ML20006A787
| ML20006A787 | |
| Person / Time | |
|---|---|
| Site: | Peach Bottom, Surry, Zion, 05000000 |
| Issue date: | 08/07/1989 |
| From: | Sholly S MHB TECHNICAL ASSOCIATES |
| To: | Ross D NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| FRN-54FR34268, RTR-NUREG-1150, RTR-NUREG-CR-3300, RTR-NUREG-CR-3428, RTR-NUREG-CR-3756, RTR-NUREG-CR-4550, RTR-NUREG-CR-4551, RTR-NUREG-CR-4840, RTR-NUREG-CR-5088, RTR-NUREG-CR-5250 54FR34268-00004, NUDOCS 9001300210 | |
| Download: ML20006A787 (3) | |
Text
gg TECHNICAL CONSULTANTS ONJN l
L I
m, Date G Bodenbaugh 1723 Hamaton Avenue-Surte K Richard B Hubbard
(
y yp.g San. lose. Caldornia 95125 Gregory C. Minor Phone. (408) 266 2716 Far (408) 266 7149 7 August 1989 Dr. Denwood F. Ross, Jr.
7 /h 0
Deputy Director for Research Office of Nuclear Regulatory Research Washington, D.C:. 2055'y Co'mmiuion U.S. Nuclear Ret ulator 5
RE:
Preliminary Comments on NUREG 1150, Severe Accident Risks: An Assessment for Five U.S. Nuclear Power Plants (Second Draft for Peer Review), June 1989.
Dear Dr. Ross:
Accident Risks An Assessment for Five U.S. preliminary comments on NUREG-1150, Seve The purpose of this letter is to convey my Nuclear Power Plants Second Draft for Peer Review), June 1989. As with the February 1987 draft of NUREG-110, detailed review and comments will not be possible without access to the underlying detailed reports (NUREG/CR-4550 series, NUREG/CR-4551 series, etc.). Although drafts of these reports are indicated as being available from the NRC Public Document Room, a.quite a number of them were not
- report (NUREG/CR-4840).g of August 1989, including specifically the external eve available as of the beginnin My principal comment based on my review to date deals with the subject of NUREG-1150's treatment of expert opinion. Pm sure you are aware that the results of NUREG 1150 could I
be quite sensitive the outcome of the expert elicitationprocess in a number of re;ards. It would be most usefulif the NRC would expedite publication of those uhderlying technica: volumes which-L describe in detail the approach used in the expert clicitation process and which contain the documentation prepared for and by the expert panels. This would allow reviewers the epportunity to become better acq,uainted with the expert elicitation process and to assess the degree of confidence one places m the outcome of that process.
More saecifically, I suggest that it would be beneficial for NRC to conduct an eneriment of sorts on tie expert elicitation process. in particular, NRC should consider selecting a particularly important issue or two and empaneling perhaps three separate expert panels (including the panel used in draft NUREG 1150) and performing the expert clicituuon, process.
This exercise would provide a useful calibration on the process to enable one to gam some
- ppreciation as to the magnitude of variability that could be introduced into the risk estimates by the expert clicitation process, I would suggest that this be done on both front end issues and back' '
end issues.
.,s 1
I have several additional comments based on my review to date. These are summarized in
" bullet" fashion below:
M$13$021o89o907 54FR34gsg pyg
2-There are a number of places in the report where conclusions are stated or 1.
implied which conilict with the results of previous analyses. A key example is the degree to which a stuck open PORV depressurizes the reactor coolant system (thus avoiding direct containment heating which could arise from pressurized melt ejection) in PWRs. Previous analyses with which I am familiar suggest terminal RCS pressures of the order of 1000-1200 psig at the time of vessel breach. Morover, in the Thil 2 accident the operators tried to depressurize via the PORV to get the core flood tanks to discharge (pressure of 600 psig, as I recall), and were unable to do so. In contrast, the hUREG-1150 analysts either assume or have more recent calculations which caused them to conclude that the terminal pressure at vessel breach under these circumstances are much lower. The hRC should make available (or at least reference) the calculations which underly these conclusions / assumptions.
