ML040050117

Phase 3 SDP Analysis: Arkansas Nuclear One Unit 1

ML040050117
Person / Time
Site:	Arkansas Nuclear
Issue date:	12/12/2003
From:	NRC Region 4
To:
References
FOIA/PA-2003-0358, IR-01-006
Download: ML040050117 (9)

Text

Phase 3 SDP Analysis: Arkansas Nuclear One Unit 1 (ANO-1)

Lack of Adequate Procedures for Manual Actions to Achieve Post-Fire Safe Shutdown Following Fire Damage in Fire Zones 98J and 99M

1. Performance Deficiency The Arkansas Nuclear One Unit 1 (ANO-1) fire zone 98-J (Diesel Generator Corridor) and fire zone 99-M (North Electrical Switchgear Room) did not meet regulatory requirements for separation of electric cables and equipment of redundant trains of systems necessary to achieve post-fire safe shutdown. The licensee did not have adequate procedures for manual actions to achieve post-fire safe shutdown following a fire In fire zones 99-J and 98-M. This condition has existed since the issue was identified as Unresolved Item 50-313;368/0106-02 in the Inspection Report 50-313;368/01 -06, August 20, 2001.

2. Fire Scenario The primary combustibles in the ANO-1fire zones 98-J and 99-M are the safety-related non-qualified IEEE-383 electrical cables routed in open cable trays that are located above numerous potential ignition sources. The height of the lowest cable tray in fire zone 98-J is approximately 6 ft. from the floor; while in the case of fire zone 99-M, the height of the lowest cable tray from the floor is about 8 ft. In fire zone 98-J, the potential ignition sources include a battery charger, 480V motor control centers, 125V DC distribution panels, wall-mounted electrical cabinets, emergency ventilation units, and an emergency chiller unit (VUC4A/C51). The potential ignition sources in fire zone 99-M include an air-cooled transformer (X6), a 120V instrument transformer (X62), 4.16kV switchgear cabinets, 480V motor control centers, a 480V load center, inverter cabinets, and ventilation units.

Other potential ignition sources, such as a power cable failure in a tray, or other electrical originated failures (in distribution panels, circuit boards, electrical wiring, internal cable fault, electrical circuit fault in switchgear cabinets, etc.) leading to ignition of the in-situ combustibles (cables), are considered in the fire scenario. Ogp

,Qtyvrk(welding-or-a limited'1 00-lbtransient-combustible-source are also possiblevbutVx areq psinoredwithin the scope-ofthis-analysis..,-

The combustible loading in fire zone 98-J consists of mostly cables in the cable trays.

According tollicensee-pr.vided information and calculations,!^th.fie.durationrinfirezofe,

~9.8-J.,ytas.Ostirpatec1.tobe,2.hours and15' minutes by considering all available.'in-situ>'

g btsiblean9 d

pZe,,WitJal4,QQbtransientecombustible sourcee The combustible loading in fire zone 99-M also consists primarily of cable insulation in open cable trays.

,Considering-all avallablein-situ combustiblessand a potential 00-lb4ransient combustible spurce~Fthe fire-durationinX99;M;W-;ag'stlmated to beL3Q-minutes..

The credible fire scenario is based on postulating that a fire develops from any one of the potential ignition sources, if undetected and unsuppressed (i.e., no immediate

[,;*...;.......'

, c;:. i2;n i

intervention from plant operators), would grow to a rate of heat release of 400 kW (or 380 BTU/s) and ignite the cable insulation of electrical cables resulting in challenging fires.

The Diesel Gene rator C ljorfnJirezone.98,J-is provided-withsmoke-detection-and-and partia overasguG;;ti watersupp essionsystemthatisactuated bycrossvor'd-smoeand-in-traylinearheat-detectorsF The ionization smoke detectors are provided at the ceiling level that only alarms in the Main Control Room (MCR). The line-type heat detectors (trade named Protectowire) are installed on the top of the cables in each cable tray, and also alarm in the MCR. TUlp na.cd-eluge.ap§dkler system provides partially ppQtetiorntahefrezqf §nd will active u

c sccessfulcross zane'-t, detection signals-frombothtpf.tbe,smok andIinear heatdetectors.-

fire hose reel and portable carbon dioxide extinguishers are located in the vicinity of the fire zone for manual fire fighting purposes when needed. Fire zone 98-J is normally unoccupied except for inspections, shift tours and maintenance activities.

