"Draft Meeting" is not in the list (Request, Draft Request, Supplement, Acceptance Review, Meeting, Withholding Request, Withholding Request Acceptance, RAI, Draft RAI, Draft Response to RAI, ...) of allowed values for the "Project stage" property.

Text

Summary of Meeting W/Dpc on 960919 at Oconee Nuclear Station Re Emergency Electrical Distribution Sys Draft Repts Prepared by NRR & Aeod.List of Attendees Encl
	ML15118A448
Person / Time
Site:	Oconee
Issue date:	09/26/1996
From:	Lebarge D; NRC (Affiliation Not Assigned)
To:	; NRC (Affiliation Not Assigned)
References
	TAC-M93550, NUDOCS 9609300313
	Download: ML15118A448 (90)
	v • d • e

9e REG(97 UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20555-0001 September 26, 1996 LICENSEE:

Duke Power Company FACILITY:

Oconee Nuclear Station, Units 1, 2, and 3

SUBJECT:

SUMMARY

OF THE SEPTEMBER 19, 1996, MEETING ON THE OCONEE EMERGENCY ELECTRICAL DISTRIBUTION SYSTEM DRAFT REPORTS PREPARED BY NRR AND AEOD (TAC NO. M93550)

On September 19, 1996, the NRC staff met with representatives of Duke Power Company (DPC) at NRC Headquarters in Rockville, Maryland, to discuss DPC's review of the draft reports of the Oconee emergency electrical distribution system that were prepared by the Office of Nuclear Reactor Regulation (NRR) and the Office for Analysis and Evaluation of Operational Data (AEOD). is a list of the individuals who attended the meeting and is the handout material that was supplied by DPC.

The purpose of the meeting was to discuss DPC's review of the two draft electrical system reports, one prepared by NRR and the other by AEOD, as requested by letter from William T. Russell, Director, NRR, dated July 8, 1996, to DPC. This meeting was intended to discuss issues raised in the draft reports in preparation for DPC's submittal of a final reviewof the reports that will be used by NRR and AEOD to finalize evaluations of the Oconee emergency power system.

The topics discussed included an overview of the Oconee electrical power system, a description of timelines for emergency power loading for the present design and following a planned modification, an overview of the open issues described in the draft reports, and a discussion of probabilistic risk assessment issues raised in the reports.

9609300313 960926 PDR ADOCK 05000269 P

PDR

-2 contains a matrix of open issues. The first column lists slide.

numbers, the second lists open issue numbers, and third column describes the open issues as defined by DPC from the draft reports. The last column of the matrix lists the corresponding issue numbers from an NRR chart listing the unresolved items from the NRR draft report (for completeness, the NRR chart is included as Enclosure 3).

The matrix is used to describe the open issues referenced at the top of some of the meeting outline pages. The corresponding meeting outline page contains DPC's response to the open issue.

Original signed by:

David E. LaBarge, Senior Project Manager Project Directorate 11-2 Division of Reactor Projects - I/II Office of Nuclear Reactor Regulation Docket Nos. 50-269, 50-270 and 50-287

Enclosures:

As stated (3) cc w/encls:

See next page HARD COPY EMAIL w/o encls. 2&3 Docket File SVarga VMcCree

-PUBLIC JZwolinski BSheron PDII-2 r/f HBerkow MPohida OGC, 0-15B18 DLaBarge JFlack ACRS, TWF LBerry GLainas EJordan JCalvo GTracy HOrnstein RCrlenjak JJaudon EMerschoff EHackett PHarmon BLeFave NSalgado GLanik ASingh To receive a copy of this document, indicate in the box: "C" = Copy without attachment/enclosure "E" = Copy with attachment/enclosure "N" = No copy OFFICE LA:PD22:

E PM:PC 42 D:P I

NAME LBERRY L

DLABA

cdw H E~r DATE q7/

6 I9 / 1(p /96

/ 1

/96

/

/96

/

/96

/

/96 DOCUMENT NAME* kG*\\0CONEE\\MEETING.SUM OFFICIAL RECORD COPY

-2 contains a matrix of open issues. The first column lists slide numbers, the second lists open issue numbers, and third column describes the open issues as defined by DPC from the draft reports. The last column of the matrix lists the corresponding issue numbers from an NRR chart listing the unresolved items from the NRR draft report (for completeness, the NRR chart is included as Enclosure 3).

The matrix is used to describe the open issues referenced at the top of some of the meeting outline pages. The corresponding meeting outline page contains DPC's response to the open issue.

David E. LaBarge, Senior Project Manager Project Directorate 11-2 Division of Reactor Projects -

I/II Office of Nuclear Reactor Regulation Docket Nos. 50-269, 50-270 and 50-287

Enclosures:

As stated (3) cc w/encls: See next page

Duke Power Company Oconee Nuclear Station cc:

Mr. Paul R. Newton Mr. Ed Burchfield Duke Power Company, PB05E Compliance 422 South Church Street Duke Power Company Charlotte, North Carolina 28242-0001 Oconee Nuclear Site P. 0. Box 1439 J. Michael McGarry, III, Esquire Seneca, South Carolina 29679 Winston and Strawn 1400 L Street, NW.

