ML13330A098
| ML13330A098 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 08/21/1980 |
| From: | Baskin K Southern California Edison Co |
| To: | Crutchfield D Office of Nuclear Reactor Regulation |
| References | |
| TASK-03-06, TASK-3-6, TASK-RR NUDOCS 8009020063 | |
| Download: ML13330A098 (14) | |
Text
Southern California Edison Company P. 0.
BOX 800 2244 WALNUT GROVE AVENUE ROSEMEAD. CALIFORNIA 91770 K. P. BASKIN MANAGER, NUCLEAR ENGINEERING AND LICENSING August 21, 1980 Director of Nuclear Reactor Regulation Attention: Mr. D. M. Crutchfield, Chief Operating Reactors Branch No. 5 Division of Licensing U.S. Nuclear Regulatory Commission Washington, D. C. 20555 Gentlemen:
Subject:
Docket No. 50-206 Site Specific Earthquake Program San Onofre Nuclear Generating Station, Unit 1 By letter from A. Schwencer, dated May 23, 1980, the NRC requested a comparative evaluation of the results of the Site Specific Earthquake Program done for San Onofre Unit 1 and the empirical evaluations done for San Onofre Units 2 and 3.
This request was provided as Question 361.65 on the Units 2 and 3 Final Safety Analysis Report. SCE's response to this question was provided on the Units 2 and 3 docket on July 1, 1980 and August 4, 1980.
Mr. Schwencer's May 23 letter also requested that the response to Question 361.65 be included on the San Onofre Unit 1 docket.
Therefore, the response to this question is being provided as an enclosure to this letter.
If you have any questions on this information please let me know.
Very truly yours, Enclosure 9 009 0 2-0J06
Responses to NRC Questions San Onofre 2&3 Question 361.65 The estimation of free field earthquake ground motion to be used in the reevaluation of SONGS 1 is based upon deterministic modeling of such motion from an earthquake occurring on the Offshore Zone of Deformation 8 km west of the plant. The estimation of free field earthquake ground motion from the same zone presented in the SONGS 2 and 3 Operating License review is based upon a largely empirical evaluation.
Since all three of these units are at the same location and are subject to the same seismic hazard a com parative evaluation of results from both methodologies is warranted.
You are requested to prepare this evaluation including sufficient discus sion and figures so that the staff may fully understand the "similarities" and differences in the results. Specific attention should be paid to include the following items.
- a. Comparative plots of predicted spectra at similar levels of damping.
- b. A comparison of the similarities and differences of these spectra at different frequency bands and a discussion of the sources of these differences and their significance with respect to uncertain ties in the prediction -of ground motion and validity of the dif ferent approaches taken.
- c. A discussion and listing of the conservatisms existing in each approach and what effect these have upon the relative differences and similarities in the resultant spectra.
- d. A discussion of the rationale as to why different approaches were used at SONGS 1 and SONGS 2 and 3.
- e. A comparison of the basic geological and seismological assumptions utilized in each approach.
Response
361.65a Introduction (Q 4)
Estimates of free-field earthquake-induced ground motions at the San Onofre Nuclear Generating Station (SONGS) Units 1, 2, and 3 site have been made using two different approaches. These approaches are the "source modeling" approach and the "empirical" approach. Analyses using the source modeling approach for SONGS Unit 1 were initiated in 1975 and are documented in the May 1978 report "Simulations of Earthquake Ground Motion for San Onofre Nuclear Generating Station, Unit 1" and the July 1979 report "Simulation of Earthquake Ground Motions for San Onofre Nuclear Generating Station Unit 1 -
Supplement 1" by Del Mar Technical Associates (TERA/DELTA).
Analyses using the empirical approach for SONGS Units 2 and 3 were initiated in 1978, and are documented in the June 1979 'report "Report of the Evalua tion of Maximum Earthquake and Site Ground Motion Parameters Associated with 7/80 Q&R 2.5-Amendment 19
Responses to NRC Questions San Onofre 2&3 the Offshore Zone of Deformation, San Onofre Nuclear Generating Station" by Woodward-Clyde Consultants (WCC) (also in the response to NRC Ques tion 361.54).
