ML20062E356
| ML20062E356 | |
| Person / Time | |
|---|---|
| Site: | Rancho Seco |
| Issue date: | 07/21/1982 |
| From: | Novak T Office of Nuclear Reactor Regulation |
| To: | NRC ATOMIC SAFETY & LICENSING APPEAL PANEL (ASLAP) |
| References | |
| TASK-AS, TASK-BN-82-75 BN--82-75, BN-82-75, NUDOCS 8208090328 | |
| Download: ML20062E356 (17) | |
Text
c o <? l O x~
' d
'DIST_RIBUTION
( M t Fil d ORB #4 Rdg AE0D Gray =FHe-
'] Ch 37 ogg NRC PDR DEisenhut JSLolz L PDR ECase ACRS TNovak OELD Docket No. 50-312 i ao n PAS Memo File MWilliams MEMORANDUM FOR: Atomic Safety and Licensing Appeal Board for Rancho Seco FROM:
Thomas M. Novak, Assistant Director for Operating Reactors, Division of Licensing, NRR
SUBJECT:
BOARD NOTIFICATION (BN-82-75) - RANCHO SECO This report presents the initial results of a program that was begun as a result of one of the Lewis Committee reconinendations following their review of NASR 1400, the Reactor Safety Study.
The Precursor Program uses Licensee Event Reports to evaluate potential nuclear plant accident precursors occurring at operating reactors.
These individual plant precursors are then suninarized to evaluate the risk (for a This particular time period) from all operating nuclear power plants.Theestimateisbetween1.7x10'pp covers th and 4.5 x 10 q period from 1969 to 1979.
per reactor year and includes contributions from three major events:
3 (1) the loss of feedwater and stuck-open relief valve at Three Mile Island Unit 2 (which actually resulted in severe core damage), (2) the loss of nonnuclear
-instrumentation at Rancho Seco, and (3) the fire in the cable spreading room at Browns Ferry 1.
This report was released as a progress report with the expectation that some conclusions may need to be changed as the report undergoes continuing peer review and public comment.
This information relates directly to issues on the probability of accidents for nuclear power reactors.
Since it estimates the probability to be much higher than past studies, it may put a different light on the issue. I would like to point out that this document does not represent an NRC position concerning the probability of accidents, and is being furnished to you as background information.
The report is a contractor's assessment of the past (1969-1979) probabilities based on the events and plant characteristics which existed during that period.
Approximately 85% of the total risk cited in the study should be mitigated by the NRC required modifications resulting from.the events at TMI-2, Browns Ferry, Crystal River and Oconee 3. (I.E.Bulletin 79-27).
In addition, the upgrading of the Auxiliary Feedwater System reliability and other generic improvements should further reduce the total risk. The attached information refers to the future work in this area.
Due to the size of this report ( 3" thick) only summary infonnation is being provided.
Copies of the report are available as noted in the attachment.
Original signed ty :
Thenas M. Evak 8208090328 e20721 Thomas M. Novak, Assistant Director
{DRADOCK 05000312 for Operating Reactors POR Division of Licensing rncinenro. Ac 9tated f.,
OFFICE )
.g.g,g...ggg.pgg.t Ve*
k sunumE )
t" DaTE )
NRC FORM 318 (10-80) NRCM Cao OFFICIAL RECORD COPY
\\
Sacramento Municipal Utility Rancho Seco, Docket No. 50-312 District l
cc w/ enclosure (s):
l
. David S. Kaplan, Secretary and Christopher Ellison, Esq.
General Counsel Dian Grueuich, Esq.
Sacramento Municipal Utility California Energy Commissiori District 1111 Howe Afenue 6201 S Street Sacramento, California 95825 P.
0>. Box 15830 Sacramento, California 95813 Ms. Eleanor Schwartz California State Office 600 Pennsylvania Avenue, S.E., Rm. 201 Sacramento County Board of Supervisors Washington, D. C.
