ML20029C121

LER 91-001-00:on 910212,use of Inadequate Control Room Filter Initiation Signal Due to Discrepancies Between Logic Diagrams.Main Control Room Doses Due to MSLB Outside Containment recalculated.W/910314 Ltr

ML20029C121
Person / Time
Site:	River Bend
Issue date:	03/14/1991
From:	England L, Odell W GULF STATES UTILITIES CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
LER-91-001, LER-91-1, RBG-34610, NUDOCS 9103260056
Download: ML20029C121 (16)

Text

_ - _ _ - _ - _ - _ _ -

GULF ST/ITES UTJLETIES COMPJINY

,# in se smm eau wa coun stinw w.m imsw om AR( A COM find fJh M M M6 t*S1 March 14 ,1991 REG-34610 File !bs. G9.5, G9.25.1.3 U.S. Nuclear Regulatory Conmission Docununt Control Desk Washington, D.C. 20555 Gentlenen:

River Berxl Station - Unit 1 Docket No. 50-458 Please find enclosed Licensee Event Report tb.91-001 for River Bend Station - Unit 1. This report is being subnitted pursuant to 10CFR50.73.

Sincerely, f -

i W. H. Odell Manager - Oversight River Bond Nuclear Group IAE/PDG/ /DCII/Jra/pj cc: U.S. Nuclear Regulatory Ccnmission 611 Ryan Plaza Drive, Suite 1000 Arlington, TX 76011 NRC Resident Inspector P.O. Box 1051 St. Franciarille, IA 70775 INPO Records Center

.1100 Circle 75 Parkway Atlanta, GA 30339-3064 Mr. C. R. Oberg s, Public Utility Ccomission of Texas 7800 Shoal Creek Blvd. , Suite 400 tbrth l Austin, TX 78757 )

l 9103260056 913314 PDR ADOCK 05000458

/ hl S PDR C

4 NMC FOMM 3E5 U t NUCLI AR 44t GUL ATORY COMMitsiON APPHOvtD OM8 NO 3160 0104

• & RPlHi& 4'10'92

' E51 W A T t b fldRD(N P(R Rt $PONtt TO COMPL Y W7 H 1Hl$

""^ ' ' " "

LICENSEE EVENT REPORT (LEN af8Mi . /'LTJJ'4'O,u"iefs'r',*u.'ri' i

"%i ft a'R ?.

AND RE POftT5 MANAGt ME NT SR ANCH 18 6306 US NUCLtAR Nt GUL ATORY COUV;1510N W ASHINGTON. Oc ?oM6 AND T O THE P Ar t HWOftK R E DUC f lON I'ROJt c T 0150 c104 OF F nct Of MAN sGE ME N T AND IRl"JCa T Wt A5HiNGTON. Y3 DC 7W'

'DOCEff NUMetR m PAGt 4 ps.CILir( N AME til RIVER BWD STATION ,

oIsl0loto1 458 11 1 lor l 115 TITLt (4)

Use of Lmproper Control Rxrn Filter Initiation Signal due to Discrepancies between lecie ningrnm OTMAR 5 ACitifies INv0LytD let IVtNT DAf8151 Lt R NUMet R i6i RIPDR f D All 17.

p3g y,, j gay ,gan F.<,tevi auts DOchil NUMBL A'5!

MONTM DA* YEAR YEAR 'N'La *'

, tm 0I6101010 1 I I

~

~ ~

0l2 1 l2 9 1 911 0l0 l l 0l0 0l3 1 14 9\ 1 THIA REPORT is suewtito PuRSVANT TO THE MgQUIRIMINTS Of 10 C8R l ICh.ca one a more P' **e faapawa' It il 0 1 6 l0 1 0 1 0 i i 1 QPE R ATING MODE m 20 402tti 20 06tel no 73ud2Hal 7 3 filbi lo Mientil to F3i.i;3nel 73.711el R M 40L4.h ti< 'I X oo, o i 8 g___ n ommne _

..u.m ,_

,>.nioo,

_ =;,g gg,,

n et.mn,.o x so n,..ano io n,.nin.. n Ai uw

? ro ai.imle so namn., so ni. nan.,i.n.,

no .o.i.m n.i .o n,.us a,,o .o n,.u Im (ICE Nlti CONT ACT S OR THIS LlR n2) gaug fittPHONE NuuBtR L.A. England - Director, Nuclear Licensing 510 14 3181114 1114151 COMPLETE ONE LINE FOR ( ACH COMPONENT # AILUR4 DitCRietD IN THf$ REPORT (13i

^ "C ' ' CAu$t $ v $T E M COM*0NENT gh "','gORTA CAUSE 548109 COMPON E NT I TO PR 7 pp l' I I I I I I I I i l i I i I l l l l l l 1 I l l I l ! l SUPPLEME NT AL REPORT 1%PlCT E D (14 MONTM CAv YEAR BUOM:55 ON q vu ,,., aane. enecno svowsv0N cA rt, ~l] NO l 1 1 A. T R AC T a -, w i m u, , . ,,-..-,,,<m .*, eu.....--.~ no On February 12, 1991, a discrepancy was discovered between the logic diagrams (LSKs) and elementary diagrams (ESKs) for the initiation signal of the main control room ventilation (HVC) charcoal filtration system.

The initiation signal configuration to actuate at reactor water level 1 instead of level 2 rendered the HVC charcoal filters unable to perform their design function to mitigate the effects of a main steam line break (MSLB) outside of the containment. Immediate action was taken to start one train of HVC filtration to maintain the plant within the control room habitability analysis. GSU's investigation of this event concludes the following:

1) Additional discrepancies were identified; however, the safety impact of this event was limited to the HVC filter initiation signal.