Ms. Karen E. Long Washington, DC 20005 Assistant Attorney General North Carolina.Department of Mr. Robert B. Borsum Justice Babcock & Wilcox P. 0. Box 629 Nuclear Power Division Raleigh, North Carolina 27602 Suite 525 1700 Rockville Pike Mr. G. A. Copp Rockville, Maryland 20852 Licensing -

ECO50 Duke Power Company Manager, LIS 526 South Church Street NUS Corporation Charlotte, North Carolina 28242-0001 2650 McCormick Drive, 3rd Floor Clearwater, Florida 34619-1035 Dayne H. Brown, Director Division of Radiation Protection Senior Resident Inspector North Carolina Department of U. S. Nuclear Regulatory Commission Environment, Health and Route 2, Box 610 Natural Resources Seneca, South Carolina 29678 P. 0. Box 27687 Raleigh, North Carolina 27611-7687 Regional Administrator, Region II U. S. Nuclear Regulatory Commission Mr. J. W. Hampton 101 Marietta Street, NW. Suite 2900 Vice President, Oconee Site Atlanta, Georgia 30323 Duke Power Company P. 0. Box 1439 Max Batavia, Chief Seneca, South Carolina 27679 Bureau of Radiological Health South Carolina Department of Health and Environmental Control 2600 Bull Street Columbia, South Carolina 29201 County Supervisor of Oconee County Walhalla, South Carolina 29621

ATTENDANCE RECORD PURPOSE: To discuss NRR and AEOD Draft DATE:

9/19/96 Reports of the Oconee Emergency Electrical Distribution System with Duke Power Company NAME (Please Print)

AFFILIATION David LaBarge NRC/NRR/PD 11-2 Jim Stoner Duke Power/NGD/ESE George Ridgeway Duke Power/ONS/OPS P. M. Abraham Duke/NGO/PRA H. T. Grant Duke/ONS/ESE W. H. Knight, Jr.

DPC/ONS/OPS Mike Barrett DPC/NGO/PRA Paul Newton DPC/GO Ed Burchfield Duke Power/ONS David Coyle Duke Power/ONS J. W. Hampton Duke Power/ONS M. S. Tuckman Duke Power Jeff Rowell Duke Power Clay A. Little Duke Power Chris Schaeffer Duke Power Greg Robison Duke Power Reed Severance Duke Power/ONS/MSE Paul Harmon NRC, Region II Nancy Salgado NRC, Region II/Region I - Oconee George Lanik NRC/AEOD/RAB Amarjit Singh NRC/ACRS Herb Berkow NRC/NRR Bill LaFave NRR/DSSA/SPLB Ed Hackett NRC/OEDO Victor M. McCree NRC/OEDO Brian Sheron NRC/NRR/DE Marie Pohida NRC/NRR/DSSA/SPSB John H. Flack NRC/NRR/DSSA/SPSB Gus Lainas NRC/NRR/DE Michael Bailey Duke/ONS/Compliance Jose A. Calvo NRC/NRR/DE/EELB Hal Ornstein NRC/AEOD Johns Jaudon NRC/Region II S. Varga NRC ENCLOSURE 1

MeNud Emergcyi.

Pvower Rpr luke-'NRC Manag ement-Meetn September 19, 1996 rn C

(D N)

IMI

Agenda

Overview of ONS Power System

Overview of the Open Issues

Overview Summary

Questions and Answers

PRA Related Issues 2

(NORTH OCW ILLE)

(CTAL)

(CINTRAL)

KO A

EW9 W9y S OVlm CAV B V

233 KV RED W1S AO NWR NRCOS

__~~r IV*4 AD

~1 v v o~R 1 RED JS

] E~14 I?

28 23 55 5a5 6

GIN_

_ _ __2 A UXIIL IARY E L V W r rY E L L V B S P O V E R SSSSEA A

(nL7rV 3T CT3 T

IT

~

~ ~ ~

II

'2 TC200 lisp 699 LA181 CC see~19 7

&A11 I

TiTIN 1818 IA II2ZA 219 H C A L S A

S 0 J 9 C S L A C S T A R TU P S C J 9 C E T

R P S L C EMERGENC POWERE DISTRIBUTIO

Overview-----

ofONS Power System

Oconee Generator through the Normal Auxiliary Transformer

230kV Switchyard through the Startup Transformer

One Keowee Unit is aligned through the Overhead Path and Startup Transformer

One Keowee Unit is aligned through the Underground Path and Transformer CT4

The Dedicated 100 kV Path from a Lee Combustion Turbine through Transformer CT5

Central Switchyard through Transformer CT5

Oconee Unit Startup Transformers can be shared between the Oconee Units (4160V and 6900V) 4

Oconee uclear Station 3 Units LOOP + Keowee in Standby Emergency Power Supply Loading (Current Design)

External Grid Trouble Protection System Initiation Keowee Emergency Start "E" Breakers Close Emergency shutdown loads aligned to Keowee via SI Overhead5Power Path 0510 15 in 5 Seconds O/H ACB Closes PCB 9 Closes O/H Path Energized

Oconee Nuclear Station 1 Unit LOCA/LOOP + 2 Units LOOP Only + Keowee in Standby Emergency Power Supply Loading (Current Design)

Oconee LOCA/LOOP Unit Engineered Safeguards Initiation External Grid Trouble Protection System Initiation "SK" Breakers Close Keowee Emergency Start "S" Breakers Close Emergency shutdown loads aligned to Keowee via Underground Power Path Sin 11 Seconds 0510 15 Load Shed Initiated Oconee LOOP Only Units (2)

External Grid Trouble Protection System Initiation Keowee Emergency Start "E" Breakers Close 0

5 15 in 15 Seconds 0/H P ACB Closes PCB 9 Closes 0/H Path Energized

AV Oiverview of the Open Issues

NRC cover letter question on Oconee plant response to a 3 unit LOOP event

Open issues addressed by changes in testing, hardware, procedures and technical specifications

Open issues addressed by current programs or additional information 7

P_ ant Response Following a3 Unit LOOP (Open Issue #1)

A LOOP-induced overcooling event which results in ES actuation is not expected to occur for the Oconee units

>> Steam loads following a LOOP should not result in a significant overcooling event

>> Operator action to throttle emergency feedwater will limit the extent of overcooling 8

o.

