNL LLNL* Normalized Normalized
. No. Attenuation Equation Weight Weight Weight Acceleration:
G16 Trifunac (1977) + Gupta-Nutt11 0.25 0.291 0 D21 Nutt11 (1983) 0.24 0.280 .395 D13 Campbell (1982) 0.175 0.204 .288 G53 Weston (1983) 0.075 0.088 .123 D12 Campbell (1981) 0.063 0.073 .104 n D14 Natt11 (1979) 0.055 0.064 .090 Total 0.858 1.000 1.000 Spectral Velocity: i RS6 Trifunac-Anderson (1777) + Gupta-Nutt11 (1976) .263 .286 0 1 I RS3 Newmerk-Hall + Campbell (1982) (101) and Nutt11 (1983) .225 .244 .342 RS3 Newmark-Hall + Nutt11 (1983) .200 .217 .304 (110) R55 SEP-RS5 .088 .096 .134 R54 SEP-RS4 .075 .081 .114 R32 ATC + Nutt11 (1983) .070 .076 .106 Total .921 1.000 1.000
- Calculated by weighting all four attenuation experts equally 9
>~
c .- - - i - ,_ t' ic p f . q__ . . . , . . : .> _o.
' .-o
(- . , .,;
. . .a< .
c s - s b' s s g ,. TABLE-3
SUMMARY
OF' CHANGES
- TO LLNL HAZARD RESULTS (At;170 cm/sec2 Ground Acceleration and 10 cm/see Spectral Velocity)
.I. ,,a wt
,m EFFECT
~
Activity ah_ min Parameter Attenuation Rates 4.5 5.0 Combined
.75/.90 .72/.88
. Peak .66/ .86 -
.36/.68 Accelerationif'.75/.90 . .72/.88 -
.42/ .72 .23/.57
- Spectral. ;63/484 .62/.93 .76/ .90 -
.39/.70
- r. Velocity (9hz) .'63/.84 .82/.93 -
.64/ .85 .53/.66 d s ..
lR Spectral
.55/.80 .62/.84. 1.0 /1.0
.34/.67 L:
-Velocity (Shz). 55/.80
- n
.62/.84 -
1.0/1.0 .34/.67 c 4 1
- 3. .
V:
* - Effects shown,he: probability reduction f actor / ground motion reduction p . factor. Thus the'.first number is-the factor by which the probability should
-be multiplied (at'a given ground motion level) to represent the effect shown 7at the top.of the.. column. . The second number is the factor by which the ground, socion level shoald be multiplied (at a constant probability level)
', _ .to represent the effect shown at the top of the column.
1
,k ( -
N < 9 N <
,f "y g - ,
y,
-)
~
9 , - g<.,. .- l l i l ;$ " '
^_ '
1 .
. -a
~
.s..
s *
'?. -
(: , ..!' u l
'.~
80* 75* 70* 65* 60* l 1 I I t l l \ \ l t y- SO'
'~
f ~ - l \ \ Zone 22 I \ \
~
! y \ \
l \ ht Zone 20 y ~~~ \
- as-y 1 F~ -
,j L i r 7 l
l 1 1 .g ' \ r v 2-
) - 40*
)- d -
f f$ Y \ l } } Figure 1 Zones for Seismicity Expert 1 Damesa Moore . 1
I
.~
80* 75* 70* 65* 60* i i i l l l \ \
~
r
\ 3 _. 50*
1 ~~~ \ _\ \ \
~
I
\ \ ~
l zona
, s2 x \
l \ \ l \ \ I 45* (
}- ~f_ l r 4.
l? \ g
\
[' i \ N I I I
)
L Zone 311
~
#N 1
/
1 1 i I ,
! \
40* l ( ' t--l- - b __.-
- % ~~~
lW t 5 a ~/
\
1 l
~
I f \ \ Figure 2 Zones for Seismicity Expert 2 Dames & Moore
80* 75* 70* SS* 60* l i ! l \ l \ \ t , 50*
~
l I k - T,
~
'T'- l T \
l \ \ l ,
\ \
l f \ pp i
*~
~
\
\
, Zone 6
/ ) T Zone 7
\
$ / {
t , 78- - r I,YI _l/ l lf 1 Y \ A \
- I \ \ \
Figure 3 Zones for Seismicity Expert 3 Damesa Moore i--mmmmum m
,. u
.. 'i
=
3' 1
~
_b? I 80* 75* 70* 65* 60* E1 l
- 1 I i l \ q l
r
\ \
- 50*
k t n
~-T___ '
~\ \ .