2) Following modification of the HVC initiation signal to actuate on level 2, operation of RBS without ~ compensatory measures has been analyzed and found to be acceptable.

3) The analysis of plant operating history indicates that this condition did not pose a significant hazard to control room personnel.

Therefore, this event did not adversely affect plant safety or the health and safety of the public.

NRC Ferni 366 to N

6 4 4. n --

g,y r u. % .u ovoma no-i.= - , _ , , , , , , , , , , ,

. wa .m, Q"a&',@e%?@/is, .#*?' A'.".'."4 UCENSEE EVENT REPORT (LER)

- TEXT CONTINUATION- Y o'."#1"C&iM'.W17^.'M'"i

!?f"n'i v Mf!

10 M,f.M.ci.a*4m""at.?W!E..'d...?%W o w mio ..aw.xr.0?!?

F.Cluty e aug n, - 00Cuitasvu.IR W tem W u..a $ . p. g. t,i .

.... esiuse.% .w.

" '" ~ " ' ' '

RIVPR DWD STATION 015 l 0 l o j o l4 l 518 9hl -

l0l1.l ol0 ol 2 0F 115 raxt n . w . w s wu nn

.. REPORTED CONDITION On February 12, 1991, during the review of Generic Letter 89-010, a discrepancy was discovered between the logie diagrams (LSKs) and elementary diagrams (ESKs) for the initiation signal of the main ,

control room ventilation (HVC) charcoal filtration system. Tha LSKs

- identified reactor water level 2 (-43 inches) as the initiation signal, while the'ESKs showed the initiation to occur at level 1 (-143-inches). The LSKs were used for the main control room habitability analyses, while the.ESKs reflected the as-built condition of the l

plant.

The initiation signal configuration of the main control room charcoal filtration units on level 1 instead of level 2 placed the charcoal

-filters in a-state-in which they were unable to perform their design 4 function to mitigate the offects of a main steam line break (MSLB) -- i outside of the containment. In-addition, an analysis of River Ber.d'.s operating history hasishown that for a postulated main steamline break outsic'e containment using the technical specification li 't for iodine spiking, the 30 rem regulatory limit for control' room dose would.be exceeded. :Therefore, this event is reportable as operation prohibited by Technical Specification 3.7.2 (10CFR50. 73 (a) (2) (3 ) (b) ) , and as a

. condition that--alone could have prevented the fulfillment of a_ safety function requiredxto mitigate the consequences of an accident (10CFR50.73 (a) (2) (v) ) .

SUMMARY

Following discovery of the above reported condition, immediate--action was taken by Operations to start one train of HVC filtration to maintain the_ plant within the existing habitability analysis. In addition, a probabilistic safety assessment (PSA) was developed to demonstrate that operation of one HVC filter tr ain was -acceptable. -

Modification. Request (MR) . 91-0011 was prepared to change the HVC initiation;from level 1 to 1cvel 2 and was placed on the forced outage list. This modification was implemented during the forced outage that began on February 27, 1991. Alternative initiators for the HVC fi-1ters were reviewed to verify the adequacy of reactor water level 2-The-investigation of this condition identified problems-with the existing control room dose calculation for a MSLB and with design documentation for loss of coolant accident (LOCA) reactor water lavel

. artuations. The dose calculation demonstrated that reliance on level l 1 initiation was not adequate to ensure that control room doses g remained within regulatory limits (30 rem). However, rio other control j room habitability analyses were impacted by this condition. In .

l addition, no other LOCA-related analyses were impacted by the level

' I/ level 2' discrepancy.

W ,. m '

l

ta gg) U S NUCtt4m Kf GVtaf 04Y Coaswisseoh i o... . w,,

UCENSEE EVENT REPORT (LER) '

'8,'s"*'!,',0'd80 E',*0"'nti's,'%c 'l'?,' .*0A'."i!

. TEXT CONTINUATlON _ 1 ""li,O'."d'o'."??CiMf',M!,'/,*, */jd"i 0!Nf2 l U.N.'/a".'o!!"."A'Ma'm*j;?We'2 09,1i 08 MANaGlutNT AND tv0G4 f.usae NCTON C 70603 9 ACskt FY NAh,4 lie DOCKl'pavMata(2' 45m NUNittB 16e 9A04 tt

"'a -" t!!:' .  ;'jg RIVER DEND STATION

.- ,. - - - we ,

015101010l4l5l8 911 -

0l0 l1 -

0l0 0J3 OF 1 l5 e rixt w .mu u nn The design documentation review revealed that the ESKs and LSKs in question have alwcjs differed in their HVC filter initiation levels.

This discrepancy was identified in 1984 during the development-of River Bend's-Technical Specifications. However, the impact of the discrepancy was not analyzed. -

Other components were discovered to have level 1/ level 2 discrepancies between thLir ESKs and LSKs. However, a review (' the , functions of '

these components determined that no satety impact was created by the discrepancies.

The root causes of this event include the following:

1) Disagreement between GSU and Stone i Webster Engineering 1 Corporation (SWEC) concerning which documents are appropriate for design-control.

2) Misunderstanding of the definition of LOCA signals (level 1 and level 2).

3) Failure to incorporate LSK level 2 HVC filter laitiation signal into ESKs.

4) Inadequate review of HVC filter ESK/LSK discrepancy when originally identified in 1984.

To resolve this condition, the following corrective actions have been taken or are.plannea

1) MR 91-0011 has been implemented to change the~ initiation signal from level 1 to level 2.

2) The main control room dcme calculation methodology has been revised.

3) ' The control room habitability during previous 5 years of operatior.

has been analyzed.

4) Appropriate ESKs will be revised _per MR 91-0011,

5) Technical Specification Table 3.3.2-1 will be revised to be consistent with MR 91-0011. This is provided by License Amendment-

' Request - (IAR) 91-05 (Reference 19).

