ML20155G320

Safety Evaluation of TS NEDC-30936P, BWR Owners Group TSs Improvement Methodology. GE Analyses Demonstrated Acceptability of General Methodology for Considering TS Changes to ECCS Instrumentation Used in BWR Facilities

ML20155G320
Person / Time
Issue date:	09/26/1988
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20155G292	List: ML20155G288 ML20155G295 ML20155G303 ML20155G314 ML20155G320 ML20155G372 ML20155G390
References
NUDOCS 9811090020
Download: ML20155G320 (7)

Text

-, ~. -. - - . - . . .-. - -- . - . . --

l .

1 -

l CCP : 6 1999 SAFETY EVALUATION BY THE OFFICE Of NUCLEAR REACTOR REGULATION l REVIEW 0F BWR OWNERS' GROUP REPORT NEDC 30936P l ON JUSilflCATION FOR EXTENDit,'G ON-LINE TEST INTERVALS AND ALLOWABLE OUT-0F-SERVICE TIMES FOR BWR EMERGENCY CORE COOLING

-SYSTEM INST ^' ~NTATION 1.0 SU m ARY The staf f has reviewed the General Electric Company (GE) Topical Report NEDC-30936P "BWR Owners' Group Technical Specifications improvement Methodology (with Demonstration for BWR ECCS Actuation Instrunentation)

Part 1" (Ref.1), issued by the BWR Owners' Group to support the proposed extensions of ECCS surveillance test intervals (STis) and allowable out-of-service times ( A0Ts) for ECCS test and repair. The staff has concluded that the analyses presented in the GE report are generally acceptable for resolving these issues subject to the limitations ar.d conditions presented in this report.

2.0 BACKGROUND

Or. July 8,1983, the staff issued Generic Letter 83-28 (Ref. 2) recuesting that all licensees of operating reactors, applicants for an operating license, and holders of construction permits respond to the generic issues raised by the analyses of the Salem ATWS events. Item 4.5.3 of this generic letter requested licensees and applicants to review existing STis f or reactor protection system (RPS) required by technical specifications (TS) to assure that current and proposed intervals for such testing are consistent with achieving high RPS availability. Furthermore, in late 1983, the staf f issued NUREG-1024, " Technical Specifications - Enhancing the Safety Impact" (Ref. 3), which recommended that TS surveillance requirements and action statements be reviewed to assure that they have an adeouate technical basis and do, indeed, minimize risk. PRA analysis may be used as a basis for TS changes, in late 1983, 8WROG formed a TS Improvement Connittee. The committee subsequently established the BWR TS Improvement Program for develuping PRA analysis to identify improvements to STis and A0Ts specified in BWR TS.

The BWROG commissioned GE to perform generic analyses and apply the generic results to the individual BWR plants. At first, two GE topical reports were l issued as a result of this effort: (1) NEDC-30844 (Ref. 4) which analyzed a representative BWR plant and provided a technical basis for ensuring that the currer.t RPS STis meet the recommendations of Generic Letter 82-28 Item 4.5.3 and (2) NEOC-30844 which established a basis for extending the current RPS Sils and A0Ts. These reports used reliability analyses with fault tree l modelling to estimate the RPS failure frequency.

9811090020 981105 "'

PDR TOPRP EMVOEPE C PDR '

ff//410030 -

-. . - ._- -- - - - - - - . - -~-.

I SEP 2 61998 i

Subsequently, as noted above, the tnird GE report HEDC-30936P, on TS I improvement analysis for ECCS actuaticr instrumentation was submitted for our evaluation. This submittal presents a methodology similar to that used in the other two GE reports.

tiEOC-30936P provides a PRA methodology for determining the impact due to ECCS TS changes for BWRs. A demonstration case, using a sample ECCS ,

instrumentation STI and A0T, is analyzed for each BWR product line, in addition, the GE report provides a set of guidelines for licensees on how I

to perform plant-specific evaluations of TS improvement analysis for STis ani 40Ts related to ECCS actuation instrumentation.

