05000261/LER-2003-002
| Docket Number | |
| Event date: | 06-05-2003 |
|---|---|
| Report date: | 07-31-2003 |
| 2612003002R00 - NRC Website | |
FACILITY NAME (1) PAGE (3) DOCKET NUMBER (2) LER NUMBER (6 05000261 H. B. Robinson Steam Electric Plant, Unit No. 2 I . DESCRI PIT CN (F EVENT At approxi nit el y 19:49 hours on June 5, 2003, during a pl armed source check of cont ai invent radi at i on noni tor R 11 [ I11 I K 14] with a routine cont ai Intent pressure relief i n progress, the cont ai Intent pressure relief i sol at i on val ves, V12-10 and V12-11 [ VA VL: I � , would not close aut onati cal 1 y. The val ves were closed by use of the control smi tch [JMI-E] i n the control room t o stop the pressure relief of the cont ai mutt at the ti ne of the source check aut onat ic isolation fai 1 ure. The source check should have caused the val ves to close aut onat i cal 1 y by the initiation of a cont ai anent vent i 1 at i on i s ol at i on s i gnal . The penet rat i on was i s ol at ed at 20: 47 hours5.439815e-4 days <br />0.0131 hours <br />7.771164e-5 weeks <br />1.78835e-5 months <br /> by the use of a closed and de- act i vat ed aut onatic is of at i on val ve ; specifically, valve V12-10 was a closed and de- act i vat ed, in accordance with Techni cal Speci fi cat i ons Ii ni t i ng andi t i ons for Clyrati on (Iia) 3. 6. 3, (Int ai anent I s of at i on Val ves, aqui red Action R 1.
The cause of the fai 1 ure was deterni ned be a fai 1 ure of the control smi tch that controls the val ves for the cont ai anent pressure relief penet ration. The control watch was repai red at approxi nat el y 14:13 hours on June 6, 2003, and the system was restored to operable status at that ti ne. A pl armed and mull toyed gaseous rel ease from the cont ai anent was i n progress at the ti ne of this event using the cont ai anent pressure relief system Pb rel ease li nits were exceeded. If pl ant condi ti ons had requi red i sol at i on of the penet ration, al arts and i ndi cat i ons i n the control room would have alerted the operat ors t o the condition and the applicable operating procedures di rect the operators to nanual ly isolate the penetration.
A root cause investigation was conpl et ed for this event . The results of that investigation were reviewed by the PI ant Nucl ear Safety Ommi t tee (MC) on Jul y 9, 2003, i n accordance with the gxlated Final Safety Analysis Itport ( 'ISAR) inapt er 17 requi renent for the PI 8C review of reportable events. That investigation concl uded that the smi t ch failure caused the event, but al so that the des i gn of t hi s cont ai Invent isolation circuitry had apparently been unintentionally nodi fi ed, in about 1980, such t hat the smi t ch fai 1 ure al one could have prevent ed the aut onati c isolation of this penetration.
The (Int ai mnent Pressure and Vacuum � ef Sys t em [VA VII is provided to control vari at i ons i n cont ai anent pressure with respect to at nos pheri c pressure. These vari at i ons are due to changes i n at nos pheri c pressure and 1 eakage i n cont ai anent AID!, Ertl:flu OCCA 11 .1/1/11I FACILITY NAME (1) PAGE (3) DOCKET NUMBER (2) LER NUMBER (6 05000261 H. B. Robinson Steam Electric Plant, Unit No. 2 fromthe Instilment Air and Penetration Pressurization System. The ant ai nment Pressure and Vacuum 11l i of System i ncl udes separate 6 inch 1 i nes penetrating the containment [Ni, each equi pped vii th tyro qui ck- closing, t i ght- s eat i ng, 125 psi ai r operated butterfly val ves, one inside and one outside cont ai nncnt. These val ves are designed t o fai 1 closed on loss of control signal or control ai r, and are closed during nornal pl ant operation, except as requi red for cont ai nncnt pressure control .
The butterfly val ves are protected by debris screens, located inside cont ai nnent and attached to the inboard pressure and vacuum rel i of val ves, which vii 11 ensure that ai rborne debris vii 11 not interfere vii th their ti ght closure. The pressure relief line discharges to the pl ant vent through a H gh-Efficiency Particulate Air (1-1PA) filter and charcoal filters [MI FL'I] . These filters are provided for renoval of i odi ne and particulate radioactivity fromthe vent ed ai r. gyrat i on of the pressure and vacuum rel i of 1 i nes i s =molly control 1 ed by the pl ant operator. A narrow range pressure transmitter [ IK PI'l] continuously indicates containnent pressure in the control room Separate hi gh and 1 ow pressure al arm [ I K PA] are actuated by this transmitter to al ert the operator to overpressure and vacuum conditions.
