ML20138J833

WPPSS Nuclear Project Unit 2 TER on IPE Back End Analysis

ML20138J833
Person / Time
Site:	Columbia
Issue date:	09/30/1996
From:	Meyer J, Hanry Wagage SCIENTECH, INC.
To:	NRC
Shared Package
ML17292A805	List: ML20138J827 ML20138J833 ML20138J839
References
CON-NRC-05-91-068-46, CON-NRC-5-91-68-46 SCIE-NRC-237-95, NUDOCS 9702130116
Download: ML20138J833 (33)

Text

y; , :t c o'

c. r :.

L s 4

1 6

APPENDIX B s

WASHINGTON PUBLIC POWER SUPPLY SYSTEM - NUCLEAR PROJECT.NO'2 TECHNICAL EVALUATION REPORT .

(BACK-END) f i-I

+

,f

^~

-' ./

w'- ,

G~ o/>

SCIE-NRC-237-95 .

1 1

t s

WPPSS Nuclear Project Unit 2 l Technical Evaluation Report ,

on the Individual Plant Examination Back-End Analysis  !

-l 4

H. A. Wagage J. F. .Meyer I

Prepared for the U.S. Nuclear Regulatory Commission Under Contract NRC-05-91-068-46 September 1996 SCIENTECH, Inc.

11140 Rockville Pike, Suite 500 Rockville, Maryland 20852 l

s I

l I

y- . - .,-

.A s TABLE OF CONTENTS P.Aat .

E. EXECUTIVE S UMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-1

1. INTRODUCTION .......................................1

, 2. TECHNICAL REVIEW .....................................4~ r 2.1 Licensee s lPE Proceu .................................4 ,

2.1.1 - Completeness and Methodology . . . . . . . . . . . . . . . . . . . . . . 4

. 2.1.2 Multi Unit Effects and As-Built /As-Operated Status . . . . . . . . 5 2.1.3 Licensee Participation and Peer Review ................ 6 >

2.2 Containment Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2.1 Front end Back-end Dependencies . . . . . . . . . . . . . . . . . . . . 7 2.2.2 Containment Event Tree Development .................7 2.2.3 Containment Failure Modes and Timing . . . . . . . . . . . . . . . . 8 2.2.4 Containment Isolation Failure .......................10 2.2.5 System / Human Response . . . . . . . . . . . . . . . . . . . . . . . . . . 1I 2.2.6 Radionuclide Release Categories and Characterization . . . . . . 12

~

2.3 Quantitative Assessment of Accident Progression and Containment Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.3.1 Severe Accident Progression . . . . . . . . . . . . . . . . . . . . . . . . 14 2.3.2 Dominant Contributors: Consistency with IPE Insights . . . . . 15 2.3.3 Characterization of Containment Performance . . . . . . . . . . . 16 2.3.4 Impact on Equipment Behavior . . . . . . . . . . . . . . . . . . . . . . 17.

2.3.5 Uncertainty and Sensitivity Analysis ..................17 2.4 Reducing Probability of Core Damage or Fission Product Release . . . 19

. 2.4.1 Definition of Vulnerability ............... .........19 ,

2.4.2 Plant Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

. 2.5 Responses to CPI Program Recommendations . . . . . . . . . . . . . . . . . 20 2.6 IPE Insights, Improvements, and Commitments . . . . . . . . . . . . . . . . 21

3. CONTRACTOR OBSERVATIOhS AND CONCLUSIONS . . . . .. .. .22...

4. REFERENCES .......................................23 l

APPENDIX: IPE EVALUATION AND DATA

SUMMARY

SHEET l

WPPSS Nuclier Project Unit 2 ii September 1996

___. - -. . -. =-. -- .- - - . . . - . _

.s ,

2 E. EXECUTIVE

SUMMARY

SCIENTECH,Inc., performed a review of the back-end portion of the Washington Public Power Supply System's WPPSS Nuclear Project Unit 2 (WNP-2) Individual Plant

. Examination submittal.

]

E.1 Mant Garacterization WNP-2.is a boiling water reactor of General Electric BWR 5 design. De rated thermal i power is 3,323 MWt which is produced by 764 fuel bundles. The reactor pressure vessel 4 (RPV) operates at a saturation pressure temperature and pressure of 549'F and 1020 psia.

The design pressure of the RPV is 1250 psig. De vessel is protected from overpressure by 18 safety relief valves (SRVs). Each main steam SRV is piped to a submerged quencher in the wetwell. The main steam SRV tail pipe vacuum breakers are located in <

the drywell.

. WNP-2 employs a Mark II pressure suppression containment design. The primary containment is a free-standing steel vessel which is divided into two major regions, the drywell and wetwell. The drywell has the shape of a truncated cone with a cap and it j houses the reactor and its associated primary system. ' The wetwell has the shape of a ,

!- cylinder with a round bottom. The drywell floor serves as a pressure barrier between the l

, drywell and the wetwell. There are 99 downcomer pipes which penetrate the drywell floor and provide a flow path for steam and gas from the drywell to a pool of water  !

(suppression pool) in the wetwell in rn accident situation. There are nine wetwell-to- l

. drywell vacuum breakers which provide a flow path for noncondensible gases from the  ;

wetwell air space back to the drywell. Three vacuum breakers between the reactor building and the wetwell are provided to limit the external force resulting from the

! condensing steam within the containment.

De pedestal region is recessed relative to the drywell floor. There are two 8-foot by 6-foot sumps cast into the pedestal floor. The sumps each have 3/8" thick stainless steel covers and normally contain water to a depth of about 17 inches (490 gallons each). If the drywell sprays are used, the water collects in trenches in the drywell floor. Water in

. the trenches drains via two 4-inch drain lines to the Floor Drain Processing sump in the pedestal. De sumps have drain lines which are routed beneath the surface of the suppression pool before exiting the containment. The drain lines are closed as part of containment isolation. There are no downcomers in the pedestal region.

$ E.2 IJcensee s IPE Process De general apprc,ach used in the quantification of the containment performance for the

. WNP-2 used the following four analytical steps. First, the selection of the actual

~

sequences or sequence cutsets were placed into each individual group based on their functional characteristics and the status of systems which are important to the containment performance assessment. . When each Level I sequence which had a frequency greater than  ;

WPPSS Nuclear Project Unit 2 E1 September 1996 l

• , l 1

IE-9 per year had been assigned to a group, the group was processed with its associated  ;

~l initiator specific containment event tree (CET). ,

Second, a set of CETs was developed to model accident progression and provide a description of the possible outcomes or containment damage states, which can result from each of the specific plant damage state (PDS) identified by the Level I analysis. Each of these CETs consisted of thirteen top events. The time frame for the Level U analysis was assumed to extend for 40 hours4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> after accident initiation. Though these CETs were developed for sach PDS, they also tend to correspond to initiator type. For WNP-2, CETs were developed for the following six PDS groups:

l

• Short-term station blackout (SBO) (power restored within four hours, hatteries remain available to provide DC power throughout an accident) j = Long-term SBO (power not restored within four hours so that battery depletion 2

results in loss of DC power) j - Transient

• Anticipated transient without successful reactor scram (ATWS) l l

• Small LOCA (break sin less than 4 inches, primary system will not depressurin unless automatic or manual depressurintion is initiated) j

• Large LOCA (break sin more than 4 inches, primary system will depressurin) t-

[ Third, quantification of the CETs to provide the estimated frequency for each individual

! sequence was accomplished by the insertion of the appropriate conditional probabilities at

each of the CET branch nodes. Final quantification was the result of propagation of each

initiating plant damage state and its associated occurrence frequency through its respective

CET and accumulating these frequencies for each defined source term group or release l category. ,.

i The CET branch probabilities were calculated in one of two ways:

s

- From fault trees developed to identify each of the individual functional failures which are im wrtant to resolution of the node

. ' Split fractions which could be assigned to each branch node. ,

4 Fourth, a set of criteria was developed to be used as the basis for grouping containment event tree end states into a limited, but complete, set of unique release categories. The following sequence characteristics were used to daracterin the release categories:

- Containment is/is not bypassed l

. WPPSS Nuclear Project Unit 2 E-2 September 1996

'k 1m. . . _ , . .-.. _,,

i

• Containment is/is not isolated ,

• Fission products are/are not scrubbed

- Time of containment failure (early/ late)

• Containmer;t failure mode (size: large/small; location: drywell/wetwell/ vent)

A set of simple logical rules was developed to use the above characteristics to consistently sort and accumulate the frequency contribution from each sequence into one of 26 defined source term bins.