20003 827 7th Street, Room 424 Sacramento, California 95814 Docketing and Service Section Office of the Secretary U.S. Nuclear Reg'llatory Comission Business and Municipal Department Sacramento City-County Library Washington, D. C.
20555 828 I Street Resident Inspector / Rancho Seco Sacramento, California 95814 c/o U. S. N. R. C.
14410 Twin Cities Road Herald, CA 95638 Dr. Richard F. Cole Atomic Safety & Licensing Board Panel U.S. Nuclear Regulatory Comission Washington, D. C.
20555 Regional Radiation Eepresentative EPA Recion IX Mr. Frederick J. Shon 215 Fremont Street Atomic Safety and Licensing Board San Francisco, California 94111 Panel U.S. Nuclear Regulatory Comission Mr. Robert B. Borsum Washington, D. C.
20555 Babcock & Wilcox Nuclear Power Generation Division Elizabeth S. Bowers, Esq.
Suite 220, 7910 Woodmont Avenue Chairman, Atomic Safety and Bethesda, Maryland 20814 Licensing Board Panel U.S. Nuclear Regulatory Comission Thomas Baxter, Esq.
Washington, D. C.
20555 Shaw, Pittman, Potts & Trowbridge 1800 M Street, N.W.
Washington, D. C.
20036 Herbert H. Brown, Esq.
Lawrence Coe Lanpher, Esq.
Hill, Christopher and Phillips, P.C.
1900 M Street, N.W.
Atomic Safety and Licensing Board Washington, D. C.
20036 Panel U.S. Nuclear Regulatory Commission Helen Hubbard Washington, D. C.
20555 P. O. Box 63 Sunol, California 94586 4
e Sacramento Municipal Utility District I
Atomic Safety and Licensing Appeal Mr. Robert H. Engelken, Regional Administratot Board Panel U. S. Nuclear Regulatory Commission, Region V U.S. Nuclear Regulatory Commission 1990 N. California Boulevard, Suite 202 Washington, D. C.
20555 Walnut Creek, Cali fornia 94596 Alad S. Rosenthal, Chairman Atomic Safety and Licensing Appeal Board U. S. Nuclear Regulatory Commission Washington, D. C.
20555 Dr. Jchn H. Buck Atomic Safety and Licensing Appeal Board U. S. Nuclear Regulatory Commission Washington, D. C.
20555 Christine N. Kohl Atomic Safety and Licensing Appeal Board U. S. Nuclear Regulatory Commission Washington, D. C.
20555 California Department of Health ATTN: Chief, Environmental Radiation Control Unit Radiological Health Section 714 P Street, Room 498 Sacramento, California 95814 i
i T
l
!I
l
> = - -
NUREG/ CR,-2497
. y.
, (. - - -..-.
ORNL/ NSIC-182/V1
~ Dist. Category AN, RG', IA Contract No. W-7405-eng-26 Engineering Technology Division PRECURSORS TD POTENTIAL SEVEE CORE DAMAGE ACCIDENTS: 1969-1979 A STATUS REPORT J. W. Minarick C. A. Enkielka Science Applications, Inc.
Manuscript Completed - April 1982
~
Date Published - June 1982 Prepared for the U.S. Nuclear Regulatory Commission Of fice of Nuclear Regulatory Rese' rch a
Under Interagency Agreements DOE 40-551-75 and 40-552-75 1
NRC FIN No. B1583 Prepared by the Oak Ridge National Laboratory Oak Ridge, Tennessee 37830 operated by UNION CARBIDE CORPORATION for the DEPARTMENT OF ENEGY
REPORT
SUMMARY
~.
The Accident Sequence Precursor study involves the re.v4ew, of Licensam.. -.
7 Event Reports of operational events that have occurred at light-water power reactors to' identify and categorize precursors to potentially sig-nificant accident sequences.
Accident sequences considered in the study are those that could lead to severe core damage.
Accident sequence pre-cursots of interest are events that are important elements in a chain of events (an accident sequence) possibly leading to core damage.