1 LROG will submit additional analyses of the remaining ECCS actuation instrumentation as an addendum (Part 2) to NEDC-30936P Part 1 of the TS l improvement methodology. The addendum will include evaluation of Sils and ACis for the reinaining ECCS act':ation inst umentation, and plant-specific verification of the apr'scability of tr e generic results.

4.0 NC ACT10ft The stef f engaged the services of Brookhaven National Laboratory (8til) to review the data at d rnethodclogy in the GE report. This review was to determine the validity of the risk measurc used, assess the adequacy of the f ault tree analyses, supporting data, and reliability calculations ustr.5 the WAMBAl' (Ref. 6), WAMCUT (Ref. 7), and FRAf4 TIC III (Ref. 8) coriiputer codes. A technical evaluation report (TER) (Ref. 9) was issued by BNL presenting the details and results of this review.

S.0 COMMitiTS Ofi THE TOPICAL REPORTS AND THEIR RESULTS:

y3ral Methodology G

Trie rneir. cbjective of the GE report is to provide a methodology for evaluating the safety impact due to proposed changes in ECCS TS require-rnents in cornparison with the current ECCS TS requirements. Core damage frequency (CDT) is used as th' risk rneasure to quantify the safety impact due to ECCS TS changes. Since ECCS is used to inject water into the core for coolitig when feedwater is lost, changes its ECCS TS requi ements would IrrpaCl the availability of the water injection function. The loss of the wdter irjtClion function may lead to core darnage; therefore, water ir.Jectiors availability may be used to bound the impact on CDF due to ECCS TS cnonge:..

The Cf analysis evaluates the changes it. the water injection failure f requency as a result of the changes in the ECCS TS, r amely changes to th* ECCS STI and the A01 for repairing or testirig ECCS actuation instrumeri-tation cr,annels. A brief discussion of the rnethodology used to calculate t he water injection f ailure f reque..cy f ollows

Y l

SEP 2e 199 3-The cdiculation of water injection failure frequency depends on two sets of parameters. The first set consists of initiating events which eventually call for water injection. The second set consists of water

{

injection unavailabilities which are probabilities that the injection water is unavailable given the demands for injection. Depending on each initiating event, the number of components that are needed for water i.njection would va ry. Therefore the water injection unavallobility for a given' initiating event may differ from that given another initiating event.

I for each initiating event, a function fault tree was developed in order to quantify the water injection unavailability per demand. The function f ault tree models the logical relationship of the f aults that contribute to the water injection unavailability. For preliminary scoping analysis, the tirre-independent coraputer code WAf2AM is used to evaluate the functior, fcult tree, estimate the water injection unavailability and scope the tf fect of TS changes on wa g'r injection unavailability. If the increase in wster injecticn unavailat 11 ties cue to the TS changes is greater than I percent, which is the accep dnce criteriori proposed ly GE, then mort realistic time-etpe.edent and yses are required. lhe code WAMClli is used to obtain the minimum cut sets frum the function fault tree for various LC0 conditions, and the computer code FRANTIC !!! 15 ustd to deteroine the effect of the TS changes.

[pplica tion Guidelines The GE report presents a set of application guidelines by which the Gi methodology con be applied for plant-specif ic TS improvement analys1:,.

The guidelines call for an initial consideration of two baseline cases and two scoping Cases.

The first baseline case is to use WANDAM to calculate the water injection unavailability assuming that all channtis of thc ECCS actuation instrumen-totivn cf interest are available. The first scoping case is based t,n the first baseline cese with the exception that all channels are assuned tu be unavailable. If the difference of the results (water injection unavailabilities) of these two cases exceeds the acceptance criterion of I percer.t. then the second baseline arid scoping cases are required.