To determine a representative source team for each bin which had an occurrence frequency greater than the assigned cut off value, a representative sequence was selected from each bin and used to define a MAAP simulation which would provide an estimate of the fission product release.

The Supply System staff performed all aspects of the IPE. His included the CET

- preparation and quantification, phenomenological MAAP analyses, primary containment structural analysis, and sensitivity analysis. The Supply System hired a consulting firm, Individual Plant Evaluation Partnership (IPEP, a partnership between Tenera, L.P., Fauske and Associates, Inc., and Westinghouse), to assist in preparation of the initial IPE. The '

IPEP provided 1) training in PRA and special topics like Human Reliability Analysis and

2) peer review of the IPE products as they were generated. To ensure an independent assessment of the conservatism that existed in the original IPE, the Supply System hired a consultcnt, NUS Corporation, to recommend model improvements and eliminate data conservatism. Revision 1 of the IPE was the result ofimplementing NUS's recommendations.

The WNP-2 IPE received a multi-tiered review in terms of technical review, peer review, independent in-house review, and management review. IPEP reviewed all quantification results and phenomenological issue position and results for the original analysis. NUS reviewed the same for the baselined Revision I analysis.

I By performing finite element analysis by using the ANSYS computer program, the IPE team determined the containment failure pressures to be 148.0,121.0. and 103.5 psig at

707,3407, and 6007, respectively. Failure was equally likely to occur at the upper cone / upper cylinder junction or in the wetwell airspace.

The IPE team found that in most cases, MAAP results showed that the failure pressure of 121 psig at 340Tould be reached first. Therefore, in the IPE study the containment failure pressure was assumed to be 121 psig.

E.3 Back-End Analysis The IPE team used the Modular Accident Analysis Program (MAAP) 3.0B, Revision 7.02

> and 7.03, to provide an integral approach for modeling of plant thermal hydraulic response and fission product transport during severe accidents. They compiled plant-specific data WPPSS Nuclear Project Unit 2 E-3 September 1996 l

'a

4 , .!

?: ..

l a.

i ,

I derived from plant design documents and drawings into the MAAP parameter file. ney l analyzed different severe accider. events (SBO, transients, N!WS, and LOCA) using the

_ MAAP code. In addition to MAAP results, they used research results in the open literature, IDCOR task reports, Shoreham and Limerick PRAs, NUREG reports, and )

engineering judgment in understanding physical processes and developing event treesc l 4

De IPE team used the MAAP code to provide integral analyses of accident sequences ,

4 from their inception until several hours after containment failure. Because MAAP 3.0B l 4 did not. include models for DCH, hydrogen detonation, and steam explosion, they l considered these as separate effects. l

\

EA Comedannent Perfersamace heprovement Issues  !

j

In Generic Letter 88-20, Supplement No. 3, NRC recommends that licensees consider insights from the CPI Program, specifically for Mark H containments: additional containment heat removal capability (e.g., hardened vent or improved reliability of  ;

~

i suppression pool cooling). Also, Supplement No. 3 requests Mark H licensee

- consideration of insights directed at the Mark I BWRs and delineated in Generic Letter 88-20, Supplement No.1: (a) alternate water supply for drywell spray / vessel injection; (b) enhanced RPV depressurization reliability; and (c) improved EOP and training. De licensee appears to have responded to all of the above recommendations.

i After reviewing the containment heat removal systems, the IPE team concluded that the j reliability of those systems were acceptable and no further changes were warranted. I 4 After performing a sensitivity study to investigate the benefits of installing a hardened

' containment vent, the IPE team concluded that once core damage occurred, hardened i venting had very limited benefits in reducing offsite consequences. t The licensee response to the insights directed at the Mark I BWRs were as follows:  ;

(4 Altemare water supply for drywell symy/ vessel injection. The IPE team noted that the i' IPE assumptions for the availability of.HP injection water inventory were limited by the

! assumed capable capacity of the CST, rather than by hardware (tank) availability. The

! available inventory was found to be twice the value assumed for the IPE, thus no additional benefits were expected from an altemate water supply.

(b) Enhanced RPV depressurization reliabiliry. The IPE team found that during most

, WNP-2 accident sequences, the reliability of RPV depressurization was dominated by human error. His included the failure to initiate manual depressurization to delay core melt during a short term SBO, or following failure of all HP injection systems in non-SBO sequences so that LP injection could be initiated. Significant improvement in the reliability of the operator performing this action was not expected to be gained. The only i

feasible actions for enhancing the probability of successful depressurization during most s

WPPSS Nuclear Project Unit 2 E-4 September 1996

G >

accident sequences were limited to confirmation that they were addressed and practiced in the operator training program.

(c)ImprovedEOP ad training. The EOPs implemented at WNP-2 had been based on i Revision 4 of the BWROG EPG since 1990. The corresponding operator training was '

also conducted according to the EOPs as based on Revision 4 of the EPGs.

E.5 Vulnerabilities and Plant Improvements The WNP-2 IPE team identified no vulnerabilities in the WNP-2 design or operation. For WNP-2, vulnerability screening was based on:

• sequence groups with frequencies greater than IE-6 per year that would require modifications per NUMARC 91-04 guidelines, total CDF must be within the NRC's safety goal of IE-4 per year, and

. sequences that indicate a plant specific feature that is an outlier to comparable BWR PRAs.

. During the course of the WNP-2 IPE, the IPE team identified one back-end plant improvement and two back-end procedure improvements. A hardware modification of the air supply to the inboard MSIVs and the containment vent valves for backup from the containment nitrogen system was investigated. The modification would improve long-

term decay heat removal by providing redundant air supply to the valve's solenoids. This  ;

modification was found to be marginally cost effective and, therefore was not l recommended for further plant evaluation. The following were the back-end procedural '

I improvements:

1

• SBO Depressurization. The IPE team recommended that the longer coping time q be evaluated for incorporation in the SBO emergency procedures, if the vessel was maintained at pressure with the benefit of depressurizing once fuel melt started, but before vessel breach.

• DrywcFWetwellBypass. The IPE team found that the Omega seal design separating the drywell and wetwell air spaces was a passive design with very low failure rate. However, its failure resulted in very large consequences. Therefore, the IPE team recommended an evaluation of the costs and benefits of periodic inspection and maintenance of the Omega seal be made.

1 l

l WPPSS Nuclear Project Unit 2 E5 September 1996 2

, l

's. .,

l 4

E.6 Observations SCIENTECH observed the followirig on the WNP-2 IPE:

• De IPE team responded to ai! the recommendations of the CPI Program which' are l

relevant to WNP-2. l

• De WNP-2 IPE team identified no vulnerabilities in the WNP-2 design or operation.

1

• . ne IPE team identified one back-end plant improvement and two back-end procedure improvements.

l

• During the WNP-2 analysis, failure of a penetration was assumed to be incapable of i removing enough energy to prevent containment failure so all sequences involving l

- failure to isolate were described with a combined source term, that from the  :

penetration and that from the associated containment failure.

• The licensee participated in performing and reviewing the IPE.

4

• The submittal demonstrated that the IPE team well understood the severe accident l phenomenology important for WNP 2. j l

• The IPE team adequately considered the role of operator actions on severe accident f progression at WNP-2.

• The IPE team was considering several IPE insights but they were not included in the  ;

submittal. The licensee was working with the appropriate BWROG committees in -

formulating the accident management procedures based on the recognized insights.

• The IPE team did not use a screening criterion that compared the fission product release categories of.WNP 2 to BWR-3 or PWR-4 release categories of WASH-1400.

However, they identified a release category with large releases (STG-10) which had a frequency of 1.05E-6 per year, Overall, the WNP-2 IPE appears to have achieved the objectives of GL 88 20.

