ML20134J601

Forwards Review of Final Accident Sequence Precursor Analysis of Operational Event at Plant,Unit 1 Reported in LERs 445/95-003 & 445/95-004 on 950611

ML20134J601
Person / Time
Site:	Comanche Peak
Issue date:	11/12/1996
From:	Polich T NRC (Affiliation Not Assigned)
To:	Terry C TEXAS UTILITIES ELECTRIC CO. (TU ELECTRIC)
References
NUDOCS 9611150302
Download: ML20134J601 (18)

Text

._. . .

Mr. C. Lance Terry November 12, 1996 TV Electric

• Grcup Vice President, Nuclear Attn: Regulatory Affairs Department P. O. Box 1002 Glen Rose, TX 76043

SUBJECT:

REVIEW OF PRELIMINARY ACCIDENT SEQUENCE PRECURSOR ANALYSIS OF EVENT AT COMANCHE PEAK STEAM ELECTRIC STATIC 1, UNIT 1

Dear Mr. Terry:

Enclosed for your information is a copy of the final Accident Sequence Precursor analysis of the operational event at Comanche Peak Steam Electric Station, Unit 1, reportad in Licensee Event Report Nos. 445/95-003 and -004.

l This final analysis (Enclosure 1) was prepared by our contractor at the Oak Ridge Nationai Laboratory, based on review and evaluation of your comments on the preliminary analysis and comments received from the NRC staff and frcm our independent contractor, Sandia National Laboratories. Enclosure 2 contains our responses to your specific comments. Our review of your comments employed the criteria contained in the material which accompanied the preliminary I analysis. The results of the final analysis indicate that this event is a precrrsor for 1995.

Please contact me at (301) 415-1038 if you have any questions regarding the enclosures. We recognize and appreciate the effort expended by you and your i staff in reviewing and providing comments on the preliminary analysis. I Sincerely, l ORIGINAL SIGNED BY: i l

Timothy J. Polich, Project Manager Project Directorate IV-1 Division of Reactor Projects III/IV Office of Nuclear Reactor Regulation Docket No. 50-445

Enclosures:

1. Final analysis of Licensee Event Reports 445/95-003 and 004

2. Response to licensee comments  ;

cc w/encls: See next page i

DISTRIBUTION:

hkatl OGC PUBLIC JRoe ]! '

I GHill (4) ACRS TPolich (2) JDyer, RIV PDIV-1 r/f CHawes (2) CGrimes EAdensam (EGAl) i WBeckner SMays P0'Reilly l Document Name: CPAE00. ASP OFC PM/PD4-1 LA/PD4-1  %

NAME TPoliqh/cf CHawes OUlb DATE it hhb6 U/(?96

/ h COPY YES/N0 YES/N0 OFFICIAL RELORD COPY l

l 9611150302 961112 i

PDR ADOCK 05000445 S PDR

a uru p *, UNITED STATES g j 2

NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20666 4 001

\*****/ November 12, 1996 Mr. C. Lance Terry  ;

TU Electric  :

Group Vice President, Nuclear  ;

Attn: Regulatory Affairs Department  ;

P. O. Box 1002 l Glen Rose, TX 76043

SUBJECT:

REVIEW OF PRELIMINARY ACCIDENT SEQUENCE PRECURSOR ANALYSIS OF EVENT -

AT COMANCHE PEAK STEAM ELECTRIC STATION, UNIT 1

Dear Mr. Terry:

1 Enclosed for your information is a copy of the final Accident Sequence  !

Precursor analysis of the operational event at Comanche Peak Steam Electric l Station, Unit 1, reported in Licensee Event Report Nos. 445/95-003 and -004.

This final analysis-(Enclosure 1) was prepared by our contractor at the Oak Ridge National Laboratory, based on review and evaluation of your comments on the preliminary analysis and comments received from the NRC staff and from our independent contractor, Sandia National Laboratories. Enclosure 2 contains our responses to your specific comments. Our review of your comments employed the criteria contained in the material which accompanied the preliminary analys is. The results of the final analysis indicate that this event is a precursor for 1995.

