ML20207H864
| ML20207H864 | |
| Person / Time | |
|---|---|
| Issue date: | 08/03/1999 |
| From: | King T NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Ralph Beedle, Bischoff G, Chexal V, Drake A, Joseph Kelly, Lochbaum D, Subalusky W AFFILIATION NOT ASSIGNED, C-E OPERATING PLANTS OWNERS GROUP, ELECTRIC POWER RESEARCH INSTITUTE, INSTITUTE OF NUCLEAR POWER OPERATIONS, NUCLEAR ENERGY INSTITUTE (FORMERLY NUCLEAR MGMT &, UNION OF CONCERNED SCIENTISTS, WESTINGHOUSE OPERATING PLANTS OWNERS GROUP |
| Shared Package | |
| ML20207H869 | List: |
| References | |
| NUDOCS 9908160114 | |
| Download: ML20207H864 (15) | |
Text
{{#Wiki_filter:[ p ym k. . ~ UNITED STATES g [ - NUCLEAR REGULATORY COMMISSION
.o
- WASHINGTON, D.C. 20666-0001
..,,.,o August 3, 1999 i
Mr. William T. Subalusky . Vice President and Director - Govemment Relations Division
= Institute of Nuclear Power Operations .
700 Galleria Parkway ! Atlanta, GA 30339-5957 I
Dear Mr. Subalusky:
SUBJECT:
REQUEST FOR REVIEW OF DRAFT REPORT TITLED "HIGH PRESSURE
- SAFETY INJECTION SYSTEM RELIABILITY,1987-1997 l Attached for your review and comnkent is the draft report on the high pressure injection (HPI) system reliability at U.S. commercial pressurized water reactors. This study is part of our '
onDoing evaluation of the.l reliability of risk-signifi. cant safety systems and their associated components. Using data obtained from licensee event reports and the common cause failure (CCF) database, the study provides an estimate of the HPl system unreliability based on ! unplanned system demands between 1987 and 1997, in addition, the report identifies dominant contributors to system unreliability, compares HPl unreliability calculated using operating experience with.unreliability calculated using data published in probabilistic risk assessments (PRAs) and individual plant examinations (IPEs), evaluates significant trends, and provides insights on failure detection methods. Major findings obtained from this study are summarized below:
- Based on 1987-1997 operating experience data, there were no failures of the entire HPl system in 224 unplanned system demands. The arithmetic average of HPi system unreliability for 72 pressurized water reactors is 4.5E-4. The industry-wide arithmetic average of HPl system unreliability calculated using data extracted from PRA/IPEs is 5.7E-4. This is consistent with the HPl system unreliability based on operating experience from 1987-1997. .
.- Individual plant HPl system unreliabilities among the 72 plants vary by a factor of 50,
- from 6E-5 to 3.5E-3. The variability reflects the diversity found in HPI system designs rather than a difference in performance among plants. Plants with both high head and ) ,
' intermediate head HPl trains have lower system unreliabilities and plants with only high I head or only intermediate HPl trains have relatively nigher system unreliabilities.
For plants with both high head and intermediate head HPI trains, the unreliabilities calculated using IPE data are significantly lower than the unreliabilities calculated using 7 g.
-1987-1997 operating experience. Some of these IPEs may have optimistically .
9y0 evaluated HPl unreliability. This difference is attributed to the dominating contribution f&G7
. _.. WC FILF REIHER C9P 9908160114 990803 "
y PDR ORO EPSINPO PDR i
! \.
b5-M-r/ i
, .n * 'f
]
i I W. Subalusky 2 (greater than 90%) of the refueling water storage tank (RWST) suction path failure probability to the overall system unreliability for the operating experience based analysis for plants in this class. The failure probability of the RWST suction segment may not have been properly incorporated in some IPEs of plants in this class.
