ML20203J769
| ML20203J769 | |
| Person / Time | |
|---|---|
| Issue date: | 07/30/1986 |
| From: | Grimsley D NRC OFFICE OF ADMINISTRATION (ADM) |
| To: | Sholly S MHB TECHNICAL ASSOCIATES |
| Shared Package | |
| ML20197F561 | List: |
| References | |
| FOIA-86-504 NUDOCS 8608050419 | |
| Download: ML20203J769 (3) | |
Text
l fDR-OIfo UNITED STATES
[p Kfc,]o o
NUCLEAR REGULATORY COMMISSION
[
W ASHINGTON, D.
C.
20555 o
q O
u k....+ #
JUL 30 1986 Mr. Steven C.
Shvily Associate Consultant MHB Technical Associates 1723 Hamilt,on Avenue, Suite K IN RESPONSE REFER can Jose, CA 95125 TO FOIA-86-504
Dear Mr. Sholly:
This is in regard to your request, pursuant to the Freedom of Information Act, to which the NRC assigned the above number.
This is a partial response to your requeet.
We will notify you upon completion of search for and review of any additional records subject to your request.
_X__
The staff has completed the search for and review of records subject to your request, and thic is the final response to your request.
The NRC has no records subject to your request.
_. X ___
Eecords subject to your request are available for public inspection and copying at the NRC Public Document Room (PDR). 1717 H Street, NW, Washington, DC 20555, as noted on the enclosure (s).
The PDR accession number is identified beside each record description.
,v__
Records subject to your request are being made available for public inspection and copying at the NRC Public Document Room (PDR), 1717 H Street, NW, Washington, DC 20555, in the PDR file folder under the above number and your name.
These records are listed on the ericlosure(s).
We are enclosing a notice that provides information about charges and procedures for obtaining records from the PDR.
Sincerely,
/
l Dc ae H.
Grimsley, Director j
Division of Rules and Records i
Office of Administration l
l Enclosure (s):
As stated l
86o8000419 860730 PDR
%tt SO4
e T
Re:
F01A-86-504 APPENDIX A 1.
05/31/83 NUREG/CR-2800- Supplement 1 -
"Prioritization Information Development,"
(see write-up for Issue C-8)(235 pages) Accession No.
8306020131 2.
12/31/83 NUREG-0933 "Prioritization of Generic Safety Issues,"
(see individual write-ups for (a) Issue C-8 BWR Leakage, 4
(b) Issue II.E. 4.3 - Containment Integrity Check and (c) Issue II.E.4.4, (4) - Containment Purge-(777 pages)
Accession No. 8401230657 s
e i
I
. - ~. _ _ _ _ - - - - -
-. ~.
APPENDIX B FOIA-86-504 1
1.
04/17/81 Memo from M. Ernst to L. Rubenstein re : Proposed Position Regarding Containment Purge / Vent Systems 2.
07/02/85 Memo from W. Dircks to the Commission, subject:
Staff Responses to Commission Questions Regarding loss of Containment Isolation l
i
-s JUL 0 2 885 JUN 3 01986 MEN FANDUM FOR:
Chairman Palladino Comissioner Roberts Commissioner Asselstine Commissioner Bernthal Comissioner Zech FRCH:
William J. Dircks ExecutiveDirectorforiperations
SUBJECT:
STAFF RESPONSES TO COMMISSION QUESTIONS REGARDING LOSS OF CONTAINMENT ISOLATION In a inemorandum dated April 23, 1985, from John C. Hoyle, SECY, to William J. Dircks, EDO, responses were, requested to several questions asked by the Comission related to the issue of loss of containment isolation.
Responses to each of those questions are enclosed.
O [ntt Mitiam ),y; William J. Dircks Executive Director for Operations
Enclosure:
Respo..se to 4/23/85 Questions cc:
w/ enclosure SECY OPE 1
OGC l
l l
l I
l hrA-7i~u tub 3 f-
.22
/Y//
STAFF RESPONSE TO COMMISSION'S REQUEST REGARDING LOSS OF C NTAINMENT ISOLA OVESTION 1.
What has been the historical record of nuclear power plants for those situations when containment integrity was not r,aintained?