The fire-zone-99-M (North'ElectricalSwitchgear, Room) -is only protected -witha smoke,,

detection system..The smoke detectors -are'provided at-the-ceiling -level which alarm in the.

¶MCR.,There -are no fixed automatic fire suppression systems in-this switchgear. roonp, A

fire hose reel and portable carbon dioxide extinguishers are also provided in the vicinity of the fire zone for manual fire fighting purposes when needed. Fire zone 99-M is normally unoccupied except for inspections, shift tours and maintenance activities.

In the fire scenario development, it is assumed that a credible fire starts from a specific ignition source (e.g., transformers, electrical cabinets, 4.16kV switchgear, 480 V motor control centers, 480V load center, 125V DC distribution panels, or cables) and has sufficient flame spread (i.e., flame height and radius) to ignite a cable tray closest to the i ignition source. TheSPLB.fire-hazards and fire modeling analyses (see ADAMS Accession #ML021490005, #ML021990405) ppstuilated-that energetic electrical faults in -

electrical cabinets, producing a fire with heat release rate (HRR) of 400kW.or.greater, can..

lead to fire growth and subsequent fire damage-to target -cables dependingop thp ventilation conditions in the compartment; -Two different cases of fire compartment conditions were considered to define the fire damaging scenarios for the two fire zones: (1) vent open and closed in Fire Zone 98-J, and (2) door open and closed in Fire Zone 99-M.

The CFAST (Consolidated Model of Fire Growth and Smoke Transport) computer code was used to model the fire growth of fire involving electrical cabinet and equipment that would lead to a challenging fire in the fire zones 98-J and 99-M.

f scenarios a range of HRR curves-rom 200 kW~to-500,kW-wereused as-inputto thebr. 1 CFASTfire modeling analyses because no-direct-data on-the burningof the-specific -

electricLnitioources at full or.-intermediate scale were available.-As documented in EGN~

/C-4527ie selected range of HRR curves were developed for electrical cabinetL

-scale fire tests conducted on electrical cabinets in a large (e.g., actual control room size) enclosure. Inthe ir-rnodeling -analyses,-the fire-was -assumed-to-develop-with a "t2fast fire-growth-rate' due to.the, electripally.eregized fire eLorL* ep t is also assumed that there is complete combustion and an ample supply of oxygen for the v

fire with the given HRR.

In modeling the fire growth and damage potential, results of SPLB fire modeling analyses show-that fires with HRR of 400kW-could -damage the overhead-cables in fire -zone 98-J with.open -vent conditions, and in fire-zone 99-M with.closed -door conditions.. In the fire

.scenario for fire zone 98-J with open vent conditions, t.

425 °.io.approximatelyA 9 ;minutes jpAhecaspof irpe zQne -99.-M -with losed doo r..,

comtnd tWessoke

-reaches425--9Flinaapproximately'10-nInutes.fF The limiting temperature of 425 F was used in the fire modeling analyses because this temperature condition can cause failure of non-IEEE-383 rated cables. Th&,results-of he~,

fire..rodeling analysesralso-indicate thatfires with l RRmfO200,kW:Iand,30O 'kW.Jnthe.,t,4y,1, fir..,znestend-to.becomeyentilatonlimited.and..,deaywo 200kW could.only-result in-a-maximum smoke-layeritemperature-of-305-<F-in-about-one h tore

.withopen-entconditionsrThis resultindicatesthat-tAhI oy.haables ay-rema,,,n.,damagedinfire. zone,98 Jfornan-hoursundertheq.