Lant Response Following a 3 Unit LOOP (Continued)

(Open Issue #1)

For a LOOP-induced three unit ES actuation:

>> The significant load from a subsequent ES actuation would be an additional HPI pump per unit

>> Duke's loading calculations include the motor driven EFW pumps in the initial LOOP loads

>> This condition is bounded by the current loading analyses for LOCA/LOOP events 9

Open Issues Addressed by Changes to:

Hardware Testing Procedures Technical Specifications 10

Modification to Delay Loading of the

'Keowee Uits (Open Issue #2)

Duke previously committed to a modification to delay loading of Keowee until voltage and frequency reaches approximately 90 percent of the nominal ratings

Voltage and frequency relays will be placed on each Keowee unit

>> add permissives to overhead ACBs (2 of 3 logic)

>> add permissives to underground SK breakers (2 of 3 logic)

Increased delay in loading ECCS equipment is still bounded by the current LOCA analysis 11

OCONEE NUCLEAR STATION Emergency Power System 230/525KV Keowee Lee Steam Station Switchyards Hydroelectric Combustion Turbine IIIIIIII IIIIIIIIIIIIIII IIIIIIIIIiiIIIIIIiiiIIII IG enerators G enerators

4 Double Circuit 230KV Lines

3 525KV Lines Red Bus C

Yellow Bus Keowee............

Overhead IIlIIIIIIIINIIlIIUIIIIIIIIlIIIIiIIII lIIIIIIIIIIII IIIIIIIIIII LJn ACB 1 ACB2 oil.........................

C 81 27/59 27/59 81 OveKr K2 head ACB3 ACB4 I

Keowee Underground Feeder CT4 Ii',

~

ilISTANDBY S~i I~

S2 S~

SL2BUSES SB1 SB2 UNIT TO UNIT 2 TO UNIT 3 1iT CT1 MAIN FEEDER MAIN FEEDER BUSES BUSES F F UNIT 1 N2E2 S

MAIN FEEDER S2 BUSES MFBI MFB2

Oconee sear Station 1 Unit LOCA/LOOP + 2 Units LOOP Only + Keowee in Standby Emergency Power Supply Loading (New Logic)

Oconee LOCA/LOOP Unit Engineered Safeguards Initiation Keowee Voltage & Frequency - Nominal External Grid Trouble Protection System Initiation "SK" Breakers Close Keowee Emergency Start "S" Breakers Close Emergency shutdown loads aligned to Keowee via Underground Power Path in ~ 16 Seconds 05 10 15 Load Shed Initiated Oconee LOOP Only Units (2)

Keowee Voltage & Frequency - Nominal External Grid Trouble Protection System Initiation O/H ACB Closes Keowee Emergency Start "E" Breakers Close Emergency shutdown loads aligned to Keowee via Overhead Power Path 0

15 in ~ 16 Seconds PCB 9 Closes

V ol tage and Frequency Protection for Keowee (Open Issue #5)

The emergency power system is protected against credible governor or voltage regulator failures

As an additional defense in depth, Duke concurs with the NRC that there are potential benefits from additional frequency/voltage protection

Additional over and under voltage and frequency protection will be placed on each Keowee unit during emergency operation

> 2 of 3 logic with time delay

>> Logic trips and blocks overhead ACBs and underground SK breakers

>> If actuated, logic must be manually reset 14

Voitage and Freq uency rotecti

'Lee (Open Issue #6)

As an additional defense in depth, Duke concurs with the NRC that there are potential benefits from additional voltage/frequency protection

Over and under voltage and frequency protection will be placed on the dedicated Lee path at Oconee

> 2 of 3 logic with time delay

>> Logic trips standby bus Lee breakers (SL) 15

Emergency Power Syste ing and Analysis (Open Issues 3 and 4)

Duke believes that the existing overlapping tests, supported by calculations, are adequate to ensure proper operation of safety equipment

Enhancements are being implemented to increase design margins 16

Emergency Power System Testing and Analysis (Continued)

(Open Issues 3 and 4)

A Pre-operational integrated ES test was performed on each ONS unit as noted below:

>> Simulated LOCA/LOOP

>> Keowee started from standby and accepted Oconee Unit ES loads on underground path while accelerating

>> The test verified that ES pumps reached rated flow and that ES valves and other components attained their ES state within the required timeframe

In addition, a separate test step-loaded a Lee CT with individual loads 17

Emergency Power Syst ing and Analysis (Continued)

(Open Issues 3 and 4)

To provide additional assurance, Duke will perform a one-time integrated ES test to demonstrate the adequacy of the current testing program

The following slides provide the preliminary scope of the integrated test 18

Emergency Power System Testing and Analysis (Contnued)

(Open Issues 3 and 4)

Overview of Integrated ES Test Keowee portion

Both Keowee units will be emergency started by an ES signal

Underground Keowee unit will accept ES loads of shutdown ONS unit

Simulated loads will be applied to the underground Keowee unit to simulate LOOP loads from two Oconee units

Overhead Keowee unit will start and remain in standby Lee portion

A Lee gas turbine will energize the standby buses prior to the initiation of the test 19

Emergency Power System Testing

criu and Analysis (Continued)