1 WPPSS Nuclear Project Unit 2 E-6 September 1996

-. - , _ _ _ _ _ . _ _ _ -__ - ___J

l l.

l 1.- INTRODUCIlON ,

i l 1.1 - Review Process His technical evaluation report (TER) documents the results of the SCIENTECH review of the back-end portion of the WPPSS Nuclear Project Units 2 (WNP-2) Individual Plant '

Examination (IPE) submittal. [1,2) This technical evaluation report complies with the

. requirements for IPE back-end reviews of the U.S. Nuclear Regulatory Commission <

i (NRC) contractor task order, and adopts the NRC review objectives, which include the i

following:

1 i

- To help NRC staff determine if the IPE submittal provides the level of detail  !

requested in the " Submittal Guidance Document," NUREG-1335.

r

- To help NRC staff assess if the IPE submittal meets the intent of Generic ';

Letter 88 20.

1 i

• To complete the IPE Evaluation Data Summary Sheet. .

SCIENTECH made a presentation on its preliminary review findings of the Back-End portion of the WNP-2 IPE submittal to NRC in June 1995. Based in part on these findings, the NRC staff submitted a Request for Additional Information (RAI) to

Washington Public Power Supply System (WPPSS) on August 7,1995. WPPSS '

responded to the RAI in a document dated October 20,1995. [2] This TER is based on the original submittal and the response to the RAI.

Section 2 of the TER summarizes SCIENTECH's review and briefly describes the WNP-2 IPE submittal, as it pertains to the work requirements outlined in the contractor task order.

Each portion of section 2 corresponds to a specific work requirement. Section 2 also outlines the insights gained, plant improvements identified, and utility commitments made as a result of the IPE. Section 3 presents SCIENTECH's overall observations and conclusions. References-are given in section 4. De appendix contains an IPE evaluation and data summary sheet.

1.2 Plant Garactedsstion WNP 2 is a boiling wate'r reactor of General Electric BWR 5 design. The rated thermal

. power is 3,323 MWt which is produced by 764 fuel bundles. The reactor pressure vessel

~

(RPV) is a low alloy steel vessel with a stainless steel interior clad. De RPV is 6.69 inches thick including the cladding with reinforcement in the area of the vessel where the jet pumps reside to a total thickness of 10.13 inches. The bottom head is reinforced to a thickness of 8 inches to accommodate the 185 control rod drive penetrations. De RPV has an intemal diameter of 20 feet,11 inches, and an intemal height of 72 feet,11 inches.

WPFSS Nuclear Project Unit 2 1 September 1996

., ,m . ,, - _ _ . _ _ _ _ _ , _

._ _ . . . __ _ . ._ _. . _ . _ _ _ . _ . ~ . . _ _ . _ . _ . _ _ _ _ _ _ _ _ . _ _ _ . __.__

l l*. ,

i I L I ne total mass of UO in 2 the core is 349,900 pounds with a normal enrichment of i 2.6 percent. He UO2fuel ellets P are clad in ziresloy tubing with a thicknes's of L 0.035 inches. Fuel assemblies contain 61,900 pounds of zircaloy. The RPV contains l

. 185 control rods with 14.86 pounds of B 4C and 224 pounds of stainless steel in each.

l l The'RPV operates et a saturation pressure temperature and pressure of 549'Fnd 1020 psia. The design pressure of the RPV is 1250 psig. The vessel is protected from

overpressure by 18 relief valves. Each safety relief valve (SRV) has two set points - a

relief set point and a safety set point. In the relief mode, the SRVs lift in banks at

l pressures between 1076 psig and 1116 psig. In the safety mode, SRVs lift at pressures  !

l between 1150 psig and 1205 psig. Each main steam SRV is piped to a submerged .

quencher in the'wetwell. He main steam SRV tail pipe vacuum breakers are located in  !

the dryviell.

! RPV venting can also be accomplished via the power conversion system (PCS)if a mam i steam line can be opened (or remains open through the transient). This vent path routes steam from the RPV to the main turbine condenser. The main steam isolation valves ~

(MSIVt) are capable of opening at containment pressures less than or equal to 54 psig. l l The MSIVs are air to-open and spring-to-close valves; and at containment pressures >

l exceeding 54 psig, the control air has insufficient pressure differential to open the MSIV i closure springs.

i WNP-2 employs a Mark Il pressure suppression containment design. The primary 1- containment is a free-standing steel vessel which is divided into two major regions, the

. drywell and wetwell. The drywell has the shape of a truncated cone with a cap and it

! houses the reactor and its associated primary system. The wetwell has the shape of a ,

cylinder with a round bottom. The drywell floor serves as a pressure barrier between the l

, drywell and the wetwell. There are 99 downcomer pipes which penetrate the drywell floor and provide a flow path for steam and gas from the drywell to a pool of water j (suppression pool) in the wetwell in an accident situation. There are nine wetwell-to-p drywell vacuum breakers which provide a flow path for noncondensible gases from the j wetwell air space back to the drywell. Three vacuum breakers between the reactor

building and the wetwell are provided to limit the extemal force resulting from the condensing steam within the containment.

j De pedestal region is recessed relative to the drywell floor. There are two 8 foot by 6-foot sumps cast into the pedestal floor. He sumps each have 3/8" thick stainless steel j

covers and normally contain water to a depth of about 17 inches (490 gallons each). If F the drywell sprays are used, the water collects in trenches in the drywell floor. Water in 3 the trenches drains via two 4 inch drain lines to the Floor Drain Processing sump in the pedestal. De sumps have drain lines which are routed beneath the surface of the

. suppression pool before exiting the containment. The drain lines are closed as part of containment isolation. There are no'downcomers in the pedestal region.

4 1

4

- ~ WPPSS Nuclear Project Unit 2 2 September 1996

't I

.L. -

s ,

e De physical dimensions of the steel containment are as follows (section 4.1.2 of the submittal):

• . Diameter of the cylindrical portion at the base of the cone is 86 feet. .

= Diameter at the top of the cone is 39.5 feet, and then narrows to 32 feet to cany

the head.

i

• Bottom head has an inside height of 21.5 feet with ellipsoidal head of 2:1 ratio, ,

and varies in plate thickness from 7/8 inches to-1.5 inches. .

- Top head has an inside height of 15.5 feet,15/16-inch thickness and is bolted with a flanged joint.

4 - Drywell steell is 99 feet high. ,

1

= Suppression chamber is 72 feet high. ,

i

= Overall steel height is 171 feet.

e

= Vessel shell plate thickness varies from 1.5 inches at the lower head to 3/4 inch at the drywell conical section.

A reinforce concrete biological shield wall surrounds the primary containment. There is a 2.25 inch gap filled with insulation between the containment and the biological shield which serves to allow interference between the two structures. The concrete used in the l WNP-2 biological shield wall and basemat is a basaltic variety.

Normal suppression pool water depth is 31 feet which gives a downcomer submergence of 11 feet,10 inches.

The primary containment can be vented either from the wetwell or the drywell.

Containment venting would be accomplished through 24 inch containment purge lines.

The emergency operating procedures (EOPs) direct operators to vent containment if internal pressure rises above 39 psig. He operator is instructed to vent the wetwell region first and to vent the drywell only if the wetwell vent path fails to open. The butterfly isolation valves in the 24-inch purge lines are designed to be able to open against containment pressure as high as 49 psig.

i l

WPPSS Nuclear Project Unit 2 3 September 1996

• ~r .,

i

2. 'IECHNICAL REVIEW 3

2.1. IJeemsee's IPE Process 211 Completeness and Methodology.

t The WNP-2 IPE submittal contains a substantial amount ofinformation regarding the recommendations of Generic Letter (GL) 88-20, its supplements, and NUREG-1335. The'

> submittal appears to be complete in accordance with the level of detail requested in

- NUREG-1335. De methodology used to perform the IPE is described clearly in the

. submittal. De approach taken, which is consistent with the basic tenets of GL 88-20, l appendix 1, is also described clearly along with the team's basic underlying assumptions.

The important plant information and data are well documented and the key IPE results and findings are well presented.