Please contact me at (301) 415-1038 if you have any questions regarding the enclosures. We recognize and appreciate the effort expended by you and your  ;

staff in reviewing and providing comments on the preliminary analysis. l 1

Sincerely, 4L4/PR Timothy J. Polich, Project Manager Project Directorate IV-1 Division of Reactor Projects III/IV Office of Nuclear Reactor Regulation Docket No. 50 445

Enclosures:

1. Final analysis of Licensee Event Reports 445/95-003 and 004

2. Response to licensee comments cc w/encls: See next page l

l .

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

Mr. C. Lance Terry TU Electric Company Comanche Peak, Units 1 and 2 cc:

Senior Resident Inspector Honorable Dale McPherson U.S. Nuclear Regulatory Commission County Judge P. O. Box 1029 P. O. Box 851

, Granbury, TX 76048 Glen Rose, TX 76043 Regional Administrator, Region IV Office of the Governor U.S. Nuclear Regulatory Commission ATTN: Susan Rieff, Director .

611 Ryan Plaza Drive, Suite 400 Environmental Policy l Arlington, TX 76011 P. O. Box 12428 Austin, TX 78711 Mrs. Juanita Ellis, President Citizens Association for Sound Energy Arthur C. Tate, Director 1426 South Polk Division of Compliance & Inspection Dallas, TX 75224 Bureau of Radiation Control Texas Department of Health Mr. Roger D. Walker, Manager 1100 West 49th Street Regulatory Affairs for Nuclear Austin, TX 78756-3189 Engineering Organization Texas Utilities Electric Company l 1601 Bryan Street, 12th Floor '

Dallas, TX 75201-3411 Texas Utilities Electric Company l c/o Bethesda Licensing 1 3 Metro Center, Suite 610 Bethesda, MD 20814 George L. Edgar, Esq. ,

1 Morgan, Lewis & Bockius i 1800 M Street, N.W. l Washington, DC 20036-5869 l

i l

LER No. 445/95-003,-004 l LERNo. 445/95-003,-004 I

! Event

Description:

Reactor trip, Auxiliary Feedwater (AFW) pump trip, second AFW l pump unavailable ,

Dete ofEvent: June 11,1995 l

l Plant: Comanche Peak 1 Event Summary While at 100% power on June 11,1995, Comanche Peak I experienced a control power supply failure resulting in both main feedwater pumps (MFPs) tripping and operators subsequently initiating an anticipatory reactor trip. Flow from one of two motor-driven auxiliary feedwater pumps (MDAFWP) was initially unavailable and the turbine-driven auxiliary feedwater pump (TDAFWP) started on low-low steam generator level but tripped on overspeed. The conditional core damage probability (CCDP) estimated for this event is 6.5 x 10 5 Event Description While at 100% power on June 11,1995, Comanche Peak I experienced a control power supply failure resulting in both MFPs tripping and operators subsequently initiating an anticipatory reactor trip. Slave relay testing was under way when a non-safety related inverter transferred from its normal inverter ac power supply to its altemate power supply. The alternate ac power supply was deenergized as required by the test procedure at the time, so associated loads were deenergized The specific cause of the transfer is not certain but it may have been caused by an electrical transient in a static transfer switch control circuit. less of the power supply caused a spurious 'MFP oil pressure low" signal when auxiliary relays in pump supervisory mstrumentation deenergized and actuated. This change caused the condensate pumps to trip; loss of the condensate pumps caused both MFPs to trip. Operators then initiated a manual reactor trip in anticipation of an automatic one.

The MFP trips caused an auto-actuation of the MDAFWPs. MDAFWP 1-02 (Train B) started and supplied water to steam generators (SGs ) 3 and 4 (Fig.1). MDAFWP l-01 (Train A) was aligned to its test header at the time and was not immediately available to supply water to the SGs. The TDAFWP started on low-low SG level but tripped on overspeed, caused by a failure of the governor valve to control turbine speed. The govemor valve stem was found to be corroded and binding against the valve packing. Operators realigned MDAFWP l-01 from the test header to its norms! configuration, and the pump supplied cooling to SGs 1 and 2 within about 8 min.

