ML20133H188
| ML20133H188 | |
| Person / Time | |
|---|---|
| Issue date: | 05/07/1985 |
| From: | Pomeroy P Advisory Committee on Reactor Safeguards, RONDOUT ASSOCIATES, INC. |
| To: | Okrent D Advisory Committee on Reactor Safeguards |
| References | |
| RTR-NUREG-CR-3756 ACRS-CT-1811, NUDOCS 8508090221 | |
| Download: ML20133H188 (3) | |
Text
_ _ _
r_
- g jg1 /[
ROxmur Associates. INCORPORATED
'Y P.O. Box 224, Stone Ridge, New York 12484 , 9/)R did M3 7 t'ay 1935
~~ 'd Prof. David Ukrent Advisory Comittee on Reactor Safeguards ;3 U.S. M lear Regulatory Commission
? r-Was'aington, DC 20o05 ,
Dear Dr. Okrent:
In accordance .vith your request transmitted by Dr. Savio, I have reviewed the Yankee Atomic Electric Company (YAEC) evaluation of the Lawrence Liver-more Report entitled "Se hmic Hazard Characterization of the Eastern United States: Methodology and Interim Results for Ten Sites" (NUREG/CR-3756). My com ents are as follots:
- 1) The YAIC evaluation study was extremely complete and carried out in extreme detail as one would expect from the principal authors (Drs. Allin Cornell cnd Thomas O'Hars).
- 2) In the Executive Sumary and throughout the text, the YAEC team stresses "the total lack of documentation associated with expert input in
- JEG/ Cit-373.>. " The Executive Sumary states "This is clearly a major wea'< ness of that study." That may well be YAEC's conclusion but many responsible scientists would not a: ee with that conclusion. In an ear-lier study, Lawrence Livermore was criticized just as severely for atteg ting to gain documentation from the experts. In my opinion, the YAEC report indicates a lack of understanding of the basic purpose--to elicit expert opinion regardless of whether or how that opinion is com-pletely documented. If complete or partial documentation is forced in the study, it can, in fact, force the experts to fall back to positions that can be doc'nented rather thaq their real opinions. This is a matter of philosophy bt.t not a " major weakness".
- 3) The YAEC study indicated a problem with the way in which the magnitude frequency relation was treated at the high magnitude M I believe that thisproblemwaspresentedtotheseismicitypanelandhr.operly resolved.
It is not clear, however, that a truncated model or a smoothly decaying model is necessarily preferable. The expert's preference is still the important factor.
- 4) The YAEC position is that the Chiburis catalog should be used for evaluating seismicit.y parameters in New England. Many people, including many of those involved in the later EPRI study do not share that view.
ORIcIhr, 8509090221 850507 le PDR q ed g__ g g g
r 4
U .
Prof. David Okrent 7 May 1935 Page 2
- 5) The YAEC document is critical of the intensity-magnitude conversion chosen by the experts and the lack of documentation of the expert's
. choice. All of these relationships are arguable and one could perform a major critique of the YAEC line of reasoning and their choice of the WGC
. relation. Again, however, it is the individual expert's opinion of the appropriate conversion that is important and, whatever that was, it should be included as Lawrence Livermore has done.
- 6) The YAEC study argues that the Wiechert maximum likelihood procedure should be used to properly determine a and b values. This again demon-strates that the YAEC team misses the point--the purpose was not to require ecch expert to evaluate the seismicity parameters using the YAEC preferred technique but rather to obtain the expert's best estimates of the parameters, however they might have been obtained.
- 7) The YAEC critique then takes the seismicity parameters specified by each expert and compares the result of using those parameters to deter-mine the total number of small events with the "true" number of small events in the Northeast in the past 12 years. This involves the assump-tion of a magnitude-intensity conversion and an assumption that the "true" number of events, as determined by m is equivalent to the true number of events (i.e. hoa dees m or mc bNa,tetothemagnitudeinthe r
magnitude-intensity relationship h The YAEC team uses m b
--is that mb or "real" m or what? Hoaever,disregardingallthat,theYAECresultk9 b
that "the number of events 2 m h=2.5 in m 3=3.0 in the past 12 years predicted by the er.perts is much higher than that observed." This is a highly . interestir.g study but again misses the point, namely, do the experts believe tnat their best estimates are correct?
- 8) The final hazard curves are presented showing how the hazard can be lowered by " correcting" the expert's opinion as discussed on page 13 in the text and using instead the YAEC preferred methodologies. Of course, when this is done, the mean results agree closely with earlNr VAEC ana-lyses.
I would like to make an overall comment. The YAEC study is hignly useful to document the areas of uncertainty in this type of analyses and to specify the particular beliefs of the YAEC experts. The usefulness of the document, however, is completely overpowered by the er.tremely critical nature of the comments. The critical reader is left with two thoughts:
- 1) The authors seek both in their language and in their science, to discredit in any way possible the results of the Lawrence Livermore study.
'l Prof. David Okrent 7 May 1935 Page 3
- 2) The authors seek to convey the impression that, had only the experts known of the YAEC preferences and the reasons for those preferences, the results vould have been vastly different. Most of the experts were only too aware of the YAEC preferences but, because of their experience and background, chose quite different values.
I would like to thank you for the opportunity to comment on the YAEC evaluation and I look forward to evaluating the final Lawrence Livermore report and the EPRI reports.
Sincerely yours, e , /
44' La my Paul W. Pomeroy P'JP:gla
.