6)' Appropriate engineering and design coordination reports (E&DCRol will be incorporated into the LSKs.

7) Future RBS shielding and dose calculations performed by SWEC will be reviewed by GSU. This cross-checking will ensure that SWEC assumptions are realistic based on plant configuration.

cc .- m ,....

1 8#.C . A:366 'a W 8 00UCLt A2 GletukafMT CoMuitsloh 1 ea.... 2 n 1 LICENSEE EVENT REPORT (LER) f.8.,'nW,*o'.fWen"o"'El'i, %Cf".f# '

,%".'n'!

TEXT CONTINUATION iT4*.!>'. "d'u'If3fM'.'A !!/CO,' '21"s' "Wf!

)

O, W

• ' l o war:,'a".'Ji"."a'aC ~ !ClG,GTON DC 70603' 40sutNY ANO svoog T,nasmas

.:Cm t, u mi m oocaat nu=esa ui _ o i,v . .., .... .

" b' W." 3*.D RIVFR-BfND STATIOfJ -

0 l510 { 0 l0 l 4l5 l 8 9l1 -

010l1 -

0l 0 0l 4 0F 1p me w . , e ,< we s- m v mi GSU's safety assessment of the use of the level 1 initiation signal revealed the following:

1) When calculations were based on the Technical Specification limit for lodine spiking, calculated doses exceeded the regulatory limit-of 30 rem.

2) However, calculations based on measured data at RBS using methodology guidance in Standard Review Plan (SRP) 15.6.4. yielded calculated doses well below the regulatory limit of 30 rem.

Therefore, GSU concludes that this condition did not represent a

-significan't hazard to control room personnel throughout its operating history.

OVERVIEW

The remainder of this report addresses the investigation, corrective actions and safety assessmerm for this event. The Investigation section of this report will focus initially on the reported condition described above and the actions taken to address it. However, several additional relevant issues were identified in the investigation of this erant. These issues will also be addressed. In general, the investigation of this event-followed two paths. One path was review of the design analyses for the main control room dose calculations.

The other path was review of the design documentation. The root causes of. this Levent invo.lved both of these areas.

INVESTIGATION' Fol~ lowing the initial identification of the discrepancy between the design-analysis and as-built configuration of the initiation signal,_

Stone and Webster Engineering Corporation (SWEC) was contacted regarding the main control room dose calculation (neference 1). SWEC stated-that HVC filters were assumed to initiate on level 2,

'immediately fol)owing the MSLB. Their bases for the level 2 initiation were'LSK-22-9.1 and LSK-27-30 (Reference 2). These identify'the level transmitters responsible for the level 2 signals,

'and show th initiation logics. GSU verified SWECs interpretation of the LSKs.

Review of the as-built circuit drawings (Reference 3 and 4) revealed

-that the HVC filters receive <a level 1 initiation signal. Following confirmation of level 1 ao the as-built initiation signal, a condition report was prepared on February 12, 1991 to document that the HVC charcoal filters would not have performed their intended function to L mitigate the offects of a MSLB outside containment and to initiate j- formal investigation of this issue.

! .C.-..u,.

j;.;,,..u '

. . m a m . -, o . u, . .

.....m.

LirENSEE FVENT R; PORT (LER) f'",',o'di,R,%"'!2",,'*Jf

,, *?,' ,* 'u'aj TEXT CONTINUATION n,%"',%'.7,%'M@rN'.W.'s". ','h'J / ' .' "'j@

iA o.P..'.','iT,JJf"".M. L*~,'):P.,i',~?%%. i"#, :idi

.~.a. ~o . .. - ove x h .v . ... '

oou n u .. o, a....,,, ,,,,,,.

,,.. n..., .. . w.

RivrR BDD STAT'ON o p [o [o lo l4 [5 [ 8 9 l1 -

0 l 0l1 -

DJO 0[5 051 l 5 mm , . .. ..e .w.

The following immediate actions were taken in recponce to this eventt

1) operations initiated one train of the main contral room filters.

This action ensured that the plant was operating within the bounds of the dose calculation (Reference 1) by eliminatir:9 the time delay to reach the actuation setpoint in the event of a MSLH outside ccntainment.

2) Modification Request (MR) 91 Oui 2 was initiated and approved.

This MR changes the !!ve ilicer inliiation signal from level 1 to level 2. This MR was added to the forced outage list. Note that River Bend Station entered a forced outage on February 27, 1991.

The MR has been implemented.

3) Engineering Analysis provided a probabilistic sh*cty analysis (PSA) of a MSLB outsido containment with a concurlsat loss of the operating IIVC filter train (hence, the standby IIVC filter would not automatically initiate to protect control room personnel).

The PSA demonstrated that for 720 hours0.00833 days <br />0.2 hours <br />0.00119 weeks <br />2.7396e-4 months <br /> of filter run time, the probability c' +hese concurrent events in 1.60-10 or four orders of magnitude b .ow the NRC's guideline of 1.0E-6 for large releases (Re' conce 5). Therefore, there is no safety significance of continued operation with one filter train in nezvice.

Following the immediate actions to address aafety end operability concerns, the investigation continued to focus on the twofold path of design analyscs and design documentation.

Denign Analyses Analysis of Level 1 Initiation Signal The main control room habitability doses due to MSLB outside containment were calculated in PR(c)-420-1 (Reference 1). This calculation was intended to rueet the requirements of SRPs 6.4 and 15.6.4 (References 6 and 7) . One assumption, not explicitly stated in y this calculatien, was provided by LSK logic diagrams LSK-27-30A, LSK-22-9.1D, and LSK-22-9.1E (Reference 2). This assumption was that the main control room charcoal intake filter would actuate on a LOCA signal produced by a low water level 2. This initiation signal la produced by the postulated accident, and tha intake filter would, therefore, be in operation prior to intake of the post-accident releasas, llowever , the as-built configuration was such that the actua usn would occur on level .. Since the postulated MSLB may noc produce a Jow water level 1 indication, there would be no resultant LOCA Jr.itiation signal to actuate the main control room intake filter.