Additional Event-Related Information The licensee event report (LER) provided additional information conceming the thermal-hydraulic effects of having only one AFW pump available immediately after a plant trip. Plant safety analyses assume for a " Loss of Normal Feedwater Flow" transient that the TDAFWP or both MDAFWPs provide a flow rate of at least 860 gpm to the SGs. During this transient, only one MDAFWP was initially available, providing a reduced flow rate to the SGs. However, the LER indicated that the reduced flow rate was adequate to remove plant decay heat from the SGs because of the early manual trip afthe reactor and because initial water levels in the SGs were greater than the assumption used in the FSAR analysis.

Because sufficient heat removal capability was available, the thermal expansion of the reactor coolant system inventory did not fill the pressunzer completely.

I ENCLOSURE 1

F LER No. 445/95-003, 004 l

Modeling Assumptions This event was modeled as a reactor trip with the TDAFWP failed and flow from MDAFWP l-01 initially unavailable.

Basic event AFW-TDP-FC lC was set to "TRUE" (failed). (Table I provides a description of the basic event names.)

It was assumed that if the remaining AFW pump had failed, operators would have attempted to recover the system by realigning MDAFWP l-01 (as they did). Recovery of MDAFWP l-01 was incorporated into the models using the l methodology described in Reference 4. This methodology suggests a nonrecovery probability of 0.1 when *[fjailure appeared recoverable in the required period from the control room, but recovery was not routine or involved substantial stress." A similar nonrecovery value was estimated by assuming that nonrecovery as a function of time was lognormally distributed with a median response time of 8 min and a recovery window of 30 min. Assuming a burdened-recovery j error factor of 6.4, the probability of nonrecovery within 30 min is approximately 0.1, which is the same value as  ;

obtained using Ref. 4. Consequently, the nonrecovery probability for MDAFWP l-01 was incorporated by setting the i probability for event AFW-MDP-FC-1 A equal to 0.1. In addition, because AFW is required without delay during ATWS l sequences, a new event, AFW-MDP-FC-AA, with a nonrecovery probability of 1.0 was substituted for l AFW-MDP-FC-1 A in the ATWS model. Because it was assumed that the entire 30 min would be dedicated to recovery of AFW-MDP-FC-1 A, the system nonrecovery, AFW-XHE-NOREC, was set to 1.0.

Because main feedwater apparently could not have been recovered without correcting the inverter problem, restarting the condensate system, and restoring a feedwater pump to senice, the feedwater system was assumed not to be recoverable (MFW-XHE-NOREC = "TRUE*). l l

The failures in this event increase the potential significance of failure to trip /ATWS sequences. To model potential  :

reactor trip failures more accurately, the reactor trip model was modified (as shown in Fig. 2) to account for recoverable versus nonrecoverable RPS failures.

'Ihe event trees for Comanche Peak assume that conditions requiring a reactor trip will first result in an automatic reactor trip demand and, if the automatic trip fails, a manual reactor trip demand. During this event, once operators recognized thrt a loss of main feedwater flow had occuned, they initiated a manual reactor trip. Because of the operators' quick response, consideration was given to the potential impacts of the early reactor trip on ATWS sequences. The Comanche Peak Final Safety Analysis Report (FSAR) indicates that 1 to 1% min may elapse between a loss of feedwater and an automatic reactor trip. The additional 1 min of response time available to operators during postulated ATWS sequences in this event was not believed to materially affect the event sequences or probabilities, and no related model changes were indicated.

Analysis Results The CCDP estimated for this event is 6.5 x 104 The dominant core damage sequence (sequence 20 on Fig. 3) involves

+

a successful reactor trip, l

. failure of AFW l

+ failure ofMFW, and

=

failure offeed-and-bleed cooling.

The second highest core damage sequence (sequence 21-8 on Figs. 3 and 4) involves l

failure to successfully trip, i

successful control of reactor pressure, and

=

failure of AFW.

I 2

l

_ _ _ . - _ _ _ _ _ _ _ _ . _ _ _ _ . _ _ _ _ . _ . - _ ~ _ _ . _ _ _ . _ .__ _ _ _ _ ____.. _ _ _._ _ _. _ . .