Characterize these occurrences by cause, nature, r
size and duration of the opening, and whether the vent path was into the auxiliary building or elsewhere.
~
RESPONSE.
In late 1963, the staff contracted with the Pacific Northwest 1.aboratory (PNL) to perform a reliibility analysis of containment isolation systems.
The primary sources of inforr.ation used by PNL in conducting the study were the y
licensee event reports (LERs) and the containment integrated leak rate tests (ILkis), reports.
Ine results of this study are published in NUREG/CR-4220,
~
Reliability Analysis of Containment Isolation Systems.
A review of the PNL report indicates that there are four major types of containment isolation failure.
Only those events that appear to have potential for direct release paths to the environs are included in the following discussion.
Type 1 involves containment isolation valves inadvertently left c;en during power cperation.
The PNL data base contains a large number of single valve failures but the failure of core than one valve is required to breach containment integrity.
However, there are two events reported where such a breach of containment integrity occurred.
Or.ewasreportecforOc$ nee 1in 1973, where three isolation vahes were lef t open on the 6-incn :LRT
QUESTION 1. (continuec) pressurization line.
The other event was reported for Palisades in 1979, where two valves were left open on the 3-inch purge bypass line.
The cause of the events were primarily procedural breakdown and operator error.
The exact duration of these events is not known bbt is estimated to be approximately 12 to 18 months for each event.
Type 2 involves holes being drilled through the contai cent liner in conjunction with certain activities without being detected until an ILRT was performed.
San Onofre 1 reported the detection of a hole during its 1977 ILRT and Surry 1 reported the detection of several holes during its 1980 ILRT.
The causes of these events can only be attributed to procedural error.
The exact duration is not known but is estimated to be approximately one year.
The sile= mi t.e hales were less than an inch in diameter.
d' Type 3 involves excessive leakage through either containment penetrations or centain:r.ent isolation valves in series.
The PNL data base contains a large nunber of events of excessive leakage.
Due to insufficient reported information, it is difficult to conclude h'w many of the events would have resulted in a o
breach of containment integrity.
Since most of the leakage rates exceeded the casuring capacity of instruments, the exact size of cost leak paths are not known.
The causes of such events are usually seal or seat deterioration and/or foreign material contamination.
In 1982, the Office of Inspection and Enforcament issued an Information Notice in which an extensive review was performed on the leak sge problem of MSIVs in BWR plants.
Numerous events of excessive leakage of both MSIVs in series were found.
I
QUESTION 1. (continued) 3-Type 4 involves containment air locks where both doors were kept open during plant operation as a result of inproper interlock design and/or operation.
A total of 75 such events were found in the PNL data base.
The duration of the :4ents vary from a few inutes to a few cays.
Although the cpening is unquestionably large, the duration is normally very short ccmpared 3
to the total reactor operating years.
m O
J l
r F
t l
l L
I OUESTION 2.
What is the staff's position allowing containment purge / vent while the plant is operating?
r.
RESPONSE.
The staff position on containment purging / venting has evolved over the past ten But, for the past several years, the main conditions of the position years.
have remained unchanged.
The elements of the current staff position being implemented on both NTOL and operating plants are as follows:
(1) Purging / venting should be minimized during plant operation since it is inherently safer to operate with ventilat!on system isolation valves
?
ciesea ratner than open since open lines require valve action to provice ccntainment integrity.
(2) Purging / venting will be permitted for safety related reasons only.
Purging / venting needs should be identified and justified.
Safety related reasons previously accepted include pressure / vacuum control, unscheduled maintenance on safety related systems and inerting/deinerting of BWR plants.
(3) The smallest purge / vent system should be used whenever possible, r
i
Q'JESTION 2. (continued) (4) Purging / vent system isolation valves must be qualified to close under LOCA conditions.
Unqualified valves must be locked closed.
(5) The amount of radioactivity released'during the time required to close the system isolation valves should be shown by analysis not to cause the total offsite dose to exceed 10 CFR Part 100 guidelines for alternatively do not perform the dose analysis and limit the total purge / vent time to 90 hours0.00104 days <br />0.025 hours <br />1.488095e-4 weeks <br />3.4245e-5 months <br /> / year.
(6) Purging should not be done for either temperature or humidity control.