p>pi,9§ateQ, conditions,,

pe_*"Based on results of the fire hazards analysis, SPLB fire protection staff also postulated a fire scenario involving a lube oil spill fire resulting from a breach or leak in the lube oil system for the emergency chiller chilled water pump located in fire zone 98-J. The fire modeling analysis for this scenario indicates that a single gallon of lube oil spill, if ignited, could form a pool fire with a diameter of approximately 1.5 feet, flame height of 6 feet, and burn duration of 7.5 minutes. The flame height of the postulated pool fire is sufficiently high to impinge on the cable insulation of the non-IEEE-383 rated cables on the lowest cable tray that is located about 6 feet from the floor. It was concluded that the turbulent diffusion flame impingement on the cables would cause potential ignition and flame spread along the cable trays, and thereby further increases the HRR in the fire zone 98-J.

3. Assumptions (a) Fire Barriers - In fire zone 98-J, the walls and doors are 3-hour rated fire barriers "A-r

,ppe-hour..rated -barrier surroUnds the Red train AC-.instr menta.tio,:power supply.-cables, whil~e>,therRed train power.cablesin.fire zone,98-J ar,e jpQteoteL The Red train redundant cables are not separated from the Green train cables by a minimum of 20 feet distance free of intervening combustibles. In-fire zone-99-M-the-Red-train-cables are-not-.,

protected with one-hour -rated barrier;and are-not separated from the.Green -train cables by-,a minimum of-20 feet distance free of intervening combustibles.,-As such; the cablesin-both fire zones -could be damaged by a floor based fire., -

(b) Automatic Fire Supression - The East portion of fire zone 98-J is protected by a cross-zoned, pre-action deluge system that is actuated by cross-zoned smoke and in-tray linear heat detectors. Periodic surveillance is performed on the cable tray detection system and room smoke detection system to ensure that the suppression system remains operable.

The sprinkler heads in the corridor are open heads, and water will be available as soon as the sprinkler valve opens in approximately 5 seconds (according to manufacturer's information). T.e.,suppression system-response time-is-assumed-to beapproximately -7t.

Mjnutesjecause the actuation -time for the lire-typ heatdet tio-s~st-rtf tosnse'a-temperature. of 9.0QF-was estimated to-be less-than 7 MinutesTherefore, the raceways that required more than 10 minutes to sustain damage can be assumed to be protected by the suppression system. wT~bprobability.of pre-action-sprinkler-system-being-unavailable, is,,assumed to be 0.05 for the normal operating state based on the EPRI-database (EPRIL.

FIVE report, page 10.3-7). Th uptay.ilability value include the consideration for.failure of thze system to operate on demand and-the system being out of service-at the time of-a-fire

(due to shut control valve, etc.). Irrfirz 9;z n-o*luio'aFire s esystem sa-provjdeda---

(b) Fire Detection and Manual Fire Suppression - Both fire zones 98-J and 99-M are equipped with ionization detection systems that will detect fires in the incipient stages and provide alarm conditions in the Main Control Room (MCR). Alarms1r~rith'e-ifbi2atidrr dfttgjo system :would esult irithsdi tchofa rfsop-ataort investigate-any-ofthetwo fire-zonesT-he central fire brigade lockeris located one elevation-above theafire:'zonesR.*

98-,4and.99 M,-and.therefore;-the 'travel time-of the'firetbrigadeirom'the locker toithe fire,.-

sge.s,,consjdered to be-reasonably.short,.7,,

Based on repqt.fire,.drills performedi-oirezdi'1 OON-,Iwhich is 'adjacentitozfire-zone;99-A 0minutes. There are two access points to the fire zones, which are easily accessible by the fire brigade response team. Based on these factors, i_ wa-.sasrrj tatany fire scenario requiring greater than 20-minutes to sustain cable damage may be suppressed by-the fire brigade.