Vacuum rel i ef can be acconpl i shed vii thout regard to at nos pheri c conditions. In the event of pressure bui 1 dup, the operator vii 11 be gui ded by at nos pheri c conditions, and by the cont ai nnent part i cul ate and radi o gas noni tor [ Ili IL: RI] i n rel i evi ng the overpressure. Mnual operation of both these 1 i nes i s overridden by aut onati c cont ai nncnt vent i 1 at i on i s ol at i on and cont ai nncnt hi gh radi oact i vi ty si gnal s I JM.
ant ai nnent pressure relief isolation val ves V12-10 and V12-11 are designed with a connon control switch located on the Itact or- Turbi ne (Inge Board (RIEB). The V12-10 and V12-11 RPM control switch has "(Ten" and "nose" switch positions and a spring return to center nechani sm The val ve circuitry is al so designed vii th position i ndi cation avai 1 abl e in the control room on the Rita The val ve circuitry is designed such that the air-solenoid val ves [IA VA VII ISM for V12-10 and V12-11 are energized to open the val ves when the operator places the control svii t ch in the "(Ten" position. The control circuit for these val ves was i nadvert ent 1 y nodi fi ed, i n about 1980, such that the "goen" svii tch cont act was electrically paralleled vii th the aut onati c containment isolation circuit functions and the close function of the control s vii t ch. T Therefore, when the "(Ten" svii tch contact renai ned closed due to the control svii tch fai 1 ure, both ai r operated solenoid val ves renai ned cont i nual 1 y energi zed, keepi ng the val ves open. The AID!, Ertl:flu OCCA 11 .1/1/11I FACILITY NAME (1) PAGE (3) DOCKET NUMBER (2) LER NUMBER (6 05000261 H. B. Robinson Steam Electric Plant, Unit No. 2 U.S. NUCLEAR REGULATORY COMMISSION 00 2003 - 002 - presence of an automatic isolation signal was not able to electrically interrupt t he continually energi zed path to the solenoid valves.
A 1 etter to the NC dated Mrch 11, 1980, i n response to Three M1 e Isl and short term 1 essons 1 earned requi renents, stated, "The second category of switch i s that which uses a single set of contacts for both t rai ns. Nbrnal 1 y, this moul d be unacceptable as the switch would not be single fai 1 ure proof. Itmever, when the switch is wired is series with the Train `A: and Train `If containment isolation relay contacts, the logic system vdti ch control s the isolation val ve is single failure proof and, t heref ore acceptable. " A stated previ ous 1 y, a nodi fi cation i n the 1980 tine-franc inadvertent1y ci rcunvent ed this design requi renent .
The autonati c isolation functi on is tested in accordance with Technical Speci fi cat i ons ('IS) Surveillance aqui renent (SI) 3.6.3.5, with a requi red frequency of 18 tient hs, correspondi ng to the periodicity of scheduled refueling outages. Ti s SR which i s conduct ed by Qterati ons Survei 1 1 ance Test (CHI) - 163, was successful 1 y conpl et ed during the 1 as t refueling out age i n Nbvenber 2002. Therefore, it is likely that the switch failure occurred after that ti ne, al though the specific ti ne and date of fai 1 ure i s not known. The switch fai 1 ure was di s covered on June 5, 2003, due to the perfornance of the cont ai mtent radi at i on noni for source check coi nci dent al 1 y with a cont ai nnent pressure relief.
A revi ew of cont ai Invent penetrations was conduct ed to determine if another set (i nboard and out board) of cont ai Invent i sol at i on valves are designed si ni 1 ar t o the V12-10 and V12-11 control circuit. � The results of the review indicate that the control circuitry for the cont ai nnent vacuum rel i ef isolation valves, V12-12 and V12-13, were nodi fi ed by the sane nodi fi cation i n about 1980 and the sane design deficiency was al so introduced into that control circuitry.
PI ant control circuits t hat ut i 1 i ze the sane COCO Ski t ch Mdel ( 404S- 3- 2-1- 2- 2- Y- IT2) were reviewed to deterni ne the i npact of switch failure. The accept ance criteria applied in the revi ew mas that the RPM cont rol switch failure di d not adversely affect acci dent mitigation. � The results of this revi ew concluded that the function of the switch in other control circuits is either non- safety- rel ated, or i f safety- rel at ed, i s bounded by the single failure analysis for that system AID!, Ertl:flu OCCA 11 .1/1/11I FACILITY NAME (1) PAGE (3) DOCKET NUMBER (2) LER NUMBER (6 05000261 H. B. Robinson Steam Electric Plant, Unit No. 2 A review of radi at i on monitor control circuits concluded that no single svii tch controls mil ti pl e devices and that no s vii tch failure could disable the associated aut onat i c actuation funct i ons.