Such pr e-cursors might be infrequent initiating events or equipment f ailures that, when coupl ed with one or more postul ated event s, could result in a' plant condition leading to severe core damage.
A nuclear plant has sa f e ty sy st e= equipnent for =itigating accidents er off-normal initiating events that =ay occur during the course of plant operation.
These saf ety systems are buil t to high quality and are redun-dant; none thel e s s, they have a definite probability of failing or being in a f ailed state when required to operate.
This report uses LERs and other plant data to calculate the unavailability of plant s a f e ty ry s t e= s.
It then uses these calculated safety system unavailabilities and the expected average frequency of initiating events (los s of f e edwater, loss of offsite pow er, l o s s-of-c ool ant accidents, and steam line breaks, al so determined when possible f rom the precursors) to evaluate the end results of saf ety system unavailability for two situations:
1.
Safety sy s t em f ailures without initiating events.
Given an LER-reported f ailure of a saf ety system or partial f ailures in several sy s t em s, the report use s expected initiating event occurrence rates to deter =ine the nu=ber of initiating events that will challenge the f ailed and backup saf ety systems during the period the s af e ty gr st em is f ail ed.
It multiplies the challenges by system failure probabilities, using event l
trees to evaluate the likelihood of the overall event sequence occurring.
2.
Initiatine event occurrences.
Although standby saf ety systems are ideally alway s avail abl e, there is a statistical probability that these systems will fail when called on to mitigate czpected accident or transient initiating event s.
Therefore, the report calculates the like-lihood of severe core da= age occurrence for each LER-reported initiating event based on expected response (f ailure probabilities) of the saf ety sy s t em s.
Failed or degraded safety rystems' existing at the time of the initiating event are accounted for in the calculations.
The study ef fort has been divided into several
- t. asks, which are de-scribed in detail in later sections of this report.
These tasks include l
(1) selection of LERs for detailed review as precursors; (2) in-depth re-l view of those LERs; (3) identification, description, and categorization of events considered to be precursors; (4) selection of precursors consider'ed significant; and (5) subsequent analysis of the precursors to determine if any trends or unique relationships exist among them.
For this study, LER ev ent s w er e selected as precursors if they met, one of the f o11 ewing require =ents:
1.
The event involved the f ailure of at least one function required to mitiga te an initiating event of interest.
l l
ri
~
- )
2.
The event involved the degradation of more than one func. tion required
~~~
l, to mitigate an initiating event of interest.
3.
The event involved an unusual actual initiating event (-e..g.,ra -to t a l- -
l' stuck-open primary relief valve, or another loss of off site power, a i
infrequent event).
Approximately 19,400 LERs concerning events that occurred during 196P 1979 were screened for accident sequence precursors according to the L
above requirements.
Of these, over 500 LER: (~3%) were selected for de-tailed review.
All LE2s selected for detailed review were subj ected to an in-depth ev al ua ti on, which included t
1.
a review of the accident sequence (if there was one) as described in the LER, j.
2.
a review of the de sign of systems in the reactor plant reporting the LER to determine the impact of the f ailure on the operation of these sy s t em s, and 3.
a review of the plant accident analyses to determine the extent to which aff ected systems would be required to f unction f or dif f erent of f-normal and accident conditions.
As a result of this detailed review, 169 events were selected as ac-cident sequence precursors.
For each of these events, four items were prepared:
a sheet describing the event, a categorization sheet including event-and reactor-specific inf ormation used in subsequent a naly s e s, and two event trees. The first event tree de scribe s the actual occurrence as reported in the LER and identifies the potential for severe core damage stemming from the actual event. The second event tree describes a postu-lated sequence of events that conid have been aff ected by the actual re-ported failures.
A set of these four items for each of the 169 events is included in Appendix B.
The f ailure information contained in the precursors was used to esti-mate initiating event f requencies and function f ailure-on-demand probabil-ities.
This inf ormation was used, in conj unction with the precursor event trees, to determine a measure of the probability of severe core damage associated with each event sequence.