The general approach used in the quantification of the containment performance for the WNP-2 used the following four analytical steps (section 4.0.1 of the submittal):

First, the selection of the actual sequences or sequence cutsets were placed into each individual group based on their functional characteristics and the status of systems which are impocant to the containment performance assessment. When each Level I sequence which had a frequency greater than IE 9 per year had been assigned to a group, the group l was procesred with its associated initiator specific containment event tree (CET)

Second, a set of CETs was developed to model accident progression and provide a .

description of the possible outcomes or containment damage states, which can result from each of the specific plant damage state (PDS) identified by the LevelI analysis. Each of these CETs consisted of thirteen top events as listed in section 2.2.2 of this report. The j time frame for the Level II analysis was assumed to extend for 40 hours4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> after accident .

! initiation. Though these CETs were developed for each PDS, they also tend to correspond to initiator type. For WNP-2, CETs were developed for the following six PDS groups:

• Short-term station blackout (SBO) (power restored within four hours, batteries remain available to provide DC power throughout an accident)

- Long-term SBO (power not restored within four hours so that battery depletion

[ results in loss of DC power) p h

• Transient

= Anticipated transient without successful reactor scram (ATWS) s, 4

. Small LOCA (break size less'than 4 inches, primary system will not depressurize 4

unless automatic or manual depressurization is initiated) 1 WPPSS Nuclear Project Unit 2 4 Sepumber 1996 v,--

i i

Large LOCA (break sin more than 4 inches, primary system will depressurin) l . Third, quantification of the CETs to provide the estimated frequency for each individual j sequence was accomplished by the insertion of the appropriate conditional probabilities at

'each of the CET branch nodes. Final quantification was the result of propagation of each

! initiating plant damage state and its associated occurrence frequency through its respective CET and accumulating these frequencies for each defined source term group or release category.

l The CET branch probabilities were calculated in one of two ways:

l .

From fault trees developed to identify each of the individual functional failures

which are important to resolution of the node Split fractions which could be assigned to each branch node.

~

Fourth, a set of criteria was developed to be used as the basis for grouping containment i

, event tree end states into a limited, but complete, set of unique release categories. The i following sequence characteristics were used to characterin the release categories:

i.

Containment is/is not bypassed  ;

} .

. Containment is/is not isolated i j - Fission products are/are not scrubbed l Time of containment failure (early/ late)

{

=

Containment failure mode (sin: large/small; location: drywell/wetwell/ vent)

. A set of simple logical rules was developed to use the above characteristics to consistently i sort and accumulate the frequency contribution from each sequence into one of 26 defined source term bins, i.

To determine a representative source team for each bin which had an occurrence i frequency greater than the. assigned cut-off value, a representative sequence was selected from each bin and used to define a MAAP simulation which would provide an estimate of 4

the fission product release.

4 212 Multi-Unit Effects and As Built /As-Onerated Status.

I I

Revision I to the IPE used plant data and system design as it was at the end of 1993 with I few exceptions. The IPE team performed walkdowns for the purpose of ensuring the following:  ;

• The as-built system was consistent with the flow diagram used in the fault tree

- development. '

WPPSS Nuclear Project Unit 2 5 September 1996

i

+ The system lineup during normal operation was consistent with assumptions made in the fault tree development and described in the system notebook.

All local vulnerabilities (high room temperature and humidity) were accounted in the fault tree development.

• All support systems were accounted in the fault tree development. I l

Walkdowns were performed in accordance with an engineering standard, IPE System Walkdown, developed specifically for the IPE program. Walkdown Checklists that were filled out during the walkdown were part of the second level documentation retained at the Supply System.

2M Licensee Particination and Peer Review. l The Supply System staff performed all aspects of the IPE. This included the CET preparation and quantification, phenomenological MAAP analyses, primary containment  !

structural analysis, and sensitivity analysis. The Supply System hired a consulting firm, j Individual Plant Evaluation Partnership (IPEP, a partnership between Tenera, L.P., Fauske j and Associates, Inc., and Westinghouse), to assist in preparation of the initial IPE. The  !

IPEP provided 1) training in PRA and special topics like Human Reliability Analysis and

2) peer review of the IPE products as they were generated. To ensure an independent assessment of the conservatism that existed in the original IPE, the Supply System hired a consultant, NUS Corporation, to recommend model improvements and eliminate data l

. conservatism. Revision 1 of the IPE was the result ofimplementing NUS's  !

recommendations. .

The WNP-2 IPE received a multi tiered review in terms of technical review, peer review, independent in house review, and management review. IPEP reviewed all quantification results and phenomenological issue position and results for the original analysis. NUS reviewed the same for the baselined Revision 1 analysis.

1 1

The Supply System established an independent In house Review Team which was composed of individuals who were not associated with the preparation of the IPE but were i I

experienced and knowledgeable of WNP-2 systems, safety evaluations, design basis safety analyses, reliability methodology, and/or training in these functional areas.

The In house Review Team met for several half-day sessions _ Each session was conducted to a) respond to concems/ questions raised at the previous session, b) have a member of the IPE team introduce a topic, e.g., system failure tree modeling, c) present the results in the appropriate section of the IPE Report, and d) respond to questions or record comments for resolution by the next meeting.

The In-house Review Team and IPEP reviewed the WNP-2 IPE Report, Revision 0. They conducted reviews at various stages of report completeness to ensure resolution of I

WPPSS Nuclear Project Unit 2 6 September 1996

,4

l t

comments addressing technical correctness and issues to report content and ci,arity. A list of outstanding inues raised during the review that did not significantly impact the results, but which the Supply System plans to address in the long term as maintaining the IPE, included the following (page 5.0-3, reference 1):

i De assumption that containment failure in the wetwell region (1/3 of the failure

, probabilities) would always lead to injection failure is too conservative. A more l realistic evaluation would lower the TW sequences contribution to CDF. This evaluation will be done as part of the configuration control of the IPE.'

2.2 Containment Analysis l 211 Ergal-and Back-end Dependencies

, The IPE team decided to manage Level 1/2 dependencies in the CETs and their associated l fault trees by developing initiator specific CETs. They binned each of the individual l Level 1 sequences, which represent a contribution to overall CDF of at least IE-9 per  !

E year, into one of several PDS bins. Each bin had the characteristics to represent its  :

members throughout the Level 2 analysis. The grouping was based on the similarity i between their Level I characteristics. Each bin, therefore, had unique characteristics

regarding containment condition before/during core degradation, reactor coolant system  ;

4 condition during core degradation, and containment safeguards system performance. The  !

IPE team binned accident sequences into six PDS bins as described in section 2.1.1 of this  !

report.

I It appears that the front end back-end dependencies have been accounted adequately.

J

. 1 212 Containment Event Tree Develonment. J l l i Using a CET for each of the six PDSs obtained from the Level 1 analysis, the IPE team determined the containment failure mode and source term release probabilities. These CETs consisted of the following top events (section 4.5.1.1, page 4.5-5 of the submittal):

l '

= Depressurization (DPR) l

• Power recovered prior to. vessel failure (PWRVF)

'

• High pressure vessel injection (HPI)

• Low pressure injection (LPI)

Power recovered after vessel failure but before containment failure (PWRCF)

!

• Injection recovered before containment failure (ICF)

, Debris cooled (DC)

Shell failure (SF)

RHR recovered (RHR)

=

Vent recovered (VNT)

• Recovery of PCS as a containment heat sink (PCS) i Containment failure mode (CFM)

WPPSS Nuclear Project Unit 2 7 September 1996 9

'A.

- , ,- a

. -. . . - - . . - - -. - . . ~ . - - . .

O g Suppression pool bypass (SPB).

The IPE team developed simple fault trees to relve each node for each accident sequence. The solution of the fault tree provided the relative likelihood with which each

, CET node branch represented the expected accident progression path. The fault tree solution returned the values of (section 4.5.1, page 4.5-2)

1.0 or 0.0 (true or false) if the outcome could be derived from information which was explicitly defined in the (Level 1) PDS description, or by the outcome which had been defined for events which had occurred previously during the cor.tainment accident sequence.  !

a probability, or split fraction, if the node could not be resolved definitely because there was uncertainty in either the behavior of plant equipment which must respond dudng the accident, or because there was uncertainty associated with the phenomena which influenced the outcome from the event represented by the CET .

node.

d The WNP-2 CETs are consistent with the CETs of other PRAs for similar plants.