~
- Common cause failures of pump trains (pumps failing to start or run) and injection headers (motor operated valves), and independent failures in the shared suction path from the RWST (only for plants with both high head and intermediate head pump trains) l are the leading contributors to system unreliability, 1
- The frequency of unplanned HPl demands and the number of HPl failure events show a statistically significant downward trend as a function of calendar time and as a function of plant age. That is, the number of HPl demapds and the number of reported HPl failure events per year per plant have gone down with calendar time. In addition, older i
_ plants tend to have less HPI demands and less reported failure events per unit time compared to new plants. In accordance with our peer review' process prior to the distribution of the final report, we want to provide you with the opportunity to review and comment on the draft report. We are specifically interested in your comments regarding the technical adequacy of the data analysis and the appropriateness of the findings from this study. In order to meet our planned publication schedule, we would appreciate receiving your comments within 30 days from receipt of this letter. l
- A copy of this draft report and letter are being placed in the Public Document Room at 2120 L l- Street NW, Washington, DC 20555.
1~
- If you have any questions related to this report, please contact Mr. Sunil Weerakkody of my staff at 301-415-6374 or e-mail at sdw1@nrc. gov.
Sincerely, t Thomas L. King, Director l Division of Risk Analysis and Applications Office of Nuclear Regulatory Research l
Attachment:
As stated - cc w/o att: l A. Thadani, RES L M; Federline, RES
- S. Col; ins, NRR R. Zimmerman, NRR B. Sheron, NRR -
l i W. Kane, NRR l L.
Mi(*g
,0 k UNITED STATES g
j NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20666-4001 g s, .,,,./ August 3,[999 Mr. Andrew Drake, Project Manager Westinghouse Owners Group Westinghouse Electric Corporation Mail Stop ECE 5-16 P.O. Box 355 Pittsburgh, PA 15230-0355
Dear Mr. Drake:
SUBJECT:
REQUEST FOR REVIEW OF DRAFT REPORT TITLED "HIGH PRESSURE i SAFETY INJECTION SYSTEM RELIABILITY,1987-1997 Attached for your review and comment is the draft report on the high pressure injection (HPI) system reliability at U.S. commercial pressurized water reactors. This study is part of our ongoing evaluation of the reliability of risk-significant safety systems and their associated . components. Using data obtained from licensee event reports and the common cause failure (CCF) database, the study provides an estimate of the HPl system unreliability based on unplanned system demands between 1987 and 1997. In addition, the report identifies dominant contributors to system unreliability, compares HPI unreliability calculated using operating experience with unreliability calculated using data published in probabilistic risk assessments (PRAs) and individual plant examinations (IPEs), evaluates significant trends, and provides insights on failure detection methods. Major findings obtained from this study are summarized below:
- Based on 1987-1997 operating experience data, there were no failures of the entire HPI system in 224 unplanned system dsmands. The arithmetic average of HPl system unreliability for 72 pressurized water reactors is 4.5E-4. The industry-wide arithmetic average of HPl system unreliability calculated using data extracted from PRA/IPEs is 5.7E-4. This is consistent with the HPl system unreliability based on operating experience from 1987-1997.
= Individual plant HPI system unreliabilities among the 72 plants vary by a factor of 50, from 6E-5 to 3.5E-3. The variability reflects the diversity found in HPI system designs rather than a difference in performance among plants. Plants with both high head and intermediate head HPl trains have lower system unreliabilities and plants with only high head or only intermediate HPl trains have relatively higher system unreliabilities.
- For plants with both high head and intermediate head HP! trains, the unreliabilities calculated using IPE data are significantly lower than the unreliabilities calculated using 1987-1997 operating experience. Some of these IPEs may have optimistically evaluated HPl unreliability. This difference is attributed to the dominating contribution
A. Drake 2 (greater than 90%) of the refueling water storage tank (RWST) suction path failure
. probability to the overall system unreliability for the operating experience based analysis for plants in this class. The failure probability of the RWCT suction segment may not have been properly incorporated in some IPEs of plants in this class.
- Common cause failures of pump trains (pumps failing to start or run) and injection headers (motor operated valves), and independent failures in the shared suction path from the RWST (only for plants with both high head and intermediate head pump trains) are the leading contributors to system unreliability.
