ML20137E148
Text
,
. i- l (
f N (q, f
6 a a._ ( Y i L. A S' {(
%, UNITED STATES e[#" '%
,I c NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20555 F' q ,([ '
d , ,
k.....,/ September 1, 1982 r t . . .i;L _ . . ., i. u ~. :;
dd'r U1989
- v. en MEMORANDUM FOR: Chairman Palladino Comissioner Gilinsky i Comissioner Ahearne Comissioner Roberts Comissioner A sel ine A
FROM: Dennis Rath un, A ing Director Office of Policy Evaluation
SUBJECT:
ACRS INTEREST IN SEISMIC REANALYSIS As requested by Comissioner Ahearne in his memorandum of June 22, we have looked into the background of seismic issues raised in the June 8th ACRS letter to the Chairman on Midland. ACRS has expressed similar concerns about Perry Nuclear Plant Unit 1 (July 13) and the Clinch River Breeder Reactor Plant site (July 13). Moreover, such issues had been raised earlier by ACRS concerning Wolf Creek Generating Station, Unit 1 (May 11, 1982) and -- to a varying degree -- six other sites as.far back as 1977.
In its recent report (NUREG-0875) on the NRC Safety Research Program the~
. ACRS has noted that-
", a topic not specifically identified is the matter of how severe earthquakes may unpredictably and simultaneously impact a nuclear power plant in a large number of different ways, leading not only to a transient and/or possibly concurrent'small LOCAs, but also'to an indeterminate-set of failures in non-safety systems, possible failure of some engineered safety features, a loss of most of the information normally available to the control room, and the existence of spurious information."
We understand that the ACRS Subconnittee on Extreme External Phenomena will be meeting again on seismic safety margins on October 20 and 21 in Los )
Angeles, California. You may wish to discuss with the Comittee its appar-ently evolving generic concern (highlighted further below) when ACRS later l meets here in November.
e N Our review included some inquiries of the ACRS staff to clarify certain N@ parts of the record and reading relevant ACRS memoranda on Midland and these '
8 other sites, and pertinent sections of the following documents: 1 as gg 1. USAEC Safety Evaluation Report for Midland Plant, Units 1 and 2 I ggn (November 12,1970); 1 nu.n gg o m.;
- 2. NRC Safety Evaluation Report for Midland Plant, Units 1 and 2 (NUREG-0793, May 1982); . n j )
' Y 'i.~ f!
$$$ m ContactI Cookie Ong,UN'.Op ^
? k b [QR.WJ "kf T .W
,1 P D if CT' -
63-43302 bL --2 (M d . ,
JC$' '[83 0 NO? 5 Gii Ti? {";%~. S.Wnff;nn ,
e
- - . (
. The Comissioners 3. Transcript for ACRS Subcomitte Meeting on Midland Plant, Units 1 and 2 (May 20-21, 1982);
- 4. Transcript for 266th General Meeting of ACRS (June 4,1982);
- 5. Transcript for ACRS Subcomittee Meeting on Extreme External Phenomena (August 11,1982);and
- 6. ACRS Coments on the NRC Safety Research Program Budget for Fiscal Years 1984 and 1985 (NUREG-0875), July 1982.
We highlight below and in the Enclosure parts of that public record that ap-pear relevant to Comissioner Ahearne's questions. OPE has not yet considered the relative merits of those excerpts regarding Midland or any other specific site.
Comissioner Ahearne's first question:
"In particular, in what ways have the SSE or the related seismic analyses changed since the CP was approved?"
As reflected in the Enclosure, seismology analyses for a site can change in the span of 12 years between the CP and OL stages - particularly in view of -
the publication of standard spectra in 1973, a continually expanding potential
, data base and an advancing state-of-the-art o.f methodglogy As a ,re.sult, SSEs, and/or associated estimated ground motions at frequencies covering structures and safety equipment and components may be subject to substantial increase.
Moreover,recurrenceintgrvalsfortypicalSSEswereinitiallyghought(vaguely) years.
to be of the order of 100 years are now considered as low as 10 to 104 Accordingly, some members of ACRS are significantly concerned with seismic mar-gins. Dr. Okrent noted on August 11, 1982 that the ACRS believed that it was important that there be considerable assurance that the combination of the seismic basis event and the procedures used to establish the design margin be such that the seismic risk represent an acceptably low contribution to the overall risk. On May 20, 1982, he had stated, "It seems to me that, among the things are relevant ... are not only what should be the safe shutdown earthquake (flor design basis purposes but also what level of safety do we think is being achieved with r.egard to the seismic source of risk and is this level of safety adequate 7" At the August 11, 1982 ACRS meeting Dr. Okrent stated, "The point I'm trying to make is, this question of meeting the regulations is a little bit fuzzy. One's knowledge of what the frequenpy~of.the:SSE, or at least one's opinion on the frequency of the SSE, Secause I'm not
-sure it's knowledge, has shifted. And so there is more importance in being assured that the margins exist in everything you need for safe shutdown ...