Both approaches provide estimates for free-field instrumen tal ground motion parameters at the SONGS site due to a postulated major earthquake of about M 7 on the Offshore Zone of Deformation located 8 km west of the SONGS site.(a)
In brief, the "source-modeling" approach is a computer method of rigorously incorporating rupture physics and wave mechanics into a procedure for modeling strong ground motions. Recorded past earthquakes were used to provide constraints on the mechanics of the model which was then applied to the existing seismologic and geologic conditions at the SONGS site.
The "empirical" approach is a method of incorporating recorded historical data in a statistical model for estimating ground motion. A careful selection of earthquake ground-motion data was made for Ms 6-1/2 recorded at sites that reasonably represent the SONGS site conditions.
These data were regressed with distance to develop response spectra at a closest distance of 8 km and extrapolated to Ms 7 using published relationships and data on the effects of magnitude on ground-motion parameters.
Use of the two approaches to estimate free-field ground motions was moti vated by a need to demonstrate the conservatism of the SONGS Units 2 and 3 design response spectra more expeditiously than the pace of the NRC and NRC consultants' review of the source modeling work which had been initiated for the purpose of seismic reevaluation of SONGS Unit 1. In Ques tion 361.33, which was transmitted by NRC letter dated March 10, 1978, the NRC requested that the Applicants provide on the San Onofre 2 and 3 docket an assessment of the "representative ground motion to be expected at the site" from "the largest potential earthquake which can be associated with the 'Hypothesized Offshore Zone of Deformation'."
Further, the NRC requested that this assessment be performed independently of the NRC/USGS geosciences licensing decision (construction-permit stage), which had been based on extensive consideration by the NRC and the USGS personnel and on extensive investigations by the Applicants and their consultants.
In evaluating the methodology to be utilized in assessing instrumental ground-motion at the site, the Applicants considered the ground motion modeling work which was being performed for and reviewed by the NRC on the
- a.
The estimate of free field instrumental ground motions has been made for SONGS Unit 1 is based on the existence of the OZD as postulated by the NRC and USGS, notwithstanding the Applicants' position (with respect to SONGS Unit 1) that the geologic evidence indicates that:
(1) the hypothesized OZD is made up of short discontinuous segments, and (2) structures capable of producing large earthquakes are distant and capable of producing less than 1/4 g at the SONGS site. For the limited purpose of licensing SONGS Units 2 and 3, during the CONSTRUCTION PERMIT PROCEEDINGS, Applicants stipulated to the geologic model, as postulated by the NRC and USGS.
7/80 Q&R 2.5-Amendment 19
Responses to NRC Questions San Onofre 2&3 San Onofre Unit 1 docket.
The Applicants considered at that time, and continue to consider, the results of this modeling study to constitute an appropriate basis for evaluating the design spectrum for licensing of either or both projects.
At the time the Applicants selected a methodol ogy for assessing instrumental ground motion to support the preparation of a response to the NRC Question 361.33, it was clear that the pace of the NRC and NRC consultants' review of the modeling study was too slow to sup port the licensing needs of San Onofre Units 2 and 3. Therefore, the Applicants elected to base the response to Question 361.33 on the empirical approach (since performed by Woodward-Clyde Consultants, reported in June 1979) independent of the San Onofre Unit 1 modeling study. The studies were carried out separately and independently, and were cross reviewed to ensure high technical quality. The assumptions for each study were made independently, and incorporated mutually exclusive data sets.
The subsections that follow address the five parts of the NRC Ques tion 361.65.
For ease of reference, the title of each subsection is fol lowed by the number of the part of Question 361.65 being addressed.
361.65b Assumptions, Bases, and Conservatisms (Q 2, 3, 5)
The general assumptions and the bases for the two approaches are discussed in this section, according to the following considerations:
geology and earth structure faulting process magnitude stress drop proximity (distance/attenuation) focusing judgment and experience Some of the more important points are summarized in table 361.65-1.