(Open Issue 3 and 4)

Preliminary Test Description

Maintenance work on affected systems completed

Voltage and frequency protection modifications completed

All AC power sources available

ES systems appropriately aligned for the test 20

Emergency Power System Testing and Analysis (Continued)

(Open Issue 3 and 4)

Preliminary Test Description (continued)

Initiate ES signal

Open switchyard breakers for the startup transformer to simulate LOOP

Load the Keowee underground unit with the ES loads followed by loading simulated LOOP test loads at the appropriate time

This same test will be repeated with a Lee CT as the source 21

E ergency Power System Testing and Analysis (Contin ed)

(Open Issues 3 and 4)

Preliminary Test Acceptance Criteria

Keowee underground unit/Lee gas turbine energizes the main feeder buses

Affected motors start and run until stopped

ES MOVs perform a full stroke 22

Emergency Power System Testing and Analysis (Conti (Open Issues 3 and 4)

Risks/Costs

Decay heat removal will be briefly interrupted

Other Oconee Units above 2000 F will enter LCO for the underground path being out of service

The total cost of the test will be approximately one million dollars 23

Emergency Power System Testing and Analysis (Continued)

(Open Issues 3 and 4)

Duke believes this test will demonstrate the adequacy of our current testing program and provide added assurance that the emergency power system can perform its intended safety function.

24

.Procedures for Paralleling Auxiliary and Startup Transformers (Open Issue #11)

Operating procedures have a precaution about the paralleling of the transformers for an extended period

In addition, cautions have been placed in the body of the procedures for transferring auxiliaries 25

£o 0

0 Procedures for Voltage Regulator Op eration. and Mai.ntenance (Open Issue #45)

Maintenance procedures associated with calibration of the Keowee voltage regulator have been developed

Training on the voltage regulator has been provided to the maintenance technicians

The voltage regulators are currently being calibrated

Keowee Emergency Start abnormal procedure will be revised to include instructions to ensure proper operation of the voltage regulator 26

Procedures for Monitoring the SSF Batteries (Open Issue #42)

Power to the SSF battery chargers can be restored by

>> the Unit 2 MFB, once re-energized

>> starting the SSF diesel generator

SSF EOP requires operators to verify battery, charger, and inverter operation

Actions to verify SSF battery charger operation are being incorporated in the Oconee Unit 2 Loss of Power Abnormal Procedure 27

Procedures for Recovery of Switchyard Battery Chargers (Open Issue #12)

Loss of power abnormal procedures will contain instructions for the restoration of the switchyard battery chargers 28

Tech Spec for Alignment of Lee for Underground Path Maintenance (Open Issue #23)

Technical Specifications allow 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months for outages of the underground path

Underground power path unavailability is minimized by scheduling Keowee unit maintenance with the unit aligned to the overhead path

Given the importance of the underground power path, Duke will propose a Technical Specification to require energization of the standby buses from Lee for outages of the underground power path in excess of 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months 29

Tech Spec for Lee Step Load Test (Open Issue #7)

The proposal to remove the Lee step load test from the Technical Specifications was based on the existing overlapping tests

Duke believes integrated testing is preferred when practical

Therefore, the planned supplement to the Technical Specification 3.7 rewrite will include the Lee step load test in the Technical Specifications 30

Open Issues Addressed by:

Current Programs Additional Information 31

Periodic Testing of the Emergency Power System (Open Issue #44)

Duke believes that the Oconee emergency power system is at least as reliable as a plant with diesel generators, as is shown by the Keowee PRA

The emergency power system is tested periodically in accordance with the Technical Specifications and Selected Licensee Commitments

Operator training and periodic testing ensure that the operators obtain the necessary experience on operation of the emergency power system

Periodic testing and maintenance ensures that the reliability of the emergency power system is maintained at a level commensurate with the Keowee PRA 32

Load Run Test of Keowee (Open Issue #29)

The monthly synchronization test of Keowee is not similar to the monthly load run test required in the standard technical specifications

In preparation for conversion to the standard technical specifications, a Load Run Test per Reg Guide 1.9 was conducted in August of this year

Preliminary analysis of the test results indicate that grid generation is similar to the monthly load runs required by the standard technical specifications

The results of this analysis will be used in the conversion to standard technical specifications to determine if a monthly load run test is necessary 33

rl, S Monthly Start of the Keowee Units (Open Issue # 9)

On average, 4 emergency starts of Keowee occur each year due to the following tests

>> Keowee Emergency Start Test

>> Degraded Grid Functional Test

>> EPSL Functional Test

>> Proposed Keowee Load Rejection Test

Additional emergency start tests could cause unnecessary wear on the machinery.

34

Ae Monthly Start of the Keowee Units (Continuedt)

(Open Issue # 9)

The monthly start test of Keowee is performed by the normal start circuit

Most components that are needed to function during an emergency start also function during a normal start

Failure of these components would be detected by grid generation and the monthly normal start of Keowee

Components not covered by normal start are exercised during the tests listed on the previous slide

Duke believes this testing is adequate and additional testing is not warranted 35

Partial Load Rejection Test for Keowee (Open Issue #10)

The 5MW load rejection performed in the EPSL Functional Test is not intended to verify Keowee's ability to perform following a partial load rejection

Keowee governor control is verified during the following load rejection tests Degraded Grid/Switchyard Isolation Test Proposed Keowee Load Rejection Test

Keowee stability following a partial load rejection was demonstrated during a one-time test

Additional partial load rejection tests are not considered necessary 36

Periodic Testing of Keowee ACBs (Open Issue #18)