/
t
. . .' (
The Comissioners -:$-
That doesn't mean, I think, that one shouldn't know that there are these possible situations and let expert opinion look at this and say, yes, I think this is okay, or no, in the same way you looked at things related to fires and made judgments and in some cases made very expen-sive fixes being required, even though there were differing opinions."
Commissioner Ahearne's second question:
"Is it the Staff's position that the plant must withstand greater than the SSE7" Based on the transcript of the ACRS Subcomittee, the Staff emphasized that NRC regulations (Appendix A to 10 CFR Part 100) do not require 'a probability for an SSE and that the Staff has not placed major reliance of probabilities in view of associated uncertainties. Furthermore, the Staff would have dif-ficulty in implementing recommendations made by the ACRS in their recent OL letters without specific criteria as to the events which should be considered and. guidance as to backfit criteria. Mr. Knight (NRR) stated:
"The more recent letters I believe are asking a significantly different question. That is, regardless of where your SSE falls with regard to our best judgment, our best exercise of technology today, we ought to be able to demonstrate that an event of more severity, albeit less prob-able, could still be tolerated. It is that somewhat nebulous require-
- ment that I think is going to give us a great deal of difficulty in-. - -
trying to implement it. You are in the posture again of going to the utility and saying, well,you've got to do something better without being able to explicitly tell them what their goal is ... I'm certainly not --
on behalf of the Staff, we're not throwing up our hands. We are openly trying to say that we see problems in trying to get on this recomenda-tion, and we are certainly searching for all the advice we can get as to what was the point of the recomendation and what is viewed as an ade-quate reponse, so we can be on target as much as possible."
He emphasized that regarding what some have already tagged as the SSSE or super-SSE, significant policy questions were involved, e.g.,
- 1. Commission safety goal;
- 2. Backfitting; and
- 3. Modification to regulations.
Mr. Knight believes NRC must be able to provide sufficient guidance to a util-ity, developing the ground rules " prior to laying the requirement on a utility
... a very significant effort that could evolve here that would have no founda-tion for acceptability until we started examining the end product. That is really an unacceptable product."
~
Enclosufe:
As stated cc: (see page 4) m
. . l (
The Comissioners f cc: Leonard Bickwit Samuel Chilk Ray Fraley O
ae I
1
. (
ENCLOSURE CHANGES IN NRC ANALYTICAL APPROACHES AND CRITERIA AFFECTING MIDLAND SSE At Midland's construction permit (CP) stage in 1970, the Staff's seismological consultant (U. S. Coast and Geodetic Survey) recomended and the Staff concur-red, that peak horizontal ground acceleration of 0.12g, resulting frorn a Modi-fied Mercalli intensity (MMI) VI earthquake (maximum historical earthquake within 150 miles of the Midland site), would be adequate for representing the ground motion from the maximum earthquake likely to affect the site. ,
Accord- ,
ingly, a peak acceleration of 0.12g was used to anchor a modified Housner response spectrum for design purposes. The horizontal design spectrum was a Housner spec-trum modified in the period range of 0.2 to 0.6 seconds (1.67 to 5 hertz). How-ever, in its initial review the last few years, the Staff concluded that this might not be an adequate representation for the SSE at the OL- stage.
^~ ---
Since the issuance of the CP for Midland, the following have occurred:
- 1. In late 1973 Regulatory Guide 1.60 (Design Response Spectra for Seismic De-sign of Nuclear Power Plants) was published.
- 2. In looking at more recent sites in the central United States.(US)',athere have been a range of peak elevations essentially between 0.12g and 0.19g.
- 3. In its initial OL review, the Staff disagreed with the applicant's position expressed in the FSAR that, based upon the Paleozoic geology, the Michigan Basin was a separate tectonic province from the central stable region. The Staff concluded from this review, based on the geology alone, that the Michigan Basin could not be separated out from the rest of the Central
/
l
. . l ( l
. Stable Region, which would include approximately one-third to one-half the surface area of the United States. As presented in the FSAR, the applicant maintained that the Midland site lies within the Michigan Basin Tectonic Province. The Michigan Basin is a regional structural basin that underlies the southern peninsula of Michigan and parts of ad-joining states.