l \ \ 5 Zone 19 ~ f \ Z
\ \
If ono 16 N x
/y \
\
45 3 J, - M D,W-
\ 5
- l. j p Zone 18 j
~
f ,
~
I
. 1 D,,, k, '
Zone 231 2 dO' (1 -1 - c ,
-fw _ s _ _
m - 1 ; != I, a'
/ '
\
\ ?
R 4 5 b % \ \ \ k
- T-d
_ n n
= 5
'E
!i Figure 4 Zones for Seismicity Expert 4
$ =l
- i!
oamC3& Moore . , W
80* 75* 70* 65* 60* 1 i i l \ l \ \ _ t i 5- 50*
'- ,-\ .
\
. l \ \
l , \ \ l ,
\
pl l MY\4
- 45' v f f\ \
Zone 1
. t 3 1
- 1 I i
? \
\ - 40' r )
~%' l
{
/
MI9
, \
_l I I I L Figure 5 Zones for Seismicity Expert 5 Dames & Moore <
.x __
~ *
- 80* 75* 70* 65* 60*
i l l \ \ l \ \
- 50*
{ -
, ; \ \ ~
i \ \
! \ \
y \ \
\
\ ,- 45*
2 i y r
/ ,
n N 4 1
,s( l - . i D Mg,l Zone 4 l 1
g ,
, 1 - to'
- % ~~~
_-l- _ _ l_ l t (* / I \ 1 l 4 \ _I %i i 1 L Figure 6 Zones for Seismicity Expert 6 D.ames & Moore
~
P *
.. 80* . 75* 70* 65* 60*
1 1 ,1 I i l \ \ { ,y- 50*
- ---- -- ,_ a-
~
; \ \
l \ \ l \m'
\ \
\ \
I
- ts*
\
n ' J Zone 24k-s, t y> NZone 15 do"
- L ____ -
1 -
-{ , ~ ~ -
l $"k I l
- I l \ \
Figure 7 Zones for Seismicity Expert 7 Qames& Moore , n
.- so- 'e 65* 00
- t ,
, ; ) \
< i
-. ; \
r > i
~
r
, 45' l
?
l ) Zone 22 Zone 2
,( ~
l ( go.
-t -
T
-}- s t
- Zone 4,
~
,ft i
3 t -- 1
- I~ l ' \
\
J ,
\
t
\- -- ^
- d g
\ \
l 1 ' \
\
Figure 8 Zones for Seismicity Expert 10. omm aMoom
E ~ r
,- 80* TS* 70* 65* 80* _
I I \ \ \ l \ \ so-1_ - + -
-r-- , i g i l l i \
\
. \
/ \
; e 9 t .
1' - 45* t
- I -
-{
l ?"
/ W \
,7 Zone 1 }
- p- If! 3 j a 1 k i 40*
J . ( " l
~~-
W Ws /
/
1 f Zone 5 l l l
- I I I i l
Figure 9 Zones for Seismicity Expert 11 Dames & Moore
, -,,.e...-, ,.--,._...,.,---...-,,~,_.,-.r.,-.. . _ _ , . . - . , . - . . . . , -
..,-------__m, , , - .
.- ' 4 . F'. 80* rs' ro* es* 's o
- 1 l .
I Zone ta ( 7 Zone 17 i y I 1 j a , f i i
- I p ~
- g. 45*
J
.- I ,-
~
f ' Zone 16- ' [ p V l t - _ I
~ ) 4o*
I~ a . - I i \ one Z 3 4-I , I f
"^
1 !/ =' l e _v .s. 2 I l s
- J
[ [l l l i I I
' 1
.~
.0* l I% ~ .
1 l f l ' l i I 1 l Figure 10 { Zones for Seismicity Expert 12 l Dames & Moore .
