ML17331B017

Summary of 931008 Meeting W/Util & Westinghouse in Rockville,Md Re HRA Portion of Plant Ipe.List of Attendees Encl

ML17331B017
Person / Time
Site:	Cook
Issue date:	10/20/1993
From:	Wetzel B Office of Nuclear Reactor Regulation
To:	Office of Nuclear Reactor Regulation
References
NUDOCS 9310280067
Download: ML17331B017 (20)

Text

Docket Nos.

50-315 and 50-316 October 20, 1993 LICENSEE:

Indiana Michigan Power Company (IMPC)

FACILITY:

D.

C. Cook, Units 1 and 2

SUBJECT:

MEETING

SUMMARY

OF OCTOBER 8, 1993 On October 8, 1993, representatives from American Electric Power Service Corporation (AEPSC) and Westinghouse met with members of the NRC staff in Rockville, Maryland, to discuss the Human Reliability Analysis (HRA) portion of D.

C. Cook's Individual Plant Examination (IPE).

A list o'f attendees is enclosed (Enclosure 1).

The licensee's overhead slides used during the meeting are also enclosed (Enclosure 2).

This meeting was a followup to a September 2,

1993, conference call between the NRC and the licensee in which the NRC asked AEPSC personnel several questions regarding their application of the Technique for Human Error Rate Prediction (THERP) methodology in the HRA.

The meeting was a working-level meeting and concentrated on the licensee's assumptions and methodologies used in developing its HRA.

Some of the specific'issues discussed included "sensitivity studies performed by the licensee, dependencies between various

errors, assumptions used when calculating times, use of performance shaping factors and use of pre-initiators such as valve restoration and post-test maintenance testing.

Currently, the NRC is reviewing the licensee's methodology for acceptability and will determine whether another meeting or conference call on the issue is warranted.

Original signed by

Enclosures:

1.

List of Attendees

- 2.

Slides Beth A. Wetzel, Acting Project Manager Project Directorate III-1 Division of Reactor Projects - III/IV/V Office of Nuclear Reactor Regulation cc w/enclosures:

See next page Office Name Date LA:PD III-1 CJamerson 10 Z>93 PM:PDIII-1 BWetzel 10/

93

'PDIII-1 WDean 10/N/93 l

93l0280067 931020 PDR ADOCK 05000315 PDR

Indiana Michigan Power Company cc:

Regional Administrator, Region III V.S. Nuclear Regulatory Commission 799 Roosevelt Road Glen Ellyn, Illinois 60137 Attorney General Department of Attorney General 525 West Ottawa Street

Lansing, Michigan 48913 Township Supervisor

,Lake Township Hall Post Office Box 818 Bridgman, Michigan 49106 Al Blind, Plant Manager Donald C.

Cook Nuclear Plant Post Office Box 458

Sridgman, Michigan 49106 V.S. Nuclear Regulatory Commission Resident Inspector Office 7700 Red Arrow Highway Stevensville, Michigan 49127 Gerald Charnoff, Esquire

Shaw, Pittman, Potts and Trowbridge 2300 N Street, N. M.

Washington, DC 20037 Mayor, City of Bridgman Post Office Box 366

Sridgman, Michigan 49106 Special Assistant to the Governor Room 1 State Capitol Lansing, Michigan 48909 Nuclear Facilities and Environmental Monitoring Section Office Division of Radiological Health Department of Public Health 3423 N. Logan Street P. 0. Sox 30195 Lansing, Michigan 48909 Donald C. Cook Nuclear Plant Mr. S. Brewer American Electric Power Service Corporation 1 Riverside Plaza

Columbus, Ohio 43216 E.

E. Fitzpatrick Indiana Michigan Power Company c/o American Electric Power Service Corporation 1 Riverside Plaza

Columbus, Ohio 43216 July 1993

DISTRIBUTION FOR

SUMMARY

OF OCTOBER 8 1993 MEETING w enclosures 1

& 2:

/Docket1'RC

& Local PDRs PDIII-1 RDG File B. Wetzel B. Jorgensen, RIII DATED:

October 20, 1993 T. Hurley/F. Miraglia 12/G/18 J.

Partlow 12/G/18 J.

Roe J. Zwolinski L. Harsh W.

Dean B. Wetzel OGC E. Jordan, MNBB 3701 E, Lois, NLS 314 C. Ader, NLS 324 H. Drouin, NLS 324 J.

Flack, NLS 324 H.P. Rubin, 10/E/4 J. Schiffgens, 10/E/4 E. Rodrick, NLS 324 J.