2,21 Containment Failure Modes and Timinn.

The IPE team considered the loading on the containment both from pressure and,

. temperature. Steam overpressure loading was considered to occur while decay heat was 1 transported to the suppression pool in the absence of supprersion pool cooling. This loading occurred before vessel failure while the core was covered with water and the decay heat was being transferred by water / steam flows by either RHR, via the SRVs, or through a pipe break inside containment. Temperature and pressure conditions in both the drywell and the wetwell were approximately those of saturated water at the pool bulk temperature.

Overtemperature loading was considered to occur after core cooling was lost, the core i melted, and vessel failed and the debris was located in the pedestal without a continuous water supply to the pedestal. Decay heat was transferred to the pedestal concrete by conduction and to the drywell by radiation and convection. Pressure rise was not

significant because the concrete was basaltic which released so little non-condensible gases. There was a limited increase in containment temperature at vessel failure. The upper regions of the drywell reached temperatures of 400 to 900'F d

Steam overpressure loading was considered to occur if the pedestal floor collapsed because of concrete ablation before the containment shell failed elsewhere. The core ,

debris would relocate into the pool and the decay heat would be transferred directly to the water. The pressure and temperature conditions in the drywell and the wetwell would tend to retum to equilibrium with the pool bulk temperature saturation values.

WPPSS Nuclear Project Unit 2 8 September 1996 T'

W

7 ,

i The IPE team found that fast pressure risc. events such as vessel rupture, very large LOCAs, ex-vessel steam explosions, and hydrogen combustion would occur at low i frequencies. In these cases, they assumed that gross failure would occur because of the

uncertainty associated with the dynamic response of the containment to acute temperature

and pressure loading.

l There is a 2.25 inch gap between the containment shell and the biological shield wall. )

The gap is filled with compressible insulation material consisting of polyurethane flexible

- foam sheets. Because of the relative compressibilities of polyurethane, containment shell  :

and the4iological shield wall, polyurethane was neglected in the analysis. I The containment failure pressure and location were determined without considering the biological shield wall. By performing finite element analysis by using the ANSYS computer program, the IPE team determined the containment failure pressures to be 148.0, 4

121.0. and 103.5 psig at 707,3407, and 6007, respectively. Failure was equally likely

to occur at the upper cone / upper cylinder junction or in the wetwell airspace.

The IPE team found that in most cases, MAAP results showed that the failure pressure of 121 psig at 3407 would be reached first. Therefore, in the IPE study the containment l failure presure was assumed to be 121 psig (page 4.3-6, reference 1). ,

The WNP 2 electrical penetrations were supplied by Westinghouse Electric Corporation and Conax Buffalo Corporation. The Westinghouse penetration assemblies were of two types. The first, a modular type, used potting compound to seal the cables in the modules. l The modules were clamped to the header plate and used ethylene propylene and silicone O ring for sealing. For these modules, Westinghouse performed thermal endurance tests i over the temperature range of 707 to 3927. The tests showed that no leakage would '

occur at the highest test temperature and a mean life of 129 hours0.00149 days <br />0.0358 hours <br />2.132936e-4 weeks <br />4.90845e-5 months <br /> could be. expected. [3]

The second, a canister type, did not contain any exposed organic material. The seal was

provided by ceramic bushings which could withstand temperatures up to 1100 7.[4] i j Therefore, the canister-type penetrations were not considered to have significant leakage potential.

j -

Tests were performed elsewhere on the Conax penetrations similar to those used at l

WNP-2. [5] The results showed that the leak integrity would be maintained during a I severe accident condition. i It appears that the IPE team has considered all the containment failure modes applicable to WNP-2.

214. Containment Isolation Failure

, 'Ihe IPE team analyzed the possibility of containment isolation failure by using it as a top -

event in the CETs which was' quantified by using a fault tree. They subdivided faults I

! WPPSS Nuclear Project Unit 2 9 September 1996  !

T

y .o o

l which contribute to failure for containment isolation by initiator so that if there were a j

] difference in conditional failure probabilities, they could be correctly accommodated.

j They considered general issues: 1) the failure of a penetration to isolate on demand,  !

j - either because the isolation system failed 'to actuate or one of the active penetrations failed 1

because of local valve faults and 2) failure of the wetwell vacuum breakers to remain

. closed during the sequence (page 4.5 29 of the submittel). The failure data was taken  !

from results of the Level I systems reliability analysis:

Group 1 - 9.3E-4 (Main Steam Isolation Valves)

Group 2 - Valves less than 0.75 inches so will have no effect

< Group 3 - Normally closed valves, > 2 inches, will have no effect Group 4 '- 3.5E-6 (Reactor Building Closed Cooling and Fuel Pool Cooling)

Group 5 - 1.6E-4 (Sample lines, of no concern)

Group 6 - 1.0E-3 (Residual Heat Removal ststem)

! Group 7 - 3.9E 3 (Reactor Water Cleanup system)

The IPE team considered that there wts some probability that the operating staff could

, manually isolate a failed penetration, either by closing the failed valve or by closing the

! non-isolation valves downstream. The likelihood of successful isolation given failure of *

the automated system was assumed to be 0.5 except in SBO sequences.

! The flow paths from the suppression chamber to the reactor building comprised two normally closed check valves in series. The check valves had magnets in the disk to hold j them in the closed position. If they were open during normal operation, the leakage i should be detectable with the radiation monitoring system; therefore, they were assumed to be closed at the onset of the event. Because a pressure increase increased the seating force, the failure mode of concem was rupture for which the probability of occurrence was L found to be negligible.

i

! The frequency of containment isolation failure for WNP-2 was calculated as 2.2E-7 per i year (page 4.8-7, reference 1).

211 Svstem/ Human Resoonse.

. As noted in section 2.2.2 of this report, the IPE team developed simple fault trees to resolve each node of the CETs for each accident sequence. In the fault trees, they used a general value of 0.1 for the probability of operator failure to initiate operation of a system i following recovery of required support systems. He following are the operator actions modeled in the IPE. Exceptions to the above assumption are noted in the discussion as identified in the submittal.

. De IPE team considered that there was some probability that the operating staff i could manually isolate a faile'd penetration, either by closing the failed valve or by j

closing the non-isolation valves downstream. De likelihood of successful isolation

]

. WPPSS Nuclear Project Unit 2 10 September 1996 l

l .

L

a -- m. e, t .* , ,

I i

l given failure of the automated system was assumed to be 0.5 except in SBO ,

sequences,  !

I g

Operator failure to align and initiate high pressure injection following recovery of l

'any needed support systems. l

1 i

• Operator' failure to align and initiate low pressure injection following recovery of l any needed support systems.

I'

• Operator failure to restore RHR pumps to operability following the failure of a support system. ,

I

• Operator failure to vent the containment before its failure, because though directed to open the vent at a containment pressure of 39 psig, the operator had a limited

window of opportunity for completing the task. This was because if the 4

. containment pressure reached 49 psig before the valves were opened, the forces .

exerted by the differential pressure across the disk would exceed the capabilities of the valve operator, i.e., it would not open. -

l i <

-* Operator failure to recover the needed support systems to vent the containment l ,

before its failure. De vent required containment air, which in tum required  !

restoration of offsite power. Here would be a need for manual alignment and  !

starting of the compressors and cooling systems.

• Operator failure to recover PCS before containment failure once offsite power is recovered.

2,2,s Radionuclide Rela == Cateoories and Characterivation. 'l l

The IPE team analyzed the PDSs using the CETs. De endpoints from the CETs I

, represented the outcomes which resulted from complete severe accident sequences from  !

the initiating event to the release of radionuclides to the environment. These endpoints 1 were binned into radionuclide release categories using the following fission product

- release characteristics:

1 Contdnment Bypars. The Level 1 analysis contained no containment bypass sequences above the cutoff frequency,' and this characteristic was used for the sake of completeness.

Consdament /solmion. During the WNP-2 analysis, failure of a penetration was assumed to be incapable of removing enough energy to prevent containment failure so all sequences involving failure to isolate were described with a combined source term, that

. from the penetration and that from the associated containment failure.