E
,.s i
80* 75* 70* p5e goe N i i i g g l \ \
'- I - 50*
~
-- j _ - _ .
l \ \ ( Zone 12 1
- 1 \ \
l , , \ \ 1
/ Zone 11 \
- 4s*
- - ~- r
^
j \ .,- { 4- - l ' Zone 10 j
& )
b y ( h y W V h e , ,
~~~ ~~
L __ - _ -
/ \
1 l l
- I \ \
\
\
Figure 11 Zones for Seismicity Expert 13 L Dames & Moore !
,,----.e,, -ww.--e -e*, ,e-rv ris i'we- e---eweSt- w' vwwa' *eee t---ew-wm e w-w t-e--~m=+-*------e---'--e--r----- ----- - - - -
l r .S ~
-2 10 - 6 . . i e 4 e 6 i e a . .
. THIS STUDY- .
AVERAGE OF
. . *BEST ESTIMATES * .
-3
~
10 - C LLNL - BEST ESTIMATE g . .
$ THIS STUDY-g -
MEDI AN (BASE CASE) - o x W -4 u.10 -
, O^ - -
~
b a g . . e l 0 . . 4 E a . . 3 2 - 4 10 10 ' , t e e , , , , . . O e o o o oo o o o e o o eoo P-
- es o o o o o ooo
- y e o e eoo ~
PEAK ACCELERATION (cm/sec2 ) Figure 12 Hazard Curves for Median and Average of "Best Estimates" ~ Damesa Moore .,
......,g
*.o *.
a, . G
.. I e
E as a y E e o O r # e ie a 4 4 i gie e a i . . . o co
. c e
A E r M
@~nnnw o
~ ~etenee _ g Io.
. 000000 , , C
- ts C
(0 o.
. to a ll
. E .a
.w V.
g
~ W 6 O O u z
< 03
.j' > G gg
. O E
~
a ~
,, , ( 03 E W
.. p z g o=Q
- ~
~
g e-enuw L "
-n-e-- - W G "g"
(000000) o
. O o
~
4 .
- co a w F E -
e w
- s
- CD g
i , il , . , , t, , , , , , , , e o o
- o. o o, o e (goes/mo) NOl1YH373DOY m
w
=
l' o o lE es a
- o E
a o O
.6 4 i i , i i . , i i . ,- . i o
- c)
- e. / _
- ", /
o. E V
^ C r ..- A e ;- -o o 2" -
o O eMMe,n mmmmme d _ g cccccc
.o E
~
w
- O n
E y a _ m .E o _ z e
. < W F- m T- .m N O .C
. c)
J v
< h CC
> r z o o,w O o ,
E >
- w m
g
. O o
O.
, - (O l , _
v
/ ES 2:
enne,~ e w mmmmmm 3
/ cccccc
\ /
a v LL C e e a a,, , , . . !, , , , , ,E , O o , 1 o (088/Wo) A. LIDO 73 A h 7
m .
.,e a... .*.
, e- .!I
- Ci 3
es e O E
=
O O! l 4 e e 6 6 4 6 g6 6 / 4 4 6 6 6 O r; / C3 e - 3 a C. E q)
- w. g f mm h '
C _ 53 0 8
~ ~ ::
w es e=s e w tu
~
C
'- O.
. mmmmmm o cememe -
11 n
- E
- \ .
m E
.ac w
8
*=
e
.e
* - m e o E z =#
. F. tu os m I. '
. 5 N
.c
- m
. <: v E >
>=
Cy "g 3 O Q. e
- m >
. E h
. o
, e n
/.
f nn E3 EEEEEE - C u yx e c c e e, w
~
e e 9 - a co m g
/
m.. o E
]I- i
, . . . . . / ,,i , , , , , , , ,
O o " o , v-(DG3/WD) A.LIOO75A e
F f .. , l... *
- W
. t
. k:
-2 10 , , , , , , , , , , , , _ g
" ~
BASE C ASE -
-3 !