I O l

l l

l LER No. 445/95-003, 604 l

Definitions and probabilities for selected basic events are shown in Table 1. The conditional probabilities associated I

with the highest probability sequences are shown in Table 2. Table 3 lists the sequence logic associated with the  ;

! sequences listed in Table 2. Table 4 describes the system names associated with the donunant sequences. Mmunal cut i sets associated with the domment sequences are shown in Table 5. i Acronymas I

ac attemating current AFW auxiliary feedwater l ATWS anticipated transient without scram $

CCDP conditional core damage probability j CST ^=te storage tank FSAR final safety analysis report LER licensee event report l l MDAFWP motor-driven auxiliary feedwater pump l MFP main feedwater pump -l SO steam generator I l

t SWS service water system l TDAFWP turbine-driven auxiliary feedwater pump l

l References

! 1. LER 445/95-003, Rev.1," Loss of Both Condensate and Both Feedwater Pumps Due to Failure of Non-Safety Related Inverter Resulted in a Manual Reactor Trip," August 14,1995.

l 1

I i 2. I.ER 445/95404, Rev.1," Allowed Outage Time Was Exceeded on Turbine-Driven Auxiliary Feedwater Pump l

Which Tripped on Overspend," September 8,1995.

3. Texas Utilities Generating Company, Comanche Peak Steam Electric Station FinalSafety Analysis Report.

4. M. B. Sanison et. al., Alethods improwments incorporatedinto the SAPHIRE ASP Afodels, NUREGICP-0140, l

Vol.1, Proceedings of the U.S. Nuclear Regulatory Commission, Twenty-Second Water Reactor Safety Information Meeting, April 1995.

l l

l 3  ;

i I I l.

LER No. 445/95-003,-004 l

l

\

l l

l l --

l From CST

.l _ . _ [ ';

I l !ri !

l 4 A~~N-- i l!!l

--c< A- From Mrw . . J. . ._ t-,_I . .!!l

.-- N N '

N N. _ Sii e i MDAFWP es Tem A .1 - - -f I y l !n!

o

--*< N i l!!!

db From MFW . _ !J. . . . .;i j ' ii

, s.-

SO 2 r

i j; h -

l X .l _. J 1 l 'n!

--O< W * & i lllI From MFW ._ _ J._

--N '

@ McArwp l N >9t- _ Y')

so l ,

I Tom B

' t-Do-M- ,w r, From 8W8 )

)[ """"'D< M i ~~~ !J() !

m ,

N @>^ x A---N--

y ,

From MFW . _ J. _ . . pls f ' ,l TDAFVP s.-

To CST SG4 i

Fig.1 Auxiliary feedwater system for Comanche Peak (Source: Texas Utilities Electric Co., Comanche Peak Steam Electric Station Final Safety Analysis Report) i a

4

LERNo. 445/95-003 -004 Failure to Trip

' i Non-recoverable Recoverable RPS Failures Failures O

V RPS-NONREC Operator Fails to Recoverable RPS Failures Manually Trip the Reactor l i l RPS-REC RPS-XHE-XM-SCRAM Fig. 2 Fault tree modeling recoverable and nonrecoverable failures for the failure to trip.

5

5

- -- ,,,,, ,,,,, ,. imaa. == --

La sermnes , pournerer puunsawise- em neuemme amo - = ~*me seat = ==

evewr everuns emme sacros cemmen sescoalesi tesesYun sensowt mmac Esso

. es.vames wr arw arw rosrv pouv.aus set pas esoOCL oOOLDONN 8BSr DFW 8 *^

E 1 OK n 2 OK 3 OK

$a 4 OK I

g 3 5 CO

.c i

I g 7 CD

<a p . CD

. OK 10 OK g 11 OK 12 OK v ,

I g 13 CD

." , 14 OK h 15 CD 16 CD 17 on

, 18 OK I

is CD 20 CD

{

21 T ATVWB N 2

Transient Event Tree P Comenche Peak Stoem Electric Staton t v.

ASP PWR Class B G

'T1 N

k Asmtym1ED N

resteMW APW EaEftGEpsCY POmf RCS N N'4WG fESEmL FYWEAt#E h PenseUBE P0ft teat E unsTso Avvis e0RA M OM2MGED UggeS TWs fes0WL IECWICULAftDIG PORV-A RHR HPR

" 8" E ATW8 RC8 PRESS AFW-ATWS BORATION COOLDONN Is I

o. -

c

$ t g- 21-1 oK s

b o

21-2 OK 9.

k 21 4 OK B

S q 21 4 CD a m E

21 4 OK f

E 4 21 4 CD R

sE n 21-7 CD 8

k 214 CD w

2'.