The paragraphs that follow highlight the information summarized in that table.
GEOLOGY AND EARTH STRUCTURE Source-Modeling Approach The properties used in the analysis were determined by seismic surveys, as documented in the TERA/DELTA report of July 1979.
7/80 Q&R 2.5-Amendment 19
K*.
Responses to NRC Questions San Onofre 2&3 Empirical Approach The data set used consisted of recordings obtained from sites having local geology generally similar to that of SONGS (i.e., from sites having deep, very stiff sedimentary soils).
FAULTING PROCESSES Source-Modeling Approach The earthquake rupture was modeled as a shear crack that spreads from the hypocenter over a specified zone of rupture. Random processes are included in the source modeling to approximate perturbations or irregularities in actual earthquake rupture. The results of repetitious simulations were used in developing the spectra presented in the July 1979 TERA/DELTA report.
Empirical Approach The data set included records from thrust and strike slip type earthquakes.
Most of the records were from the 1971 San Fernando earthquake that occur red on a thrust fault.
MAGNITUDE Source-Modeling Approach The seismic moment was selected to yield a magnitude Ms 7, i.e., the hypo thesized earthquake has a 130 centimeters of offset over a length of 40 kilometers along the OZD.
Empirical Approach Question 361..54 directed that the calculated Ms 6-1/2 ground motions be extrapolated to an Ms 7 earthquake. The extrapolation was made in a rea sonably conservative manner to an Ms 7 earthquake as described in the response to Question 361.54.
STRESS DROP Source-Modeling Approach The stress drop was extracted from fault offset, fault configuration, and earth structure, and was estimated to be 100 bars. This is conservative when compared to the low stress drop (less than 30 bars) for the 1933 Long Beach earthquake which occurred on the Newport-Inglewood Zone of Deformation comprising the north end of the hypothesized OZD.
7/80 Q&R 2.5-Amendment 19
Responses to NRC Questions San Onofre 2&3 Empirical Approach The stress drop during the 1971 San Fernando earthquake, which strongly influences the data set used for the empirical analysis, is estimated to have been about 200 bars. This value significantly exceeds the stress drop of less than 30 bars estimated for the 1933 Long Beach event.
There fore, stress drop was conservatively incorporated in the analysis.
PROXIMITY Source-Modeling Approach The attenuation of ground motions with distance was determined by simulat ing wave propagation from the rupture surface to the site. Located 8 kilo meters from the closest point on the fault, using site specific earth properties.
Empirical Approach The data used in the analysis were recorded at distances greater than the distances of SONGS from the OZD. The extrapolation of the results to the design distance is considered conservative. Furthermore, it was shown in the response to Question 361.62 that the definition of distance used in the analysis was conservative.
FOCUSING Source-Modeling Approach From the different rupture configurations tested, the extreme configuration that were investigated (focusing) was used in the analysis of instrumental ground motion for the SONGS site.
Empirical Approach The response to Question 361.56 addressed the focusing effect.
It was shown in that response that the strong-motion data used in the empirical analysis were recorded under conditions of focusing that would tend to increase measured ground motions.
Therefore, focusing effects were con servatively incorporated in the analysis.
JUDGMENT AND EXPERIENCE Source-Modeling Approach The pervasive assumption upon which the earthquake model is based is that much of the scatter in near-field recordings is due to processes that can 7/80 Q&R 2.5-Amendment 19
Responses to NRC Questions San Onofre 2&3 be adequately modeled deterministically with stochastic perturbations.
Particular assumptions or mechanisms used in the model are summarized in table 361.65-1.
To apply the source-modeling approach, it is necessary to set several para meters, some of which are not well known from crack mechanics.
Other than the gross characteristics of an earthquake rupture (hypocentral locations, rupture extent and static stress drop), the dynamic stress drop is the only parameter that has been calibrated directly from strong-motion data.