The fire-induced core damage frequency equation for the fire zones can be defined as follows:

FCDF = F* Sf P

P2 P3 where F = Fire ignition frequency of ignition source Sf = Severity factor for a challenging fire P1 = Probability of automatic fire suppression system being unavailable P2 = Failure probability of manual suppression by fire brigade P3 = Conditional core damage probability, with or without recovery actions 4: Fire lanition Frequencies The various ignition sources in the fire zones 98-J and 99-M respectively, and their associated fire ignition frequency estimates, as calculated using the EPRI FIVE methodology and listed on the Ignition Source Data Sheet (ISDS) for the two ANO-1 fire zones, are shown below:

Fire Zone 98-J Generic WFL WF ISDS Ignition 7 Ignition Sources Fre y

Frequency Electrical Cabinets 1.9 E-2

,z' 1

1 1.01 E-1 1.9 E-3 Battery Charger 4.0 E-3

""2 9.52 E-2 7.6 E-4 Ventilation Subsystems 9.5 E-3 2

1.12 E-2 2.1 E-4 Fire Protection Panels 2.4 E-3 2

2.33 E-2 1.1 E-4 Emergency Chiller Pump 2

2.80 E-3 5.3 E-5

Fire Zone 99-M Ignition Sources Electrical Cabinets Transformers Generic Frequency WFL WF, ISDS Ignition Frequency 3.8 E-3 1.5 E-2

/ 0.25.: 1.0 7.9 E-3 2

2.044 E-2 3.2 E-4 Ventilation Subsystems 9.5 E-3 2

5.60 E-3 1.1 E-4 The ISDS ignition frequency estimates of each identified ignition source were derived based on adjusting the generic fire ignition frequencies by a location weighting factor (WFL) and an ignition source weighting factor (WF,). The generic fire ignition frequencies used in this analysis were based on the EPRI database (EPRI Fire PRA Implementation

/

Guide, pages 4-7 & 4-8, Table 4.2). A6comparison of.the genericf ire ignition-frequency I

.tgs-,

garnstthe;NRG:updated fire-,'tsatab'a§'e;tHoglifn~RES),showed:that Ahe.generjpfgue~rq,,,stm~ew,~t-qlyirnbe go r,

Vith the exception of the electrical cabinets, all of -the above. listed.sources.were..

considered.as uPlantWide " components and therefore, were assigned a W.FL= 2.(i.e.,..

&.rpe r.fgunits:per~site).

For fire zone 98-J in the auxiliary building, the electrical cabinets were assigned WFL = because of the number of units per site divided by the number of auxiliary buildings. For fire zone 99-M (which is a switchgear room), the electrical cabinets t7 were assigned WFL = 0.25 because of 2 units per site divided by 8 switchgear rooms.

The weighting factor,WF, for plant-wide components is obtained by dividing 0

Copponents in the specified room by the total number of components inte n firer zone 98-J, WF = 0.1 01 is derived for the electrical cabinets by dividin in the corridor by the total number of 1452 cabinets in the auxiliary building. Similaf;¶e F,

factors for the other ignition sources in both fire zones 98-J and 99-M were derived from plant-specific data (as provided in ANO-1 licensee response package). In fire zone 98-J, there are 4 ventilation subsystems, whereas there are two ventilation units in fire zone 99-M. In fire zone 98-J, there are 2 fire protection panels, whereas there are none in fire zone 99-M. In fire zone 98-J, there are no transformers, whereas there are two transformers in fire zone 99-M. In fire zone 98-J, there are 2 battery chargers, whereas there are none in fire zone 99-M. Based on licensee Calculation 85-E-0053-47, the total number of ventilation subsystems is 357, the total number of fire protection panels is 86, the total number of transformers is 98, and the total number of battery chargers is 21 (increased from 19 due to recent modifications).

The ignition frequency of the emergency chiller pump was derived using the generic frequency for the ventilation subsystems because there was no plant-specific ignition frequency data. The WF, factor for the emergency chiller pump was based on a single ventilation subsystem in the fire zone. Therefore-,WFr=-1/357

&-.Ewas~useddinr, deriving the ignition frequency for the emergency chiller pump. The emergency chiller unit in fire zone 98-J is a standby component and its operability is demonstrated on a monthly basis by a surveillance test with duration of less than 30 minutes. The test is performed by Operations personnel who are trained fire brigade members.

