ML20248E839
| ML20248E839 | |
| Person / Time | |
|---|---|
| Issue date: | 05/19/1998 |
| From: | Rossi C NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | Gillespie F, Holahan G, Lainas G NRC (Affiliation Not Assigned) |
| Shared Package | |
| ML20248E842 | List: |
| References | |
| NUDOCS 9806040006 | |
| Download: ML20248E839 (6) | |
Text
pf%9
,[
k UNITED STATES y
I *j NUCLEAR REGULATORY COMMISSION
- 3 WASHINGTON, D.C. 20666 0001
,s
- /
May 19, 1998 MEMORANDUM 'IO: Frank P. Gillespie, Director, DISP:NRR Gary M. Holahan, Director, DSSA:NRR Gus C. Lainas, Acting Director, DE:NRR Jack W. Roe, Acting Director, DRPM:NRR R. Lee Spessard, Director, DRCH:NRR Lawrence C. Shao, Director, DET:RES l
Thomas L. King, Director, DST:RES l
James T. Wiggins, Director, DRS:RGN-l Johns P. Jaudon, Director, DRS:RGN-l!
John A. Grobe, Acting Director, DRS:RGN-Ill Arthur T. Howe:1, Director, DRS:RGN-IV FROM:
Charles E. Rossi, Di Safety Programs Division Office for Analysis and Evaluation of Operational Data -
SUBJECT:
SPECIAL REPORT - HIGH-PRESSURE CORE SPRAY SYSTEM RELIABILITY 1987-1993 (INEEL-EXT-95-00133), JANUARY 1998 i
i Attached for your use and irJormation is the final report on the reliability of the high-pressure core spray (HPCS) system. The study is based on the operating experience from 1987 through 1993 as reported in licensee event reports. This is the fifth in a series of system reliability study reports whicn focus on using operational data to determine the reliability of risk significant systems in U.S. commercial reactors. The results are compared with Probabilistic Risk Assessments (PRAs) and Individual Plant Evaluations (IPEs). Insights from an engineering j
analysis of the data are also included. Earlier drafts of this report were provided to the Office of l
Nuclear Reactor Regulation (NRR), the Office of Nuclear Regulatory Research (RES), and the regions for review and comment. The suramary of resolutions to comments is attached. This report will be published as a NUREG report.
jl The study covers only the eight U.S. commercial boiling water reactors that have an HPCS system. Consequently, the operating experience data, including demand counts, failure counts, and run times, for computing HPCS unreliability estimates are limited. However, there is sufficient data to reasonably estimate the reliability of the system and its associated i
uncertainties, but information regarding dominant contributors and trends are less robust and f
could change as additional experience is obtained.
S
\\
pdk CONTACTS:
Steve Mays, SPD, AEOD (415-7496)
Tom Wolf, SPD, AEOD (415-7576)
. g 9806040006 980519 ce g @ f((
~
eg #
y5 p,
/1 4 PDR ORG NEXD l,
y1 p-
j I
Multiple Addressees 2
)
Notable observations and findings made from the limited data include:
The n iean observed operational mission HPCS unreliability was 0.075, including recoury. None of the actual demands to operate involved a loss of offsite power requiring the Division 111 emergency diesel-generator to energize the bus, nor did any last long enough to require pump suction transfer to the suppression pool. Only one demand failure was observed during 29 operational demands and accounted for 67 percent of the total system unreliability. This failure in the injection subsystem occurred as a result of the system being in a maintenance-out-of-service condition 'vhen the system was demanded. The other failure used in the unreliabi'ity estimete had less impact on the estimate, accounting for only 7 percent of the total system L reliability. This failure occurred during routine surveillance testing when the HPCS pump failed to start as a result of a failed over-frequency relay, For 24-hnur missions typical of those considered in PRAs and (pes, the operational data-e based unreliability estimate is 0.23. This is somewhat hgher than the equivalent unreliability estimated from the PPNIPEs. The unreliablity of the injection subsystem estimated from the operational experience is a factor of tr.c higher than that estimated from the PRA/IPEs. The difference in the estimates is primarily at+ributed to a factor of 50 difference in the average hourly failure rates used in calculating the HPCS injection pump failure to run (FTR) probability.
The PRA/IPE data appear to use generic FTR data for all pumps rather than plant specific l
(or system specific generic) pump failure data. The operational experience data primarily j
consisted of runs of one hour or less. The total operating experience run time was 316 hours0.00366 days <br />0.0878 hours <br />5.224868e-4 weeks <br />1.20238e-4 months <br /> and no run failures were reported during this time. The operational experience failure rate was estimated from this limited data and assumed to remain constant for the typical 24-hour PRA/IPE mission requirement. Thus, the operating experience estimate for FTR may be pessimistic. Additional data is necessary to ascer'.ain whether the I
differences between the reliability based on the operating experience data and the PRA/IPEs are real or an artifact of the limited available data.