A review of the docuttent at i on associated iii th the nodi fi cat i ons to the cont ai Intent pressure and vacuum rel i of control circuits in the 1980 ti nefrane concluded that the nodi fi cat i on design failed to identify the applicable design requi renent pertaining to the single control svii t ch. At the ti ne of the nodi fi cat i on, engi neeri ng procedures were 1 ess detai 1 ed and the design basis requi renents were nore di ffi cul t to identify and retrieve. Si nce that ti ne, design basis documents for H 11 Ibbi ns on St eam El ect ri c II ant (1111SHI, Unit N). 2, have been created, and the testing and design gui dance provi ded i n the engi neeri ng procedures have been i nproved. Therefore, it is concluded that the current design barriers effect i vel y prevent recurrence of this type of nodi fi cat i on error. This, coupled with the pl armed and conpl et ed ci rcui try design reviews, provi des assurance that other errors of this type are not being introduced.
II. CASE CF EVENT
The cause of the event was determined to be the control svii tch failure coupled vii th the design deficiency of this contai=lent isolation circuitry that had apparently been unintentionally nodi fi ed, i n about 1980, i n a nanner such that the switch fai 1 ure al one could have prevented the aut onat ic is of at i on of this penet ration.
III . ANALISI S CF EvENT LCA 3. 6. 3, Itqui red Action R 1, establishes the appropri ate conpens at i ng action and conpl et i on tint for inoperability of two cont ai nnent i sol at i on val ves i n a penetration. Itqui red Action R 1 requi res the penetration to be i sol at ed by use of at 1 east one closed and de- act i vat ed aut onat i c val ve, closed nanual val ve, or blind fl ange, vii t hi n 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />. As stated i n the event description, the penetration was i sol at ed vii t hi n 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> by the use of a closed and de- act i vat ed aut onat ic is of at i on val ve ; speci fi cal 1 y, val ve V12-10 was closed and de- act i vat ed.
Val ves V12-10 and V12-11 are nornal 1 y closed. These val ves are opened peri odi cal 1 y to allow cont ai nnent pressure relief, and are designed t o close aut onat i cal 1 y on receipt of a cont ai nnent i sol at i on or cont ai nnent hi gh radi at i on signal . If an acci dent had occurred resulting in release of radioactivity into contaimtent during the ti ne that AID!, Ertl:flu OCCA 11 .1/1/11I FACILITY NAME (1) PAGE (3) DOCKET NUMBER (2) LER NUMBER (6 05000261 H. B. Robinson Steam Electric Plant, Unit No. 2 U.S. NUCLEAR REGULATORY COMMISSION 00 2003 - 002 - the control switch for the valves was nal funct i oni ng, and if a cont ai invent pressure relief had been occurring at that ti ne, it is possible that unavai 1 abi lity of the aut onat i c closure function could have del ayed or prevent ed val ves V12-10 and V12-11 from aut °mat i cal 1 y closing. This could have resulted i n an uni nt ended, although noni toyed, rel ease.
For the purpose of assessing the risk i npact of this event, i t was assured that the condition nay have existed since the 1 as t refueling outage. This i s approxi nat el y 28 weeks, and esti mat i ng approxi nat el y 7.4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> of pressure relief per meek results in a rough est i mate of 208 hours0.00241 days <br />0.0578 hours <br />3.439153e-4 weeks <br />7.9144e-5 months <br /> of ti tint when cont ai nnent pressure relief operati ons were t aki ng pl ace. The average annual effect of the condition on Large Earl y al ease Frequency (Min nay therefore be esti nated by assuning the risk existed for a fracti on of the tine equal t o 208 / 8760 or 0.024.
A cal cul at i on was perforned to est i Hate the 1 i kel i hood of the failure of operators to nanual 1 y position the control smi t ch to isolate f � V12-10 and V12-11. Based on the as s impt i ons t hat t here would be one half hour of avai 1 abl e ti ne before a 1 arge fracti on of the cont ai nnent vol une could be rel eased through these val ves, that the action would nornal 1 y be acconpl i shed mi t hi n 5 ni nut es of i ndi cat i ons t hat the action i s needed, and t hat the action requi res only 1 ni nut e t o acconpl i sh, a probability of failure of 0.001 was estimated for this CPI action, i . e. , "an action to be taken in response to di rect cue. " The probabilistic safety assessncnt analysis currently esti mates a core danage frequency of 4. 32E-5 / year. A fracti on of this core damage frequency i s associ at ed mi t h acci dent sequences which are expect ed to result i n earl y fai 1 ure or bypass of the cont ai nnent . Cbnt ri but i ons from t hese sequences would be unaffected by post ul at ed fai 1 ures of V12-10 and V12-11, and their contributionss t o core danage frequency coul d be subtracted before eval uati ng the consequences of the smi t ch malfunction. For si npl i ci ty, this correction was not made. This simplification results in a conservative estimate of the risk Therefore, the increase i n 1 arge earl y rel ease frequency associated mi th this event can be calculated: � 4.32 E-5 / year * 0.024 * 0.001 = IE-9 / year.