- The frequency of unplanned HPl demands and the number of HPI failure events show a statistically significant downward trend as a function of calendar time and as a function of plant age. That is, the number of HPI demands and the number of reported HPl
- failure events per year per plant have gone down with calendar time. In addition, older plants tend to have less HPI demands and less reported failure events per unit time .
compared to new plants. in accordance with our peer review process prior to the distribution of the final report, we wbnt
- to provide you with the opportunity to review and comment on the draft report. We are specifically int:srested in your comments regarding the technical adequacy of the data analysis and the appropriateness of the findings from this study. In order to meet our planned l publication schedule, we would appreciate receiving your comments within 30 days from receipt of this letter.
A copy of this dratt report and letter are being placed in the Public Document Room at 2120 L L Street NW, Washington, DC 20555. If you have any questions related to this report, please contact Mr. Sunil Weerakkody of my staff at 301-415-6374 or e-mail at sdwi@nrc. gov. l Sincerely, L c LL.\ Thcinas i.. King, Director Division of Risk Analysis and Applications Office of Nuclear Regulatory Research
Attachment:
As stated i_ cc w/o att: L A. Thadani, RES l M. Federline, RES S. Collins, NRR R. Zimmerman, NRR B. Sheron, NRR W. Kane, NRR
js
. 1 p t. . UNITED STATES y
j: NUCLEAR REGULATORY COMMISSION O 2. . WASHINGTON, D.C. 20666-0001
% .... /
4 August 3,1999 Mr. J.J. Kelly, Manager i L B&W Owners Group Services ' l Framatome Technologies, Inc. !. P.O. Box 10935 Lynchburg, VA 24506-0935
Dear Mr. Kelly:
SUBJECT:
REQUEST FOR REVIEW OF DRAFT REPORT TITLED "HIGH PRESSURE
. SAFETY INJECTION SYSTEM RELIABILITY,1987-1997 Attached for your review and comment is the draft report on the high pressure injection (HPI)
L
- L system reliability at U.S. commercial pressurized water reactors. This study is part of our l
ongoing evaluation of the reliability of risk-significant safety systems and their associated l components. Using data obtained from licensee event reports and the common cause failure l (CCF) database, the study provides an estimate of the HPI system unreliability based on l unplanned system demands between 1987 and 1997. In addition, the report identifies l dominant contributors to system unreliability, compares HPI unreliability calculated using , l operating experience with unreliability calculated using data published in probabilistic risk ;
- assessments (PRAs) and individual plant examinations (IPEs), evaluates significant trends, !
! and provides insights on failure detection methods.: l Major findings obtained from this study are summarized below: - a Based on 1987-1997 operating experience data, there were no failures of the entire HPl system in 224 unplanned system demands. The arithmetic average of HPl system E unreliability for 72 pressurized water reactors is 4.5E-4. The industry-wide arithmetic average of HPl system unreliability calculated using data extracted from PRA/IPEs is 5.7E-4. This is consistent with the HPi system unreliability based on operating ; experience from 1987-1997.
- Individual plant HPl system unreliabilities among the 72 plants vary by a factor of 50, from 6E-5 to 3.5E-3. The variability reflects the diversity found in HPl system designs iather than a difference in performance among plants. Plants with both high head and intermediate head HPl trains have lower system unreliabilities and plants with only high -
head or only intermediate HPl trains have relatively higher cystem unreliabilities.
- For plants'with both high head and intermediate head HPl trains, the unreliabilities calculated using IPE data are significantly lower than the unreliabilities calculated using 1987-1997 operating experience. Some of these IFEs may have optimistically i evaluated HPI unreliability. This difference is attributed to the dominating contribution l.
t
f b
l 1 c l J.J. Kelly 2 (greater than 90%) of the refueling water storage tank (RWST) suction path failure probability to the overall system unreliability for the operating experience based analysis l for plants in this class. The failure prooability of tne RWST suction segment may not have been properfy incorporated in some IPEs of plants in this class.