N zu co x

2 ATWS Event Tree P Comanche Peak Steam Electric Ststion t w

ASP PWR Class B 3

.e k

l t

i LER No. 445/95-003,-004 l

l l Table 1. Definitions and Probabilities for Selected Basic Events for LER 445/95-003,-004 MWhd l Event Base Current for this name Description probability probability Type event l

l IE-LOOP Loss of Offsite Power Initiating 8.5 E-006 0.0 E+000 IGNORE No l l Event l IE-SOTR Steam Generator Tube Rupture 1.6 E 006 0.0 E+000 IGNORE No Initiating Event j i IE-SLOCA Smallloss of Coolant Accident 1.0 E 006 0.0 E+000 IGNORE No l Ir.itiating Event IE-TRANS Transient Initiating Event 5.3 E-004 1.0 E+000 Yes AFW-MDP -CF-AB Common Cause Failure (CCF) of 2.1 E-004 2.1 E-004 No Motor Driven Pumps l AFW-MDP-FC-AA AFW Motor Driven Pump A 4.0 E-003 1.0 E+000 TRUE Yes Fails During ATWS AFW-MDP-FC-1A AFW Motor-Driven Pump A 4.0 E-003 1.0 E-001 Yes

, Fails AFW-MDP-FC-1B AFW Motor-Driven Pump B 4.0 E-003 4.0 E-003 No Fails i

AFW-PMP CF-ALL AFW Serial Component 2.8 E-004 2.8 E-004 No Common to all Trains Fails

! (i.e. CCF) l l AFW TDP-FC-IC AFW Turbine-Driven Pump 3.2 E-002 1.0 E+000 TRUE Yes l

Fails AFW-XHE-NOREC Operator Fails to Recover AFW 2.6 E-001 1.0 E+000 TRUE Yes System AFW-XHE-NREC ATW Operator Fails to Recover AFW l.0 E+000 1.0 E+000 No System During an ATWS AFW-XHE XA-SSW Operator Fails to Align Suction 1.0 E-003 1.0 E-003 No to Service Water Symm (SSW)

HPI-XHE-XM-FB Operator Fails to initiate Feed 1.0 E-002 1.0 E-002 No

and Bleed Cooling l MFW-SYS-TRIP Main Feedwat.r System Trips 1.0 E+000 1.0 E+000 No MFW XHE-NOREC Operator Fails to Recover Main 2.6 E 001 1.0 E+000 TRUE Yes i

Feedwater __

l 8

l

___ _ _ _ _ _ _ .__..m . _ . . . . . . _ _ . -. _ - _ _ _ . - _ . _ _ _ _ _ . _ . _ _ _ . _ . . . _ . . . _ . . _ . . . . . . _

4 LER No. 445/95-003,-004 Table 1. Deflaitions ud Probabinties for Selected Basic Events for LER 445/95-003,-004 Medined Event Base Current for this masse Description probabuity probability Type event PPR-SRV CC 1 Power Operated Relied Valve 6.3 E-003 6.3 E-003 No (PORV) 1 Fails to Open on Demand I

! PPR SRV-CC-2 PORV 2 Fails to Open on 6.3 E-003 6.3 E 003 No I Demand l

l RPS-NONREC Nonremverable RPS Trip 2.0 E-005 2.0 E-005 NEW Yes i

Failures RPS-REC Recoverable RPS Failures 4.0 E@5 4.0 E-005 NEW Yes l RPS-XHE-XM-SCRAM Operator Fails to Manually Trip 1.0 E-002 1.0 E@2 NEW Yes

the Reactor l

l t

l l

1 a

1 9

LER No. 44565-003,-004 1

Table 2. Sequence Conditional Probabilities for LER 44555-003,-004 Conditional core Event tree damage Percent name Sequence name probability Contribution l (CCDP)