(7) An estimate or goal of yearly purge / vent time should be established.
1 SP I
e l
6-QUESTION 3.
To what extent have the licensees complied with the staff's purge position?
RESPONSE.
I s
The industry has, for the most part, complied with the above staff position.
For lit 01.s, the staff position is implemented during the normal review process.
For the 65 plants with operating licenses in 1930, the review was conducted under Generic Issue B-24. 'A~t the present time, there are approximately 10 operating plants for which total resolution of the purge and vent activities has not been achieved.
All plants comply with the key element of the position, which is to limit the purge / vent activities for safety related reasons.
The
~
ramsir.ir.; piar.ts, hcwaver, have not as yet satisfactorily identified and justified the spe:ific. reasons for purging.
The staff is continuing its efforts to cbtain this~information.
l l
The yearly time for purge / vent operations in the operating plants varies l
widely due to fundamental plant layout differences.
EWR plants typically purge 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> per year cr less because the need to enter containment to perform t
i r.aintenance or surveillance is much less than for most PWRs.
PWRs have significantlymoresafetyrelatedequipmentlocatedinsidethepri$ary containment.
Therefore, the need to purge more frequently is greater.
This l
l l
QUESTION 3.
7-dependence is seen with the anticipated purging practices of the newer EWR Mark III containment design.
This design contains a substantial increase in safety related equipment within containment.
Purging times in the order of 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> are expected.
E There are, however, some plants that would not be able to continue to operate without purging almost continuously unless major modifications to the plant systems were made.
The staff is currently pursuing the costs and benefits associated with purge time reduction fo'r these selected plants.
Table I contains a summary of the past purge system usage for the older PWR cperating plants. 'It is worth noting that approximately two-thirds operate th: pur :/:.:r.t :y: tem less than 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> / year.
Also included in the table are the applicant's commitments for the recent NTOLs to limit purging.
Althcugh it would appear that the NTOLs have increased the average purging time, cne rust understand the difference between the data bases.
The NTOL data is extracted from the Technical Specifications of each plant.
As a result, these values can be considered as upper limits.ather than the actual plant usage values listed for the operating plants.
O
-g-TABLE 1 PURGE USAGE PWR CRY CENTAINMENTS (H0JRS/ YEAR)
DATE
<100 100-1000 1000-3000
>3000 1950 20-11 3
8 1981 25 6
2 9
STOLs 4
3 3
0
.g.
QUESTION 4 TheCommissionfurtherrequestedstaffthdevelopprobabilitiesonlossof containment isolation as a result of such factors as failure to close isolation valves, and/or the existence of unknown pre existing holes.
Specifically, the staff should provide analysis of and values for "S" (the WASH-1400 conditional probability of containment failure resulting from inadequate isolation of containment openings and penetrations) considering containment type and historical' development.
RESPONSE.
the PNL report ~in the response to Question 1 includes analysis of containeent isolation unavailability based on the LERs and ILRT reports reviewac by PNL.
Table 2 contains the summary results of this study of containment unavailability.
These estimates are dependent on several assurptions (particularly on event duration times) necessitated by the type of data available in the LERs and ILRT reports.
The cata reviewed by PNL revealed that most primary containment integrity problems for EWRs are most often related to valves failing to close or excetsive valve leakage.
Leakage past MSIVs was especially frequent (much rore than PWRs).
. l TABLE 2.
Summary Results for Containment Unavailability Unavailability leak Area A.
Maior Leakage Data Summary Airlocks 5x10 5 5000 sq inches large Leaks 10 3 to 10 2 2 to 28 sq inches B.
Mipor Leakage Data Sum.T.ary(a)
PWR ILRT Results 0.05 0.006 so inches (b) 0.12 0.06 0.07 0.60
~
EWR ILRT Results O.16 0.006 so inches 0.14 0.06 0.04 0.60 In acdition, PNL has conservatively estimated from the local leak rate test data of valves and penetrations for both PWR and EWR plants that containment leakage will exceed, by up to a factor of 10, the Technical Specification limits design pressure) appro(ximately 30 percent of the time. typically 0.1 to 1.0 vo
~
Ial[]hevaluesinpartBofTable2donotresultinleakageratesthatare considered risk significant, but do exceed Technical Specification, limits.