There was only one failure in 29 unplanned demands and one failure in 267 tes's that were used in the computation of the system reliability over the seven year period of this
- tudy. From this limited data, no trends over time for the reliability would be expected to be observed. None were observed in the statistical analyses of the unreliability versus calendar year and in the unreliability versus plant age.
These findings are discussed in more detail in the report. Graphical and tabular displays, along l
with. specific discussions, are included. Specific failures and failure mechanisms are identified l
and characterized.
l RECOMMENDATIONS:
Due to the limited amou.it of operational data on the hPCS, no specific recommendations have been idantified as a result of this study. However, we do suggest the followls.g:
L
Malti le Addressees 3
NRR should examine the potentialimpact of license changes where the HPCS FTR e
probability can have a significant impact on the decision making criteria. Examination of the data sources for FTR estimates and their applicability to the particular application should be established where appropriate.
'nspectors should be made aware of the FTR differer.ces so that plant activities which e
could adversely impact HPCS FTR probability can receive relatively higher attention in HPCS inspections. Also, data relating to actual run times that could improve our ability to estimate FTR probabilities for HPCS could be identified for AEOD and NRR use.
Further examination is required to understand the differences between failure to run estimates for long term missions from PRA/IPEs and those extrapolated from the limited operating experience. AEOD will actively pursue acquisition of appropriate FTR data from LERs and industry sources such as the industry's Equipment Performance information and Exchange (EPlX) system. This information will be included in the reliability database presently under development. AEOD will use this data in future updates to HPCS system reliability evaluations to better assess the relationship between PRA/IPEs and operating experience.
Attachment:
1.
Resolution of Comments on Draft High Pressure Core Spray System Reliability 1987-1993 Report.
2.
High-Pressure Core Spray System Reliability 1987-1993 (INEEL/ EXT-95-00133), May 1998 cc w/att.:
V. K. Chexal, EPRI Pat Lewis, INPO Debbie W. Queener, NOAC Distribution w/att.:
Public DHickman BSheron, NRR File Center JRosenthal RZimmerman, NRR RRAB R/F JStolz, NRR AThadani, DEDE SPD R/F RWessman, NRR JMitchell, OEDO TTMartin MMayfield, RES MCunningham, RES FCongel REmrit, RES JTappert, NRR KRaglin RSavio, ACRS SCollins, NRR MKnapp, RES DOCUMENT NAME:
H:\\TRW\\3TRANSLE.RPT
'See previous concurrence To receive a copy of this document, indicate in the box: "C" s Copy without a%chment/ enclosure "E" s Copy with attachment / enclosure
- N" a No copy i
OFFICE BRANCH C
BRANCH C
C: BRANCH E
D:SPD lp NAME TWolf:nll*
SMays*
PBaranowsky*
CRospriyJ DATE 5/8/98 5/12/98 5/12/98
{,/j9/98
OFFICIAL RECORD COPY
I t
I 1
ATTACHt4ENT1 l
RESOLUTION OF COMtIENTS ON DRAFT REPORT HIGH PRESSURE CORE SPRAY SYSTEM RELIABRJTY 19871993 Summerlos of the comments recewed on the draft high pressure core sprey (HPCS) system rollebuty report along with their associated resolutions follow. The final wwsion of the HPCS report Wr:: these moolutions.
1.
Comment: From AEOD: Add perspecove regarding the overat risk imped of the observed operating experience HPCS failure rate being greater then that assumed in PRMPEs.
I From the Regions: The report discusses the 11PCS ir(ection subsystem. It seems that the PRMPEs underestimate the NiaHhood of problems cousing i
unretsbety as indicated by the operating experience. Doesn1 this mean that the PRMPE approach should be modified? Are there going to be any actions in this regard?
From the Regions: The failure estimates assodeled with the operadng experience reviews em accompanied by signl8 cont uncertainty bands and as a reeut, it is undear as to what condusions shouW be mede regen 2ng HPCS system retabully given the uncertainty spans approulmetely three orders of l
magnliude.
From NRR: The study states that contresulars to HPCS unredebWty calculated from the PRMPE informadon are not consistent wNh those from the operaung exportance. The pnmary reason for this difference isins almost two orders of magnitude disparity between tw PRMPEs everage hourly fanure rate for the HPCS motor pump and the operadng experience. The much larger operating exportance fature rate is particularly surprising in light of Table C-1, Point Estimates and Con 6dence Bounds for HPCS Fature Modes which documents 0 faRures to run of the irjection subsystem in 297 demerwh and 316.1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />. I may be misinterpreting the data but the results appear suspect.
Resoludon:
Text danges have been incorporated into the Execueve Summary and L
- w body of tw report to indicate the Emitsuons assodated with the speralty of data. AEOD wR actively pursue acquiring the -gC:
data for use in future updates to this report to better evaluele the reladonship tutwoon PRMPEs and operaung exportance. One effect of having additional data should tw the nomming of the uncatainty bounds, especiasy for comparison with #w 244wur mission emos used in tw PRMPEs. Another effect wm tw more con 5dence in the estimeled FTR value and whether the differences between the reliabihty based on the operating experience data and the PRMPEs are real or an artifact of the limi.,d avallAie Jata, i
==
O 2.