5: Conditional Core Damage Probability (CCDP)

In the various fire scenarios considered (i.e., each scenario initiated by a different ignition source), conditional core damage probabilities (CCDPs) were calculated for the two fire zones iaqltleANi-!.1,d'PEE EfireiskmodelfortwcasesBa),3, ooperator1ret'bvdty' actiens V

'dtevn~)orj actiormediterand;be~sretit~f~<_p opra^rctiris-torerover-thei: mergencyzific FX.dwaterEFW),systemand-otherjequire.dactions for-safetshutdownlAn.both cases, they.

CP.2Mcalculaiajqrjs were.perf-ormed-for two conditions: (i) one Red equipment-train-is;i zvailable to-perfom-mitigating functionsrand>(ilbothLFDd EL!,

uQ~~egi sains,.

ugavailabletdue tothe-severechallenging fire.4lrth'e^;eV'eht-that both-redundanta-eqwp;menttran' irafie7zntatEraffe'Cted by-firevthe-CCDPs 'VWuld bedoiniat6d by

• ,ope.Fajq.qLo-risloachieve safe -shutdown outside "f thiemain controlsro0.M.

test

,imateheCCDPs ioopwetrecoveryaofonss (gI£)A~le1Calculations.O2-E-0004-JQ.cl~ g Q2,XE.-00.-02)4 thefollowing operatorrevryfl actions iatnre

-dited(i~e.,IsettoogicalTRUE inthe cutsets!of the riskvhodel).

1.

Operator fails to isolate ICW after automatic SW isolation fails on ES

2.

Operator fails to start and control EFW pump P-7A manually when offsite power is available

3.

Operator fails to start and control EFW pump P-7B from control room when offsite power is available

4.

Operator fails to open breaker locally at Al from the unit auxiliary transformer and close the breaker from startup transformer SUT1

5.

Operator fails to manually close breaker 152-308 or 152-408 for EDG recovery

6.

Operator fails to de-energize CV-2646 and CV-2648 (with consideration of hot-short probability for CV-2646 or CV-2648)

As shown in the ANO-1 Calculations 02-E-0004-01 and 02-E-0004-02, e,

s.z;

.c,1Q;!qqultedjor~the wo-fire zones.for the different fire scenarios ar.e provided below:

Fire Zone 98-J CCDP with No Operator CCpPwith Operator East Area Recovery o1vryq L/Ww All Redundant Cable 1.13 E-2 1J18 E-4 4.

Trains Failed Red Train Protected 8.10 E-3

'197E Fire Zone 98-J CCDP with No Operator CCDP with Operator West 1 Area Recovery Recovery.

All Redundant Cable 5.38 E-4

..1.39 E-4.

Trains Failed Red Train Protected 5.38 E-4

-1.39,. E-4,.

~11/_j C CP

Fire Zone 98-J West 2 Area All Redundant Cable Trains Failed Red Train Protected Fire Zone 99-M All Redundant Cable Trains Failed Red Train Protected CCDP with No Operator Recovery 2.49 E-3 5.38 E-4 CCDP with No Ope at r Recovers g1v/cyIe 5.76 E-2 f

i t

7.96 E-3 p

(CCDJ CIDT57 Cooerator 4,ecovery

.1A85iEN'-

CGE)PftlhOpator Recok iI

-Z>.

I

6. Integrated Assessment of Fire-Induced Core Damage Frequencv The fire-induced CDF estimate for fire in the fire zones 98-J and 99-Mijthno.operator,.

• eroveactions s

below?,,;&.,

fi e SA s

M Fire Zone 98-J F1 S

P1 P2 P3 FcDF Ignition Sources Electrical Cabinets 1.9 E-3

0.