This probability is an estimate of the chance of severe core damage given the precursor event occurred in the manner it did.
These probability measures were then used to rank the pre-Fif ty-two precursors with probability measures of 110-8 were cursors.
selected as significant.
The probabilities of severo core damage associated with the precur-sors were also used to estimate the frequency of severe core damage per reactor year for the years 1969-1979. This point estimate is be tween 1.7 x 10-8 and 4.5 x 10-8 per reactor year and includes contributions from three maj or event s:
(1) the loss of f eedwater and stuck-open relief valve at Three Mile Island Unit 2 (which actually resulted.in severe core dam-age), (2) the loss of nonnuclear instrumentation at Rancho Seco, and (3) the fire in the cable spreading room at Browns Ferry 1.
These nu=bers are compared with other estimates f rom FRAs and f rom the TWI-2 event alone in Fig. 1.
xii
1 w
CRNL.oWG82-5603 ETD -
2 UPPER 95%
pogg5an p
7
--y-CONFIDENCE f~
10 '
BOUND 5
ASP PO:NT ESTIMATE j
BETWEEN THESE VALUES <
I c-(SEVERE CORE DAMAGE)
O 2
I
/
,e
=
ESTIMATE SASED ON
>.,g_2 TMI-2 EVENT (SEVERE Ey CORE DAMA3E!
3*
5
~
AIF TASK FORCE IL#
~ ^
GERMAN RISK FUEL MELT)
STUDY (COR E 10 "
M E LT) g LOWER 95%
POISSON CCNFIDENCE 5
g WASH-1400 SOUND ICORE MELT) 4 F
Fig. 1.
Comparison of. ASP results with other core damage estimates, Subsequent analyses of the information included in the selected pre-i.
cursors resulted in the folicwing additional conclusions:
1.
Many of the initiating event frequencies and function f ailure-on-demand probabilities developed from operational event information agree reasonably well (within a f actor of 10) with the Recetor Safety Sc:dy2 median results.
2.
A variation in the rate of occurrence of significant precursors per plant as a function of plant age cannot be justified.
3.
Differences do not appear to exist in the number of significant pre-cursors observed between plant types and among reac cor vendors, archi '
t e c t!-e n g ine e r s, and plant power ratings.
4.
Approximately 38% of all significant precursors involved human error."
These analyse s did not involve extre=e statistical sophistication b'ut were first a tte= pts to de te rmine if trends were discernible in the se-1ected events.
Changes =ade in reactor plant operation after the TMI-2 rili
1 accident (particularly the potential use of high pressure inj ection fol-loving auxiliary feedwater-system failure and the ability to proyide. flow...__.
from at least one anziliary feedwater pnmp during a loss of ac power in expect'ed to reduce this estimate considerably in later years.
PTRs) are For reference, highlights of this study are summarized in Table 1.
Table 1.
Accident Sequence Precursor study highlights Period covered 1969-1979 Total number of LERs searched 19,400 Number selected for detailed review 529 Nc=ber selected as precursors 169 Nc=ber of significant events 52 l
A point e s t ima te of the frequency of severe core damage calculated from precursor inf ormation for the years 1969-1979 lies be tween 1.7 x 10-8 l
and 4.5 x 10-8 per reactor year.
Reasonable agreement exists be tween ASP and Recctar Safety Study initiat-ing event f requencies and function f ailure probabilities.
No variation with plant age can be demcastrated in the number of signifi-cant events.
No apparent dif f erence s exist between plant type s and among vendors, architect-engineers, and plant power ratings.
E Reference 1.
U.S. Nuclear Regulatory Commission, Recctor Safety Study: An Assessment of Accident Risks in U.S. Commerciat Suctect Pouer Plante, WASH-1400 (NUREG-75/014) (October 1975).