Lane, NLS 013 R. Hernan, 14/C/7 ACRS (10)

B. HcCabe, EDO T. Kobetz, RIII E. Schweibinz, RIII

October 8

993 D. C. Cook Units and uman e iabil't Anal sis ethodolo Hect I'A<<d ENCLOSURE 1

arne Beth Wetzel Bill Dean Hary Drouin John Flack Erasmia Lois Charles Ader John Schiffgens Ed Rodrick John Lane Gordon Arent Jeb Kingseed Doug Halin Jenna HcNanie Rick Bennett Selim Sancaktar Or ani at o

Acting Project Hanager, PDIII-1/NRC Acting Project Director, PDIII-1/NRC RES/NRC RES/NRC RES/NRC RES/NRC SPSB/NRC RES/NRC RES/NRC AEPSC AEPSC AEPSC AEPSC AEPSC Westinghouse

ENCLOSURE 2

Donald C. Cook Nuclear Plant Human Reliability Analysis Methodology Overview of HRA Results

- Importance Ranking

- Consistency Methodology

. Diagnosis

- Recovery Factors

- Performance Shaping Factors Dependence Modeling Stress vs. Event Path Valve Restoration after Test and Maintenance Rick Bennett - October 1993

Action Overview - HRA Importance RaQcing

( >10% )

Reduction Achievement Probability

(%)

(%)

Refill CST 1.7 Restore RCS Inventory - RRI 1.0 Post LOCA Secondary Cooldown -OA6

.9 Trip RCP (Loss of CCW,ESVU).18.7.

1423 1.30e-2 31302 5.30e-5 1466 6.80e-4 1495 6.23e-4 Station Blackout AC Recovery

.7 XHR Standby Air Compressor in Service 22 120 3.05e-2 1.08e-3 HP Switchover LP Switchover SGRT Cooldown - OAI Long Term Cooldown - OA3 Primary Pressure Relief - PPI 0

13718 6.31e-6 415 28 62 1.05e-4 2.1IM 1.46e-3 504 5.31e-5

Overview - MULConsistency Range

-.>10'imple 10-'0'2 8

11 10-'otal 43 26 Most Calculated Probabilities Below 10~ are hand calculations Ifall hand calculated accident actions were 10~ or greater, CDF would increase 45%

Overview - MW. Consistency m ari n fHih Lw l

-. High Values Non-EOP items Short Action Time Low Values Simple EOP Steps Non-emergency alignments Long Action Time ther nsi tenc n

Switchover HP 4 LP - similar actions Factor of 10 due to more time for HP Valve Restoration - Air 8r Motor better due to Control Room indication - more later Sensitivity-Human Errors Ranked 2 to 8 Increased by 10 CDF Increases by 23%

Methodology - Diagnosis EQM Handbook Event Driven EOPs Diagnosis of Initiating Event Cognitive Behavior - little data Diagnosis by team Cook Nuclear Plant Application Symptom Driven EOPs Diagnosis of Action Requirements and Initiating Events More Rules/ Fewer Decisions Recursive Procedures Handbook. Noted Symptom driven EOPs may make diagnosis negligible Column format should decrease error factor of 2-3 Performance Shaping Factors Required

Methodology - Diagnosis Recovery n r li FaB To Diagnose if FaB procedure use yz FaB to respond to @arms

'aB procedure use by FaBing diagnosis assuming correct instrument reading

("cognitive")

gr FaB diagnosis assuming some incorrect meter reading (reading error, high dependence) rima Bleed A F e Exam I

Procedures E-0 step 15

- no aux feed and low steam generator level F 0.3 (Status Tree) Red - Loss of heat sink Alarm Response Low-low steam generator level

~ ~

FAILURE OF THE OPERATOR BLEED I

FEED COOLING ACTION O'BF)

PBF FAULT TREE OPERATOR FAILS TO DIAGNOSE THE NEED f'R BLEED t FEED COOLING OPERATOR FAILS TQ ESTABLISH BLEED L FEED COOLING FAILURE TO DIAGNOSE DEPRESS.

NEED VITHIN 30 HIN.

OPERATOR FAILS TQ RESPOND TQ HULTIPLE ALARHS VITHIN 30 HIN.

2QF3 PRESSURIZER PORVS FAIL TO OPEN OPERATOR ERROR FAILURE OF HIGH PRESSÃ INJECTION (2r CHP OR SI PUN 23 PBF-DIAG-HN-HE PBf'-SGALARH"HE SUB-PORVP SUBWP2

~PEND 'O THE ATTACHMEHT TO AEP>NRC-10B2P

~

~

~

~

7of 27 Mean HEP Descriptions (error'ype) gable h-2 Reference)

Task h1-DIAGNOSIS TABLEM PBF HEP CALCULATIONS Performanct. Shaping Factors HEP I

S~

Dependence Others Fault Tree ra tire~

hdditlonal Notes (Equlpmeng hctlons, Indications, Locations) a) Failure to diagnose within30 minutes of a compelling signal 1) glren success ln reading any of.the multiple low SG

~

lerel and high RCS tan perature and pressure indicators (saturation meter also arallable) 2)

glren failure ln reading any of the multiple low SG

'eret and high RCS temperature and pressure indicators 2.7RA3 (3) a) Dlagnosb!

5 2.7E43 (3)

High (Success)

Ngh 0.01 5.0E41 (Failure) 0,01 6,75E45 PBF-DfhG-MN-Bt(l q(a1) <<1KEP(a1) ~PSFI2 q(a2a) <<PKP(a2a) >>

PSF+1]I2 b) Reading (1 5

of2 types of analog meter)!

2~3.75843 (51) 0.01

'.8M4 Combined!

