ML20246G925

Provides Addl NRC Audit Insights Re Plant Risk Review for Info Helpful in Refinement of Response to Generic Ltr 88-19. NRC Recommends That Util re-examine Plant Emergency Operating Procedures Used in Recovery from Transient

ML20246G925
Person / Time
Site:	Browns Ferry
Issue date:	07/10/1989
From:	Black S Office of Nuclear Reactor Regulation
To:	Kingsley O TENNESSEE VALLEY AUTHORITY
References
GL-88-19, NUDOCS 8907140281
Download: ML20246G925 (10)

Text

a

,', ,- July 10, 1989 D$ck;t bos. 50-259', 50-266' Distribution and 50 296 Docket f ues MSimms BSheron NRC PDR GGears EGoodwin Local PDR DMoran SWeiss e DCrutchfield TDaniels RBarrett l

Mr. Oliver D. Kingsley, Jr. ADSP Reading NMarkisohn EChelliah Senior Vice President, Nuclear Power BDLicw OGC BFN Rdg. File Tennessee Valley Authority SBlack BGrimes 6N 38A Lookout Place RPierson EJordan

, 1101 Market Street BWilson- ACRS(10)

Chattanooga, Tennessee 37402-2801 WSLittle GPA/CA

SUBJECT:

ADDITIONAL NRC STAFF AUDIT INSIGHTS PERTAINING TO THE BROWNS FERRY RISK REVIEW - GENERIC LETTER 88-19 An audit of the draft Probabilistic Risk Assessment (PRA) for Browns Ferry was conducted by the NRC staff in Knoxville, Tennessee on November 1 and 2, 1988.

By letter dated March 29, 1989, the staff forwarded one audit finding involving a potential single failure vulnerability to the Tennessee Valley Authority (TVA I or the licensee) for review and evaluation. By letter dated June 15, 1989, TVA provided its response to this issue. The staff is currently reviewing your response.

Additional insights were developed by the staff as the result of its audit.

We are forwarding these insights since they should be helpful in your

. continued refinement of the Browns Ferry PRA and your efforts to respond to Generic Letter 8819 (Individual Plant Examination for Severe Accident Vulnerabilities - November 23,1988). No specific response to this letter is required. However, the staff notes that the remarks for PRA sequence 4 of the enclosure recommends that TVA re-examine the plant Emergency Operating Procedures (E01s) used in recovery from a transient with 1-3 stuck open relief valves in which the operators must manually start the RHR pumps and align the RHR system valves to establish pool cooling. These procedures may be reviewed ,

by the staff for corrective action and clarification changes during the  !

upcoming NRC staff EDI followup inspection at Browns Ferry.

Any questions should be directed to Gerald E. Gears, Browns Ferry Project Manager at 301-492 0767. Original signed by Suzanne Black, Assistant Director for Projects TVA Projects Division 8907140281 890710 Office of Nuclear Reactor Regulation PDR ADOCK0500g9 P

Enclosure:

As stated g cc w/ enclosure: 8{

See next page 4

• SEE PREVIOUS CONCURRENCE II 0FC :NRR:TVA/LA* :hRR:TVA/PM* :NRR:TVA :hRR:ADSP* :TVA:/,D

..___,......______:________.__.:......__{h__:.._____.....:...../TP*______:

g. g :

NAME :MSimms :GGears:as :EGoodwin > :SWeiss :RPierson  : 'k  :

6/30/89 / / i >89

~  :

DATE : 6/29/89  : 6/29/89 :7/f/89  : f>/29/89 0FFICIAL RECORD COPY

1 Docket Nos.'50-259, 50-260 Distribution

-and 50-296 Docket File MSimms BSheron NRC PDR GGears EGoodwin Local PDR DMoran SWeiss DCrutchfield TDaniels RBarrett Mr. Oliver D. Kingsley, Jr. ADSP Reading NMarkisohn EChelliah Senior Vice President, Nuclear Power BDLiaw OGC BFN Rdg. File Tennessee Valley Authority ' SBlack BGrimes 6N 38A Lookout Place RPierson EJordan 1101 Market Street BWilson ACRS(10)

Chattanooga, Tennessee 37402-2801 WSlittle GPA/CA

SUBJECT:

ADDITIONAL NRC STAFF AUDIT INSIGHTS PERTAINING TO THE BROWNS FERRY RISK REVIEW - GENERIC LETTER 88-19 An audit of the draft Probabilistic Risk Assessment (PRA) for Browns Ferry was conducted in Knoxville, Tennessee on November 1 and 2, 1989.