Comment: From the Regions: Two plants in the study, LaSalle and Perry, no longer have automatic transfer capability of the HPCS suction from the condensate storage tank to the suppression pool. At LaSalle, the HPCS suction was permanently shifted to the suppression pool by installing a blank flange in the condensate storage tank suction pipe. This was predicated by concoms of piping integrity due to rnicrobiologically induced cormsion of the CST suction line. At Perry, HPCS suction was realigned from the condensate storage tank suction to the suppression pool due to tomado misslie protection concoms.
Resolution:
The text in section 3.2.1 "PRA Comparison Unrelist24ty" has been modified to reflect these conditions. Checks with the other plants having an HPCS system determined that this condition was unique to just these two sites. These changes had a negligible affect on the estimated reliabilities as computed based on the PRA/IPEs or the operating experience.
3.
Comment: From NRR: From Table C-1, the 95th percentile value for the average FTRI is 0.0009. This value is rnost likely a rate (/hr), although it is not specified in the report. Using the Poisson process point estimate upper bound [ Chi Square 95th percentile for zero failures /(2 x the number of pump hours)]
yields a value of 0.00948/hr. Thus, the upper bour'd value of 0.0009 should be checked.
From NRR: The operational data yielded an average FTRI of 1.6E-3/hr. This average rate is close to the point estimate 95th percontile of 0.0006/hr (assuming the point estimate 95th percentile is an hourty rate). The value 1.6E-3thr is an average, which implied that it is a mean of a distribution. If this is the case, this distribution is skewed towards a range which is predicted to be a rare occurrence (i.e., greater than or equal to the point estimate 95th percontile). Thus, the distribution associated with the 1.6E-3/hr average value l
should be cheded.
Resolution:
Table C 1 has been revised to show the value in question is 0.000 l
rather than 0.0009. This was a typographical ermr which had not been l
identified during the report review prior to releasing it for comments.
The demand data listed for the failure to run information shows that the values are in hours, thus indicating that the resultant Ested probabilities are /hr rates.
4.
Comment: From RES: In 1995, as part of the NPAR program, we issued NUREG/CR-5462
- Aging Study on Boileg Water Reactor High Pressure injection Systems." you may find the information in this report useful and complementary to the subject AEOD study. Recommend adding a reference to NUREG/CR-5462 in the INEEL report.
Resolution:
No report changes were made. A comparison of the information contained in each report indicated similarities in dominant failed componer.ts and detection methoos. As there was nu new pertinent information obtained, the report was not referenced. When the HPCS
)
l
system reliability repor11s updated, NUREGICR 5462 will be induded in the review of and considered for indusion in the references.
5.
Comment: From AEOD: Obtain ?, needed surveillance frequency or test data, or l
provide a better explai,ation on page 43 regarding why five surveillance test failures were not used.
~
Resolution:
No report changes have been made. This comment refers to five surveillance test failures induded in subsection "Other Surveillance Test Failures" of section 4.1.2
- Factors Affecting HPCS Reliability." The introductory paragraph to this subsection states that "these were at used to estimate system reliability because the periodicity of the surveillance test was unknown'or the number of tests could not be readily estimated from the data available." Thus, demand data needed for indusion of these failures in the unreliability estimate is not readily
~
available. Attempts to obtain data which would permit the indusion of these failures in the reliability otimate were beyond the scope of the statement of work for this study und, therefore, no attempt was made to obtain the needed additionalinformation.
i I
6.
Comment: From the Regions: On pages B-6 and B 9, there are 2 LERs listed for LaSalle Unit 2 events where the LaSalle Unit 1 docket number (373) is referenced.
LER 37389009 and LER 37389011.
Resolution:
No report changes were made. LER reporting guidance pomwis submittal of an event which affects multiple units in one report a.x! this report is normally to be assigned to the lowest docket number. In these cases both events affected both units. They were property reported via Unit 1 LERs but each event was counted g- -yspi'iy against both units.
7.
Comment From AEOD: Explain why only 1987 to 1993 data were used.
i From the Regions: The study covers 1987-1993 which is relatively earty in l
the life of most of the plants equipped with HPCS. For your information, on March 31,1995, Clinton experienced an event where the Clinton HPCS diesel I
generator failed in a destructive manner. The generator inboard bearing failed due to lack of lubrication and self destructed. While this event actualty occumed during a test run of the machine, conditions were such that it would have occurred during a valid demand, had one existed. Although this failure occumed outside the report data collection period, it was a significant failure that made the HPCS emergency power supply inoperable for a significant period of time.
Resolution:
No report changes were made. The cutoff date for indusion of data in this report was 1993 to keep it consistent with the other already issued 1
system reliability reports. Events which occurred after this cutoff date will be considered for inclusion in this study when the system reliability studies are updated.
_ _ _ _ _ _ _ - - _ - -