The high radiation signal produced by released noble gases drawn into l

m..,...

1 a

• v.eocam..-o c - h ,,,,oa, ... .. ,,.. m.

i......,

LICENSEE EVENT REPORT (LER) 2.'E*,',o'#R U*o*'Td'j',,'%'f",?' ,*o'!.'.y TEXT CONTINUATION ETi.o'.'F

. ' ' ." fan @'.'d!'e'.*, f,'MM".' "'Mf!

T,'t,','aW"."at?,%*.,'uminJM".? m.H 08 WahAGewthf Ah0 pvDGif mapa NGioh DC 70603 heittf. * *ws lie l Mat t Numpla (3' Lla tvusta i.i ,agg a3,

, RIVER BLUD STATT # "" "MD 4'J.?.E v w -, e , v.ea.< e c , sma e o 151o l0 l0 l4 l 5j0 9l 1 -l 0 l0 l1 ~

0 l0 0l6 0F 115 om i

the main control room Jccal intake radiation monitor would then actuate the intake filter. This process could take as long as 6ti i seconds following the arrival of the post-accident radioiodine release at the main control room local intake radiation monitor. As a result, the released iodines would be drawn into the main control room before

' intake filtration could become available. This raised the potential for a 30-day thyroid dose to main control room operators above the 30-rem regulatory limit imposed by General Design Criteria (GDC)-19 and the guidance in SRP 6.4.

Analysia of Dose Calculu lon Assumptions An analysis of the MSLB outside of containment was performed using calculation PR(c)-4?0-1 (Reference 1) to validate all assumptions.

During this proceps, it was discovered that the value for the wind dispersive factor (X/0) was non-conservative for this event. The ,

calculation assumed the maximum X/Q for River Bend meteorological conditions. This X/Q was associated with a windspeed of 0.54 m/sec.

At this windspeed, the radioactive release does not reach the lfVC intake prior to the start-up of the IIVC filters.

For the MSLD, the !!VC filters start-up in 60 seconds (30 seconds for level 2 initiation, plus 30 seconds for dampers to reposition). It is also approximately 62.5 9 eers from the steam tunnel blow-out panels to the IIVC intake. Therol,,,re, if windspeed exceeds a critical value of 3.04 m/sec, the release will reach the intahe prior to filter ini' lion. This could increase control room thyroid dose.

The : e, the critical windspeed and its associated X/O should have bee, 3 in-PR(c)-420-1 instead of the maximum X/O and its associated wind.,t.ed. Therefore, the original analysis which was correctly based on level 2, used a non-conservative assumption in establishing the windspeed.

SWEC has performed a preliminary re analysis of the main control room radiological consequences of a MSLB outside containment (Reference 8).

This re-analysis is predicated on a low water level 2 indication occurring 30 secondo postaccident, with a resultant LOCA isolation signal that actuates the main control room intake /rocirculation filtration-system and causes full filtration to be attained within an additional 30 seconds. That is, the HVC filters start-up 60 seconds (total) following the MSLB. The calculated dose consequences are based on design. iodine activity mass concentrations scaled up to the preaccident iodine spike Technical Specification limit of 4 uC1/gm I-131 dose-equivalent; a liquid release flashing fraction of 42.9 percent nnd a steam release lodine carryover factor of 2 percent; X/O values ccrresponding to the main steam tunnel blowout panel release point, which are based on a critical (bounding) windepeed resulting in a plume'trrival time at the local intake occurring just prior to the

~

.ae on s.ss n,

u,w,, u. m. m .1Usfo.,c -

'a' *

, om. .w LICENSEE EVENT REPORT (LER) fj,'s't'!,0,c,*,fW,y,to*'!;j;*$,1,' ,g*fA' ,TNJ'!

. TEXT CONTINUATION M*,',",,'o',';16,'"nUlJ3',11!,'e'!','!M",'

UN..O'Jf"."atW,c",'!R*M*JJ*# 'Mi MR Of MAhA0lME NT AND $0D011.W A&MINGT0h DC 70h03 P ACis,t v HAMS tu DOCA ti NUM.8. Qi 48 A NUM.lR ISI P A00 (3l

'ia " W. $*.T:

RIVER IED STATIW o p jo jo jo]4l5 lC 9 l1 -

0l0l1 .-

0 l0 0 l7 or 1 l5 rw w , . e , M wv , m on attainment of full filtration; a profiltration ventilation flow rate of 2000 cfm; and a postfiltration ventilation scheme which includes filtered recirculation.

On these bases, the calculated main control room thyroid dose due to the postulated MSLB outside containment was determined to be 17.2 rom (which is below the GDC-19 and SRP 6.4 regulatory lintit of 30 rem) .

Therefore, with the filtration initiation signal corrected to level 2 via installation of MR 91-0011, continuous operation of one train of La flitration system is no longer necessary, based on the above

..nalysis. This analysis will be formali::cd by SWEC in Calculation G13.18.9.5*26, Rev. 0 (Reference 9), which will supersedo Calculation PR(c)-420-1.

Note that the USAR, Chapter 15 evaluates four standard accidents which require analysis of main control room habitability. The MSLB outside containment was discussed above. The remaining three accidents are (1) LOCA, (2) control rod drop accident, and (3) the fuel handling accident. These accident analyses were reviewed to determine the impact of the level 1/ level 2 discrepancy on them.