This indicates that the risk associ ated with the control switch malfunction was negligible.
AID!, Ertl:flu OCCA 11 .1/1/11I FACILITY NAME (1) PAGE (3) DOCKET NUMBER (2) LER NUMBER (6 05000261 H. B. Robinson Steam Electric Plant, Unit No. 2 U.S. NUCLEAR REGULATORY COMMISSION 00 2003 - 002 - As stated previously, no release 1inits were exceeded. If plant conditions had requi red i sol at i on of the penet ration, al arm and i ndi cat i ons i n the control room moul d have al ert ed the operat ors t o the condi t i on and the appl i cabl e operat i ng procedures di rect the operat ors to mutually isolate the penetration. This event i s bei ng report ed in accordance with 10 GR 50. 73( a) (2) (v) (C) and (I), any event or condition that could have prevented the ful fi 1 1 rent of the safety function of structures or system that are needed to control the release of radioactive noteri al or ni t i gat e the consequences of an accident.
1 V. CICRREC11 VE AC110% The corrective act i ons for this event t hat have been pl armed or conpl et ed include the fol 1 ovi ng:
V Procedures used to operate the ant ai mtent Pressure and Vacuum Itl i ef System have been revised. The revised procedures requi re i sol ati on of the penetration by use of a closed and deact i ved aut omit i c val ve, and subsequent verification that the switch is not in a failed state such that it mould prevent the aut omit i c i sol at i on function prior to use of the system for contai mtent pressure or vacuum relief. After verification of switch operability, the closed and deactivated val ve i s restored and the system i s used to acconpl i sh the needed function of cont ai anent pressure or vacuum rel i ef.
V The defective switch has been repl aced.
V A t enporary nodi fi cat i on mill be i npl enent ed, which mill rest ore the design basis requi renent that the switch fai 1 ure vi 11 not prevent the safety function. Ti s i s a pl armed action with a scheduled completion date of Sept enter 30, 2003.
V The control switch for the cont ai anent vacuum rel i of i sol at i on val ves (V12-12 and V12-13) will be repl aced and the renoved switch will be eval uat ed for any signs of degradation. This is a pl armed action with a scheduled conpl et i on date of Ctt ober 25, 2003.
V A review of the safety-related conponent control miring di agram vi 11 be perforned to veri fy that redundant conponent s are des i gned mi th i ndi vi dual AID!, Ertl:flu OCCA 11 .1/1/11I FACILITY NAME (1) PAGE (3) DOCKET NUMBER (2) LER NUMBER (6 05000261 H. B. Robinson Steam Electric Plant, Unit No. 2 control svii t ches i n each ci rcui t, and that fai 1 ure of i ndi vi dual conponent control switches al one will not cause a failure to fulfill a safety function.
Ti s is a pl armed action vii th a scheduled conpl et i on date of Cet ober 25, 2003.
V It will be confi rned that functional testing verifies operability of the i sol at i on function or procedural gui dance vii 11 be est abl i shed that verifies operability of the isolation functi on when a source is applied to containment radi at i on noni t ors 14-11 and 14-12. Ti s is a pl armed action vii th a scheduled conpl et i on date of Qt ober 25, 2003.
V A permanent nodi fi cat i on on the cont ai invent vacuum and pressure relief control circuits will be installed to restore the design such that the control switch is vii red i n series with the Train ' AV and Train `If cont ai Invent i sol at i on relay cont acts. This is a pl armed action with a scheduled completion during the next refueling out age, current 1 y pl armed for Apri 1 and My of 2004.
V. AE4I 11 CNAL I NFCRIVAII CN A Failed Cbnponent I nf ornat i on:
The failed snitch i s a CEMD Smi tch Mdel Pb. 404S-3-2-1-2-2-Y-132. The cause of the switch failure has been deterni ned to be age-related degradation and cyclic fatigue.
11 Previous Si nil ar Event s :
A review of recent (past 3 years) events at FERSEP, Uni t Pb. 2, for conditions that could have prevent ed the ful fi 11 nent of a safety function was conducted.
There were no events found.
AID!, Ertl:flu OCCA 11 .1/1/11I