2.58844 q(a2b) IKP(s2bF 0 q(a1)+q(~)'q

~

(a2b) b) Failure to respond to SG low-low leret annunciator (under ANDgate with Ala) (30 minutes into transient, lt b assumed that 7-10 annunctators nre alarming - conserratlre) 13E41 (45) 0.1 2.6EA2 PBF<GALARM-HE

Methodology - Diagnosis Performance Shaping Factors hc&iiix STRESS LEVELAPPLICATION Extremely High (threat stress, e.g., ATWS, Station Blackout; Loss of Indicators during Loss of 250 VDC-early in transient) (conservative w.r.t. recommended value of 5) 10 Typical Transient - Step-by-step, Moderately High Stress 5

(conservative w.r.t. recommended value of 2 - Assumes Experienced Operators during off-normal conditions-Appropriate for diagnostic activities early in the transient)

Typical Transient - Longer Time Frame for Response -

2 Step-by-step, Moderately Kgh Stress - Appropriate for activities performed after the initial diagnosis since there would be expected to be only a limited amount of confusion in the control room at the Cook Nuclear Plant Very Low Task Load Optimum Stress Level - for activities performed in a non-transient situation where the operator is very familiar with the activity

Methodology - Diagnosis Performance Shaping Factors hct~iy OTHER CONDITIONS General value for response for operators who are well-trained in the appropriate procedures (this also applies to the failure to foHow procedures once entered).

0.1 Medium Time Frame for Response Absence of a Direct Indicator of Change 0.1 Availability of Multiple Supportive Indicators 0.1 Failure to Enter Detailed Procedures at Cook Nuclear Plant by Following the Initial Steps of E-0 due to Emphasis Placed on Following Procedures (applies typically to the diagnosis phase of the transient).

0.01 Selection of Wrong Control when a Relatively Long Time Frame is Involved.

1.0 Selection of Wrong Control when Controls are Very Clearly Marked with Mimic Lines 0.1 Special Value for Cases where the Operator is Especially 0.01 Well-Trained in a Specific Diagnosis - At the Cook Plant the operators are especially well-trained on entrance into the EOPs (e.g., Steam Generator Tube Rupture)

Memorized Procedure 0.1

~

~

t~ O

~ Os fe

~ TTZS

)

~

~

~

CPT<<SIR ra<<Z 10 <<sSSOSZ SsC

<<Z0 tf<<<<ZZZ 0 tZZZ SZZLO<<

OT<<sfR SIZZ 10 SZTS<<S0s

<<ZZS 0 ITZZ SSSZO0 SIFT

~i%VS tJ0 10 OOI O%04TR SSSS<<

tSSSK O'OSI~

OP R Sf IMPS

~

01 OC<<SIR tk<<S fs4JIC 10 aau5 KITS

~0 0STTSIZ SISSSfsf ISIZZ<<SC KAS0O CSSSR tSZS SIJCITZI VS0 OCZ<<SST eaLT0sSZS EFSZAITÃ ta<<Z tkSSIK 10 10 00TSSTZ t<<SSft IVIZZ<<SC

<<ZZS KT00<<

Vf0 OCO<<SST S000SIR tJ<<S 10 XZ<<f tuCr00s a VROUW tSZIJC SASK 10 tfSSfft IVIZZ<<JC Vsl OCCWJST g

5 g

tel OZSSIOS SVSKS00 tTS

~KZ<< ~ TOVs

~ sClZSS OTS>IR SIZZZIS VSOC SVITOS 10

~IKSTI.T OVSI

~SS tOTVS CFTSSSIR SSLZZT

~f10~

VSOO SOSIKS. 10 O ause CCZS STa4I SSSZ ~

SVsss SOOsssSC

~CL R IT0sf 3

Q 2

I I

FIGURE 2"2 HUMAN ERROR r OR PRIMARY BLEED AND FEED (PBF>

PAGE 88 SHEET I OF' 14I5< Ie"OCT"9I

Methodology - Dependence Modeling Qgnn~rl "Dependence is not assumed among separate steps of procedure.

Exceptions when conditions clearly indicate the need.

Parallel Systems with Similarities rima Iced 8r FR-H.1 Steps 18 to 23 Each step called out by SRO RO Performs R Replies to SRO Self verification is force on RO SRO checks RO (dependence could be assumed)

However, verification is not credited at this point Verification is credited by checking action ~~

by checking pressure and temperature trends t re Contr 1Air Effectively, faBure of control air recovery is noted by continued unavailability of control air.

Methodology - Stress vs. Event Path Generally - Stress for a particular action assumed independent of event path At time of entry, reactor conditions are simBar Similar or same procedure Supported by operator interviews and simulator observation Largest potential stress difference is in diagnosis Few Action Important in Many Events

..Primary Bleed dh Feed Switchover Refill CST

Methodology - Valve Restoration "Plant" Generic Numbers are used for Typical Manual Operated Valves and Typical Airor Motor Operated Valves A common, Level 1, tagging system is used Each tag is recorded A verified Operator verification of restoration Common tagging system - individual verification sheets Conservative Values Chosen to Use One Value Airand Motor Operated valves have added recovery factor of control room indication

l