.05 0.5 9.2E-3' 3.3E-7 Battery Chargers 7.6 E-4 0.05 0.5 9.2E-3 1.3E-7 Ventilation Subsystems 2.1 E-4 0.08 0.05 0.5 9.2E-3 3.OE-9 Fire Protection Panels 1.1 E-4 0.12 0.05 0.5 9.2E-3 3.OE-9 Emergency Chiller Pump 5.3E-5 0.08 0.05 0.5 1.4E-2 1,

-s -7 Total CDF Fire Zone 99-M Ignition Sources Electrical Cabinets F,

3.8 E-3 3.2 E-4 1.1 E-4 S.

P1 (9 1.0 0.10 1.0 0.08 1.0 P2 P3 4.7E-7 FCDF 1.3E-5 9.3E-7 2.6E-7 Transformers Ventilation Subsystems 0.5 0.5 0.5 5.8E-2 5.8E-2 5.8E-2

7ill, estimate of 9.2E-3 (summed over all portions of fire zone 98-J) was used in the risk analysis of the fire scenarios involving these ignition sources.Lthecaseo;theiei..gX scenario involving the emergency chiller pump, the pool fire with a flame height of 6 feet and burn duration of 7.5 minutes was postulated to impinge on the cable insulation of the non-IEEE-383 rated cables on the lowest cable tray that is located about 6 feet from the floor. It was concluded that the turbulent diffusion flame impingement on the cables would cause potential ignition and flame spread along the cable trays, and thereby further increases the HRR In the fire zone 98-J. It is likely that a fire from this pool fire may damage both equipment trains at the same time. Therefore, the CCDP estimate of 1.4E-2 (summed over all portions of fire zone 98-J) was used in the risk analysis of this-fire For the fire scenarios in fire zone 99-M involving ignition of the electrical cabinets, transformers, and ventilation systems, it is assumed that both equipment trains would not be available to perform mitigating functions because the Red train cables are not protected with one-hour rated barrier, and are not separated from the Green train cables by a minimum of 20 feet distance free of intervening combustibles. Therefore, the CCDP estimate of 5.8 E-2 was used in the risk analysis of the fire scenarios in fire zone 98-M.

7: Incremental Fire-induced CDF

.T baseline CDF~ conroing ns4gr the fire scenarios in the fire zones-98-J and-99-M.

perator..recovery. actions scalculated.by-assuming the manual suppression iitywof.O.

,,;his value is used for the manual suppression capability because it is considered to be appropriate for the entire population of fires, including severe fires, arising from an ignition source. The fire protection SDP methodology, which uses the entire population of fires as the basis to derive an ignition frequency, also uses the probability value of 0.1, in gSfl

_j jteitureppbalityofinondegraded manuals

.,suppression capability..,hebaselinezGDFwithcredit of.operator recovery actionsifor.the comrpg7..case nal yes~.Arplho~w.a.

Fire Zone 98-J Ignition Sources F.

P1 9

P3 Electrical Cabinets Battery Charger Ventilation Subsystems Fire Protection Panels Emergency Chiller Pump 1.9 E-3 0.05 0.1 7.6 E-4 0.05 0.1 2.1 E-4 0.05 0.1 1.1 E-4 0.05 0.1 4.6E 4.OE-9 4.6E-4 1.OE-9 4.6E-4 5.OE-1 0 4.6E-4 3.OE-1 0 5.3E-5 0.05 0.1 5.4E-4 1.E-10

-5

'e Total CDF

Fire Zone 99-M Ignition Sources Electrical Cabinets Transformers Ventilation Subsystems F.

P1 P2 P3 FCDF 3.8 E-3 1.0 3.2 E-4 1.0 1.1 E-4 1.0 0.1 1.3E-3 4.9E-7 0.1 1.3E-3 4.1 E-8 0.1 1.3E-3 1.4E-8 Axe-'

I.

qE-(4

>I a,

Total CDF Therefore, the incremental CDF changes due to taking credit for operator recovery actions for fire scenarios in fire zones 98-J and 99-M are estimated as follows:

A.

Fire zone 98-J: (4.7E-7)- (6 B.

Fire zone 99-M: (1.4E-5)

(5.5E-7) =4E fire scenarios in fire zone 98-J is 4.6E-7, and the significance characterization is GREEN.

The change in CDF due to taking credit for operator recovery actions for fir cenarios in fire zone 99-M is 1.3E-5, and the significance characterization is

_L