From studies of past earthquakes it appears that dynamic stress drop (initial slip velocity) is independent-of magnitude and fault type, probably because this parameter is related to the fracture strength of earth constituents in the gouge zone of faults. Parameters such as rise time, rupture velocity, and random irregularities are assigned values based on generic knowledge of earthquakes and have been validated by modeling particular earthquake recordings.
Based on test calculations previously done and currently being performed to simulate past recorded earthquakes, the source modeling approach provides a reasonable and rational means for estimating ground motions at the SONGS site, taking into account the specific geologic and tectonic conditions.
Results generated by the earthquake model are expected to be reliable estimates of hypothesized ground motions at the sites because:
the model has been calibrated and validated against near-field recordings of earth quakes in the same distance range that is relevant for SONGS; the model provides a rational basis for extrapolating from past earthquakes to hypo thesized conditions at SONGS, including site-specific magnitude, distance, and earth structure; the model provides a basis for appraising the likeli hood of unusual combinations of fault rupture and waveguide effects that could cause large-amplitude shaking at SONGS; and the model includes sev eral margins of conservatism to guard against underestimating ground shak ing at SONGS, as presented in table 361.65-1.
Empirical Approach In the WCC June 1979 report, it was shown that the site-specific analysis results were conservative with respect to the results from similar analyses of a larger data set comprised of soil-site recordings.
It can also be seen in the WCC June 1979 report (figures 8 and 9) that the results of the site-specific regression analyses are conservative with respect to the data at the close distances used in that analysis.
In the response to Question 361.55 the results of the June 1979 analysis were compared with the data obtained from the October 15, 1979 Imperial Valley earthquake.
It was shown therein that the ground motions predicted from the referenced analysis were conservative with respect to the record ings obtained during the 1979 Imperial Valley earthquake.
7/80 Q&R 2.5-Amendment 19
Responses to NRC Questions San Onofre 2&3 361.65c Results and Discussions (Q 1, 2, and 3)
The mean response spectra developed from the source modeling and empirical approaches are presented in figures 361.65-1, -2, -3, and -4 for damping values of 2%, 5%, 7%, and 10%, respectively. As noted in these figures, mean instrumental peak acceleration (Zero Period Acceleration) values from the source-modeling approach is 0.33g and from the empirical approach is 0.47 g.
The response spectra designated "source modeling" in figures 361.65-1, -2,
-3, and -4 are based on the results documented in the July 1979 TERA/DELTA report. The response spectra designated "empirical" in figures 361.65-1,
-2, -3, and -4 are based on the results documented in the WCC June 1979 report and in the response to NRC question 361.54. The June 1979 report documents development of the instrumental response spectrum for Ms 6-1/2 for 2% damping. Development of the instrumental response spectrum for Ms 7 for 2% damping was done by extrapolating the associated Ms 6-1/2 spectrum using the procedure documented in the response to NRC ques tion 361.54. The procedure described in that response was also applied to Ms 6-1/2 response spectra for dampings of 5%, 7%, 'and 10% to develop the Ms 7 response spectra shown in figures 361.65-2, -3, and -4.
As expected from the evaluation of the basic features of the two approaches, the empirically derived response spectra exceed the corresponding source modeling response spectra over the range of periods considered (0 to 2 sec onds).
From the nature of their deriviations, the results of both approaches are conservative, with the levels of conservatism varying with period. None of these spectra would, however, represent a design spectrum, which, based on several physical considerations, would lie below the instrumental spectrum.
REFERENCES FSAR section 2.5.
Responses to NRC Questions 361.33, 54, 55, 56 and 62.
No FSAR change was made.
- 1. Del Mar Technical Associates and TERA Corporation, Simulation of Earthquake Ground Motions for San Onofre Nuclear Generating Station, Unit 1, submitted to Southern California Edison Company, May 1978.
- 2. Del Mar Technical Associates and TERA Corporation, Simulation of Earthquake Ground Motions for San Onofre Nuclear Generating Station, Unit 1 -
Supplement 1, submitted to Southern California Edison Company, July 1979.