4 Sequence Arrested 1> Vessel. His characteristic was important for non-bypass sequences because if injection could be restored before the core support plate failed, the core would WPPss Nuclear Project Unit 2 - 11 September 1996

,,- i

( , i i

1

be in a coolable geometry and the vessel would be saved. . As a result, the fission products would be largely contained within the primary system pressure boundary, and those which ]

4 would be released to the containment would be scrubbed by the water in the primary  !

system.

j- Drywel/ Fdlure. This attribute was imponant to the source term for sequences in which >

! the vessel failed because failure of the'drywell might result in a direct path to the reactor ,

i building or to the suppression chamber airspace which would bypass the scrubbing action normalliy provided by the suppression pool. De only possibility for fission product scrubbing was if the debris in the drywell was flooded.

Drywell failure'was possible if: ,

• - the drywell/re. actor building pressure boundary failed from overpressure.

i

= drywell floor failed as a result of molten core-concrete interaction (MCCI).

I e drywell integrity was lost as a result of failure of the penetrations or seals which serve as part of the pressure boundary between drywell and suppression chamber.  !

l l Pool Bypars. If the drywell was intact (floor, shell, and penetrations), the remaining issue l of concem was whether all of the fission products released to the environment actually passed through the suppression pool so that scrubbing could take place. The following were the ways in which suppression pool bypass could occur:

- Use of the PCS to remove containment energy - The main steam system bypassed the pool by providing an open path directly from the drywell, through the vessel, to the main condenser. The IPE team considered a release to the main condenser .

j. was equivalent to a release to the environment.

. Loss of water in the suppression pool - The IPE team considered this as a

- . possibility during gross containment failures. However, because these large

' failures were expected to also cause drywell failure, the information became i redundant.

. Time of Consdnment Fdlure (CF). This descriptive source term attribute was imponant

! because it affected the time available for fission product release mitigation by natural j removal processes, scrubbing, and retention of fission products. It was an issue for all scenarios which did not involve containment bypass or core melt arrested in vessel. The

- times defined were as follows:

! = Very early - before core damage

. Early - at or near the time of vessel failure '

. Late - significantly after vessel failure

i WPPSS Nuclear Project Unit 2 September 1996 12

,4

I

i. ,

l l

• Immediate - specifically for large LOCA sequence (A.S08) in which failure l of the omega seal led to los of vapor suppression function and J

I' subsequent containment failure.

Mode of Consdament FWim (CIM). This attribute was important because it govemed the rate of fission product release to the atmosphere and it affected the release by p

controlling the time available for fission product attenuation in the containment.

Containment failure mode is large represented a catastrophic rupture, which was defined as the loss of a substantial portion of the containment boundasy with possible disruption of the piping systems that penetrate or were attached to the containment wall. Containment failure mode is small represented a leak, which was defined as a containment breach that '

' would' arrest a gradual pressure buildup and would depressurize the containment within .

, two hours.

Containment failure mode was only considered to be an important discrimination for those I sequences which had very early or early containment failure, because late failure of

ontainment would have allowed time for effective fission product attenuation. l

}

Based on the above characteristics, the IPE team identified 17 basic source term groups l (STGs) into which CET damage states will fall, with a limited number of additional l modified states, in which the expected release was changed with a release from a failed

, hich

- penetration. These modified states were only explicitly developed for sequences w  ;

l survived the truncation value of IE-10 per year. Two cases were identified with containment failure before core damage and high pressure melt ejection. Seven cases l were identified for unisolated containment.

The IPE team did not use a screening criterion that compared the fission product release i

categories of WNP-2 to BWR-3 or PWR-4 release categories of WASH-1400. However, ,

' they identified a release category with large releases (STG-10) which had a frequency of

1.05E-6 per year.

2.3 Quantitative Assessment of Accident Progression and Containment

! Behavier I

l 11), Severe Accident Proeression. <

7 The IPE team used the Modular Accident Analysis Program (MAAP) 3.0B, Revision 7.02 l'

and 7.03, to provide an integral approach for modeling of plant thermal hydraulic response ,

and fission product transport during severe accidents. They compiled plant specific data derived from plant design documents and drawings into the MAAP parameter file. They analyzed different severe accident events (SBO, transients, ATWS, and LOCA) using the .

MAAP code. In addition to MAAP results, they used research results in the open

- literature,IDCOR task reports, Shoreham and Limerick PRAs, NUREG reports, and engineering judgment in understanding physical processes and developing event trees.

13 September 1996 WPPSS Nuclear Project Unit 2 -

i

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

. .w The IPE team used the MAAP code to provide integral analyses of accident sequences from their inception until several hours after containment failure. Because MAAP 3.0B did not include models for DCH, hydrogen detonation, and steam explosion, they considered these as separate effects.

Steam Explosions -In-Vessel. For WNP-2, the IPE team assigned a probability value of l IE-4 for the likelihood of containment failure from in-vessel steam explosions for the r sequences where the vessel stayed at low pressure. His value is consistent with that i calculated by neofanous, et. al., for pressurized water reactors. [6] For high vessel

pressure sequences, they assigned a probability vake of IE-5 which was an order of magnitude below that for the low vessel pressure sequences. Lowering the value by an order of magnitude for high pressure sequences is consistent with the results of NUREG-i 1150, E

Steen Explosions - Ex-Vessel. For vessel failures at low vessel pressure, the IPE team assumed that steam explosions in the pedestal region were likely to damage the cavity

~ drain line which would cause wetwell bypass - and potentially large radiological releases.

In addition they assumed that the high temperature melt could also fail the drain line even if no steam explosion occurred. The probability of drain line failure leading to wetwell i bypass was assumed to be 0.9. Althoesh probably conservative, this is an appropriate

assumption, based on the uncertainties involved.

1 j For vessel failures at high vessel pressure, the IPE team assumed a probability of 0.9 for structural failure of the containment due to ex-vessel steam explosions, again an

appropriately conservative value.

i

> . Direct Conts'nment Hearing. The IPE team considered DCH as an unlikely phenomenon

! for BWRs like WNP-2 based on the following: 1) the inherent inability to entrain and

' finely particulate debris due to the~ pedestal configuration that did not promote debris entrainment and 2) the presence of numerous structures in the drywell which would highly promote debris deentrainment. A value of IE-3 was assigned to the conditional probability of DCH causing early contaisiment failure. This value is relatively low considering the uncertainties involved. However, the impact ofincreasing this value on

! the radiological releases will be small since the DCH phenomena only occur when the i vessel fails at high vessel pressure -- a relatively low probability event.

! Hydrogen Combustion. Combustion was assumed to occur 100 percent of the time that 4

hydrogen reached combustible range (concentrations of 5 percent or more of oxygen and 6

percent or more of hydrogen). He IPE team considered that this assumption was

' conservative because most scenarios which yielded large quantities of hydrogen also produced an atmosphere with more than 55 percent of steam. Once a combustion event

started, it was conservatively assumed that the likelihood of detonation was the same as

( that of deflagration. Also, it was assumed that detonations would fail the containment.

i ,

WPPSS Nuclear Project Unit 2 14 September 1996

4-In summary, there was a complete discussion of the IPE team's analysis of severe accident phenomenology.

i 212 Dominant Contpbutors Consistency with IPE faninhts.

Table 1 below compares the IPE results of the WNP-2 plant and the Nine Mile Point Unit 2 plant (with Mark II containment). Compared to the Nine Mile Point 2 plant WNP-2 l was postulated to have higher conditional probability for early releases. The WNP-2 source term groups (STGs) contributing to early release were as follows: ,

J l l t Table 1. Conditional Containment Failure Probability (Percent) 1-l CDF Early Late  ;

Study per yr Failure Failure Intact l l Nine Mile Point 2 Irt 3.lE-05 7 67 25 )

WNP-2 IPE 1.75E-05 31 30 39  !

i

• STG-13P (13.5 percent of CDF): This STG was characterized by an unscrubbed

! release of fission products through a small breach in containment, beginning at the time of core damage. Initially the release would be scrubbed, but after vessel failure release would bypass the suppression pool. Typical accident sequences of this STG would involve successful injection but failure of all viable means of containment heat removal.- Containment pressure would continue to increase until a membrane tear at the vicinity of the wetwell horizontal stiffeners occurs. The energetic release of steam from this location into the reactor building basemat would lead to consequential failure

, - of all injection systems and core melt.