10 - - s
- : 2
. . =
. . 2 j
W 3 0 " ~
=
Z
< ALTERNATE # d Q ,
ATTENUATION WElGH TS ,
- W O --
M _- I 1
$ 0"4 7 g
~
> . : 3 t . . :
y . - y i, m _3 O . . C " i D. a . - g
- 5
- 10-5 ;
. . ?
f
. . S i
3 10 i t i t i ' ' t ? ? ' 5 C C C C C C C C C C C C 1 e C k O C C C C C C C C y w
- N c3 y c c P. co C3 C ~
2 PEAK ACCELERATICN (cm/sec ) q 1 Figure 16 ; Median Acceleration Hazard Curves, h Base Case And Results Without Trifunac Attenuation " Dames & Moore 3
=
s 3
- m. _
\
r... r... = _? r- e' . s E 3 re s 1.- 5 10 . . 6 a a e a 4 . 4 . F . v . . BASE CASE - r . ALTERNATE f = 5hz . ATTENUATION L-
. WEIGHTS .
i
- s. -
-3 10 _
~
h BASE CASE
~ -
- 4 is9hz
~
O ALTERNATE
~
U -4 ATTENUATION WEIGHTS x 10
'..-- , ha . .
, g .
o -
~
~
P > 5= . y -
= . .
'.. e e,
.J 4
-5 10
~
; 10 1 I I 8 I I I I I i 1 2 3 4 5 6 8 10 15 20 30
.; SPECTRAL VELOCITY (cm/sec) Figure 17 Median Spectral Velocity)(9 (5% Damping and Cases
, Base Shz) Hazard and Curves Results Without Trifunac Attenuation b
u Dames & Moore ;
.b
r .* l a SO E i 4 1
- a . -
1 : :
\
\ .
I' 10 - h
* \ EXPERT 2's SEISMICITY -
\ DESCRIPTION r
l . w i _
- u. -
ALTERNATE SEISMICITY O : DESCRIPTION e . g . s . 3 z . 3 _ . E
; - 6 -
5 m o\
\
E O.1 - x - o .
\ :
i : \ : 3 z . z .
< =
.., ., \ .
O.01 - X HISTORICAL DATA USING NUTTLt CONVERSION Y
\\
a : 2 O HISTORICAL D ATA USING - WESTON CONVERSION - O INSTRUMENTAL DATA (77-82) - I
~
T MEAN + J (INST. DATA ONLY)
- l. ME AN - 0 (INST. DATA CNLY) -
0.001 1.0 2.0 3.0 4.0 5.C 6.0
*b Figure 18 Seismicity Data, Zone 31 of Expert 2 pamesa uoors
7 s.. .. 50 - e i i ' i i - k\ . b - 10 - -
- \
EXPERT 3's SE!SMICITY - f \
". DESCRIPTION
[
. \ .
iiY
~ ~
2 W j _ ()\
- u. -
0 : : m .
\ _
y . -
\. -
2 - ALTERNATE SELSMICITY q _ 3 DESCRIPTION Z "
) -
r* --
'g -
5 N) e 3 h O.1 o
\
- \
a . 3 . . 2 -
. Key -
X HISTORICAL D ATA USING \ NUTTLI CONVERSION O.01 - V -
- O HISTORICAL D ATA USING :
WESTON CCNVERSION \ - g . O INSTRUMENTAL D ATA(77-82) g T MEAN + 6 (INST. DATA ONLY)
.l. ME AN - 6 (INST. D ATA ONLY) 10 2.0 3.0 4.0 5.0 6.0
*b Figure 19 Seismicity Data, Zone 7 of Expert 3
- ) Damesa Moore ;
j ___
(. . - .. e* e 50 . s e . s s s . . 10 -
\
I ~ g .. . EXPERT 6's SEISMICITY
, , DESCRIPTION .
. N U W q O g . .
O ~
~
.. c: .
W .
. 2 .
3 2 . O - - e- m 2 . ALTERNATE
.i y SEISMICITY DESCRIPTION a
- _if s 0.1 .- .. : 7 1 o .
g .
~-
n . 2 . 2 - Key , D X HISTORICAL DATA USING NUTTLI CONVERSION 0.01 - O HISTORICAL DATA USING T
. WESTON CONVERSION X
~
. O INSTRUMENTAL DATAC77-82) .
T MEAN + & (INST. DATA ONLY) ~
~
- 1. MEAN - & (INST. DATA CNLY) 0.001 ' '
1.0 2.0 3.0 4.0 5.0 6.0
*b Figure 20 i
Seismicity Data, Zone 4 Of Expert 6 Dames & Moore .
----o
c'.' .= .
- e'.