1) LOCA:

It is assumed that control room filtration occurs at time t = 0 for a LOCA. In reality, it takes about 3 seconds for a high j drywell pressure LOCA signal to occur and a maximum of 30 seconds for the control. room filter to actuate mechanically, for a total actuation time of 33 seconds. Ilowever, it can be reasonably assumed that LOCA releases will not enter the environment in less than-33 seconds. Also, LOCA releases are protracted, not instantaneous, llence, LOCA releases will arr've at the control room local intake after the intake 111ter is actuated and continue to enter the intake for the duration of the accident ( 30 days).

Therefore, the level 1/ level 2 discrepancy has no impact on the LOCA analysis.

2)_ Control Rod Drop Accident LOCA initiation of the IIVC filters is not assumed. This dose calculation assumes that the filters initiate due to high noble gas activity-in the intake. Initiation occurs 66 seconds after noble gases enter the intake. This time delay is based on 36 seconds for radiation monitor response time plus 30 seconds to startup the IIVC filters and reposition dampers in the ductwork.

Therefore, this analysis is not impacted by the level 1/ level 2 discrepancy, uc . au u.,

,. _ _ , . - 4 - - - _ -- - . . , _ _ _ _ . _ .

8eRC ,(khM )e64 USe M A2001VtA10ATC04apiteMe C 491 .

e 9 291Xf 8 af30S7 a

tlCENSEE EVENT REPORT (LER) ($,'oi2,%',y" M2 Min"o '#f,"$i,'ti"o?,'

i ,TM"4 TEXT CONTINUATION 17a',%,',",!6#nd.$'OP',W.,.'N',1,P 'u"'a=o OMTN

. ainutaion coMuiu o= swi oron oc nous. vo M 08 lNT AN DVt i t , vt M Of 03

. .e. . . ..M nI m.,... ,,,.,,,,,,,s- ,..,,,,

,,,. .. w. o . 6 .....n RIVfR Bl?O STATICT1 ***"' **""

ol6jojolol4l5l8 9l1 -

0 l 0l 3 -

010 0lA M 1 l5 >

tart o nia. . ,* eac e,. am4.nm

3) Fuel llandling Accident:

Same assumption and conclusion an the control rod drop accident.

As requested by GSU, SWEC examined additional accident analyses to determine which LOCA and isolation systems were used to initiate post-accident systems. For all accidents occurring inside the drywell the LOCA signal generated by high drywell pressure rather than reactor water level 2 or level 1 was used. For pipe breaks outside containment, local temperature monitoring or high system flow rates are relied on to initiate break isolation. The only accident analysis reviewed that relied on reactor water level to initiate a post-accident system operation was the analysis of control room dose ollowing a LOCA occurring either inside or outside (i.e., main steam 1 ne break in the main steam tunnel) containment.

Design Documentation Design Engineering researched the LSKs and ESKs from their current revisions back to their origintz1 vers!,ons, to determine if either set of documents had been erroneously modified (Reference 10) . A review of the IS-217 drawing history report was completed to determine if the present reactor water level initiation (level 1) of the IIVC charcoal filter trains has ever been changed from a level 2 signal. The re"iew encompassed the applicable Logic Diagram, LSK-22-9E, and Elementary Diagram, ESK-71!VC01 Sh. 1. The review concluded that the present initiation logic has never been changed. The level 1 reactor water level signal for the IIVC charcoal filters was included in the original design for RBS.

In addition, the same ESK/LSK discrepancy for the llVC filters was identified in 1984, during the preparation of River Bend's Technical Specifications (Reference 11). At that time, GSU resp *.1ded that level I was the correct filter initiation signal. Ilowe ve r , no justification was provided for this determination.

Design Engineering performed an investigation of the ESKs involved with the llVC filter initiation to determine if other plant systems might be affected (Reference 12). The ESKs (References 3 and 4) were compared to LSKs (Reference 2) for components receiving level 1 initiation signals. The following discrepancies were identified

1) Per LSK-22-14.2, unit cooler IliVR*UC1h dischargo valve ll!VN*MOV22A is required to close on level 2. Ilowever, l ESK-7ISC01 indicates that the valve closes on level 1.

. Similarly, IlfVN*MOV22B on lifVR*UC1B also closes on level 1, l while the LSK indicates level 2.

l l

CEC 8eem 3e6A 16491

g,... 2 v. me.a moummo . ,,,,,,,,,,,,,,,,,

. . .m t . .-

LICENSEE EVENT REPORT (LER)  !?M*4'?,41R,n"J' sty',,

TEXT CONTINUATION  !?"",f,",'A7,5'!n*iM',Mlls",'%cf"*!,'

• 'Z,",' $7f!

, T,"4 o was onut= ae,oevoost i

n O,f",'.'.,a'a"aE!,"at'J!,,a'.M.a"E.u.n.ovou t =oc ans 8 AC Lat y toawl ill Ducal 1 Nvuota t,' . AOS (F 489 wu.selm it!

• • • a " C'M." 4*A".D RIVIE DIND STATION 0F o l5 l0 l0 l014l5 l 8 911 -

0 h 11 -

010 Ol9 1 l_5.

ftKt (# suure aisse a secuesel use eswnuvW h4C #erm JUM es (01

2) Reactor plant campling system valves ISSR*SOV130, SOV131, SOV133, and SOV134 are shown te close on level 2 per the LSK-21-2A. Ilowever, ESK-7ISCO3 indicates that the initiation signal for these valves is level 1. This agrees with Technical Specification Table 3.6.4-1 which lists these primary containment isolation valves as Group 10, or level 1 per Technical Specification Table 3.3.2-1.

3) The Division I and II hydrogen mixing system inlet and outlet valves ICPM*MOVI A (B) , MOV3A (B) , and MOV4A(B) are shown in LSK-27-24 as level 2 1solations. Ilowever, por ESK-7ISC01 and ESK-7ISC03, those valves isolate on level 1. Technical Specification Table 3.6.4-1 lists these as Group 10 (level 1) isolation valves.