• STG 9 (8.7 percent of CDF): This STG was characterized by a release through a

small breach in containment beginnirig at or near the time of vessel failure. The

!' accident sequences of this STG would initiate by a loss of offsite power and progress L to long-term SBO or look-alike sequences in which a source of high pressure injection substitute for a third failure diesel generator.

l l The dominant containment failure mode for STG-9 resulted from a situation in which a rapidly flowing jet of hot debris impacted the shell and caused early containment 4

failure. ,

i

~

• STG-10 (6.0 percent of CDF): This STG was characterized by a catastrophic failure L of containment and a release of unscrubbed fission products around the time of vessel

! ' failure. The dominant accident sequences would involve high pressure transient sequences in which a failure of injection would occur before vessel failure. There would be an accumulation of water in the pedestal cavity as a result of operation of

- WPPSS Nuclear Project Unit 2 15 September 19%

i

drywell sprays so that when the melt would forcibly be ejected at high pressure, there -

would be an ensuing steam explosion. This explosion was postulated to result in

! severe overpressure in the cavity, collapse of the pedestal, and gross structural failure of containment.

I J

211. Characterieon of cane inment Performance. l As noted in section 2'.2.2 of this report, the IPE team developed initiator-specific CETs to characterize containment performance. De probabilities for the nodes were obtained form

either node specific fault trees developed to identify each of the functional failures, or split fractions (1 or 0) assigned to each branch node based on Level 1 system status The

- CET and point frequencies were computed using the NUPRA code.

4 i

Because of the severe environmental conditions to which the hardware would be exposed during a core melt, the following values were used in the Level 2 analysis for the

! - probabilities of failure for the systems:

= single train system - probability of failure to start,0.1 l

= two train system - probability of failure to start,0.05; probability of common cause failure to run is 100 times greater than common cause failure to run during normal operation.

De above values were intended to accommodate the operation of equipment at the following limits ofits design envelope:

. minimum NPSH pumps ,

- higher than normal temperatures in the operating fluids

. higher than normal ambient operating temperatures

- the possibility of debris being ingested by running equipment i

i ne IPE team considered that the stress and confusion levels of the operating staff would

be extremely high during and after a core melt event. So even if there appeared to be ample time to initiate a particular action, the team assumed higher non response

probabilities than if the actions were performed under normal operating conditioris. A i non response probability for recovery of needed systems was used as 0.1.

gli Imnact on EauiumW Bahavior, ne IPE team recognized that uncertainty existed in the performance of critical hardware following a core melt accident (section 4.0.3, page 4.0-4). In such an accident, hardware

' reliability was expected to be adversely affected by severe environmental stress, operation  :

WPPSS Nuclear Project Unit 2 16 September 1996

't

)

.e. ,  ;

1

- at the limit of its design envelope, and difficulty in reestablishing normal hardware alignment and operation following a complete loss of all AC, DC, or critical cooling systems and degraded operation of control, monitoring, and actuation systems. As described in section 2.3.3 of this report, the IPE team considered the above facts in using values for the probabilities of failure for the systems.

l ,

4 211 Uncertaintv and Sannitivity Analvsis.

De IPE team performed sensitivity studies by requantifying the CETs which involved ]

varying of important parameters that were likely to have the largest emet on the

' likelihood or time of containment failure and the magnitude of the source term. These studies were as follows:

Depressurization During a Shon term SBO. During a short term SBO, the primary system i would remain at high pressure and at the time of vessel failure an HPME would occur.

Because the batteries would be available, the operators would have the opportunity to ,

i open SRVs and depressurin the vessel resulting in an LPME at the time of vessel failure.

The conclusions were that depressuriution of the primtry system should be delayed as  ;

i long as possible; but once core melting had been irreversibly initiated, it would be l advisable to depressurin as quickly as possible. This had two benefits: 1) if power was restored and all injection systems became available as potential sources of core cooling and 2) if depressuriution was at least partially successful, the chances for delaying containment failure or maintaining the integrity of containment were enhanced.

]

Decay Hear Removd Systems - RHR. The IPE team performed an analysis of the

! sensitivity of source term grouping results to the availability of DHR systems. The results showed that if the RHR systems were perfectly available from the Level 2 perspective, the fraction of no containment failures (STG-1 and .2) increased from 38.9 percent to ,

46.9 percent. Correspondingly, the frequency of late failure in STG-3 decreased from l.07E-6 per year to 9.98E-9 per year. l l Decay Heat Removal Systems - Hedened Vent. The IPE team performed this sensitivity

to investigate the benefits of installing a hardened containment vent. In Level 2, it was assumed that venting was always successful except in TW sequences and the effects of the

. source term groups were investigated. The IPE team concluded that once core damage occurred, hardened venting had very limited benefits in reducing offsite consequences.

i Engrility of she Containment Shell. This sensitivity study was performed to fmd the effect ,

of the following on an HPME:

4 . the amount and velocity of corium ejected from the deep pedestal cavity which could be available to come in contact with the containment shell  ;

• . the intrinsic fragility of the shell when exposed to corium WPPSS Nuclear Project Unit 2 17 September 1996 l

1 l a

l

• u  :

5

' He results showed that as the assumptions were changed to reflect conditions where the fragility of the containment shell decreased, or less corium was assumed to be ejected from the cavity, the fraction of early containment failures did decrease (page 37-1, reference 2).

j Drywell Spmys. He FE team performed this sensitivity study to understand the value of

initiating drywell sprays before vessel failure because there were competing risks. He negative contribution was that because the drywell drained to the pedestal cavity,if the sprays were actuated before vessel failure, water would accumulate in the cavity to a sufficient depth which would enhance the occurrences of ex-vessel steam explosions. De ,

positive contribution was that if the sprays were operating at the time of an HPME, the cooling of the debris ex-vessel would be enhanced and the threat of corium ejection to the shell or corium attack on the downcomers might me reduced. De FE team found that

, more benefit could be gained if the sprays were not actuated before vessel breach.

Debn's Coola6flity. He FE team performed this analysis to investigate the effect of the ,

l debris being coolable or not coolable since there was uncertainty regarding the geometry j of the debris under HPME conditions. When the split fraction for cooled /not cooled was

changed from 43/57 to 8.5/91.5, the conditional probability of an intact containment

{ remained unchanged (page 38-1, reference 2). Here was a small increase in unscrubbed

fission products. The predicted release categories were relatively insensitive to the assumed fractions for oebric cooled /not cooled.

! 2.4 Reducing Probability of Core Damage or Fission Product Release W Definition of Vulnerability, i

As noted in section 3.4.2 of the submittal, the WNP-2 FE team identiSed no

, vulnerabilities in the WNP-2 desitn or operation (page 3.4~i t, reference 1). For WNP-2, 1

vulnerability screening was based on:

a sequence groups.with frequencies greater than IE-6 per year that would require

modifications per NUMARC 91-04 guidelines

= total CDF must be within the NRC s safety god of lE-4 per year 4

j

= sequences that indicate a plant-specific feature that is an outlier to comparable BWR PRAs.

s W Plant Imorovements.

1 During the course of the WNP-2 FE, the FE team identified one back-end plant improvement and two back-end procedure improvements (page 6.0-4, reference 1). A hardware modification of the air supply to the inboard MSIVs and the contain=mt vent valves for backup from the containment nitrogen system was investigated. The WPPSS Nuclear Project Unit 2 18 September 1996 i ,

- , . - , ~ - , .

• a 4

I

~

modification would improve long-term' decay heat removal by providing redundant air - ,

, supply to the valve's solenoids. This modification was found to be marginally cost effective and, therefore was not recommended for further plant evaluation. .The following

_ were the back-end procedural improvements:

1 - SBO Depressurisainn. The IPE team recommended that the longer coping time .

L be evaluated for incorporation in the SBO emergency procedures, if the vessel was j maintained at pressure with the benefit of depressurizing once fuel melt started, but before vessel breach.