- 3. Kanamori, H. and Anderson, D. L., "Theoretical Basis of Some Empiri cal Relations in Seismology," Bulletin of the Seismological Society of America, Vol. 65, pp. 1073-1095, 1975.
7/80 Q&R 2.5-Amendment 19
Responses to NRC Questions San Onofre 2&3
- 4. Seed, H. B., Murarka, R., Lysmer, J., and Idriss, I. M., "Relation ships of Maximum Acce.lerations, Maximum Velocity, Distance from Source and Local Site Conditions for Moderately Strong Earth quakes," Bulletin of the Seismological Society of America, Vol. 66, No. 4, pp. 1323-1342, August 1976.
- 5. Seed, H. B., Ugas, C., and Lysmer, J., "Site-dependent Spectra for Earthquake-Resistant Design," Bulletin of the Seismological Society of America, Vol. 66, No. 1, pp. 221-243, February 1.976.
- 6. Woodward-Clyde Consultants, Report of the Evaluation of Maximum Earthquake and Site Ground Motion Parameters Associated with the Offshore Zone of Deformation, San Onofre Nuclear Generating Station, prepared for Southern California Edison Company, June 1979.
7/80 Q&R 2.5-32 Amendment 19
FREQUENCY, cps c
so 0110' a 03 CD oc x
x 4 2 tgo7 a
S ONFR RESPNSE PECRA FR DMPIN
=
.0 V,
V%
7 to Figure Source 7/8Modeling A
d t
1XI X-J 4,4 S/N ONOFR NULAXEEATN TTO COMARSONOFTH MNFEANISRMNA RESPONSE SPECTRA FOR DAMPING =0.02 Figure 361-.65-1 7/80 Amendment 19
FREOUENCY, cps to a
to s
I I 5
0 5
C) 0-? 015 0
1o 00 0
943DV no 00 100 400 30c So Source SS 4
4 II X
8*
8a
.8x 3
\\>
X 0
.0 0
a 06 4
a 00 PERIOD, sec SAN ONOFRE NUCLEAR GENERATING STATION COMPARISON OF THE MEAN INSTRUMhENTAL RESPONSE SPECTRA FOR DAMPING =0.05
---Figure 361.65-2 7/80 Amendment 19
- g 0
FREQUENCY, cps c
AD D
0 to a 10 1
1
's C 7 0 t 03 02 0 *
'000 r
.oc 7 ' 4
)
odelin o
/V AA aDamp Mgd=i0.07 S%
Ar ~A
.0 4;03 04.8D
&4
- 4.
8 3*
- 81 PERIOD, sec SAN ONOFRE
,NUCLEAR GENERATING STATION Units 2 & 3 COMPARISON OF THE MEAN INSTRU 3ENTAL RESPONSE SPECTRA FOR DAMPING = 0.07 Figure 361.65-3 7/80 Amendment 19
FREQUENCY, cps OD
?.
sc PL 0
IV 03
- 2.
C V~y~~r 2
izjO%
~
o 44, v44 se r
0
> -7r
'<<:A 4C
-7
>Y>.
~
/
Empirical 10 0
Z4
-V~~~-
Modeling 2
L AA 1Y1
.0' 01 3
04 06 0'
.5
.0 S
2
~
6N RESPONSEmpin SPCRAFRDAPN1001 Fiue 6.6 7/80.02.end
.6.ent3
TABLE 361.65-1 Bammary of Assumptions, Bases, and Conservatisms for Source Modeling and Eapirical Approach"e EMPIJRI CAL APPROACH PATAM RS Assu ptions "asis Corne-vatian and Val2dity Assumptions Basis Cor.srvatism and Valiv Vertically stratified,
-ieldmeasurements Actual properties wre Grouede motgo e
be.vations durin 9 ata set ileed coitod aul Pr cess s rt ica s ri e-lip Generic a
rthquake Resut slowca past eartr iq.eokes tT.go, of reeording obt e
rupro eaga tin s ad eaemlations simu lti os e re usede edh Muraka Lya er froo teshern U.S. i pee f
s s
matn f
p attenuation adriss ef197fc.
and hardimng tocay geo through the idealized e
q uas used than for Seed. Uges, Lyser generally aimi o
a representation of the Imperial valley model-
[1976o)
eartr.
mne.