- Common cause failures of pump trains (pumps failing to start or run) and injection headers (motor operated valves), and independent failures in the shared suction path from the RWST (only for plants with both high head and intermediate head pump trains) are the leading c.ontributors to system unreliability.
- The frequency o. unplanned HPl demands and the number of HPl failur;. '.a/ents show a statistically significant downward trend as a function of calendar time and as a function of plant age. That is, the number of HPI demands and the number cf reported HPl failure events per year per plant have gone down with calendar time. In addition, o' der plants tend to have less HPl demands and less reportei ' ' are events per unit time compared to new plants.
In accordance with our peer review process prior to the distribution of the final report, we want to provide you with the opportunity to review and comr.ient on the draft report We are specifically interested in your comments regarding the technical adequacy of the data analysis and the appropriateness of the findings from this study. In order to meet our planned publication schedule, we would appreciate receiving your commenM within 30 days from receipt of this letter. A copy of this draft report and letter are being placed in the Public Document Room at 2120 L Street NW, Washington, DC 20555. If you have any questions related to this report, please contact Mr. Sunil Weerakkody of my staff at 301-415-6374 or e-mail at sdwi@nrc. gov. Sincerely,
/,
Ch.w, a k l Thomas L. King, Dire)ctor Division of Risk Analysis and Appt ations Office of Nuclear Regulatory Researcc
Attachment:
As stated cc w/o att: A. Thadani, RES M. Federtine, RES S. Collins, NRR R. Zimmerman, NRR B. Sheron, NRR W. Kane, NRR
.,- @ MQ p UNITED STATES 4
g { NUCLEAR REGULATORY COMMISSION o g WASHINGTON, D.C. 20666 0001
%,***<<j August 3,1999 i l
ABb Combustion Engineering Owners Group l clo Gordon Bischoff, CEOG Project Manager M.S. 9615-1932 - 2000 Day Hill Road i Windsor, CT 06095 l L
Dear Mr. Bischoff:
SUBJECT:
' REQUEST FOR REVIEW OF DRAFT REPORT TITLED "HIGH PRESSURE )
SAFETY INJECTION SYSTEM RELIABILITY,1987-1997 ) l- Attached for your review and comment is the draft report on the high pressure injection (HPI) , system reliability at U.S. commercial pressurized water reactors. This study is part of our ! ongoing evaluation of the reliability of risk-significant safety systems and their associated ; !. components. Using data 'obtained from licensee event reports and the common cause failure ! !- (CCF) database, the study provides an estimate of the HPl system unreliability based on unplanned system demands between 1987 and 1997. In addition, the report identifies dominant contributors to system unreliability, compares HPl unreliability calculated using operating experience with unreliability calculated using data published in probabilistic risk assessments (PRAs) and individual plant examinations (IPEs), evaluates significant trends, l .and providec insights on failure detection methods.
' Major findings obtained from this study are summarized below:
. Based on 1987-1997 operating experience data, there were no failures of the entire HPl l system in 224 unplanned sydem demands. The arithmetic average of HPI system l
unreliability for 72 pressurized water reactors is 4.5E-4. The industry-wide arithmetic average of HPl system unreliability calculated using data extracted from PRA/IPEs is 5.7E-4. This is consistent with the HPl system unreliability based on operating ) experience from 1987-1997, I a Individual plant HPl system unreliabilities among the 72 plants vary by a factor of 50,
. from 6E-5 to 3.5E-3. The variability reflects the diversity found in HPl system designs L rather than a difference in performance among plants. Plants with both high head and l , intermediate head HPl trains have lower system unreliabilities and plants with only high y head or only intermediate HPl trains have relatively higher system unreliabilities.