• DrywelFFetwellEypars. The IPE team found that the Omega seal design I

separating the drywell and wetwell air spaces was a passive design with very low o failure rate. However, its failure resulted in very large consequences. The'refore, l the IPE team recommended an evaluation of the costs and benefits of periodic

inspection and maintenance of the Omega seal be made.

j 2.5 Responses to CPI Pmgrun Recomunendations j In Generic Letter 88-20, Supplement No. 3, NRC recommends that licensees consider j insights from the CPI Program, specifically for Mark II containments: additional

! containment heat removal capability (e.g., hardened vent or improved reliability of  ;

4 suppression pool cooling). Also, Supplement No. 3 requests Mark II licensee  !

consideration ofinsights directed at the Mark I BWRs and delineated in Generic Letter l

!' 88 20, Supplement No.1: (a) alternate water supply for drywell spray / vessel injection; (b) erahanced RW depressurization reliability; and (c) improved EOP and training.

After reviewmg the contr,inment heat removal systems, the IPE team concluded that the i reliability of those systems were acceptable and no further changes were warranted (page l 32-2, reference 2).

De IPE team performed a sensitivity study to investigate the benefits ofinstalling a hardened containment vent. In Level 2,it was assumed that venting was always

! successful, except in TW sequences in which loss of DHR initiate containment failure, and

, the effects of the source term groups were investigated. He IPE team concluded that i l cace core damage occurred, hardened venting had very limited benefits in reducing offsite

consequences.

i The licensee response to the insights directed at the Mark I BWRs were as follows (pages

32-3 through 32-5, reference 2)

(4 Alternme weer supplyfor drywell spmy/ vessel injection. he IPE team noted that the

, IPE assumptions for the availability of HP injection water inventory were limited by the I assumed capable capacity of the CST, rather than by hardware (tank) availability. The available inventory was found to be twice the value assumed for the IPE, thus no

- additional benefits were expected from an alternate water supply, WPPSS Nuclear Project Unit 2 19 September 1996 1

l.

l

.* a (b) Enheced RPV depressurization reliability. The IPE team found that during most WNP-2 accident sequences, the reliability of RPV depressurization was dominated by human error. His included the failure to initiate manual depressurization to delay core melt during a short term SBO, or following failure of all HP injection systems in non- l SBO sequences so that LP injection could be initiated. Significant improvement in the I reliability of the operator performing this action was not expected to be gained. He only fossible actions for enhancing the probability of successful depressurization during most accident sequences were limited to confirmation that they were addressed and practiced in i the operator training program. l

\

(c) 1mpmved EOP and tmining. He EOPs implemented at WNP-2 had been based on l Revision 4 of the BWROG EPG since 1990. He conosponding operator training was 1 also conducted according to the EOPs as basad on Revision 4 of the EPGs.  !

l De licensee appears to have responded to all the relevant recommendations stemming  :

I 1 from the CPI Program.

1 l

3.

2.6 IPE Insights, Impmvements, and Commitments j 1

Section 63, page 6.0-5, of the submittal notes the following [1]

l A major benefit in performing the IPE is a greater awareness of the plant response to severe accidents, the systems and components most important to prevention and mitigation of the accident, and the importance of operator actions in the prevention

and mitigation. Most of these insights have been recognized and are currently in i the process of incorporation into a revision of the BWROG Emergency Procedures Guidelines. De Supply System is actively working with the appropriate BWROG l committees in formulating the accident management procedures. Therefore, these recognized insights are not included in this report.

However, the submittal noted other back-end insights observed during sensitivity analysis.

These insights are described in section 23.5 of this report.

As described in section 2.4.2 of this report, during the course of the WNP-2 IPE, the IPE team identified one back-end plant improvement and two back-end procedure improvements.

l De licensee made no commitments.

l WPPSS Nuclear Project Unit 2 20 September 1996 i

l'  !

i . ~

l o' -o I

3. CONTRACIOR OBSERVA110NS AND CONCLUSIONS t ,

i SCIENTECH observed the following on the WNP-2 IPE:

i

• The IPE team responded to all the recommendations of the CPI Program which are

' relevant to WNP-2.

• The WNP-2 IPE team identified no fulnerabilities in the WNP-2 design or operation. ,

1 t'

• ne IPE team identified one back-end plant improvement and two back-end procedure improvements.

• During the WNP-2 analysis, failure of a penetration was assumed to be incapable of removing enough energy to prevent containment failure so all sequences involving failure to isolate were described with a combined source ter:n, that from the penetration and that from the associated containment failure.

• The licensee participated in performing and reviewing the IPE.

4 4

• He submittal demonstrated that the IPE team well understood the severe accident i

I phenomenology important for WNP-2.

. The IPE team adequately considered the role of operator actions on severe accident progression at WNP-2.

l l

• The IPE team was considering several IPE insights but they were not ircluded in the i

- submittal. The licensee was working with the appropriate BWROG committees in )

formulating the accident management procedures based on the recognized insights. {

j

• The IPE team did not u'se a screening criterion that compared the fission product I l

j release categories of WNP-2 to BWR-3 or PWR-4 release categories of WASH-1400.

1 However, they identified a release category with large releases (STG-10) which had a frequency of 1.05E-6 per year.

Overall, 'the WNP-2 IPE appears to have achieved the objectives of GL 88-20.

3

)

i l l

1 1

i l

. l l

l l

WPPSS Nuclear Project Unit 2 21 September 1996 j

. l

• 6

4. REFERENCES ,

1. Washington Public Power Supply System, WPPSS Nuclear Project Unit 2:

' Individual Plant Examination Report in Response to Generic Letter 88-20, Rev.1, July 1994.

1 2. Washington Public Power Supply System, WPPSS Nuclear Project Unit 2 -

Response to Request for Additional Information on Individual Plant Examination, October 1995.

3. Westinghouse Document No. 75-7B5-BIGAI R2, Predicting the Thermal Life of Modular Penetrations, May 1975.

1

4. Westinghouse Document No. PEN-TR-80-100, Qualification Tests for 5 kv ,

Medium Voltage Penetrations for Susquehanna #1 and #2, September 1980.

I 5. D. Gauss, Severe Accident Testing of Electrical Penetration Assemblies, NUREG/CR-5334, November 1989.

6. T. G. Theofanous, B. Najafi, and E. Rumble, An Assessment of Steam Explosion-Induced Containment Failure. Part I: Probabilistic Aspects, Nuclear Science and

]

Engineering, 97, pp. 259-281,1987.

4 P

4 WPPSS Nuclear Project Unit 2 22 September 1996

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

o .s l l

A PPENDIX ,

IPE EVALUA110N AND DATA

SUMMARY

SHEET 1 BWR Back-End Facts

]

i .. .  !

l Plant Name .l 4

J WPPSS Nuclear Project Unit 2 Containment Type 4

MarkII.

Unique Containment Features The pedestal area under the reactor vessel is recessed below the drywell floor area.

. This allows most of the molten corium to accumulate in the pedestal area during 4

accident scenarios where vessel failure would occur at low RCS pressure. Therefore, '

the primary containment shell is protected from direct corium attack following vessel a breach.

Unique Vessel Features i

None Number of Key Plant Damage States 6

Ultimate Containment Failure Pressure 121 psig (median value)

Additional Radionuclide Transport and Retention Structures None

. Conditional Probability that the Containment Is Not Isolated a'

0.013

. Important Insights, including Unique' Safety Features ,

See section 2.6 of this report.

WPPSS Nuclear Project Unit 2 A1 September 1996 4

- = . . - . _. . . . . - . - - .

.o - g I

l Implemented Plant Improvements .

The IPE team identified the following plant and procedure improvements related to the l

back-end analysis (see section 2.4.2 of this report for details): hardware change to supply containment nitrogen system nitrogen to MSIV and vent valves; procedure changes on 1) vessel depressurization during an SBO and 2) periodic inspection and maintenance of the Omega seal in order a lower the possibility ofits failure causing drywell/wetwell bypass.

C-Matrix Information provided in the submittal is not sufficient to generate a e-matrix.

4 9

4 I

I l

1 t

J WPPSS Nuclear Project Unit 2 A2 September 1996 4

-- e , -