Automatic closing circuits were installed on the Keowee underground ACBs

This function is periodically tested per the ONS Technical Specifications

In addition, overfrequency close permissives were installed on the Keowee overhead and underground ACBs

A proposed Technical Specification amendment has been submitted for a periodic surveillance of the close permissive 37

P_ eriodic Load Rjection Testi f

Keowee (Open Issue #20)

Proposed Technical Specifications for periodic load rejection tests were submitted in February of this year

Periodic load rejection tests are currently performed

>> Degraded Grid Functional Test

>> EPSL Functional Test

>> Proposed Keowee Load Rejection Test

The above tests result in approximately 4 Keowee load rejection tests per year 38

.......... SSF Testing Issue~s (Open Issue #33)

SSF testing issues will be resolved by the following tests

>> SSF diesel generator 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months run test which was conducted during the first week of September 1996

- This test was successful in further demonstrating that the SSF could perform its intended safety function during an SSF event

>> Integrated SSF RC makeup test to validate the model used to determine the flow balance of the makeup system 39

DIG Lube Oil Temp 195

-+- Eng A Lube Oil Inlet

--a-Eng A Lube Oil Out

-A-Eng B Lube Oil Inlet 185 x*- Eng B Lube Oil Out 175 U

0 165 155 145 9/3/96 9/3/96 9/3/96 9/3/96 9/4/96 9/4/96 9/4/96 9/4/96 9/4/96 14:00 17:00 20:00 23:00 2:00 5:00 8:00 11:00 14:00 Time Diesel Started at 14:15

SSF HVAC Results from 24-Hour Diesel Run Test Maximum Allowable Room Temperatures:

-Diesel Generator Room = 125

-Pump, HVAC, Switchgear Rooms = 120

-Control Room = 95 100 95 90 85 80 Ambient r

70 HVAC Rm E

-&-Pump Rm

-E-Diesel Rm

---.- Sw itch R m

................................ A/C Return 60

....x.... A/C Supply 55 50 -

1 9/3/96 9/3/96 9/3/96 9/3/96 9/3/96 9/4/96 9/4/96 9/4/96 9/4/96 9/4/96 9/4/96 9/4/96 9/4/96 14:00 16:00 18:00 20:00 22:00 0:00 2:00 4:00 6:00 8:00 10:00 12:00 14:00 Diesel Started at 14:15 Time

SSF DJW HX TEMP CHANGE DURING 24 HOUR TEST 105 10 0 95 ILL 0

90 o-Hx A Inlet Hx A Outlet

-+-

Hx B Outlet 85 80 9/3/96 9/3/96 9/3/96 9/4/96 9/4/96 9/4/96 9/4/96 9/4/96 9/4/96 15:00 18:00 21:00 0:00 3:00 6:00 9:00 12:00 15:00 TIME Diesel Started at 14:15

Integrated Testi of the SSF'RC Makeup System (Open Issue #8)

An integrated test of the SSF RC makeup system is being developed for each Oconee unit

Makeup water will be collected at each reactor coolant pump to verify that balanced flow is received

The entire flow path will be tested except for the portion between the seal injection point and the RCS, which has a negligible pressure drop

In addition, our planned test will not require the Oconee unit to be placed in natural circulation 43

Keowee ACBs in-the Maintenance Rule (Open Issue #19)

The Keowee emergency power source is included in the Maintenance Rule

Reliability of the Keowee ACBs will be monitored by the impact on the associated Keowee unit or emergency path 44

<,uede ejowee oad ejecton Testi m

the Main-tenance'Rul~e (Open Issue #21)

The Keowee emergency power source is included in the Maintenance Rule

Failure of a Keowee load rejection test will be monitored by the impact on the associated Keowee unit availability 45

Lee Combustion Turbines in the Maintenance le (Open Issue #16)

Lee combustion turbines are tracked in the maintenance rule and are considered unavailable if:

>> Less than two Lee CTs are available to supply power within one hour

>> Dedicated path from Lee is not available within one hour

> A Lee CT is lost while supplying power to CT5

Duke's programs are adequate to monitor the reliability of the Lee CTs 46

Dual Keowee Unit Maintenance (Open Issue #17)

The Keowee PRA recommended that a formal review of severe weather be performed for dual Keowee unit outages that are expected to last longer than one shift

Operations procedures for scheduling dual Keowee unit maintenance were revised to considered severe weather potential

Additional Technical Specifications are not warranted 47

Periodic CIeaning of Keowee trainers (Open Issue #24)

Keowee generator cooling water strainers are currently cleaned on a frequency of every other month

No clogging of the strainers has occurred except for one isolated occasion in 1983

Clogging of the strainers will not trip the Keowee unit in an emergency, but will be alarmed in the Keowee Control Room

The strainer cleaning process is contained in a Keowee procedure

Based on the Duke surveillance program, procedures, and operating experience, addition of the periodic strainer cleaning program to the Technical Specifications is not warranted 48

O erator Procedures and Traini g Upgrades (Open Issue #40)

Following the Oconee Unit 2 1992 LOOP event, Oconee recognized that operator procedures and training needed to be upgraded

Corrective actions were initiated to prevent the recurrence of the problems experienced during the 1992 LOOP event

All of the corrective actions associated with the 1992 LOOP event have been completed 49

Operator Procedures and Training Upgrades (Continued)

(Open Issue #40)

Keowee Emergency Start abnormal procedure was written to address potential problems during an emergency start of the Keowee units

Oconee Loss of Power abnormal procedure has been revised to include failure contingencies