The second baseline and scoping cases are run with WAHDAM by doubling the time-dependent una.ailabilities of all contonents fur the first boseltne dnd scoping Cases. The ef fect of doubling is equivalent to assuming thut the LCO conditfun occurs at the end of a surveillance interval. The differtu e of the results of the second baseline and scoping cases is again compared with the acceptance criterion of I percent. If the difference is larger thar j 1 percent, then a third scopir.g case is run using the component 6navo11obilit) date for the second baseline case and using a mure realistit assumption that c l suf f icient nebtr of channels are undvailuble fur a L C0 condition.

l.

SEP 2 61980

.t.

If the differente between the results of the third scoping and the second baseline cases exceeds 1 percent, then time-dependent analysis using FRANTIC 111 ,i is needed. In order tu use FRANTIC 111, W!iCUT is used at first to 1 generate the cut sets including the instrumerJath,. of interest at LC0 condition. The cut sets, together with the inforaation on testing and maintenance, are pst into FRANTIC 111. The I percent acceptance

]il i

criterion is also applied to the difference of the results of the FRANTIC }

Ill run with all cut sets included and the FRANTIC 111 run with instrumenta+ ion cut sets included. An iterative process may be needed to aetermine the croposed TS changes which would satisfy the 1 percent acceptance Criterion.

Sample Analyses To demonstrate the feasibility of the plant-specific guidelines, GE considers the impact of A07 and Sil changes on the water injection una va i la b i li ty. Specifically, changes in the A07 and STI are considered far the RPV low-pressure perm' sive signal for LPCI (for each of the four ,

BWF product line grcups, which are BWR-6 solid-state, BWR-3/4, BWR-5/6 1 re lay, and BWR-?). Generic fault trees are prepared for each of the four i BW product line groups. Generic fault trees contain system fault trees for water injection systems, essential support systems, ana function fault trees for the water injection function. These generic fault trees may be u:,ed as a baseline for rncdifications to plant-specific configurations. The Sample analyses are performed for each of the four BWR product line groups.

GE proposes that the STI for the in:,trutncntation function analyzed (kPV l Iow-pressure permissive) be increased from 31 days to 92 days and the -j stinted A0T be increased from I hour to 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. j>

lhe results of the sample analyses indicate that f or all BWR product line i groups, the A0T and $11 for the instrurrsntation function could be increased with little ef fect on water injection unavailability, for example, the 3 e

results f or a BWR-6 solid-state plant using WAMBAM indicate that for the 4 dominating f eilure frequency initiating event is about 1.2ofx loss 10' g/yr for the second baseline andf offsite power, the third scoping cases. The results also indicate that there is rio measurable ,

increase in water injection unavai16bility during the LC0 period of the <

RPV low pressure permissive trip function. Ar additional scoping analysis >

with the STI increased to 18-inoriths produces a negligible increase in ,

una v a i la b il i ty. A confirming analysis using fl! ANTIC 111 with the A0T -

increased to 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> and the STI increased to 100 d6ys also produces a ,

negligible increase in unavailability. -:

Staff findi m ,

I The staff reviewed the TER from DNL and, on this basis, endorses the TER.

[

BHL finds that the inethocolouy and data used in the GE repor* is generally  :

ddeQuate for assessing the risk frnpact due to TS changes. However, BNL }

iderd if ied several important issues which need to be addressed if the A appIltat tun guidelines in th< ct recurt are to be applied to a specific  :

plant for 15 changes. This is discussed as follows:

)

i i

~

SEP 2 6 W

-S-GE used the developed methodology to analyze a single selected ECCS TS requirement on the RPV low-pressure permissive for LPCI. liowever, the ECCS action stotements include a combination of channel failures from both the low-pressure permissive and the discharge line low differential pressurt permissive. It is important not to miss any impact which may be derived f rom the interrelationship of the channels. Therefore one should determine a minimal group of interreldted TS requirements for a rneaningful TS improve-ment study.

In addition, analysis is required to define the boundaries of an instrumen-tation channel and identify.what components should be included for TS study.

fcr example, the instrument line of the RPV pressure sensors is cornmon with some level sensors. If the channel boundary, as defined by the functional surveillance tests, includes the instrument line, then interface with the additional group of TS related to the level sensors should be considered.