The Zion risk estimates in NUREG 1150 fail to include external events. As 2.
ou are no doubt aware such estimates are available from the Zion PSS y(Commonwealth Edison PLG, September 1980, the Zion Review (Sandia, May 1984, which identi ed a seismic station blackout as the 8th most likely sequence, at 5.6 x 10' per reactor year), the SSMRP analysis of Zion and the LLNL seismic hazard estimates (NUREG/CR 3428,
(NUREG/CR 5250. It would seem that a fairly,' minimal effort would be required to includ the Zion external events m the risk estimates for NUREG 1150, and that this effort would be worthwhile for the final repor,g (The SSMRP report estimated a seismic core damage frgquency of 3.6 x 10 per reactor year on a point estimate basis, and 3.0 x 10 per reactor-year on percentile estimate of 8.0 x 10,tpe mean value would appro a median basis; presumably done with the *old" LLNL hazard curves -- i.e., year. These calcula per reactor-NU, REG /CR 3756.)
The draft NUREG 1150 containment failure analysis considers steam 3.
explosions as a containment failure mode (" alpha mode" failure). Such a containment failure probably represents only a small fraction of the total number of times a steam explosion actually occurs. It would seem plausible that there is another fraction of steam explosion events which, while not resulting directly in containment failure, would nonetheless result in impact on the containment which could weaken the containment and cause its failure from other causes (steam, hydrogen burn, etc.) at a later time during the accident arogression. There is no evidence in NUREG 1150 per se that this eventuality has been addressed (perhaps this is discussed in the unpublished underlying technical volumes).
4 NUREG 1150 considered the risk posed by fires for Peach Bottom and This evaluation does not appear to have considered spurious,
Surry.
actuation of fire suppression systems on the availability of plant equipment.
There have been several NRC notifications to industry on this issue (IE Information Notices 83-41,84 57,85 85, and 8714). This is also discussed in the Fire Risk Sgoping Study (NUREG/CR 5088), which indicated that spurious suppression events are occurring at a rate similar to actual fires (it also indicated that this could be quite significant since those areas protected by fire suppression systems generally contain safety-related equipment). It would seem appropriate that NUREG 1150 address this issue to the extent feasible or discuss the rationale for assuming that it is not significant for the plants in question.
I I
3-s Further, NUREG 1150 mentions that the Peach Bottom cabic spreading 2 or fire suppression (NUREG-1150, Vol.1, at 8-17).
f room uses CO of a CO2 suppression system would lower the Spurious actuation perhaps exceeding temperature of the cabic spreading room dramatically,ble spreading room.
l the environmental capabilities of componer.ts in the ca NUREG 1150 should consider this possibility.
Finally, it should be noted that severe accident conditions (including initiatmg events or equipment failures resuhing from HVAC failures, as well as the severe accident environment) could result in fire suppression system actuation. This should also be addressed.
NUREG 1150 does not appear to have addressed the potential for fires to be 5.
started as a result of seismic events. The Sandia Fire Risk Scop,ing Study recommends a walkdown to march for seismic / fire (NUREG/CR 5088) ties. NUREG 1150 should address whether this w interaction possibih done for Surry and Peach Bottom, and,if not, why this is not important for these two plants.
As Jim Harding and I indicated in our comments on the previous version of 6.
NUREG-1150, there may be for multiple-unit plants a non trivial likelihood of concurrent accidents at multiple unit sites. This should be addressed in final NUREG 1150 since it could significantly affect risk estimates and cost-benefit analyses of risk mancgement improvements. (This was addressed to a limid extc.nt in the Sandia Zion Review; see NUREG/CR 3300, Vol.1, l
at 4 31. li was also addressed in the Seabrook PRA.)
l I expect that as more docutaentation becomes available I will have additional comments It you or your staff have any questions regarding the above l
l NUREG 1150 (June 1989).
comments, please do not hesitate to call or write.
L i
Sincerely, l
Steven C. Sholly cc:
J. Murphy, NRC E. Gorham Bergeron, SNL G. Thompson, IRSS l.
1
't m.
..)