The impact of the above discrepancies is discussed below.

Design Engineering has reviewed the impact of level 1 isolations on IllVN*MOV22A and B (Reference 13). These valves isolate chilled water from the containment unit coolers during a LOCA. On level 1, the standby service water supply valves to the unit coolers open. Closure of MOV22A and B on either level 1 or level 2 ensures that service water flow in the containment chilled water system is limited to the unit coolers only. Therefore, even though the LSK indicates level 2 isolation for MOV22A and B, level 1 is an acceptable alternative.

A review of the !!VN chilled water system and the standby and normal service water systems to the containment /drywell unit coolers reveals that the Elementary Diagram, ESK-7ISC01, depicts the proper LOCA isolation signal initiation of level 1 reactor water level. Below is a listing of the chilled water and service water valves reviewed.

Service Water Valves _ llVN Chilled Water _

SWP*MOV 4A 6 48  !!VN

• MOV 10 2 SWi^MOV SA & SB  !!VN *MOV 127 SWP*Mcv 302A & 502B IIVN *MOV 12 8 SWP*MOV 503A & 503B IIVN*MOV 129 ffVN*MOV 130

!!VN*MOV 22A IIVN*MOV 22B flVN*MOV SA & SB C$R esem 3 eta 16491

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

Gj o

,,,A u . m.n umm.T c- .D.

. , ,,,u , . m LICENSEE EVENT REPORT (LERl ,,,eyWdia o*'ng,',,'%'f"l# ,T,l',,'Jj TEXT CONTINUATION M,'s",",J',llF'fl0in"F.M'J"*,'l#

F"* ,'

• L",' 0'RifA O M ANAGlut NT ANDASHl880f SUDQif,W 00s DC Mit413 f,i *,, .M"!A,'a"

1. ACoklif 9>AMI 41. DUCElf 88UM.t h 431 k g n ,gu,se$13 ,$1 9 AGI (31 "A- a tt.2 I;;n RIvnn EmD STATIM o p,]ololol4l5l8 9l1 _

0 l 0l 1 0l0 1l0 or 1 l5 .

nn ,, . ~.c , ., ,..,

SWEC has reviewed documentation related to the group 10 isolation valves in the reactor plant sampling and hydrogen mixing systems.

Engineering and Design Coordination Reports (E&DCRs) were submitted to revise the discrepancies in the LSKs. E&DCR P-41,086 (Reference 14) requested that LSK-27-30A for the reactor plant sampling SOVs be changed to indicate a level 1 isolation.

LSKs-21-2A and 27-30A for hydrogen mixing MOVs were also to be changed to reflect level 1 isolation per E&DCR P-41,170 (Reference 18).

Note that containment isolation of the sample valves (SSR) and drywell isolation of the hydrogen mixing system valves (CPM) on a low level 1 LOCA signal is acceptable because these valves are also isolated by redundant safety-related high drywell pressure LOCA signals.

Logic Diagrama LSK-21-2A, Sh. 1; LSK-27-24A, Sh. 1; LSK-27-30A, Sh. 1, and LSK-27-30B, Sh. I were revised by E&DCRs P-41,086 and P-41,170 to show valve isolation on a "LOCA Initiation Signal" (loa level 1) instead of on a "LOCA Isolation Signal" (low level 2). These LSK changes agreed with the as-built elementaries, the USAR, and the Technical Specifications. Also, no design basis calculations are based upon isolation of the valves due to level 1 or 2. The cross- referencing of the LSK sheet to incorporate the LOCA initiation signal was performed correctly. However, the E&DCR changes for LSK-21-2A, Sh. 1, and LSK-27-24A, Sh. 1, did not include the editorial changes to show the correct LSK signal cross-references within the signal flow diagrams, i.e., "1" should have been changed to "2" and " Isolation" should have been changed to " Initiation".

Therefore, the discrepancy between the ESKs and LSKs for these group 10 isolation valves is strictly a documentation problem, and does not affect plant operations.

Based on GSU's review of logic diagrams and associated elementaries regarding LOCA isolation and initiation signals, the problems identified are isolated to these ESK/LSK discrepancies.

Root Cause Analysis Four dir. tinct root causes of this event havo been identified, as follows:

1) The use of LSKs as design documents without follow-up verification using ESKs and other documents. This reficcts a difference of opinion between GSU and SWEC as to which documents, LSKs or ESKs, are the appropriate source documents for design work. SWEC assumes that.LSKs are source documents which form the bases for the ESKs. Hence, use of LSKs for design basis calculations would

. be approprinto. This was truo durinn River Bond's conntruction edIC Form Se6A M9) e

WaC DonM 3e6 A U S tadt1beir kleuk&TDWT CoteMit6ap=

DMWUWM i o ,2 . . -,

0,yE'*J%"'ni,"l',,'"Jf'O,' 5%'e'!

LICENSEE EVENT REPORT (LERI TEXT CONTINUATlON 'i,';"e',".,S','.'iE*J0',i!!,'/"f,'!,'!

M",' To' C',' $!',"f!

0,?",",'A"u?"."2D, o Ma=*c Minta o evoca,",'.0.as iwoto= oc muW?in%" :? 't iACitif,h&M4 Of DUCR 41 NOMSt h til Llh NVM6t h les P A06 (3)

RIV m BLIO STATICt1

" N*' ' US1I o l6 l 0 l o l c l 4l 5 l8 911 -

O p l1 -

0j0 1 l .1 0F 1 l5

,u, ea . < a m r., m v on and start-up. Now however, GSU Design Engineering considers ESKn to be the de finitive documents for configuration control, since thece repreacnt the plant as-beilt configuration. ESKc are revised and controlled moro rig >rously than LSKa. Therefore, une of only LSKs for design analyres is not adequate to provide an accurate representation of plant configuration.