Training and Job Performance Measures are established for the abnormal procedures

Keowee operators are trained and qualified

The Oconee Simulator replicates the emergency power system in a functional manner

Simulator training allows operators to perform all control room actions required by the procedures used to mitigate a loss of power event 50

S SF Activation within 10 minutes (Open Issue #13)

ONS has an aggressive program in place to ensure that the SSF can be activated within 10 minutes

SSF activation is trained and tested in the license preparatory classes and requalification classes

Unannounced SSF activation drills are performed for operations personnel while on shift

SSF activation is tested during Appendix R drills for emergency preparedness 51

Keowee Technical Specialist Onsite (Open Issue #14)

Keowee recovery actions have been proceduralized in the Keowee Emergency Start Abnormal Procedure

Keowee operators are trained on the use of the procedure to recover the Keowee units

Oconee operators are also trained on the Keowee Emergency Start Abnormal Procedure

An additional program that requires a Keowee Technical Specialist onsite prior to severe weather is not warranted 52

A.,o Minimum Keowee Voltage For ONS eer1Load Oeration (Open Issue #28)

Several Keowee voltage levels have been evaluated for different specific failure modes resulting in different acceptable values for each mode

The minimal Keowee terminal voltage analyzed for all bounding scenarios is 13.5kV

This value is verified during the various periodic tests 53

Fault Current Ratings of the RCP Breakers (Open Issue #43)

During a LOCA/LOOP, the 6.9kV startup source breakers which trip to isolate the overhead path are not overdutied for any fault

Therefore, the emergency power system is protected from faults on the 6.9kV switchgear

All 6.9kV breakers are within their interrupting capability for faults at the electrical penetration and motor termination points

For low probability 3-phase bolted faults at the switchgear, the non-startup source 6.9kV breakers are slightly overdutied

Based on our cost benefit analysis, Duke believes replacement of these breakers is not warranted 54

riutV Open PR A Issues

The following open issues will be addressed in the detailed written response

>> Lee gas turbine reliability

>> Keowee reliability

>> IPEEE submittal 55

Overview Summary

Duke believes that the independent NRC assessment was a thorough review of the design, operation, and testing of the Oconee emergency power system

Based on our analysis of the NRC recommendations, Duke believes Oconee is operating within its licensing basis

Duke believes factual clarifications are necessary before the final NRC reports are issued

Duke is committed to work with the NRC to achieve closure on all open issues 56

Emerge er Report Open atrix Open

-NRC Slide Open Issue Open Issue Chart Number Number Number A question was raised during NRC Management review of NRC draft reports. This question covered the Oconee plant response to a three 0-08 1

unit LOOP. NRC wanted to know if an overcooling event on 3 units led n/a to ES actuation and unanalyzed loading of Keowee.

1, 37, The NRC staff will review the particulars of the modification to 35 0-11 2

delay loading of the Keowee units when the details are available.

39 52, 62 A Keowee voltage regulator or governor failure could result in an out 6, 35, of tolerance voltage or frequency. Therefore, protection should be 36, 38, 0-14 5

installed such that operator action is not necessary to mitigate a 61, 70, governor or voltage regulator failure.

71 A Lee voltage regulator or governor failure could result in an out-of 7, 8, tolerance voltage or frequency. Therefore, protection should be 9, 54, installed such that operator action is not necessary to mitigate a 72, 73, governor or voltage regulator failure.

74 4, 5, Duke calcs have a small margin. These calcs need to be reviewed by 10, 53, the NRC.

If the staff cannot determine the accuracy of the calcs and 66, 67 0-16 3

overload protective devices to a reasonable degree, then a more fully integrated Keowee test needs to be performed.

Pagel

Emerge er Report Ope~n e Matrix 4, 10, Duke calcs have a small margin. These calcs need to be reviewed by 55, 66, the NRC.

If the staff cannot determine the accuracy of the calcs and 67 0-16 4overload protective devices to a reasonable degree, then a more fully integrated Lee test needs to be performed.

NRC staff could not determine what corrective actions were taken after a bus connection overheated in 1980. NRC recommends that 14, 60, 0-25 11 procedures for each ONS unit prevent paralleling of the auxiliary and 77 startup transformers except for brief power transfers.

The NRC staff believes that the operations and maintenance training and procedures for the voltage regulator need to be improved, along n/a 0-26 4

with additional testing of the system due to recent errors with the voltage regulator settings.

Operator action is required to restore power to the SSF battery 0-27 42 charger following an ONS Unit 2 LOOP event. NRC recommends that n/a Oconee control room operators be provided with instrumentation and procedures to monitor and respond to a degraded SSF battery.

For an event which automatically results in the load shedding of the Oconee switchyard battery chargers and discharging of the switchyard 15, 57, 0-28 12 batteries, the NRC staff believes restoration of the chargers should 78 be included in the emergency recovery procedure.

Page 2

Emerge er Report Open e Matrix Since the availability of the underground path from Keowee is risk significant, the staff recommends that Duke consider a Technical 18, 25, Specification to pre-align Lee during any CT4 maintenance activity, 118 as is currently stipulated for dual Keowee maintenance.

NRC staff believes that Duke should not delete the Lee gas turbine 0-30 7

step load test which energizes Oconee loads from the ONS Tech Specs.

This is the only step load test of the Lee gas turbines even though the loading is not the same as-during an accident.

Periodic testing of a LOOP would allow ONS to obtain a level of 0-32 44 safety comparable with D/G plants. Also, the operators do not routinely obtain experience with delivering power to emergency loads as occurs once per refueling interval at D/G plants.