O 2
xiv
- l 2
(
1 1
i h
I f
PRECURSORS,TD _ POTENTIAL SEVERE CORE
- e r
m -
DAMAGE ACCIDENTS: 1969-1979 A STATUS REPORT J. W. Minarick*
C. A. Kukiella*
e.
i ABSTRACT De scriptions of 169 operational events reported as Licensee Event Reports, which occurred at commercial light-water reactors during 1969-1979 and which are considered to be precursors to potential severe core damage, are presented, along with associ-ated event trees and categorizations and subsequent analyses.
The report su=marizes work in (1) the development of methods used to screen -19,400 LER abstracts f or potential precursors, (2) the initial screening of those abstracts to determine which should be reviewed in detail, (3) the detailed review of those selected LERs that yicided the 169 events, (4) the categoriz a-tion of the 169 events, (5) the calculation of function f ailure estimates based on precursor data, (6) the use of probability of severe core damage estimates to rank precursor events and esti-mate the f requency of severe core damage, (7) the identification l
of 52 event s considered significant, (8) trends analyses of those significant events, and (9) the identification of the other events of interest that occurred within 1 month of signif-icant events.
l l
1.
INIRODUCTION t
ne Accident Sequence Precursor study involves the review of Licenser Event Reports of operational events that have occurred at light-water pow-er reactors between 1969 and 1981 to identify and categorize precursors to potential severe core damage accident sequences. This progress report details this ef fort f or 1969-1979 LERs. Although Licensee Event Reports were not required until mid-1975, event reports comparable to LER: existed bef ore the inception of' the LER system and are considered to be LERs for the purpose of this study.
[The requirements of Licensee Event Reports are described in Regulatory Guide.l.16 (Ref.1).]
York on the ASP study began at the Nuclear Saf ety Information Center on June 15,1979, in re-sponse to FT-1979 Nuclear Regulatory Research Order 60-7 9-185, " Ac ci d e nt -
Sequence Precursor Study" dated June 7,1979, and subsequent orders.
ne program was initiated, in part, because of coni:1usions contained in the Risk Assessment Revieu Croup Report.* This report states "that l
unidentified event sequences significant to risk might contribute.
- .a I
- Science Applications, Inc.
4 4
e 1-1 o.---
y 7-u-y
_ + - -
.-.m-
,--,--,-,,,-..,y-w
l 1-2 w --
~~~
" ~
small increment
[to the overall risk]."*The report recomm' ends:
"It is important, in our v.iew,_ that potentially,significant (accident) se-quences, and precursors, as they occur, be subj ected to the kind of analGr '" ~ ~
7 sis contained ir TASE-1400 [Ref. 3].
Accident sequences considered in the study are those that could lead to severe core damage. Accident sequence precursors of' interest are events that are important element s in a chain of event s (an accident se-quence) possibly leading to core damage.
Such ~ precursors could be infre-quent initiating events or equipment f ailures that when coupled with one or more postulated events, could result in a plant condition leading to severe core da= age.
Note that the results achieved in this repcrt have been obtained basti on events reported in LERs and subsequently selected as precursors.
Because of the use of LERs, biases may have been introduced as a result of differences in plant technical specifications and approaches to LER re-porting and of changes in LER reporting requirements over the period of the study.
These considerations may result in the f ailure to include cer-tain events that under dif f erent ciren= stances would have been selected for inclusion.
Hew ev e r, the events selected were more serious than most, and it is expected that most of these would have been reported indepen-dently of small differences in reporting requirements.
The ASP study ef fort has been divided into the following tasks:
1.
selection of LERs deserving a detailed review as precursors; 2.
detailed review of selected LERs; 3.
identification, description, and categorization of events considered accident sequence precursors; 4.
selection of precursors that are considered significant; and 5.
analysis of precursors to determine if any trends or unique' relation-ships exist.
These tasks are described in detail in the f ollowing sections.
References 1.
U.S. Nuclear Regulatory Co= mission, Regulatory Guide 1.16, Reporting of Operating Information, Appendim A: Technicct Specifications, Rev. 4 (August 1975).
2.
U.S. Nuclear Regulatory Co==ission, Risk Assessment Revies Group Report, p. 15, NUREG/CR-0400 ( Septe=ber 197 8).