(b) Denotes that a leak rate corresponding to a 0.006 sq. inch opening existed for 5% of the ILRTs reviewed.
l i
O
QUESTION 4 (continued) For PWRs, personnel air lock leakage dominated the failure modes and accounted for approximately half of the LERs for PWRs.
Failure of one or both of the personnel air lock doors to close and latch were also reported frequently.
Excessive leakage through large purge isolation valves, often due to degradation of the seat material, was a significant type of valve failure at the PWRs.
Although most of the leak rates listed in Table 2 could be considered as excessive leakage based on Technical Specification limitations, analyses performed for the Reactor Safety Study (VASH-1400) have shown that severe consequences or significant effects on other systems only result from large leaks.
For cases where a preexisting hole did not exist, the WASH-1400 study of the Surry Power Plant utilized a leakage value equivalent to 10 times the Technical Specification limitation of 0.1 volume percent per day at design pressure throughout the radiological analysis until such time the containment integrity was lost by other threats such as overpressurization, hydrogen combustion, or basemat failure.
The overall risk was not sensitive or influenced in any significant way by this leakage value.
WASH-1400 also examined other preexisting leakage possibilities extending well beyond the 1% regime to deter.ine what size leak openings would have to exist so that 1) overpressure failure would be precluded, 2) other system effects or interactions (e.g., coolant losses or pump cavitation) could possibly occur,
1
=
)
QUESTION 4. (continued) and, 3) competition with fission product depletion, deposition, or scrubbing processes would occur.
The WASH-1400 analysts determined that, for Surry, a leak opening of about 4 inches in diameier (12.5 square inches), which is equivalent to approximately 200 volume percent per day at design pressure, would have competing effects with the natural deposition and depletion processes and with the threat of overpressure failure.
This leak size was then designated as S and it encompassed loss of containment integrity caused through either (1) failure of the containment isolation system to initially secure containment, integrity, (2) leakage resulting from systers not isolated
)
by the accident but secured sometime after operation during the accident, or p
(3) leakage resulting from missiles or other dynamic effects of the" accident.
tault tree logic was used to help explore and estimate the probability of a S leakage category existing during the accident.
For the BWR plant studied in VASR-1400, Peach Bottom 2, a similar exploration was carried out and a 1 inch diameter leak size (equivalent to appreximately 100 volume percent per day at design pressure) was determined to give rise to system effects.
l The containment isolation unavailability results for large leaks from Table 2, t
10 2 to 10 3, compare favorably with the loss of containment isolation values, 10 2 to 10 4, used in the WASH-1400 report and later PRAs.
It should be noted.that there may be a few plants that have higher containment isolation unavailabilities.
The PRA perforced by the licensee for the
QUESTION 4 (continued) 75 Ma'e Big Rock Point plant, for example, utilizes 6x10 2 as the unavailability value for containment isolation.
Containment isolation failure has been and continues to be considered in full scope PRAs.
In a large dry containment,the conditional failure probability attributed to significant leakage sizes has been considered small in ecmparison with containment failure modes attributable to overpressure failure due to non-condensible gas producti'on, hydrogen burns, or basemat melt through.
i b.
l l
l 4
O l
\\
JUL 0 21985 i
MEMORANDUM FOR:
Chairman Palladino Commissioner Roberts Commissioner Asselstine Commissioner Bernthal Commissioner Zech FROM:
William J. Dircks Executive Director for Operations
SUBJECT:
STAFF RESPONSES TO COMMISSION QUESTIONS REGARDING LOSS OF CONTAINMENT ISOLATION In a memorandum dated April 23, 1985, from John C. Hoyle, SECY, to William J. Dircks, EDO, responses were requested to several questions asked by the Commission related to the issue of loss of containment isolation.
Responses to each of those questions are enclosed.
l l
ODtG Miliam1.Dircks l
William J. Dircks Executive Director for Operations
Enclosure:
Distribution Response to 4/23/85 Questions Central file DEisenhut RRAB Rdg HDenton cc: w/ enclosure AD/T Rdg WDiecks l
SECY DST CHRON
~ WITS #85061 OPE '
ED0 Rdg Schilk #M850403A OGC MFields LMu11ey J_ields CHRON G o,.,_
.b',, G LD FCoffman pQ
Contact:
- va v // 2 01.3 Cf-AThadani A. Thadani, NRR FRowsome TSpeis 492-4705
[I Y f.
