ML20205E729
| ML20205E729 | |
| Person / Time | |
|---|---|
| Issue date: | 10/07/1985 |
| From: | Harold Denton Office of Nuclear Reactor Regulation |
| To: | Gilland B AFFILIATION NOT ASSIGNED |
| Shared Package | |
| ML20205E733 | List: |
| References | |
| NUDOCS 8510170435 | |
| Download: ML20205E729 (6) | |
Text
'
%1STRIBUTION CENTRAL FILE DEisenhut 00T 071985 AEB R/F HDenton JRead HSmith-000948 Mr. Bernard Gilland LSoffer PDR Lille Bakkegaardsvej 2, LHulman EDO R/F 3060 Espergaerde, Denmark DMuller PBrandenburg 000948 RBernero CPaul
Dear Mr. Gilland:
KPulsipher SECY (85 W )
734 WDircks
- JRShea, I am pleased to respond to your letter to Chairman Palladino of August 21, 1985.
In your letter you ask for an explanation of the difference between two estimated probabilities, one attributed to the NRC.
Estimates of the likelihood of a core melt accident have been made for many V. S. nuclear power plants, but six have been studied in special depth.
The average of the estimates for these six plants is 3.x10 4 per reactor year.
If this average was assumed to describe each of 100 reactors operated for 20 years each, then the chances of observing one or more core melt accidents would be 45 in 100, in agreement with the article quoted in your letter.
The number of U. S. power reactors is expected to increase to about 120 by the year 2000, but the average number operating during the next twenty years will be smaller.
Given an accident in which a power reactor core is severely damaged, it is not a certain outcome that the core will melt through the reactor pressure vessel.
There are many ways in wtk;lh the damage may be arrested or limited in extent to less than large scale c, ore melt.
Estimates made in PRA studies usually ignore these possibilities and treat the probability of severe core damage as the probability of large scale core melt.
An accident in which a molten core melts through its pressure vessel producing the release of a sufficiently large amount of radioactivity to the environment to cause fatal exposures is an even smaller probability.
The occurrence of fatal exposures due to accidentally released fission products would require not only a large release to the environment, but also the occurrence of moderate to low probability meteorological conditions and failure to complete evacuation or other protective measures by the surrounding population.
Thus, it is reasonable to expect for any nuclear power plant that the probability of one short term death is much lower than the probability of core melt.
The NRC staff has not assessed the accident probabilities of the proposed Sizewell plant.
However, NRC certainly expects that future plants, designed with the benefit of probabilistic risk assessment techniques, b
will have improved accident responses in comparison to plants constructed prior to the general use of those techniques.
The experierce gained by
,jl' if f
), ?
'l
} \\)
Y
,(; J b t
OS10170435 051007 i
i PDR MlUC 0510170435 PDR OFC :
NAME :
DATE :
l OFFICIAL RECORD COPY l
L
6 Mr. Bernard Gilland
-2 studying the potential accidents of existing plants will undoubtedly lead to improved safety in future plants in addition to assuring the safe operation of the existing plants themselves.
I hope this information will be of assistance to ycu.
Sincerely, pignalSism84 (p.A.Dete Harold R. Denton, Director Office of Nuclear Reactor Regulation
- See Previous Sheet for Concurrence OFC : OSI: AEB *
- DSI: AEB *
- OSI: AEB*
- DSI: AD/RP* : OSI:D
- NRR r----
..!..$.' $$...!...$..![....!...."
S $[....!.. $[."ISS...!....I".S...!.....'....
DATE 9/ ///85
$ 9/ // /85 9/// /85
$ 9/// /85 $ 9/ /F /85 9M@85 i M M, /85 OFFICIAL RECORD COPY
tr. Bernard Gilland Size 11 plant, not yet constructed, has benefited from the widespread use ofrikassessmentsfollowingtheThreeMileIslandaccident. Hence, it is not surp ising that the core-melt probability for Sizewell is believed to be significa ly lower than the average value of all U. S. plants.
Third, the S ewell estimates of early fatalities have taken into account the likelihood f reduced accident releases (" source terms"), at least in an approximat manner. You may be aware that recent research has indicated that man severe accident sequences for PWR designs are predicted to result lesser quantities of radioactivity released to the environment than had p eviously been believed. This is presently an area of great interest in the world. wide nuclear engineering community.