By letter dated March 29, 1989, the staff forwarded one audit finding involving a potential single failure vulnerability to the Tennessee Valley Authority (TVA or the license) for review and evaluation. By letter dated June 15, 1989, TVA provided its response to this issue. The staff is currently reviewing your response.

Additional insights were developed by the staff as the result of its audit.

. We are forwarding: these insights'since they should be helpful in your continued refinement of the Browns Ferry PRA and your efforts to respond to Generic Letter 88-19 (Individual Plant Examination for Severe Accident Vulnerabilities - November 23,1988). No specific response to this letter is required. However, the staff notes that the remarks for PRA sequence 4 of the enclosure recommends that TVA re-examine the plant Emergency Operating Procedures (E01s) used in recovery from a transient with 1-3 stuck open relief valves in which the operators must manually start the RHR pumps and align the RHR system valves to establish pool cooling. These procedures may be reviewed by the staff for corrective action and clarification changes during the upcoming NRC staff E01 followup inspection at Browns ferry.

.Any questions should be directed to Gerald E. Gears, Browns Ferry Project Manager at 301-492-0767.

Suzanne Black, Assistant Director for Projects TVA Projects Division Office of Nuclear Reactor Regulation

Enclosure:

j As stated cc w/ enclosure: }

See next page l 1

0FC :NRR:TVA/LA :NRR:TVA/PM :NRR:TVA :NRR: ADSP :TVA: D/IP :TVA:AD/P : i

~

M'......:--.....__..:...--....._ l NAME :MSimms  : f :EGoodwin :SWeis :RP14fs n :SBlack  :

lDATE[/1Md/89  : 11/ 39  : / /89 :4 M /89  : / /89, :

(c/Q89 0FFICIAL RECORD COPY

p. , s x .v

., . V-

0 .Q 4 4 > "Mr.bliverD,; Kings' ley,Jr. Any questions should be directed to Gerald E. Gears, Browns Ferry Project- _

Manager at-301-492-0767.

Sincerely.

Suzanne Black, Assistant Director for' Projects .

TVA Projects Division. .

Office of Nuclear Reactor. Regulation

Enclosure:

As stated

,cc w/ enclosure:

See next page:

Distribution ~

Docket File MSimms BSheron NRC PDR GGears EGoodwin Local PDR DMoran SWeiss DCrutchfield TDaniels RBarrett~

ADSP Reading NMarkisohn_ EChelliah-

~ BDLiaw.- OGC '.

BFN Rdg. File SBlack BGrines .

RPierson EJordan BWilson ACRS(10)'

. , ~

WSlittle GPA/CA'

• SEE PREVIOUS CONCURRENCE OFC :NRR:TVA/LA* :NRR:TVA/PM* :NRR:TVA :NRR:ADSP* :TVA:AD/lP* :TVA:AD/P  :

NAME :MSimms :GGears:as :EGoodwin :SWeiss :RPierson :SBlack  :

_DATE : 6/29/89  : 6/29/89  : 7/09/89  : 6/29/89  : 6/30/89  : 7/10/89  : i

)

i 0FFICIAL' RECORD. COPY' i

o Mr. Oliver D. Kingsley, Jr. CC:

General Counsel Chairman, Limestone County Commission Tennessee Valley Authority P. O. Box 188 400 West Summit Hill Drive Athens, Alabama 35611 ET 118 33H Knoxville, Tennessee 37902 Claude Earl Fox, M.D.

State Health Officer Mr. F. L. Moreadith State Department of.Public Health Vice President, Nuclear Engineering State Office Building Tennessee Valley Authority Montgomery, Alabama 36130 400 West Summit Hill Drive WT 12A 12A Regional Administrator, Region II Knoxville, Tennessee 37902 U.S. Nuclear Regulatory Commission

, 101 Marietta Street, N.W. .