-2 10 *
- i
- . a a e i e i i ..
=
=
~
BASE CASE
-3 10 --
~
c w - w o 5 Al.T E RN AT E
^ g10 - ACTIVITY R ATES
- I 1 1
t:: - N . 4 m ' C .
,v a -
" 10 I -
~
~
~
~
~
-6 10 e i e , , , , , , , ,
C C O O o e o e o ooo
- O P- C o o o o oe. ooo
" a <* e v e e eoo 2
PEAK ACCELERATION (cm/sec ) Figure 21 Median Acceleration Hazard Curves, Base Case and Alternate Activity Rates j - i Damesa Moore : l l
m
]
. . = - ' ,
en
-2 to . i 4
. 6 i a i e i i .
f 'o-3 i - BASE CASE -
@ f u 9hz ALTERNATE -
0 ACTIVITY R ATES e s 10 - c : : g . .
. 2 -
BASE CASE - f =5hz - m ALTERNATE l ACTIVIT Y RATES
- 1. a.
- i
$ 10 _
5 . : I . to i I I I I I I I I i 1 2 3 4 5 6 8 10 15 20 30 SPECTRAL VELOCITY (cm/sec) i I i Figure 22 Median Spectral Velocity (9 and Shz) Hazard Curves (5% Damping), Base Cases and Alternate Activity Rates Dames & Moore :
.'s 4 .. . *s r 'e ,
10 , E BASE CASE -
-3 *b. min
- d'0 10 7 7 y - mb. mins 5.0 -
e c w - . w o x 10 - O -
> . l 2 - .
,. =d . .
m e O - - g G. a - . B 2 _ " 10 - T
.s -
10 , i t , , , , , . , , o o o oo o o o o o e o e e-e ce o n o w o e o o o. 8oo r aao oo 2 PEAK ACCELERATION (cm/sec ) Figure 23 Median Acceleration Hazard Curves, Base Case, mb, min = 4.5 and 5.0 l ( Qames& Moore
e .. - i 4* l l
~2 to . i i i e i 6 i , , .
. DASE CASE :
[ m b. min = 4.5 f = 5h2
. e m b. min = 5.0 .
3 10 _
~ '
BASE CASE
, f = 9hz m
- b. min = 4.5 ,
-
- b, min = 5.0 ,
k - 8 U C' '.4 : - 3 : :
~
y - s . E m O ~
- 4. k ~
T --
~
~
10 l I I I I I I t I i 1 2 3 4 5 6 8 to 1s 20 30 SPECTRAL VELOCITY (cm/sec) l I Figure 24 Median Spectral Velocity (9 and Shz) Hazard Curves (5% Damping), Base Cases, mb, min =4.5 and 5.0 Dames & *Acore I I
f':; - 3 eq- *
-.(
10
~'
I O w 10
-3 \\ LLNL-SHCP MEDIAN W -
~O - LLNL-SHCP REVISED X
g BY APPLICANT (USING m b s 4.5) O LLNN-SHCP REVISED BY p APPLICANT (USING m b
- 5.0)
O 10~4 -
.T apptican7 r g (MEDIAN PSS)
,, O -
m r- E !
~
10 i N xN .
~
NxN
~N 10 ' ' ' ' ' ' ' ' ' ' '
50 100 500 1000 2 PEAK ACCELERATION (cm/sec ) Figure 25 Acceleration Hazard Curves l Dames & Moore ( i
I [.,N ']I[,: Tl [. d c 4-* _
~
-2 F i 10 >
s
. K y/~0HCP 85% Fractile W s SHCP 85% Fractile Revised x g ^
2 I SHCP Median
, tu 10 -
, N N,,kg SHCP Median Revised L' N N \
N. N
, \ *
. ' 10 - \g N i
h , N \ s NK N
< s.
s \ \ z 10' ~ ' \ z - N \
< SHCP 18% Fractile A \ \
\ \ '\
- SHCP 15% Fre':tlic '\ \ \
Revised \g \ \ g 10
-6 I
\ f ' ' ' ' '
50 100 500 1000 - PGAK ACCEL ERATiON (cm/sec2) ( - Figure 26
~
- 4. .
Acceleration Hazard Curves 1 i G O g m Dames & Moore
-mme .mA --}}