[ + For plants with both hig'h head and intermediate head HPl trains, the unreliabilities b calculated using IPE data are significantly lower than the unreliabilities calculated using l 1987-1997 operating experience, Some of these IPEs may have optimistically L evaluated HPI unreliability. This difference is attributed to the dominating contribution r
+
G. Bischoff 2 (greater than 90%) of the refueling water storage tank (RWST) suction path failure l. i probability to the overall system unreliability for the operating experience based analysis for plants in this' class. The failure probability of the RWST suction segment may not j have been properly incorporated in some IPEs of plants in this class. i
- Common cause failures of pump trains (pumps failing to start or run) and injection headers (motor operated valves), and independent failures in the shared suction path !
. from the RWST (only for plants with both high head and intermediate head pump trains) )
are the leading contributors to system unreliability. j
- The frequency of unplanned HPI demands and the number of HPl failure events show a I statistically significant downward trend as a function of calendar time and as a function of plant age. That is, the number of HPI demands and the number of reported HPl
' failure events per year per plant have gone down'with calendar time. In addition, older plants tend to have less HPl demands and less reported failure events per unit time compared to new plants.
In accordance with our peer review process prior to the distribution of the final report, we want to provide you with the opportunity to review and comment on the draft report. We are specifically interested in your comments regarding the technical adequacy of the data analysis and the appropriateness of the findings from this study. In order to meet our planned publication schedule, we would appreciate receiving your comments within 30 days from receipt of this letter. A copy of this draft report a'nd letter are being placed in the Public Document Roorn at 2120 L Street NW, Washington, DC 20555. If you have any questions related to this report, please contact Mr. Sunil Weerakkody of my staff at 301-415-6374 or e-mail at sdw1@nrc. gov. Sincerely, w, ( Thomas L. King, Director Division of Risk Analysis and Applications Office of Nuclear Regulatory Research
Attachment:
As stated cc w!o att:. A. Thadani, RES M. Federline, RES S. Cellins, NRR -- R. Zimmerman, NRR _ B. Sheron, NRR _W. Kane, NRR l l s i
l i fA MCg% \ 4 -' 4 UNITED STATES L l ,j
' 4" NUCLEAR REGULATORY COMMISSION WASHINGTON. D.C. 20666-0001 j
\*****/ August 3, 1999 l
l Mr. V. K. Chexal, Director ; Nuclear Safety Analysic Center ! Electric Power Research instituts P.O. Box 10412 Palo Alto, CA 94303
Dear Mr. Chexal:
SUBJEC' T: REQUEST FOR REVIEW OF DRAFT REPORT TITLED "HIGH PRESSURE SAFETY INJECTION SYSTEM RELIABILITY,1987-1997 Attached for your review and comrpent is the draft report on the high pressure injection (HPl) system reliability at U.S. commercial pressurized water reactors. This study is part of our , ongoing evaluation of the reliability of risk-significant safety systems and their associated components. Using data obtained from licensee event reports and the common cause failure l (CCF) database, the study provides an estimate of the HPI system unreliability based on I , unplanned system demands between 1987 and 1997. In addition, the report identifies ! dominant contributors to system unreliability, compares HPl unreliability calculated using operating experience with unreliability calculated using data published in probabilistic risk assessments (PRAs) and individual plant examinations (IPEs), evaluates significant trends, and provides insights on failure detection methods. ] l- Major findings obtained from this study are summarized below:
.. Based on 1987-1997 operating experience data, there were no failures of the entire HPl i system in 224 unplanned system demands. The arithmetic average of HPl system i unreliability for 72 pressurized water reactors is 4.5E-4. The industry-wide arithmetic average of HPl system unreliability calculated using data extracted from PRA/IPEs is 1 5.7E-4. - This is consistent with the HPl system unreliability based on operating i experience from 1987-1997.
I
- Individual plant HPl system unreliabilities among the 72 plants vary by a factor of 50, from 6E-5 to 3.5E-3. The variability reflects the diversity found in HPl system designs rather than a difference in performance among plants. Plants with both high head and l intermediate head HPl trains have lower system unreliabilities and plants with only high head or only intermediate HPl trains have relatively higher system unreliabilities.