JKudrick WHouston RRAB: DST RRAB: DST CSB:DSI g
RBernero MFields vr AThadani JKudrick l
5/20/85*
5/20/85*
5/20/85*
SEE PREVIOUS CONCURRENCE g
omesp. AD/,RS,:DSI,,
, D : DS,,I,,
AD/T:D T 0: T,,
D.
D:tgR EDO[,(,#,
,,,,,Dirc ks W
1 HD@ ton m-o
..W.H,o,u,s,,t,on,,,
RBerne,ro FRowsome TSpe,s m,..
5a9.t.85*
5awa5*
5/.20/85*
59,5 Sgg,85 5/p.85
.atus.5
STAFF RESPONSE TO COMMISSION'S REQUEST REGARDING LOSS OF CONTAINMENT ISOLATION QUESTION 1.
What has been the historical record of nuclear power plants for those situations when containment integrity was not maintained? Characterize these occurrences by cause, nature, size and duration of the opening, and whether the vent path was into the auxiliary building or elsewhere.
RESPONSE.
In late 1983, the staff contracted with the Pacific Northwest Laboratory (PNL) to perform a reliability analysis of containment isolation systems.
The
~
primary sources of information used by PNL in conducting the study were the licensee event reports (LERs) and the containment integrated leak rate tests (ILRis) reports.
The results of this study are published in NUREG/CR-4220, l
Reliability Analysis of Containment Isolation Systems.
A review of the PNL report indicates that there are four major types of containment isolation failure. Only those events that appear to have potential for direct release paths to the environs are included in the following discussion.
Type 1 involves containment isolation valves inadvertently left open during power operation.
The PNL data base contains a large number of single valve failures but the failure of more than one valve is required to breach containment integrity.
However, there are two events reported where such a breach of containment integrity occurred.
One was reported for Oconee 1 in 1973, where three isolation valves were left open on the 6-inch ILRT
QUESTION 1. (continued) pressurization line.
The other event was reported for Palisades in 1979, where two valves were left open on the 3-inch purge bypass line.
The cause of the events were primarily procedural breakdown and operator error.
The exact duration of these events is not known but is estimated to be approximately 12 to 18 months for each event.
Type 2 involves holes being drilled through the containment liner in conjunction with certain activities without being detected until an ILRT was performed.
San Onofre 1 reported the detection of a hole during its 1977 ILRT and Surry 1 reported the detection of several holes during its 1980 ILRT.
The causes of these events can only be attributed to procedural error.
The exact duration is not known but is estimated to be approximately one year.
l Tha sizas of tha holes were less than an inch in diameter.
l Type 3 involves excessive leakage through either containment penetra.tions or containment isolation valves in series.
The PNL data base contains a large I
number of events of excessive leakage.
Due to insufficient reported information, it is diff,icult to conclude how many of the events would have resulted in a breach of containment integrity.
Since most of the leakage rates exceeded the measuring capacity of instruments, the exact size of most leak paths are not known. The causes of such events are usually seal or seat deterioration and/or 1
foreign material contamination.
In 1982, the Office of Inspection and Enforcement issued an Information Notice in which an extensive review was performed on the leakage problem of MSIVs in BWR plants.
Numerous events of excessive leakage of both MSIVs in series were found.
QUESTION 1. (continued) ;
Type 4 involves containment air locks where bot'h doors were kept open during plant operation as a result of inproper interlock design and/or operation. A total of 75 such events were found in the PNL data base.
The duration of the events vary from a few minutes to a few days.
Although the opening is unquestionably large, the duration is normally very short compared l
to the total reactor operating years.
4 l
1 f
I f
e
F QUESTION 2.
What is the staff's position allowing containment purge / vent while the plant is operating?
RESPONSE.
h The staff position on containment purging / venting has evolved over the past ten years.
But, for the past several years, the main conditions of the position have remained unchanged.