Dr. Mark 0. Medford Atlanta, Georgia 30323 Vice President and Nuclear Technical Director Mr. Danny Carpenter Tennessee Valley Authority Senior Resident Inspector 6N 38A Lookout Place Browns Ferry Nuclear Plant Chattanooga, Tennessee 37402-2801 U.S. Nuclear Regulatory Commission Route 12, Box 637 Manager, Nuclear Licensing Athens, Alabama 35611 and Regulatory Affairs Tennessee Valley Authority Dr. Henry Myers, Science Advisor SN 157B lookout Place Committee on Interior Chattanooga, Tennessee 37402-2801 and Insular Affairs U.S. House of Representatives Mr. O. J. Zeringue h'ashington, D.C. 20515 Site Director Browns Ferry Nuclear Plant Tennessee Valley Authority Tennessee Valley Authority Rockville Office P. O. Box 2000 E1921 Rockville Pike Decatur, Alabama 35602 Suite 402 <

Rockville, Maryland 20852 Mr. P. Carier Site Licensing Manager Browns Ferry Nuclear Plant Tennessee Valley Authority P. D. Box 2000 Decatur, Alabama 35602 Mr. G. Campbell Plant Manager Browns Ferry Nuclear Plant Tennessee Valley Authority P. O. Box 2000 Decatur, Alabama 35602

L ,

. ENCLOSURE l

Browns Ferry Nuclear Plant (Docket Nos: 50-259, 50-260 and 50-296) l The following is an itemization of audit findings of dominant sequences of the revised Browns Ferry PRA (September 1987 Version).

Sequence 1:

A loss of feedwater transient followed by failure of the high pressure coolant makeup systems, and failure of the manual ADS blowdown to allow the low pressure reacter.. coolant

.

• makeup systems to provide coolant inventory makeup to the '

Finding: the frequency estimate seems reasonable.

Remarks:

HPCI/RCIC vacuum breakers were not modeled in the PRA because manuaT start failures of the HPCI/RCIC pumps. licensee determine Also, at this time the improved version of the rupture disks have been installed in the discharge lines of the HPCI pump and will trip at a higher pressure (about 175 psig). Credit has been given for this design feature. Also, the control rod drive has been makeup found to be a viable backup system to other high pressure (coolantCRD) sy systems.

flow requirements and CRD system flow capability.The revised PRA w Sequence 2:

i Small LOCA followed by failure to provide the required pool cooling.

Finding: The frequency estimate seems reasonable. However, it is higher than typical estimates found in other BWR PRA analyses and should be lowered by pool coolingthe improving operating procedures needed for long term RHR recovery in the mode.

Remarks:

less than 1 inch in diameter.The frequency of a small LOCA includes the frequen The licensee will look into the appropriateness of the high estimate used for operator failure to manually start the RHR pumps and to align RHR system valves to establish the pool cooling function following a postulated small LOCA event (referred to as R7 event in the PRA).

l 1

)

, ;Seque9c3 3 4 main steam isolaticn valva closuro transient followed by failure of the high pressure coolant _ makeup systems, and failure of the manual ADS to allow makeup to the low pressure the reactor. coolant makeup systems to provide coolant Finding: The frequency estimate seems reasonable.

Remarks: Sara as sequence 1.

hour period will not be affected by the initiating event.The PRA sssum The reduced capacity factor in estimating the due to the initiating recent event prolonged outage has been trken into acco frequency.

Sequence 4:

A trensions required followed by 1-3 stuck opeu relief valves and failure to provide pool' cooling.

Finding: The frequency estimate seems underestimated.

Remafks: ;The licenses should reexamine the approp'riateness of the failure probability Ferry plant. estimate used for the 1-3 stuck open relief valves for the Browns -

The ifcensee should also reexamine the plant operating procedures used in estimating the operator failure to manually start RHR pum and elign RM system valves to establish the pool cooling function following ,

tranMent the PRA). followed by 1-3 stuck open. relief valves (referred to as R7 event in .;

Sequence 6:

A pressure regulator valves closure trans'ient followed by failure of the feedwater and high pressure coolant makeup systems, and failure of the manual ADS blowdown to allow the low pressure coolant makeup systems to provide coolant inventory makeup to the reactor.

Finding: The frequene,y estimate seems reasonable. -

Remarks: Same as sequence 1. o hour period will not be affected by the initiating event.The PRA a:;sumes t The reduced capacity factor due to the recent prolonged outage has been taken into account in estimating the initiating event frequency.