- For plants with both high head and intermediate head HPl trains, the unreliabilitics calculated using IPE data are significantly lower than the unreliabilities calculated udng 11987-1997 operating experience. Some of these IPEs may have optimistically evaluated HPl unreliability. This difference is attributed to the dominating contribution j_ .
L
5 [O V.K. Chexal 2 (greater than 90%) of the refueling water storage tank (RWST) suction path failure probability to the overall system unreliability for the operating experience based analysis for plants in this class. The failure probability of the RWST suction segment may not have been properly incorporated in some IPEs of plants in this class. j
+ Coremon cause failures of pump trains (pumps failing to start or run) and injection headW .iotor operated valves), and independent failures in the shared suction path frWJ RWST (only for plants with both high head and intermediate head pump trains) 3 s ;5r leading contributors to system unreliability.
l
= The frequency of unplanned HPl demands and the number of HPi failure events show a
. statistice!!y significant downward trend as a function of calendar time and as a function of plant age. That is, the number of HPl demands and the number of reported HPI failure events per year per plant have gone dcIun with calendar time. In addition, older- ,
plants tend to have less HPl demands and less reported failure events per unit time compared to new plants. in accordance with our peer review ' process prior to the distribution of the final report, we want to provide you with the opportunity to review and comment on the draft report We are > specifically interested in your comments regarding the technical adequacy of the data analysis and the appropriateness of the findings from this study, in order to meet our planned publication schedule', we would appreciate receiving your comments within 30 days from receipt of this letter. A copy of this draft report and letter are being placed in the Public Document Room at 2120 L Street NW Washington, DC 20555. If you have any questions related to this report, please contact Mr. Sunil Weerakkody of my staff at 301-415-6374 or e-mail at sdwi@nrc. gov. Singy, _ L n- \ ThomP; L King, Director Division of Risk Analysis and Applications Office of Nuclear Regulatory Research
Attachment:
As stated oc w/o att: A. Thadani, RES M. Federime, RES S. Colhns, NRR ~ R. Zimmerman, NRR B. Sheron, NRR W. Kane, NRR g P .;..
W'fu O' ~h UNITED STATES l ]- NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20066 4 001 o% ..... f August 3,1999 Mr. Ralph E. Beedle Nuclear Energy institute 1776 i Street N.W. Suite 400 --
]
Washington, D.C. - 20006-3708 l
Dear Mr. Beedle:
SUBJECT:
REQUEST FOR REVlEW OF DRAFT RF, PORT TITLED "HIGH PRESSURE SAFETY INJECTION SYSTEM RELIABILITY,1987-1997 i i Attached for your review and comment is the draft report on the high pressure injection (HPI) ! system reliability at U.S. commercial pressurized water reactors. This study is part of our ) ongoing evaluation of the reliability of risk-significant safety systems and their associated I components. Using data obtained from licensee event reports and the common cause failure (CCF) database, the study provides an estimate of the HPl system unteliability based on unplanned system demands between 1987 and 1997. In addition, the report identifies
. dominant contributors to system unreliability, compares HPl unreliability calculated using operating experience with unreliability calculated using data published in probabilistic risk assessments (PRAs) and individual plant examinations (IPEs), . evaluates significant trends, and provides insights on failure detection methods.
Major findings obtained from this study are summarized below: e Based on 1987-1997 operating experience data, there were no failures of the entire HPl system in 224 unplanned system demands. The arithmetic average of HPl system unreliability for 72 pressurized water reactors is 4.5E-4. The industry-wide arithmetic average of HPi system unreliability calculated using data extracted from PRAllPEs is 5.7E-4. This is consistent with the HPl system unreliability based on operating experience from 1987-1997.
+ Individual plant HPl system unreliabilities among the 72 plants vary by a factor of 50, from 6E-5 to 3.5E-3. The variability reflects the diversity found in HPI system designs rather than a difference in performance among plants. Plants with both high head and intermediate head HPl trains have lower system unreliabiiities and plants with only high head or only intermediate HPI trains have relatively higher system unreliabilities.