There are various LCO conditions that arise during operation. Depending on the LC0 conditions chosen for TS study, the results of TS study may vary.

In orde+ to assess fully the irnpact of the TS changes, it is important to select the most bounding LC0 conditions for TS analysis.

With respect to a plant-specific TS improvernent analysis, it is important to review the plant-specific procedures end infortnation relating to testing, raaintenance and calibration of ECCS actuation channels. Since the plant-specific data vary greatly among the plants, each plant should use its own data and perform time-dependent combined A0T/ST) analyses to deterraine the impact as a result of the proposed TS changes. This issue was not explicitly addressed in detail.

The issues discussed above are mainly due to lack of sufficient considerations in the GE TS analyses with respect to plant systerns, instrumentation channels and plant procedures. We believe that, with respect to the proposed set of application guidelines, rnore detailed analyses, which include the issues discussed above, are needed for mednirigful plant-specific TS changes.

6.0 CONCL USI0ff On the basis of the Bill findings, we conclude that the GE analyses have demonstrated the acceptability of general methodology fur considering TS changes to the ECCS instrumentation used in the BWR facilities. With respect to plant-specific implementation of chances in the ECCS TS for a plant, we have identified a list of plant-specific issues which need to be addressed by each licensee or applicant in order to rnake any proposed ECCS TS changes acceptable.

t l

l 1

7.0 REFERENCES

1. D. Atcheson, et al., "BWR Owrers' Group Technical Specification Impruve-tra at Methodology (with Demonstration fer BWR ECCS Actuation Instrumentationi)

Part 1," General Electric Company, NEDC-30936P, November 1985.

2. NRC Letter, D. G. Eisenhut to All Licensees of Operating Reactors, Applicants for Operatilig License, and holders of Construction Permits, " Required Actions Based on Generic implications of Salem ATWS Events (Generic Letter 83-28)," July 8,1983.

3. NUREG-1024, " Technical Specifications - Enhancing the Safety impact,"

November 1983. .

4 S. Visweswaren, et al., "8WR Owners' Group Response to NRC Generic letter 83-28, Item 4.5.3," General Electric Company, NEDC-30844, January 1985.

5. W. P. Sullivan, et al., " Technical Specification Improvement Analyses for BWR Reactor Protection Systen," General Electric Company, NEDC-30851P, May 1965.

6. F. L. Leverenz and H. Kirch, " User's Guide for the WAM-8AM Computer Code," EPRI, January 1976.

7. R. C. Erdmann, F. L. Leverenz, and H. Kirch, "WAMCUT: A Computer Code for f ault Tree Evaluation," Science Applications, Inc., EPRI NP-803, June 1978.

8. T. Ginzburg, et al. , " FRANTIC !!!: A Computer Code for Tirne-Dependent Feliability Ar.alysis," Brookhaven National Laboratory and Science Applications, Inc. , April 1984 '

9. D. Ilberg, et al., " Review of 8WR Owners' Group Technical Specification Ir.iprovement Methodology," Brookhaven National Laboratory, February 1987.

5 SEP 26 19 l

TABLE 1 CONDITIONS FOR CLOSING OUT BWR PLANTS On plant-specific applications of the methooology for proposing technical speci-fication changes to the ECCS, individual licensees for plants must perform plant-specific analyses with consideration of the following items:

(1) Determine the minimal group of TS for consideration of TS changes I and provide the technical bases for the determination.

(2) Identify all equipment for consideration of TS changes to which the surveillance tests required by the TS are applicable.

(3) Determine the complete set of LC0 conditions that can arise in plant operation and select the most bounding LCO conditions for the analysis.

(4) Perform confirmatory, time-dependent, combined A0T/STI analyses, using plant-specific information on testing, maintenance and calibration.

Analyses in reference to the above items should reflect the views discussed in the staff's evaluation.

9

-