- 4. Using the tectonic province approach of Appendix A to 10 CFR Part 100 to develop the vibratory ground motion corresponding to the SSE, the NRC Staff developed two concerns in the SER: first, the size of the control-ling earthquake increased over that which wa's determined at the CP stage; and, second, given the larger controlling earthquake, concern was expressed about the use of a modified Housner spectrum anchored at 0.129 to repre-sent the vibratory ground motion at the site. TheStaffdeterminNdthat the design response spectrum was no longer a conservative representation of the ground motion as used. Consequently, the Staff asked the applicant to either use a Regulatory Guide spectrum anchored at 0.19 or develop a
- site-specific spectrum while perfonning a seismic hazard analysis.
- 5. As outlined below, however, these issues have been resolved using state-of-the-art seismological information and data analysis, including the use of a site-specific response spectrum and a comparison of the relative seismic hazard between the Midland site and five other sites within the north-eastern. Central Stable Region of the US.
t
- 6. The Staff has used the seismic hazard analysis to quantify the apparent low seismicity near the Midland site compared with other areas in the 4
i
(
. ~
Central Stable Region. The Staff has recognized that the Paleozoic geology of the Central Stable Region does not explain the fact that different areas of this large region exhibit different levels of seismic activi ty. In an attempt to quantify the low seismicity near the Midland site, the applicant performed an' analysis of seismic hazard'for Midland along with five other sites in the upper midwestern region of the US.
The output of the seismic hazard analysis is the annual probability that a given intensity earthquake will occur at each of the sites (underlined for OPE emphasis). The key elements of the applicant's methodology to estimate the seismic hazard involve the selection of the earthquake oc-currence model (including seismic source zones, rate of activity in each zone, and largest earthquake in each zone) and ground motion model (relat-
- ing size of earthquake to ground motion). . ._ . . . . _ _ . .
- 7. In its review, the Staff compared the relative probabilities for a number of sites, rather than the absolute probability at a specific site. The Staff then found that, in general, the Midland site has lower expected in-tensities than the five other sites at all exceedance probabilities.
- 8. In the Central Stable Region, the largest historical earthquake, in terms of intensity, is the Marcit 9,1937 event (MMI = VII - VIII at Anna, Ohio (Hgg = 5.0-5.3) Consider the entire Central Stable Region as one unit, MMI = VII - VIII or magnitude M gg = 5.3 would then be used by the Staff to define the SSE. However, the SER found that defining the Central Stable R.egion from the Paleozoic geology may be inappropriate when one is attempt-ing to achieve consistent descriptions for the SSE.
/
i
- 4
- 9. It is the Staff's position that magnitude is a better estimator of earth-quake source strength than intensity, which is a subjective description of damage and felt effects of buildings and the general public.
- 10. Based on its overall review of the hazard analysis results, the Staff concluded that for the Midland site the SSE should be defined by a con-servative magnitude M gg = 5.0. For purposes of seismic design, the Midland site is in a different seismotectonic province, that requires a smaller controlling earthquake compared with many other areas within the Central Stable Region. A magnitude 5.0 is larger than that of any his-torical earthquake within about 200 miles of' the Midland site.
- 11. The Staff eventually agreed with the applicant that the Parkfield, ,
- California records of strong motion would not be appropriate for _incht- -
sion in the data base in the development of a site-specific response spec-trum. If the Parkfield records had othemise been added, the site-speci-fic spectrum would have increased by 10 to 30 percent for frequencies be-tween 2'.5 and 20 hertz - frquencies (as shown below) where the site-specific spectrum exceeds the original Midland design spectrum.
- 12. Given the choice noted in 4, above, the applicant elected to develop a site-specific spectrum for Midland (original' ground surface) by searching the earthquake strong-motion data base. The acceleration records (22 sets of accelerograms) collected fell in the magnitude range of 4.9 to 5.5, at epicentral distances of 7 to 33 km. Site conditions of this data set were chosen to match the Midland site. Based on its. review, which included
/
, - +
~
. (
l
. l 1
j sensitivity tests, the Staff found thht the 84th percentile of the appli-cant's site-specific spectrum (without the Parkfield acceleration records) is a conservative representation of the ground motion expected if an mgg = 5.0 earthquake occurred close to the Midland site.
- 13. Enclosure 1 displays that site-specific spectrum and the extent it exceeds the original Midland design basis earthquake spectrum (modified Housner).
Also shown is the extent that a Regulatory Guide 1.60 standard spectrum anchored at 0.129 (MMI = VII) exceeds the other two spectra.
- 14. Because some of the seismic Category I struc.tures are founded above the original ground surface in the plant fill, the applicant was asked to as-sess the potential for soil amplification through the fill. In making
. this assessment, the applicant chose two methods to_ arrive at a. response -
spectrum for the top of the fill material: (1) developing a second site-specific spectrum matching the fill soil profile; and (2) utilizing a one-dimensional wave propagation program to assess possible amplification.