Sequence 7:

A loss of coadenser vacuum transia. t followed by the failure of the high oressure cooli.nt makeup systems, and failure of the manual ADS blowdown to .

allow thetolow makeup the pressure reactor. coolant makeup systems to provide coolant inventory '

Finding: The frequency estimate seems reasonable.

Remarks: Same as sequence 1.

hour period will not be affected by the initiating event.The PRA assumes Also, the reduced i capacity factor due to the recent prolonged outage has been taken into account in estimating the initiating event frequency.

{,

)

t Sequence 8:

. A loss of offsite power event followed by 1-3 stuck open relief valves and failure of the emergency equipment cooling water system resulting in failure of all coolant inventory makeup systems and pool cooling systems.

Finding: The frequency estimate seems somewhat underestimated. The licensee should datermine the realistic success criteria for the emergency equipment cooling water (EECW) system and then: focus attention on the actual accident vulnerability to be identified from this sequence.

Remarks:

The licensee should look into the appropriateness of the failure probability estimate used for 1-3 relief valves sticking open. The licensee should also look into the appropriateness of the use of a success criterion that 3 cutofofoffsite the loss 4 (for power all three Units) EECW system pumps are needed to mitigate event.

If the licensee finds that 2 out of 4 EECW -

trains are. sufficient to mitigate the loss of offsite power event for all 3 Bro'wns' Ferry, Units, then the licensee should appropriately reduce the high sequence frequency estimate. The reduction in sequence frequency is suggested because the sequence involves pool cooling failures and is expected to result.

in an unscrubbed release following a core damage event. We also note that the EECW system unavailability is dominated by unavailability due to test and maintenance into cerefully. and thus the test requirements of the EECW System should be looked Sequence 9:

Finding: It was shown that the sequence is illogical. The licensee agreed with the finding.

Sequence 10: '

A transient followed by 1-3 stuck open relief valves, failure of the high

• pressure coolant systems, and failure of the low pressure coolant systems to provide coolant inventory makeup to the reactor and required pool cooling.

Finding: The frequency estimate seems somewhat underestimated. Also, the licensee should consider some measures in reducing the frequency estimate of the sequence involving a transient followed by 1-3 stuck open relief valves and pool cooling failure. The above sequence is expected to result in an unscrubbed release following a core damage event.

l

Remarks:

The ifcensee should look into the appropriateness of the failure probability estimate used for 1-3 stuck open relief valves as applicable to the Browns Ferry plant. The sequence frequency estimate is also based on a success criterion that one RHR pump is needed for vessel injection and two RHR pumps are simultaneously needed for pool cooling (referred to as the R2 event in the PRA). This criterion may be overly stringent, assuming that the RHR pumps cou?d be re-started manually (without automatic actuation signals);.and considering a recovery time of at least 30 to 60 minutes available for the RHR system prior to substantial pool heateup. The future revisions to the PRA should injectionexplore a realistic and pool cooling success functions.criterion for a combination of the vessal 1

11. ATWS Sequences:

Finding: The frequency estimates seem reasonable.

Remarks:

(RPT) includes the modification made recently to the recircu design. {

The PRA assumes that successful operation of the high pressure  ;

coolant GPM.

makeup to the reactor following an ATWS event needs only about 700 The PRA also assumes that a 700 GPM flow rate capability is sufficient {

4 to prowide coolant makeup for ATWS events followed by stuck open relief valve -

scenarios. '

Thus the PRA has taken credit for the combined operdtion of the RCIC and the'CR0 system as a backup to the HPCI system for A1WS sequences involving successful operation of the standby liquid control (SLC) system.

The revised PRA should provide supporting thermal hydraulic analyses to support the above conclusion. At the Browns Ferry facility, the ATWS induced core damage frequency is dominated by SLC system failures. The SLC 1 system unavailability is in turn dominated by unavailability due to test and maintenance activities. As part of safety improvement activities for the Browns Ferry facility, the Staff previously recommended (Reference 1) a design improvement to the SLC system such as an online test capability. I licensee has decided to provide enriched boron to the SLC system. In turn, the the enriched boron will require fewer testing requirements than theIn general, conventional boron solution. ,

The future revisions to the PRA should model the impact of enriched boron on the unavailability of the SLC system. The Staff believes that enriched boron will reduce the overall SLC system unavailability.

l

12. Sequences involving pipe breaks outside the containment:

Finding: The frequency estimates seem reasonable.