TRANS 20 4.3 E-005 66.8 l l

TRANS 21-8 2.0 E-005 31.2 l

l Total (all sequences) 6.5 E-005 l

l l

Table 3. Sequence logic for Dominant Sequences for LER 44555-003,-004 {

Event tree name Sequence name logic TRANS 20 /RT, AFW,MFW,F&B l l TRANS 21-8 RT,/RCSPRESS AFW-ATWS l

l Table 4. System Names for LER 445S5-003,-004 l System name Logic AFW No or Insufficient AFW Flow AFW-ATWS No or Insufficient AFW Flow - ATWS F&B Failure to Provide Feed and Bleed Cooling MFW Failure of the Main Feedwater System RCSPRESS Failure to Limit RCS Pressure to <3200 psi RT Reactor Fails to Trip During Transient 10 l

i I

_ . . . . _ _ _ _ . -_ . _ _ _ . . . -_ _. _ _. _. . ._. m _. . . . . _ _. _ .__ _ _ .. _ -. = _ _ __ _ .--

LER No. 445/95-003,-004 Table 5. Conditional Cut Sets for Higher Probability Sequences for LER 445/95-003, 404 Cut set Percent Conditional

! Naasber Contribution Probability

• Cut sets

• TRANS Sequence 20 4.3 E-005 ~'/ /

• M 1 22.9 1.0 E-005 AN-XHLXA-SSW, AFW-XHE-NOREC, MFW-SYS-TRIP, MN XHE NOREC, HPI-XHLXM-FB 2 14.4 6.3 E-006 AN-XHLXA-SSW, AFW MIE-NOREC, MFW-SYS-TRIP, MFW XHENOREC, PPR-SRV-CC-1 3 14.4 6.3 E-006 AN-XHE XA-SSW, AN-XHLNOREC, MFW SYS-TRIP, MFW XHLNOREC, PPR SRV CC-2 l

l 4 9.2 4.0 E-006 AN-MDP-FC 1 A, AFW-MDP-FC-1B, AFW-TDP-FC-IC, . ,

j AFW-XHE-NOREC, MFW-SYS-TRIP, MFW-XIIE-NOREC, j i HPI XHE XM FB l

5 6.4 2.8 E-006 AN PMP<F-All, AFW-XHE-NOREC, MFW-SYS-TRIP, MFW-XHE-NOREC, HPI-XHE-XM-FB 6 5.8 2.5 E-006 AN MDP-FC-1 A, AFW MDP FC-1B, AFW TDP FC-lC,

AFW-XHE-NOREC, MFW-SYS TRIP, MFW XHE-NOREC, -

i PPR-SRV CC 1 i

l 7 5.8 2.5 E-006 AN-MDP FC-1 A, AFW TDP-FC 1C, AFW MDP-FC 1B, l AFW XHE-NOREC, MFW SYS-TRIP, MFW-XHE-NOREC. 1 i I PPR SRV-CC-2 8 4.8 2.1 E-006 AFW-PMP4F-AB, AFW-MDP-FC-lC, AFW-XHE-NOREC, l MFW-SYS-TRIP, MFW XHLNOREC, HPI-XHE-XM-FB '

l

( 9 4.0 1.7 E-006 AFW-PMP CF ALI. AFW-XHLNOREC, MFW-SYS-TRIP, t MFW-XHE-NOREC, PPR-SRV CC-1 1

10 4.0 1.7 E-006 AFW-PMP-CF.ALL, AFW-XHE NOREC, MFW-SYS-TRIP, l l MFW-XHE-NOREC, PPR-SRV-CC-2 l t  :

l 11 3.0 1.3 E-006 AN MDPCF AB, AFW TDP-FC lC, AFW XHE-NOREC, MFW-SYS TRIP, MFW XHLNOREC PPR-SRV-CC-1 12 3.0 1.3 E-006 AN-MDP-CF AB, AFW TDP FC 1C, AFW-XhE NOREC, MN-SYS TRIP, MFW-XIILNOREC, PPR SRV CC-2 l TRANS Sequence 21-8 2.0 E-005 4 -

~ .' f# , m ^

l 98.0 2.0 E-005 RPS NONREC, AFW MDP-FC-AA, l AFW TDP-FC-1C, AFW XHE-NREC-ATW

.2 1.9 4.0 E-007 RPS REC, RPS XHLXM-SCRAM, AFW-MDP-FC-AA, AFW-TDP-FC 1C, AFW-XHE-NREC-ATW

, Total (au sequences) 6 5 E 405 a -

^

f 11

- -. . ~ _ . .-.