3.
U.S. Nuclear Regulatory Co= mission, Recc:or Sqfety Study: An A,sassa-ment of Accident Rieka in U.S. Commercial Nuctect Pouer Ficnte, TASB-1400 (NUREG-75/014) (October 1975).
y
m
POREWORD
.--- ~..-.
This report presents the initial results of a program that was begun as a result of one' of the Lewis Committee recommendations following their review of WASH-1400, the Reactor Safety Study.
One of the committee's review findings was that more use should be made of operational data to assess' the risk f rom nuclear power plants.
The Precursor Program, per-formed at Oak Ridge National Laboratory and administered by the Nuclear Regulatory Commission, responds to this Lewis committee finding. Th e Pr e-cursor' Program uses Licensee Event Reports to evalnate potential nuclear plant accident precursors occurring at operating reactors.
These individ-ual plant precursors are then su=marized to evaluate the risk (for a par-ticular tL=e period) from all operating nuclear power plants. This r e-port, covering 196 9-197 9 LERs, is being released as a progress report with the expectation that some conclusions may need to be changed as the report i
undergoes continuing peer review and public comment. The next report (using 1980-1981 LER data) should reflect the rist from nuclear plants since the TMI-2 accident and may show what effects new procedures and equipment modifications (lessons learned) have had.
'In addition to the documentation of 169 identified precursors and preliminary trends analyses, the report estimates the frequency of severe core dosage based on the precursor information.
It is a difficult problem
- to derive a credible probability for severe core damage using limited operational experience data f rom plants that have many significant phys-ical and operational dif ferences among them. The authors of this report partially account for plant dif ference s by using generalized (functional) event trees for individual precursor evaluation, which in their quantifi-ca tion are then specialized, as much as possible, to the particular plant.
Nonetheless, simplified methods are used to determine and quantify severe core damage precursors.
Several aspects of this report are expected to i
affect the calculated results, either conservatively or nonconservatively.
The first two of the following items are expected to introduce a conserva-tive and nonconservative bias, respectively. The remaining items may in-troduce either conservative or nonconservative biases.
}
The probability of subsequent core da= age given the precursor may be e
conservative in some cases.
)
e The LER screening process may have overlooked precursors that should have been included.
The accuracy and completeness of the LERs in reficcting pertinent oper-e t
ational f ailure or initiating event s is somewhat que stionable.
1
)
e The event trees used for most precursors are generic and may not ade-i quately reflect dif ferences among plants.
Average or generic data are combined with plant-specific operational e
occurrences in calculating the probability of subsequent severe core d ama g e.
The repair (recovery) credit for system f ailure involves engineering e
judgment.
I e The method used to calculate the f requency of severe core damage is subj ect to various interpretations because of the combined use of event v
--t
-w e
4 I
j statistics and generic initiating event and function failure probabil-ities.
=. -
'Ib e use of LERs to attempt to extract severe core damage probabilities, on scale and to the detail su'ch as done in this re' port, is Enlque 'Th I ' ~ ~~
~
a full meaning and luitations of the severe core damage calenlations made in this report are not clear.
It is felt, how ev er, that the report pro-
~
vides valuable information that can help validate or supp1ement probabil-istic risk asse ssments performed on nuclear power plants.
Much of the basic data and information needed for additional ' calculations or infer-ences of reactor risk by the reader is included in the report.
As men-tiened, this report will be f ollowed by other reports that will evalua te LERs in the 198(r-1981 ti=e period and also will provide further analysis, refine =ent, and practical use of the basic data contained within this re-port.
Reader com=ents and suggestions are earnestly solicited and should be sent to the Chief, Reactor Risk Branch, Division of Risk Analysis, at the address below.
R. M. Bernero, Director Division of Risk Analysis Office of Nuclear. Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555 3
a 9
9 9
vi
'~
PREFACE
)
The work reported here was undertaken by the' Nuclear Operations Anal- ~. -.
l ysis Center at Oat Ridge National Laboratory on behalf of the Division of The FT-1982 NRC tech-i Risk Analysis of the Nuclear Regulatory Co= mission.