The NRC staff will verify with Duke that the monthly load test of 0-33 29 Keowee is similar to the standard tech spec diesel generator monthly 45 load run test.

A normal start test may not detect failures that could affect the 0-34 9

emergency start sequence.

The NRC staff therefore believes the 12, 44, Keowee units should be tested on a monthly frequency by using an 47, 75 emergency start signal to initiate the monthly start test.

The 5 MW load used in the EPSL test does not bound the maximum 036 10 partial load that might be rejected during a three unit LOOP event.

13, 49, A bounding test should be provided to assure that the single load 76 rejection limits would not be violated.

Page 3

Emerge er Report Open e Matrix The NRC staff believes that periodic testing of the ACBs 18, 22, 0-37 18 (particularly ACBs 3 and 4) and their new function and associated

101, logic should be performed.

120 18, 23, The NRC staff believes that a periodic Keowee load rejection test

102, O-38 20 (hot start test) should be performed.

119, 121 Since the validity of the SSF data in the Oconee IPE is dependent on 0-39 33 the adequacy and scope of tests, any testing issues associated with n/a the SSF should be resolved.

11, 16, The NRC staff has concerns that the planned SSF RC makeup flow 11, 16, 68, 80, 0-43 8

testing will not provide sufficient data to verify balanced makeup 83, 84 flow. The licensee should develop more detailed and meaningful 8 5, 86, acceptance criteria for these tests and submit for NRC review.

85 88 87, 88 Keowee ACB reliability goals should be established consistent with 18, 22, 0-44 19 the PRA numbers. ACB performance should be tracked and monitored 120 against these goals under the maintenance rule.

The NRC staff believes that Duke should be sensitive to-the Keowee load rejection reliability. Goals should be established at the values used in the Keowee PRA and performance monitored against these goals in the maintenance rule.

Page 4

Emerge er Report Open eMatrix Due to the importance of Lee in the emergency ac power system for all 18, 95, three Oconee units, the staff believes reliability goals should

115, 0-46 16 established for the CT5 power components, and performance should be

116, tracked and monitored under the maintenance rule.

131 The staff believes that the modifier used in the Keowee PRA for severe weather is acceptable provided the scheduling guidance for 18, 21, O-47 17 dual Keowee unit outages is included as a program in the Oconee Tech 97, 117 Specs.

Clogging of the common intake cooling water header is a dominant contributor to failure of Keowee. Thus, the NRC staff recommends 18, 26, 0-48 24 that Duke consider adding a periodic tech spec surveillance to reduce 105 the potential for clogging of these strainers.

Oconee operators may not have sufficient understanding of the system to base.operating decisions, compared to a plant with D/Gs.

Therefore, the NRC believes that operator procedures and training need to be upgraded for emergency power system operators.

Given the potential for Oconee Unit 1 RCP seal failures after 10 minutes without seal injection, the NRC staff concluded that 17, 83, 0-51 13 additional emphasis needs to be placed on operator training and 89 drills for SSF activation.

Given the importance of the technical specialist during periods of severe weather, the NRC staff believes that a Keowee technical 18, 19, 0-52 14 specialist should be stationed at the site in advance, whenever 91, 106 severe weather is predicted to occur.

Page 5

Emerge ewer Report Open e Matrix Various voltages have been shown to be adequate for proper operation of the emergency loads from Keowee in Duke's analyses. The staff will pursue the question of the minimum Keowee voltage needed for acceptable operation of the Oconee loads.

The circuit breakers supplying the RCP switchgear have insufficient 0-54 43 fault current ratings. The NRC staff feels that circuit breakers with n/a sufficient fault current ratings should be installed.

18, 92, The staff feels that the unavailability of the Lee gas turbines in 93, 94, the Keowee PRA may be underestimated due to questionable reliability 99, P-03 15 assumptions for a grid related LOOP. DPC's position is that a grid-

112, related LOOP does not fail the entire 100kV system.

113, 114 The NRC staff believes that the licensee should evaluate potential 18, 24, mechanisms for a single common cause grid degradation failing both 1,

P-07 22

103, Keowee power paths when generating to the grid. If dual unit 122 generation increases, the common cause failure increases.

In the Keowee PRA, components with different demand frequencies are 18, 27, given the same failure probability. The licensee should consider the

104, potential for differences in the failure probabilities of components

123, with widely varying test/demand frequencies.

133 Page6

Emerge r Report Open e Matrix The staff believes the point estimate provided by the sensitivity study which uses the 50% upper confidence interval values from the Chi-Squared distribution with 2 degrees of freedom is a more accurate estimation of Keowee reliability.

The AC power CDF for internal events is very sensitive to the SSF.

In light of its importance, the staff believes that the SSF failure probabilities used for the IPE update (which is now being performed by DPC) should be examined thoroughly.

The NRC staff believes that interpretation of the Keowee PRA results

107, should be based on generic data rather than plant specific bayesian-

108, updated data. Most of the plant specific data appear to be lower 132 than their generic counterparts.

There are some subtle differences associated with applying grid generation data to emergency demands. The Keowee PRA does not

108, P-12 35 provide justification for the applicability of grid generation data 134 to emergency demand situations for these cases.

The failure probability of Keowee given a plant-centered LOOP or a severe weather LOOP is equal to the failure probability of the underground path and its aligned Keowee unit. These failure probabilities were not explicitly discussed in the Keowee PRA.

The Keowee PRA was performed under a less stringent framework than P-14 37 typical plant PRAs. The staff and licensee should be sensitive to changes in operating practices, test frequencies, time in configurations, etc., and their impact on the PRA.