=- For plants with both high head and intermediate head HPl trains, the unreliabilities calculated using IPE data are significantly lower than the unreliabilities calculated using 1987-1997 operating experience. Some of these IPEs may have optimistically evaluated HPI unreliability. This difference is' attributed to the dominating contribution
F R. Beedle 2 (greater than 90%) of the refueling water storage tank (RWST) suction path failure probability to the overall system unreliability for the operating experience based analysis for plants in this class. The failure probability of the RWST suction segment may not have been properly incorporated in some IPEs of plants in this class. l
- Common cause failures of pump trains (pumps failing to start or run) and injection headers (motor operated valves), and independent failures in the shared suction path
, from the RWST (only for plants with both high head and intermediate head pump trains) l are the leading contributors to system unreliability.
The frequency of unplanned HPl demands and the number of HPl failure events show a statistically significant downward trend as a function of calendar time and as a function , of plant age. That is, the number of HPl demands and the number of reported HPl j failure events per year per plant have gone down with calendar time. In addition, older plants tend to have less HPl demands and less reported failure events per unit time i compared to new plants. In accordance with e oeer review process prior to the distribution of the fina! report, we wdnt ) to provide you with . ppportunity to review and comment on the draft report. We are specifically interested in your comments regarding the technical adequacy of the data analysis. and the appropriateness of the findings from this study. In order to meet our planned l publication schedule, we would appreciate receiving your comments within 30 days from receipt l of this letter. A copy of this draft report and letter are being placed in the Public Document Room at 2120 L Street NW, Washington, DC 20555. If you have any questions related to this report, please contact Mr. Sunil Weerakkody of my staff at 301-415-6374 or e-mail at sdw1@nrc. gov. Sincerely,
-fa w. t ]Y 3 Thomas L. King, Director Division of Risk Analysis and Applications Office of Nuclear Regulatory Research
Attachment:
As stated cc w/o att: A. Thadani, RES M. Federline, RES S. Collins, NRR R. Zimmerman, NRR B. Sheron, NRR W. Kane, NRR
y 2 arog
)p -
->t UNITED STATES l i*
j_ 2-NUCLEAR REGULATORY COMMISSION WASHINGTON. D.C. 2005dH1001 p (...../ August 3,1999 1
' Mr. David A. Lochbaum )
Union of Concemed Scientists 1616 P Street, N.W.
' Suite 310 l Washington, D.C. 20035-1495-
Dear Mr. Lochbaum:
SUBJECT:
. REQUEST FOR REVIEW OF DRAFT REPORT TITLED "HIGH PRESSURE
' SAFETY INJECTION SYSTEM RELIABILITY,1987-1997
. Attached for your review and comment is the draft report on the high pressure injection (HPI)
. system reliability at U.S. commercial pressurized water reactors. This study is part of our
*. ongoing evaluation of the reliability of risk-significant safety systems and their associated components. Using data obtained from licensee event reports and the common cause failure 1 1(CCF) database, the study provides an estimate of the HPI system unreliability based on unplanned system demands between 1987 and 1997. In addition, the report identifies dominant contributors to system unreliability, compares HPl unreliability calculated using operating experience with unreliability calculated using data published in probabilistic risk l assessments (PRAs) and individual plant examinations (IPEs), evaluates significant trends, !
and p ovides insights on failure detection methods. Major findings obtained from this study are summarized below:
.- Based on 1987-1997 operating experience data, there were no failures of the entire HPl !
system in 224 unplanned system demands.. The arithmetic average of HPl system
. unre ali bility for 72 pressurized water reactors is 4.5E-4. The industry-wide arithmetic average of HPl system unreliability calculated using data extracted from PRA/IPEs is ;
i 5.7E-4. This is consistent with the HPl system unreliability based on operating experience from 1987-1997.
= Individual plant HPl system unreliabilities among the 72 plants vary by a factor of 50, from 6E-5 to 3.5E-3. The variability reflects the diversity found in HPl system designs rather than a difference in performance among plants. Plants with both high head and intermed9te head HPl trains have lower system unreliabilities and plants with only high -
head or only intermediate HPi trains have relatively. higher system unreliabilities.