I nical monitor, F. M. Kanning, succeeded M. A. Taylor, who was technical f
monitor until his reassignment in Nov ember 1981. The work on accident sequence precursors was initiated early in 1979 with R. L. Scott as pro-assisted by the authors of this report. How ev er, when Scott ject manager, was reassigned in July 1980, J. W. Minarick became acting manager and has retained that role since.
Both Minarick and C. A. Kulkielka are Science Applications, Inc., e=pl oy e e s and perf ormed their work under subcontract to NOAC. Most of the work was perf ormed at NOAC of fiess because of the availability of relevent documents and technical support.
The training and background of these authors well qualified them for the task.
I Minarick, an electrical engineer, has had 12 years of reactor systems ex-perience, including 5 years on Admiral H. G. Rickover's staff and 3.5 Kulkielka, who receiv ed his M.S.
years with Babcock & Yilcox Company.
degree in nuclear engineering in 1979, had 2 years' prior experience with l.
the U.S. Ar=y nuclear program bef ore j oining SAI in 1979.
This status report covers the first 2.5 years of ef fort.
The work involved (1) development of selection criteria for the identification of reactor events that are precursors of potential severe core damage',
1 those (2) application of these criteria against all the licensee event reports that have been received since 1969, and (3) detailed analyses of the se-This report covers the work co=picted for LERs submitted i
f lected events.
during the 11 year period from 1969 to 1979. Although the NRC has pre-it has viously reviewed the selection criteria and the events selected, not been directly involved in the application of these criteria against This tast has been perf ormed entirely by the NOAC the existing LERs.
staff, using its best judgment in doing so.
While this j udgment reflect s many years of experience in reactor design, reactor operations, and sys-te= s ev alua tions, the proce ss is subj ective, and not all specialists will necessarily agree with every event selected and/or omitted.
This report deals only with historical data and, at this point, with minimal statistical interpretation.
The TMI-2 accident is responsible for ll Yet, could about half the core damage f requency value estimated herein.
l!
say given the conditions of early 1979 - that the frequency of a at that time was once every 4 years or once every 100 one f
IXI typ e accident same selection criteria that were used in this i;
years? Furthermore, the study would also have been applicable had the top event been severe fuel ll cladding f ailure, severe core damage, or core meltdown.
In any event, th e design and operation ma=y changes that have occurred in nuclear plant l'll since 1979 are expected to substantially reduce the future probability of l
all such events.
Continuing work on this program is expected to include:
f of the uncertainty in the core damage probability calcu-e an assesament that TMI-2 has besn lations-(a simplified approach, based on the f act
^
core damage statistic, indicates the report estimate the only true l
or two could be too low by a f actor of 2 to 3 or too large by one crders of magnitude) and vii lI L-=
i calculation of the probability of severe core damage accidents based e a i~
on the ~7500 LERs submitted in 1980 and 1981.'
' '~
Inevitably, the resul ts -of this report will be compa. red.wi,th the -deta -
in the Reactor Safety Study (WASH-1400) and other probabiliatic risk as-sessment studies.
Although the casual reader may interpret the Accident Sequence Precursor study results as inco=patible with other core damage estimates, it is quite likely that because of the statistical uncertainty, no significant difference exists.
That, of course, remains to be demon-strated.
In conclusion, I direct your attention to the various trends a naly s e s l
included in this report.
Al though the statistical precirion is not great, i
3 the trends are of considerable interest.
In any event, th e r e s ul t s pr e-sented here indicate how very important it is that the operating experi-ence be analyzed f or trends that a more casual surveillance of such exper-ience might not reveal.
Ym. B. Cottrell, Director l-Nuclear Operations Analysis Center P.O. Box Y Oak Ridge, IN 37839 l
e i
=
l l
l l
O 1
i l
l viii 1
l l
1
'S f
3
'S
} -
e h,
i d
1 y
Printed in the Unsted States of America. Available from 5
National Technical information Service
.i U.S. Department of Commerce
?