Parametric calculations h e shown that estimates of early fatalities are particularly sensitive to t estimated releases and might be reduced sharply. The NRC is pursuing vigorous research effort in the " source term" area, but the research is being evaluated at this time.
I hope this information has been o assistance to you, incerely.
Harold
. Denton, Director Office o Nuclear Reactor Regulation
\\N
'N\\
- See Previous Sheet for Concurrence 0FC : DSI:AEB*
- DSI:AEB*
- DSI:AEB*
- OSI:AD/RP
- DSI:D
.....:............:............:.................. 43.................:........................
Dylier NAME : JRead:ye
- LSoffer
- LGHulman
- RBernero
- DEisenhut : HDenton
.....:............:............:............:.........................:..............N...
DATE : 9/// /85
- 9/// /85
- 9/ // /85
- 9//!/85
- 9/
/85 : 9/
/85 OFFICIAL RECORD COPY
k Bernard Gilland 2-N\\
Sizewe}(plant, not yet constructed, has benefited from the widespread use of risk dssessments following the Three Mile Island accident. Hence, it is not surprising that the core-melt probability for Sizewell is believed to be significantly lower than the average value of all U. S. plants.
Third, the Sidewell estimates of early fatalities have taken into account the likelihood ofs reduced accident releases (" source terms"), at least in an approximate ' manner. You may be aware that recent research has indicated that many severe accident sequences for PWR designs are predicted to result iitxlesser quantities of radioactivity released to the environment than had previously been believed.
This is presently an area of great interest in the world-wide nuclear engineering community.
Parametric calculations have. shown that estimates of early fatalities are particularly sensitive to the estinated releases and might be reduced sha rply. The NRC is pursuing a, vigorous research effort in the " source term" area, but the research is being evaluated at this time.
An important point to note is that estimates of such infrequent events must be viewed as having considerable" uncertainty.
I hope this information has been assistance to you, Sincerely, i
liarold R. Dento'n. Director Office of Nuclear Reactor Regulation
' 0FC :
AEB
- AEB
- DCI: B
- DSI:AD/RP : DSI:D
......: L g(.S...:.g
.NAME :
ad:ye
- L 0:'rer
- LW n
- DMuller
- RBernero
- DEisenhut
- ilDenton
.DATE : 9/ U /85
- 9/ R /85
- 9/ \\ \\ /85
- 9/
/85
- 9/
/85
- 9/
/85 0FFICIAL RECORD COPY
y[
,'o, UNITED STATES
^y y
3.c, g
NUCLEAR REGULATORY COMMISSION 1
f, 3.h
- E WASHINGTON, D. C. 20555 l
I o $'ss 8
[i
(
[..
Y 44.)
.,/
g,
,o I
p-A
!)
EDO PRINCIPAL CORRFSPONDENCE CriNTROL FROM:
DUE: 09/12/85 EDO CONTROL: 000948 DOC DT: 68/21/95 BERNARD GILLAND FINAL REPLY:
DENMARK TO:
CHAIRMAN PALLADINO FOR SIGNATURE OF GREEN SECY NO: 85-714 DENTON DESC RnUT1NG:
DISCRCPANCY IN NEWS ARTICLE RE OPERATING REACTORS SHEA MELTDOWN PROBADIt.ITY DATE: 08/28/85 AcSIGNED TO: NRR CONTACT: DENTON Jg.uw SPECIAL INSTRI.lCTIONS OR REMARKS:
fiRR RECEIVED: 08/29/85 ACTI0ft:
DSI, R. BERNER0 ROUTIf1G: DEflT0fi/EISEfillVT PPAS 99'ca f
Bernard Gilland CORRESPONDENCE CONTROL TICKET SECY NUMBER:
85-734 LOGGING DATE 8/28/85 0FFICE OF THE SECRETARY ACTION OFFICE:
EDO AUTHOR:
Bernard Gilland AFFILIATION:
Denmark LETTER DATE:
8/21/85 FILE CODE ADORESSEE:
Palladino
SUBJECT:
Discrepancy in news article re oper reactors meltdown pro-bability ACTION:
Direct Reply... Suspense: Sept 6 DISTRIBUTION:
SPECIAL HANCLING:
None SIGNATURE DATE:
FOR THE C0fftISSION BAC RM'if get EDO 0.te J - J y.
,N Ti n i a /, ' L > g'.'.' '.* *.."..".
PD0 --- ooongg