- For' plants with both high head and intermediate head HPl trains, the unreliabilities calculated using IPE data are significantly lower than the'unreliabilities calculated using 1987-1967 operating experience. Some of these IPEs may have optimistically '
evaluated HPl unreliability. This difference is attributed to the dominating contribution O.- i
D. Lochbaum 2
- (greater than 90%) of the refueling water storage tank (RWST) suction path failure probability to the overall system unreliability for the operating experience based analysis for plants in this class. The failure probability of the RWST suction segment may not have been properly incorporated in some IPEs of plants in this class.
+ Common cause failures of pump trains (pumps failing to start or run) and injection headers (motor operated valves), and independent failures in the shared suction path from the RWST (only for plants with both high head and intermediate head pump trains) are the leading contributors to system unreliability.
- The frequency of unplanned HPl demands and the number of HPI failure events show a statistically significant downward trend as a function of calendar time and as a function of plant age. That is, the number of HPl demands and the number of reported HPl failure events per year per plant have gone down with calendar time. In addition, older plants tend to have less HPI demands and less reported failure events per unit time compared to new plants, in accordance with our peer review process prior to the distribution of the final report, we want to provide you with the opportunity to review and comment on the draft report. We are
' specifically interested in your comments regarding the technical adequacy of the data analysis and the appropriateness of the findings from this study. In order to meet our planned publication schedule, we would appreciate receiving your comments within 30 days from receipt of this letter.
A copy of this draft report and letter are being placed in the Public Document Room at 2120 L Street NW, Washington, DC 20555. If you have any questions related to this report, please contact Mr. Sunil Weerakkody of my staff at 301-415-6374 or e-mail at sdwi@nrc. gov, Sincerely,
/
L w.w L Thomas L., King, Director Division of Risk Analysis and Applications Office of Nuclear Regulatory Research
Attachment:
As stated cc w/o att: A. Thadani, RES M. Federline, RES ' S. Collins, NRR R. Zimmerman, NRR B. Sheron, NRR W. Kane, NRR
o- #
- MEMRANDUM DATED: 8/ 3 /99
- INDENTICAL LETTERS
~
SUBJECT:
REQUEST FOR REVIEW OF DRAFT REPORT TITL2D "HIGH PRESSURE SAFETY INJECTION SYSTEM RELIABILITY,1987-1997
Attachment:
As stated 1
' I cc w/o att:
' A. Thadani, RES M. Federline, RES '
S. Collins, NRR R. Zimmerman, NRR
, B. Sheron, NRR
. W. Kane, NRR Distribution w/att.:
OERAB RF ARubin, RES MReinhart, NRR PBernard, Ril DRAA RF MCunningham, RES JCalvo, NRR MParker, Rill cFile _ Center,; NSiu, RES- RBarrett, NRR - SBurgess, Rlli l Public MDrouin, RES PWilson, NRR JShackelford, RIV l JRosenthal JVora, RES DCoe, NRR WJones, RIV i JFlack . Llancaster, RES WDean, NRR Homstein GParry, NRR - MJohnson,N U MKnapp, DEDE SMalik, NRR AMadison, NRR JMitchell, OEDO - ' FTalbot, NRR JTrapp, RI
- JLarkins, ACRS LMarsh, NRR TShediosky, RI
.MMarkley, ACRS MRubin, NRR WRogers, Rll DOCUMENT NAME: S:\DRAA.OERAB\SDW1\HPIEXT.WPD i
To receive a cost of tNs document Indicate in the box: "C" = Cow wolwnci "E" = Cow wienct "N" = No copr OFFICE OERAB fjL_OERAB _ C OERAB M OERAB, NAME SWeerakkody SMays, gy) PBaranovy$ky d TKfr)d DATE # / 2 /99 C/1/99 [/h/d9 8/ $ /99 OFFICIAL RECORD COPY = _ _ _ _ _ - _ _}}