5285 Port Royal Roac, Springfield, Virginia 22161 Available from GPO Sales Program Division of Technical Information and Document Control U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Tnis recc t was preparec as an account of work sponsorec by an agency of the Unce: States G overnment Neither the Unitec States Government nor any agency theieof. no' any of tneir ernp'oyees, makes any warranty, express or impliec, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information. apparatus. product, or precess casCICsec, or represents tnat its use wouic not infringe privately ownec rights. A eference herein io any specific comtvercial procuct. p rocess, or service ey trace name, trademark, manufacturer, or otnerwise. does not necessarily constitute or impty its encorsement. recemmencation. or favoring by tne Unitec States Government or a9y agency therect The views anc opinions of autners expressec nerein co not necessardy state or reflect those ofine unitec States Government or any agency thereof 1
1
.. = -.. - - -
-.~..
=
CONTENTS
- j c.
-Pare -
- {
FOREWORD y
PREFACE vil LIST OF ACRONYMS AND INITIALISMS iz
~!
REPORT
SUMMARY
xi ABSTRACI 1-1 1.
IN'IEODU CIION 1-1 References 1-2 2.
CRITERIA FOR SELECIION OF LERs FOR DETAILED REVIEW AS PRECURSORS 2-1 Reference 2-2 3.
DETAILED REVIEW AND IDENTIFICATION OF SELECTED LERs AS ACCID %T SEQUENCE PRECURSORS 3-1 Reference 3-15 4.
QUANTIFICATION OF PRECURSORS AND IDENTIFICATION OF
'IHOSE CUNSIDERED SIGNIFICANT 4-1 4.1 De termina tion of Initiating Event Frequencies and Fur.ction Failure Probabilities 4-1 4.2 Calculation of Probability Mea sures Associated with Each Precursor 4-2 4.3 Selection of Significant Procui sors 4-23 5.
ANALYSIS OF TRENDS IN SELECIED EVENTS 5-1 5.1 Initiating Event Frequency and Function Failure Probability Comparisons 5.................................
5.2 Instantaneous Failure Rate Trends 5-2 5.3 Time Lines for Initiating Events and Function Fail ur e s 5-4 5.4 Variation in the Number of Significant Precursors as a Function of Plant Age 5-4
~
5.5 Potentia 1 Dif f erences Between Plant Types and Among Vendors, Ar chite c t-Eng ine e r s, and Plant Power Ratings 5-7 5.6 Identification of.De graded-Function Events that Occurred Yithin 1 Month of Each Significant Precursor 5-7 5.7 Percentage of Precursors Involving Human Error 5.-16 5.8 Estimation of the Probability of a Diesel Generator Failing to Start Given a Loss-of-Of f site-Power Demand 5-18 References 5-19 iii
,r-m
b
...L.
a~~
- f
'M
- 9
!4 APPENDIX A.
STANDARDIZED EVENT IREES..~..'........................
A-1
.w w
' d APPENDII B.
PRECURSOR SU.LCHARY SHEETS AND EVDiT IREES-
..... f.'. ;T.? BJP 4
Sj APPENDII C.
. SUPPORT MATERIAL FOR SELECTION OF If POTENTIAL PRECURSORS *IHAT ARE CONSIDERED N
SIGNIFICANT 6
C-1
=
APPENDII D.
SUPPORT MATERIAL FOR TRENDS ANALYSIS -
TIME ON TEST PLOTS D-1 E
APPENDII E.
ADDITIONAL SUPPORT MATERIAL FOR TRENDS ANALYSIS 17:
E-1
.j.'
M.
E E1.
h.~._
r
?
%d ON D>
~t*
z+
~
~
-4 Jqi
?-
- .9 r~
El, e
e@
c siY 4I
( Eu a
C 7
%d 7:
- 1
&d iv i
i
._