Fault Processes Vertical, strike-slip Generic earthquake Results of repetitious rupture that spreads studies and rilations simulations were used earthqeakes from both from the hypocenter at of recorded ground t o determine t h o
strike slip and thrust a speed of slightly Local mast effects of random tyne dauancego less than the shear-earthquakes using the processes that are tp alig e a v e v A 1 o c t y m
sst lete rescip-inclided in the source Irregularities in the tiork of rt-pt-ura process modeling to approximate rupture process are currently a agable.
perturbations or suitauly odeled by irre()larities in Grouied motions random perturba-actual. earthquagkne reasonably cons tiona.
r-upture.
tively extrapolated o et 7 qarthqua Magnitude The size of rupture and (K
and n character-Seismic moment and t
For the SONGS site Available data (response to 0 361 tultuoffset are Ize fault capaba9lity.
are at or eater the seichwhichthe estniate directly related to L ca magnitude the upper limits of Cant distance of seismic moment (MOb (ML) is determined capability for the 8 km from OZDy the anMd surface-wave from computed ground 07..
ground m ot iocn is magnitude (Me).
- notion, influenced by the magnitude of the earthquake.
Stress Drop Stress drop was ex-The average stress drop Stress drop used (100 Ground mot ion is Effects of high stress Results were strongt tracted from fault for inter-plate earth-bars) is greaterthan influenced by stress drop were Incorporited influenced by thera
- Offset, fault config-quakes is about 30 bars what is reasonably drop.
in the selected ata Fernando earthquake f uration and earth (Kan ri Ard Anderson expected for a large
- set, which the estimae structure and was 1975).
earthquake along the srs rpi 0
estimated to be 100 OD (note that less b
s rn This va bars for the evalub-than 30 bars occurred tion.~~~~whr in th 193 Lgnificantiicnty x
distanc froms faulto earthquake, WCC 1979).
than 30 bars estimte for an event on theOZ based on 1933Lo Beach earthqua Proximrity Attenuation of ground The earthquake modal Ground motion was Ground motions atten-Study of observed di.ta.
The definition o
motion with distance is has been calibrated and modeled at SONGS, 8 km uate with distance from distance used inth modeled -using an exact validated a g a Inst from the OZD.
fault.
The attenuation analysis was conter simulation of visco-near-field data re-reainhp-used was tiv e
(r esapons elastic waves in a
corded over the die-of the form:
t 3nts vertically stratified tance range of Inter earth.
est.
rb f ob 2M (R+Ctsjtab3 where R -
significant distance from fault i
magnitude of earthquake pocusing cecorded strong motion Siml ted ground notion Most extreme rupture Ground aotion is Effects of focusing The ground motion d daet cont i y aff ct is s g i i a t yn i u a i n ( o u -
ifluenced by focusing.
were incrporated in used were recorde focussed in the direc-Ing) is used for mSOS the selected deta under conditiono rupture focusing.
tion of
- rupture, a naIysaesa r eseuIt s.
set.
focusing that WOul te nd to increas neasured ground motion
- Thus, focusing effect were conservativel included.
Judgment/
h sie pcii Te prah euld Experience res te ae rne Val2*y rhuae reslt Ar bsedonin estimate$ for gon the most complete and re rigorous simulation of conservative wi currently a Va Il-available strong motio able.
The resulting recordings includin spectrum t.SON CS is in -th s
ob a n d ur g
reasonle agemt the 1979 Imperia w It hs8PeaCt ra I n Valley earthquake the %&me distance range for the recent Imperial Valley eartbquake.