2) Misunderstanding of the definition of LOCA signaln (level 1 and level 2)

LOCA reactor water level nignals are:

Level 2 -

Small break LOCA IIPCS/RCIC Initiation Containment Isolation Level 1 -

Large break LOCA LPCS/LPCI/ ADS initiation MSIV isolation (All levels 2 actuations occur as water level decreases on a large break LOCA)

3) Failure to incorporate the LSK level 2 !!VC filter initiation signal into the ESKa. Although LSK logic diagrams LSK-27-30A, LSK-22-9.1D, and LSK-22-9.1E show that it in the low water level 2 indication which produces the LOCA signal that actuaten the main control room charcoal intake filter; due to human error, this was not reflected in the derived ESK diagrams which provide the basis for the as-built plant design. According to these derived ESK diagrams, the main control room charcoal intake filter is actuated by a LOCA signal which is produced by a low water level 1 signal.

4) Inadequate review of the llVC filter ESK/LSK discrepancy when originally identified in 1984. Proper evaluation of the ESK/LSK discrepancy during Technical Specification development could have eliminated thin condition.

In addition, no evidence of follow-up has been found for the discrepancy reported in 1984. This was prior to initial plant start-up and the condition report program. Lack of an adequate corrective action program may have contributed to the failure to effectively disposition the discrepancy.

1 1

1 kmc9wm3e6AiH9

,,,,,,,,,,. u. m. vu.,o., c-, .oh ,,,,,,,, ,,,,,, ,,,,,,,,

. ...:8 . . m, LICENSEE EVENT REPORT (LER) "'l"'i',%',,18pdj'J'5t*ji,'o.

, ,, c?",*t,' ,TM"4 TEXT CONTINUATlON cut"',',*,',$.i!TM ',y,I,P',' !!,'/*f,'!3J 7,' Mfj '

0 W?",'t,'3"Jon**#'i,*,a'.

O. ANAGEMt NT ANO SVDGE1 W ASHim,010h. DC 706

' b' ' T " 0 8 DOCElf NUMSI A (2, g g gg g g, pg ,

,,,, sto . ,% ....

RIVER BD!D STATION 0l5l0l0l0l4l5l0 9l1 -

0l0l1 -

0l0 1l2 0F 1 l5 ruta . w.w wes ,m4wim

-CORRECTIVE ACTION The following are 630's corrective actions to address this event:

1) Modification Request (MR) 91-0011. Tnis modification changed the IIVC charcoal filtration initiation signal from level 1 to level 2 and was installed during the forced outage which began on February 27, 1991.

2) Main Control Room Dose Calculation:

The main control room habitability doses due to a MSLB outside containment have been recalculated (Reference 9). The following assumptions were used:

a)  !!VC charcoal filters initiate on reactor water level 2 at 30 seconds after MSLB.

b) Following the level 2 initiation signal, IIVC dampers reposition and filters start-up in 30 seconds (MSLB + 60 seconds).

c) Iodine spiking per SRP 15.0.4 (Reference 17),

d) Control room thyroid dose limit is 30 Rem per SRP C.4 (Reference 16),

c) Radioactive release enters control room intake at a continuous rate of 2000 scfm.

3) An analysis of the previous 5 years of plant operations with the level 1 initiation of IIVC charcoal filters was performed to ensure that main control room habitability was maintained (Reference 15).

4) Revision of ESKs and other eleiaentary diagrams as necessary to make the drawings consistent with MR 91-0011,

5) Technical Specification Table 3.3.2-1 will be revised to be consistent with MR 91-0011. This is provided by License Amendment Request (LAR) 91-05 (Reference 16).

6) E&DCRs P-41,086 and P-41,170 will be fully incorporated, as follows:

LSK-21-2A, Sh. I and LSK-27-24A, Sh. 1 should be editorially revised within the signal flow diagrams to show the correct LSK signal cross-reference from logic diagram LSK-27-308, i.e., "1" should read "2" and " Isolation" should read " Initiation".

Nmcfom 3e6A1649

g, ... A u. m.m.,or ' M.,

t ahmte arJ0W v UCENSEE EVENT MEPORT (LER) fSst,'#,d"MdM'U?,*h?".#"l# ,*D'"-d"4 TEXT CONTINUATlON R"",' *,L'.';M!,0i M'.M!,'e'O,'!dM".' M"f's

• 'b.','a"AR".*.ta*N L"3P4:2' t* fili D. MANAQt St h? AND DVDG4 7. h A6HINQ10N D4 M1bO3

. AC66 tit hA.St (13 Dr , ,suMg4 R u t 4th OptaMeln iti ,Aet (3' "n**

• nW."  ;*;*n:

RIVER DDO STATION o ls10lolol4]5 p 9 l1 -

0l0l1 -

0 l0 1l3 0F 1 l5 text o . .m.wwe re ame.mn

7) Puture RBS shielding and dose calculations performed by SWEC will be reviewed by GSO. This cross-checking will ensure that SWEC ,

assumptions are realistic based on plant configuration.

SAFETlY ASSESSMENT Main Control Room Dose Calculations Corrective Action 2 above involves developing a realistic methodology for performing a main control room habitability analysis following a MSLB outside containment. This calculation involves the use of good engineering judgement, actual plant operating parameters, and the regulatory limits provided in SRP 6.4 and 15.6.4. As previously diLcussed, this analysis demonstrates that continuous operation of the HVC filters is no longer necessary. This conclusion is supported by the agreement between GSU and the NRC (Reference 17).