The elements of the current staff position being implemented on both NT0L and operating plants are as follows:
(1) Purging / venting should be minimized during plant operation since it is inherently safer to operate with ventilation system isolation valves closed rather than open since open lines require valve action to provide containment integrity.
(2) Purging / venting will be permitted for safety related reasons only.
l Purging / venting needs should be identified and justified.
Safety related reasons previously accepted include pressure / vacuum control, unscheduled maintenance on safety related systems and inerting/deinerting of BWR c
plants.
(3) The smallest purge / vent system should be used whenever possible.
1
QUESTION 2. (continued) (4) Purging / vent system isolation valves must be qualified to close under LOCA conditions.
Unqualified valves must be locked closed.
(5) The amount of radioactivity released during the time required to close the system isolation valves should be shown by analysis not to cause the total offsite dose to exceed 10 CFR Part 100 guidelines, or alternatively do not perform the dose analysis and limit the total purge / vent time to 90 hours0.00104 days <br />0.025 hours <br />1.488095e-4 weeks <br />3.4245e-5 months <br /> / year.
(6) Purging should not be done for either temperature or humidity control.
(7) An estimate or goal of yearly purge / vent time should be established.
i O
__,__-__.-,,_y_,,7-,
,p_
s
,_s y-
~,. -
-._w-,.,
-.-y
. QUESTION 3.
To what extent have the licensees complied with the staff's purge position?
RESPONSE.
The industry has, for the most part, complied with the above staff position.
For NT0Ls, the staff position is implemented during the normal review process.
For the 65 plants with operating licenses in 1980, the review was conducted under Generic Issue B-24.
At the present time, there are approximately 10 operating plants for which total resolution of the purge and vent activities has not been achieved.
All plants comply with the key element of the position, which is to limit the purge / vent activities for safety related reasons.
The remaining plants, however, have not as yet satisfactorily identified and justified the specific reasons for purging.
The staff is continuing its efforts to obtain this information.
The yearly time for purge / vent operations in the operating plants varies widely due to fundamental plant layout differences.
BWR plants typically purge 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> per year or less because the need to enter containment to perform maintenance or surveillance is much less than for most PWRs.
PWRs have significantly more safety related equipment located inside the primary containment. Therefore, the need to purge more fraquently is greater.
This
QUESTION 3. dependence is seen with the anticipated purging practices of the newer BWR Mark III containment design.
This design contains a substantial increase in safety related equipment within containment.
Purging times in the order of 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> are expected.
There are, however, some plants that would not be able to continue to operate without purging almost continuously unless major modifications to the plant systems were made. The staff is currently pursuing the costs and benefits associated with purge time reduction for these selected plants.
Table 1 contains a summary of the past purge system usage for the older PWR operating plants.
It is worth noting that approximately two-thirds operate the purgc/'.ent :yttem less than 1000 hours0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> / year.
Also included in the table are the applicant's commitments for the recent NT0Ls to limit purging.
Although it would appear that the NTOLs have increased the average purging time, one must understand the difference between the data bases.
The NTOL data is extracted from the Technical Specifications of each plant.
As a result, these values can be considered as upper limits rather than the actual plant usage values listed for the operating plants.
i e
i
n
-t r
TABLE 1
-1
't PURGE USAGE i
PWR DRY CONTAINMENTS j
(HOURS / YEAR) l DATE
<100 100-1000 1000-3000
>3000 l
I 1980 20 11 3
8 1981 25 6
2 9
l NTOLs 4
3 3
0 i
t j
l l
I l
l
e
-9_
QUESTION 4.
The Com,nission further requested staff to develop probabilities on loss of containment isolation as a result of such factors as failure to close isolation valves, and/or the existence of unknown pre-existiae holes.
Specifically, the staff should provide. analysis of and values for "S" (the WASH-1400 conditional probability of containment failure resulting from inadequate isolation of containment openings and penetratior.3) considering containment type and historical development.
RESPONSE.
The PNL report in the response to Question 1 includes analysis of containment isolation unavailability based on the LERs and ILRT reports reviewed by PNL.
Table 2 contains the summary results of this study of containment unavailability.
These estimates are dependent on several assumptions (particularly on event duration times) necessitated by the type of dr.ta available in the LERs and ILRT reports.