• . J

. o l l

l LER No. 445/95-003,-004

• ne condissonal probability for each cut set is determmed by muhiplying the pobability of the initiating event by the pobabilities of the basic events in that nunanal cut est. De pobability of the initiasms events are given in Table 1 and begin with the designator "IE". ne probabilities for the basic events are also given in Table 1.

Basic events AFW MDP FC AA, AFW TDP-FC-IC, AFW-XHE NOREC, and MFW XHLNOREC are all type TRUE events which are not normally included in the output of fauk tres reduction payans. Dese events have be*en added to aid in understandmg the sequences to potential core da.nage associated with the event.

12 I

i l

LER No. 445/95-003,-004 l

t

{

LER No. 445/95-003,-004 Event

Description:

Reactor trip, Auxiliary Feedwater (AFW) pump trip, second AFW l

pump unavailable l

Date ofEvent: June 11,1995 Plant: Comanche Peak i l

i Licensee Comments l

Reference:

Letter from C. L. Terry, Texas Utilities Electric Company, to U.S. Nuclear Regulatory Commission, transmitting Comanche Peak Steam Electric Station (CPSES) - Unit 1 Docket Nos.

50-445 Comments on Preliminary Accident Sequence Precursor Analysis ofReactor Trip at CPSES Unit 1 on June 11,1995, TTX 96397, July 15,1996.

l Comment 1: The quantitative values presented in this analysis are, in some cases, different from the values used in the CPSES Individual Plant Examination (IPE) study. In general, the core damage frequency contributions in the Preliminary Accident Sequence Precursor Analysis are based on sosnewhat conservative assumptions and more simplified models. Therefore, the numerical values should not be considered as accurate contributions to total core damage frequency.

Rather, they should be used to determine the relative importance of various accident sequence precursors.

Response 1: Since its inception, this has been one of the primary objectives of the acddent sequence

! precursor program. Over time, efforts have been made to make the models which are employed l more realistic. However, it remains true that the ASP models are more simplified and potentially more conservative than those found in some IPEs.

Comment 2: The failure probabilities for the following events appear to be significantly higher than the values used in the CPSES IPE study.

. Operator fails to initiate Feed and Bleed cooling: a value of 1.0E-2 was used in this study versus 1.0E-3 used in the CPSES IPE.

l

• Failure of non-recoverable RPS trip; a value of 2.0E.5 was used in this study versus 1.0E-5 used in the CPSES IPE.

Response 2: Because of the sparseness of system failure events, data from many plants must be combined to estimate the failure probability of a multitrain system or the frequency of low- and moderate-frequency events (such as LOOPS and small-break LOCAs). Because of this, the 13 I-ENCLOSURE 2 l

LER No. 445/95-003,-004 1

i l

modeled response for each event will tend toward an average response for the plant class (refer to Table B.1 in NUREG/CR-4674, Vol. 21 for a listing of plants and their respective plant class).

If systems at the plant at which the event occurred are better or worse than everage (difficult to ascertain without extensive operating experience), the actual conditional probability for an event i could be higher or lower than that calculated in the analysis. Regardless, the non-recoverable l

RPS trip value and the operator failure to initiate feed and bleed value are consistent with those i values used in other probabilistic risk assessments (e g., the Sequoyah PRA, the Farley IPE, the l McGuire IPE) for all ASP plant classes for PWRs. Nevertheless, if the values for HPI-XHE- l XM-FB (operator fails to initiate feco and bleed) and RPS-NONREC (nonrecoverable RPS trip failures) are changed to match the CPSES IPE values, the CCDP would be 3.7 x 10 Even with these changes, the resulting CCDP is on the same order of magnitude and the event still meets  :

the selection criteria as an ASP event (i.e., CCDP > 10-'). ,

i l

Comment 3: This event demonstrated that the CPSES operating crew was capable of recovering the Train A Motor-Driven Auxiliary Feedwater Pump (MDAFWPI-01) within 8 minutes and safely shutting down the plant. As a result, the failure probability to realign the Train A pump from a test configuration to the operating configuration should be very low. It should be noted that the Train A pump did not fail to operate and, therefore, the recovery here did not require repairing a failed pump but rather realigning Train A to an operating configuration. Consequently, the probability  ;

of not successfully realigning the pump under the given conditions should be bdween IE 2 and 1E-4.