Remarks:

As part of the modeling of the liPCI and RCIC system unavailability, the licensee should verify the qualification requitstments of the (normally open) outboard steam supply line isolation valve in the HPCI ud RCIC systems to close against high blowdown loads resulting fram downstream postulated pipe {

breaks outside the containment, and resulting systems interactions (location l dependent failures), if any. i

l The following is an itemization of the audit findings of unavailability estimates (Septemberof1987 some selected systems provided in the revised Browns Ferry PRA Version).

i l

1. ADS system:  !

The PRA assumes that 4 out of 6 automatic depressurization system (ADS) valves are needed to depressurize the reactor following a transient and failure of the high pressure coolant makeup systems. 4 indicates that the above assumption is somewhat pessimistic.The Staff review of oth The licensee should look into applicable safety and other thermal hydraulic analyses performed for typical BWR 4 plants for a more realistic assumption than the one used in the revised PRA. If, for example, the licensee finds that 2 out then frequency estimates of sequerces 1, 3, 6, and 7, a should be reduced. ,

The reductio'n in the frequency estimates of the above sequences would be for transients such as loss of feedwater, main steam iso pressure the Browns regulator Ferry closure, facility. and loss of condenser vacuum, were implemented at above design improvement as a part of the safety improvement ac the Browns Ferry facility and TVA agreed by letter dated March 1,1988.

2. RHR system:

The system unavailability estimate for a small LOCA event and a transient with 1-3 stuck op6h eh,es ;s based on a succe's criterion that one RHR pump is needed for vessel coolant injection and two RHR pumps are simultaneously needed for suppression pool cooling (referred to as the R2 in the PRA). This criterion may be overly striegent, assuming that RHR pumps could be re-started manually; and considering the availability of a recovery time J at least 2 to a hours for the RHR system prior to a substantial pool heat up. Future revisions to the PRA should explore a realistic success criterion for a ,

combination of the vessel injection and pool cooling functions. The licensee should also look into the appropriateness of the relatively high estimate used for operator failure to manually start RHR pumps and align RHR system valves to establish the pool cooling function following a small LOCA event or a transient followed by 1-3 stuck open relief valves (referred to as the R7 event in the PRA).

The licensee indicated that, for initiating events such as a small LOCA event or a transient with 1-3 stuck open relief valves, the combination of venting the containment and providing high pressure service water to the reactor is not a viable backup method of pool cooling to mitigate the initiating events.

Staff review of other BWR PRAs indicates that the combination of venting the containment and providing low pressure coolant makeup to the vessel from the of the decay heat removal sequences resulting from poo Therefore, the licensee should explore alternate ways of cooling the core following event. pool cooling failures subsequent to a transient or a small LOCA

' " ~

g ' 3.' EECW System:

The licensee should Icok it,to the appropriateness of the use of a success criterion are neededthat 2 out of to mitigate the4loss (rather 1 out of offsite of 4event.

power for each Unit) EECW system pu power event will have impact on all three Browns Ferry Units.The loss of offsite If the licensee sufficient to mitigate the loss of offsite power event, th should requentify the relatively high (Sequence No. 8) frequency estimate.

The reduction in sequence frequency is suggested because the sequence pool cooling f'o11owing failures a core damage and is expected to result in an unscrubbed release event.

We also note that the EECW system unavailability is dor:inated by unavailability due to test and maintenance, and thus the test requirements of the EECW System should be looked into c .

4.. ,MSIV 5ystem: ,

The licensee' indicated that desigr.~.nodtfications will be provided at the Browns Ferry facility so the the main steam isolation valves (MSIVs) will close at water level L1 instead of water level L2. Future revisions to the PRA of shouldheat the-norr.a1 include sink. the overall impact of this modification on the availa The Staff previously recommended (Reference 1) the above plant Browns Ferry improvement facility. as part of the safety improvement activities for the

References:

1. S. D. Ebneter to 5, A. White (TVA), "Probabilistic Risk Assessment for Browns Ferry fluclear Plant (BFt4P)," October i,1987.

o i

_ _ _ _ . _ _ _ _ _ _ . - -