The data reviewed by PNL revealed that most primary containment integrity problems for BWRs are most often related to valves failing to close or excessive valve leakage.
Leakage past MSIVs was especially frequent (much more than PWRs).
l l
. TABLE 2.
Summary Results for Containment Unavailability Unavailability Leak Area A.
Major Leakage Data Summary Airlocks 5x10 5 5000 sq inches Large Leaks 1023 to 10 2 2 to 28 sq inches B.
Minor Leakage Data Summary (a)
PWR ILRT Results 0.05 0.006 sq inches (b) 0.12 0.06 0.07 0.60 BWR ILRT Results 0.16 0.006 sq inches 0.14 0.06 0.04 0.60 In addition, PNL has conservatively estimated from the local leak rate test data of valves and penetrations for both PWR and BWR plants that containment leakage will exceed, by up to a factor of 10, the Technical Specification limits (typically 0.1 to 1.0 volume percent per day at design pressure) approximately 30 percent of the time.
(a) The values in part B of Table 2 do not result in leakage rates that are considered risk significant, but do exceed Technical Specification limits.
(b) Denotes that a leak rate corresponding to a 0.006 sq. inch opening existed for 5% of the ILRTs reviewed.
O h
QUESTION 4. (continued) For PWRs, personnel air lock leakage dominated'the failure modes and accounted for approximately half of the LERs for PWRs.
Failure of one or both of the personnel air lock doors to close and latch were also reported frequently.
Excessive leakage through large purge isolation valves, often due to degradation of the seat material, was a significant type of valve failure at the PWRs.
Although most of the leak rates listed in Table 2 could be considered as excessive leakage based on Technical Specification limitations, analyses performed for the Reactor Safety Study (WASH-1400) have shown that severe consequences or significant effects on other systems only result from large leaks.
For cases where a preexisting hole did not exist, the WASH-1400 study of the Surry Power Plant utilized a leakage value equivalent to 10 times the Technical Specification limitation of 0.1 volume percent per day at design pressure throughout the radiological analysis until such time the containment integrity was lost by other threats such as overpressurization, hydrogen combustion, or basemat I
failure. The overall risk was not sensitive or influenced in any I
significant way by this leakage value. WASH-1400 also examined other i
l preexisting leakage possibilities extending well beyond the 1% regime to determine what size leak openings would have to exist so that 1) l overpressure failure would be precluded, 2) other system effects or l
l interactions (e.g., coolant losses or pump cavitation) could possibly occur, s
QUESTION 4. (continued).
and, 3) competition with fission product depletion, deposition, or scrubbing processes would occur.
The WASH-1400 analysts determined that, for Surry, a leak opening of about 4 inches in diameter (12.5 square inches), which is equivalent to approximately 200 volume percent per day at design pressure, would have competing effects with the natural deposition and depletion processes and with the threat of overpressure failure. This leak size was then designated as p and it encompassed loss of containment integrity caused through eit!.er (1) failure of'the containment isolation system to initially secure containment integrity, (2) leakage resulting from systems not isolated by the accident but secured sometime after operation during the accident, or (3) leakage resulting from missiles or other dynamic effects of the accident.
Pault tree logic was used to help explore and estimate the probability of a p leakage category existing during the accident.
For the BWR plant studied in WASH-1400, Peach Bottom 2, a similar exploration was carried out and a 1 inch diameter leak size (equivalent to approximately 100 volume percent per day at design pressure) was determined to give rise to system effects.
l The containment isolation unavailability results for large leaks from Table 2, 10 2 to 10 3, compare favorably with the loss of containment isolation values, 10 2 to 10 4, used in the WASH-1400 report and later PRAs.' It should be noted that there may be a few plants that have higher containment isolation unavailabilities. The PRA performed by the licensee for the l
l l
QUESTION 4. (continued) 75 MWe Big Rock Point plant, for example, utilizes 6x10 2 as the unavailability value for containment isolation.
Containment isolation failure has been and continues to be considered in full scope PRAs.
In a large dry containment the conditional failure probability attributed to significant leakage sizes has been considered small in comparison with containment failure modes attributable to overpressure failure due to non-condensible gas production, hydrogen burns, or basemat melt through.
G 9
u