This recovery, on the other hand, should have a failure probability of 1.0 for an Anticipated Transient Without Scram (ATWS) event since the pump is required to be available almost immediately.

Response 3: The non-recovery value was estimated using the methodology described by Sattison (Methods Improvements Incorporated into the SAPHIRE ASP Models, Sattison et. al., NUREG/CP-0140, Vol.1, Proceedings of the Twenty-second Water Reactor Safety Information Meeting, October 1994, USNRC).

In response to this comment, a more rigorous approach was taken. Non-recovery as a function of time was modeled as being lognormally distributed (see E. M. Dougherty and J. R. Fragola, Human Reliability Analysis, Wiley and Sons,1988), with a median response time of 8 minutes (actual response time) and a window of 30 minutes. Assummg a burdened-recovery error factor of 6.4, the probability of non-recovery within 30 minutes is approximately 0.1, the same as estimated by the Sattison approach.

Since it was assumed that the entire 30 minutes would be dedicated to recovery of AFW-MDP-FC-1 A, the system nonrecovery, AFW-XHE-NOREC, was set to 1.0.

As described in the modeling assumptions section, motor-driven AFW pump A was assumed to be inoperable for ATWS events.

i l

14 l

i LER No. 415/95-003,-004 I

Comment 4: The cut set numbers 1,2, and 3 of TRANS Sequence 20 in Table 5 have the basic event AFW-XHE-XA-SSW for which there is no defmition. This basic event appears to be the failure of the f operator to align the suction of the AFW pumps to the Station Service Water (SSW) system.

This capability exists at CPSES Units I and 2. However, it is normally in a locked closed position and it is only required when the nonnal water source of the Condensate Storage Tank (CST) is not available. Since the failure probability of the CST is very low, the alternate SSW water source was not credited in the CPSES IPE study. Nevertheless, if it is modeled, the 4 corresponding cut sets should also include the failure of CST. In cut sets 1,2, and 3, the CST j tenn is not included, but SSW alignment is included. l l

Response 4: Basic event AFW-XHE-XA-SSW is dermed in Table 1 as " Operator Fails to Align Suction to the Service Water System (SSW)." However, this event would be better described as " Operator fails to align makeup to the CST."

Comanche Peak's condensate storage tank (CST) has a voleme of approximately 500,000 gal.

Technical Specifications (3.7.1.3) require a minimum tank level of 53% be maintained.

Assuming that tank level correlates linearly with volume, this corresponds to about 265,000 gal.

The FSAR ind. cates that 282,000 gal will be dedicated for AFW operation and the IPE makes a similar assumption. This reserved capacity of the CST is good for 9 h-maintaining hot standby for 4h and then to cooling down to conditions which would permit alignment to the RHR. I i

The ASP model for AFW success requires sufficient inventory for up to 24 h of operation. A I simple calculation was performed using the Untermyer-Weills decay heat correlation to estimate  ;

the amount of CST inventory that would be required to remove this decsy heat without replenishment (

Reference:

S. Glasstone and A. Sesonske.NuclearReactorEngineering, D. van Nostrand,1%7). It was estimated that about 2.2E+9 BTUs would be rejected, requiring about 330,000 gal of CST inventory. This quantity is significantly more than the amount required by technical specifications or assumed in the IPE.

The actual level in the CST will fluctuate from near maximum (500,00') gal) to near the technical specification limit. Without further knowledge of Comanche Peak's operating practices, it is impossible to determine when sufficient inventory exists in the CST sad when it does not.

Regardless, any reasonable assumption will not gready affect the CCDP calculated for this event. I For example, ifit is assumed that the CST inventory would be inadequate 50% of the time, those cut sets containing the basic event AFW-XHE-XA.SSW would be weighted by 0.5 (specifically, cut set numbers 1,2, and 3 in TRANS Sequence 20). This change would reduce the estimated CCDP from 6.5E-5 to 5.2E-5. Hence, the event still qualifies as an ASP type event.

15