- 15. Enclosure 2 displays the 84th percentile of the site-specific spectrum for the top of the fill material, and shows the extent of excess over the original Midland spectrum.
- 16. As a result of the above spectra exceedances, the applicant is assessing seismic margins
. Based on the site-specific earthquake;
. Including structures and equipment;
. Using a screening process to identify elements and components for review for seismic; and
/
. (
- . Allowing for deviations from the Standard Review Plan for failure capacity evaluation.
The applicant has provided an evaluation determining that the use of the design spectra in conjunction with the conservative damping values identified in the FSAR provide design-margins similar to those resulting from the use of the Regulatory Guide 1.60 response spectra in conjunction with Regulatory Guide 1.61 damping values. In addition, the applicant is currently evaluating those seismic Category I structures necessary for shutdown and continued heat removal to determine seismic margins result-ing from application of site-specific spectra. Upon completion of the Staff's review of these evaluations, a supplement to this safety evalua-
~
tion report will be issued. .
- 17. At the ACRS meetings on Midland, extensive discussions centered on the use of probabilistic estimates to determine seismic hazards, e.g., in defining the Operating Basis Earthquake (0BE), determining the return period asso-ciated with an SSE, calculating equivalent levels of seismic hazard for one plant relative to other sites, and helping to make a rational seismic decision for a tectonic province that covers one-third or more of the US and has widely varying levels of seismic activity. The Staff emphasized that probability estimates were for relative rather than absolute value and for insight rather than direct use -- particularly in view of uncer-tainties.
- 18. Nonetheless, some members of ACRS in both Midland and a generic basis are concerned about:
/
- I (
. What is the probability of an SSE?
. What is the earthquake or acceleration associated with probabilities of'IO-5 or 10-6 per year?
~
Applicant, NRC Staff and ACRS seismology consultants for Midland gave their estimates (with 90% confidence) of the probability of exceedance of one in 1,000, one in 10,000, and one in 100,000 per year. Although each seismologist expressed concern over uncertainties involved in making such estimates, their estimates were:
Mr. Klimkiewicz (Western Geophysical) 10-4 approx. 0.12g 10-5 0.179 approx.
Mr. Kimball (NRC/NRR/DE/GEB) 10-4 0.12g
-5 0.16g to 0.20g 10 .
-4 approx. 0.12g -
Mr. Reiter (NRC/NRR/DE/GSB) 10 - . .- -
10-5 no answer Dr. Pomeroy (ACRS Consultant) 10-3 intensity 8 1 10-4 intensity 9 2 10-5 intensity 10
- 3 Dr. Trifunac (ACRS Consultant)* 0.05g 80% exceedance prob. 30 yr. return period 0.10g 40% exceedance prob. 100 yr. return period 0.15g 15% exceedance prob. 300 yr. return period 0.20g 8% exceedance prob. 600 yr. return period
- ACRS minutes for the meeting indicated, however, that Dr. Trifunac corrected an error in his calculations after the meeting. He amended his answer stating:
"the exceedance probabilities for peak accelerations greater than 0.03g are smaller than 5% for 50 years of exposure and the return periods are longer than 100 years."
i j l
- ~
[ ofw74. O
- . 9
- .I' ~
-}2 ,
.- , ~j I Ill } \ l \ IIII-i i i i illt I .I 10.0 ~
REGULATORY C
/ [MIDLANDg" d %7 _
.ncTR4
/ = == MIDLAND snt sncine
/ l *
RESPONSE
sncTRA
/ -
31.0 ~ f -
2 j '
~
e a
/ -
g . .
q ,4r c.12e PEAK ACCELERATION n.1 -
~
5% DAMPlNG )
~ i i i 11111 l I I III
- ; q 111 1 I I
, ,, 10.0 c.10 a.m
~
, PERIOD IM Figure 2.7 Comparison of Midland original design spectra, site-specific .
spectra, and Regulatory Guide 1.60 spectra
. i( ~
/ .
Midland SER -
s
' .s . . . - . _
- n.
. Ea ctos -ve . 2.
(
e I i i i i l li-i i l i lit i I i ltill I
10.0
,,, OG ~
DESIGN -
- SPE,CTRA
~ -
L
~
~~
MIDt.AND SITE f r
SPEC 1RC SPECTRA FOR THE TOP OF THE PLANT /
RLL MATERIAL _
I. 1.0 r
}- _
g _
g _ .._
8, g _
- / .
f -
0.10 , ~
~
~
~
i 5% DAMPtNG
- .i l l I l l llI l j I I I I II I I I I IllIl 10.0 0.10
,1.0
. 0.01
- PERIOD (sec) n, Figure 2.8 Comparison of Midland original design spectra and site-specific
\ spectra for top of fill material ld
/
't e e t .on . . a eco
,4Q