Since River Bend Station's initial criticality (10/31/85), HVC charcoal filter initiation has been provided by the following signals:

1) High drywell pressure (1.68 psig)

2) I.ow reactor water level (level 1 = -143 inches)

3) . High intake noble gac activity If an MSLD outside-containment had occurred during this period, there is a possibility that a level I signal would not have been generated (Reference 18). In that event, then a high intake radioactivity signal.would have been generated due to noble gases entering the HVC ,

intake. _ This initiation signal would have occurred 66 seconds after the radionuclide release entered the intake (36 seconds for radiation monitor response plus 30 teconds for HVC damper repositioning and ,

start-up). Per SRP 15.6.4, an iodine spike 20-times equilibrium I-131 is assumed prior to the MSLB. Two situations were addressed, iodine spiking equal to the Technical Specification limit and iodine when the plant had perforated cladding (measured data for 5/28/90). Under such circumstances, the following main control room thyroid dose results were calculated using the methods dcVeloped for Corrective Action 2:

Iodine Spiking Basis Calculated Dose Regulatory Limit Tech Spec Limit 63 Rem 30 Rom Measured Data for 8.1 Rem 30 Rom 5/28/90 (x 20)

Therefore, operation with level 1 initiation of the HVC charcoal filters exceeded the regulatory limit when design basis data was used for' iodine spiking. However, for the actual operation of River Bend, i this condition did not represent a significant hazard to control room personnel from 10/1/85 until the disecvery of the ESK/LSK discrepancy.

lCC ,eem 3e6A 4895

,,.r -

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

c omw sesa W 6 seMelp LeiptAlpay coenneitssob '

APPMCr,ID OMS WO 3t60416e

. t *Piet t 4/3042 LICENSEE EVENT HEPORT (LER) '4,'#^'.7,olf30.n"o"'s?,',n,' .!f"?.' Ca'J!

TEXT CONTINUATlON 2n' *4f,'45'2ni Of"f' M','c'#','!2 '".' OMTf! '

uitA,'s'J?".".

U ,*,o".' %'E ?*i n E's 0?!?

08 MAhsAGIW5 hi ANO $VDGI1.W Al* TING 1DN Oc 70603

1. ACIL11V haast its poc a s t Nunegg a m gg g gyggg g gg, pagg ggi RIVII BmD STATIM "'" "dP I'Y.U ol6l0l0l014l$1c 9 11 010l1 -

0 10 11 4 0' 1 l5 mt e . e s. emes maavim As previously discussed, no other !!VC charcoal filter dose chiculations or LOCA related analyses were impacted by the described condition.

In addition, the discrepancies identified between the ESKs and LSKs do not affect the operation of any components other than the !!VC filters.

MOVs and SOVs cited previously function as required on reactor water level 1. Therefore, the design documentation discrepancies found during this investigation do not have a significant impact on plant safety.

Based on the discussion above, GSU concludes the following:

1) Continuous operation of one ifVC filter train is no longer necessary base.d on revised analyses and the modification of the initiation signal from level 1 to level 2. l

2) The analysis of River Bend's operating history indicates that operation with the level 1 initiation signal did not comply with regulatory guidelines however, analyses using actual measured data from RBS indicate that this condition did not pose a significant hazard to control room personnel.

3) The level 1/ level 2 discrepancy did not affect other charcoal filter dose calculations or analyses.

4) Design documentation discrepancies identified during this

investigation do not impact the operation of any components other

, than the IIVC filters. Discrepancies concerning MOVs and SOVs have I no operational impact since the identified valves function as I

required on level 1.

Therefore, this event did not adversely affect plant safety or the health and safety of the public.

l l

unc . m..,

5WAC PDRW306A US 12 klaut A1&JV COMust48DN i ,

ie 1 86 46 e

' '"a"t'!?,c,*,,180.M"o"'#?,'10i'l*1',,' '  !?"?.' .5 LICENSEE EVENT MPORT (LER) .

TEXT CONTINUATlON !U"A*4'.Wu'A fai 'f0'.% 1l'e'.". ','!i!! '"' Olf."f!  !

• ^

0??A,'8".'o!!".".'t'a",0.,

% is" int!J'# '"?,li Of MANAQtWE NT AND DVDGIT.W ASHINQ104. DC W3

~

t ACiLit T 4AME Ill Ducast Numete ul LIR WLahetIS (61 9A05 (38 7*** "ttih" A*J o l5 l0 l0 l01415 l 8 9 l1 -

0 1011 -

0 10 1 l5 0F 1 l5 Tart e. . e. - =.c r , m. v im kEFERENCES

1) SWEC Calculation PR (c)-4 20-1, " Radiological Consequences of Main Steam Line Failure Outside Containment.

2) Lmss: 21-2, 22-9.1, 22-14.3, 27-24, and 27-30

3) ESKs 61!VC03, 7HVC01, 71SC01, and 7ISC03

4) GE Dwgst 828E534AA, and 828E535AA

5) Memo EA-PM .11-0166

6) Standard Rtn iew Plan (SRP) 6.,, .ontrol Room Habitability System".

7) Standard P.oview Plan (SRP) 15.6.4, " Radiological Consequences of Main Ste'Em Line Failure Outside Containment (BWR)".

6) P.sno EA-PI-91-275-F 0; Calculation G13.18.9.5*24-0

10) Memo ED-91-156

11) Tech Spec Job Concern Resolution Form (JCR) No. 5713A-0260.

12) Memo ED-91-133

13) Memo ED-91-144

14) Engineering & Design Coordination Report (ED&CR) : P-41,086 and P-41,170.

15) Calculation G13.18.9.5*25-0

16) License Amendment Request (LAR) 91-05

17) GSU letter to NRC, RBG-34,552, dated 2/28/91

18) Memo EA-TM-91-0164 NIC Pee 3e64 4491

- -