Examination of the internal flood indicates that Keowee or Lee power could be credited through the internal events electric power fault P-16 39 129 trees. It is assumed that the.power supplies were included but unclear what impact they had on the results.

Page 7

Emerge er Report Oen Matrnx Past vulnerabilities associated with design, testing and operation would have rendered Keowee inoperable. The unavailability due to these conditions were not included in Keowee PRA calculations since the Keowee model reflects the current condition.

Because the staff believes that the unreliability of CT5 power is potentially underestimated, the validity of the comparison of the P-18 46 98 total Oconee emergency power system to two 100 percent EDGs is questionable.

Some components that are only challenged every refueling outage are applied the same failure probability as components that are demanded 109 daily. Therefore, the results of the generic data sensitivity study are considered more robust than the base case.

Undeveloped event failure rates were calculated with Keowee plant P..21 48 specific data and used in the base case and the generic data 110 sensitivity study. Because of this, the generic data result is not a completely generic data assessment.

Since external event and severe weather event uncertainties can be P-23 26 large, the NRC staff believes that these uncertainties should be

135, taken into consideration when making decisions regarding the 136 availability of the Oconee emergency AC power system.

The licensee's IPEEE focused on catastrophic external events that result in widespread destruction of equipment, particularly the P-25 27 common turbine building. The contribution of less significant but 130 more likely external event scenarios may be misrepresented.

Page 8

Emerge er Report en e

atrix The seismic cutset frequencies in Table 3-4 of the IPEEE will need to P-26 38 be explored further; the cutsets appear to total 1.7E-04, not 3.6E-127

05.

The seismic fragility curves relating equipment failure probability P-27 49 versus ground acceleration will need to be reviewed further to assess 126 the reasonableness of their current estimates.

Page 9

PRA Related Open Issues

Lee Reliability

Keowee Reliability

IPEEE 1

Lee Reliability

Impact of Lee for various loss of offsite power initiators (01 #15)

CT5 modeling detail (01 #15) 2

Lee Gas Turbine Reliability (Open Issue #15)

The loss of power event analyzed is consistent with the assumption that all 3 Lee CTs are available

The frequency of a LOOP event which causes a concurrent power outage at Lee is an order of magnitude less than the 0.0273/yr value assumed in the Keowee PRA

Additional analysis assuming a Lee failure probability of 0.5 produced only a small increase in CMF (5.3 E-7) 3 CT5 Modeling Detail (Open Issue #15)

Oconee IPE model included Lee run failure (for dual Keowee maintenance)

Recovery value applied for other situations

Oconee PRA is currently being revised

The new model contains specific events for Lee start and run failures, CT5 maintenance events, and human errors 4

Keowee Reliability

Both units generating to the grid (01 #22)

Frequently versus infrequently tested/demanded components (01 #25)

Failure rate calculation when no failures have been observed (01 #31)

SSF failure probability (01 #32)

Generic data results versus plant specific data (01 #34)

Justification for use of the grid generation data (01

35)

Impact of various loss of offsite power initiators (OI

36)

Keowee Reliability (continued)

Keowee PRA data assumptions (01 #37)

Internal flood analysis (01 #39)

Consideration of past unavailability in the Keowee PRA (0I #41)

Comparison of Oconee emergency power system to diesel generators (01 #46)

Demand frequencies for Keowee data (01 #47)

Generic data sensitivity study for undeveloped events (01 #48) 6

Dual Keowee Unit Grid Generation (Open Issue #22)

Percentage of the time Keowee can generate is limited by the local water flow and rainfall

Dual Keowee unit generation is not expected to increase significantly above the value in the Keowee PRA

Protective devices exist in the switchyard and at Keowee to isolate Keowee from any grid disturbances

Multiple failures of these protective devices are necessary to cause the postulated failure of Keowee.

Additional analysis using a range of values for the common cause failure probability indicates substantial CMF margin 7

Failure Rates for Infrequently vs Frequently Used Components (Open Issue #25)

The Keowee PRA combines component demands from power generation and emergency operation to determine failure rates

This is consistent with the professional standards on data analysis and generic failure rates

Plant-specific failure data is not large enough to provide a statistically significant distinction on failure rates based on frequency of demand 8

0Chi-S uare Van_

__te

_ncertan (Open Issue #31)

Duke believes that the method suggested by the NRC staff is conservative

Also, Duke believes the method used in the Keowee PRA provides a best estimate of the failure rate for equipment which has no observed failures

NUREG/CR-4407 identifies the method used in the Keowee PRA as a more realistic point estimate for cases of zero equipment failures

The Keowee PRA result is not sensitive to the method suggested by the staff (increase <3%)

9 SSI Faiure c

Poabi Litie (Open Issue #32)

The LOOP induced CDF would increase from 1E-06/yr to 6E-06/yr if the SSF were assumed unavailable for the year

The SSF failure modes consist of human error, maintenance unavailability, and hardware failure

Failure probabilities for the components that influence the core damage risk were assigned using operational data from 1988-1993 10

v. ? Generic Versus Plant Specific Data (Open Issue #34)

Duke believes that the use of plant specific data in the Keowee PRA is more meaningful for determination of Keowee reliability than ignoring it

Available generic data was combined with the plant specific data using the Bayesian update technique

Therefore, the base case results represent the combined effect of the generic and plant specific data

An analysis of the PRA models using the generic data instead of the Bayesian updated data was performed and the Keowee PRA results were not appreciably altered Justification For Use of Grid Generation Data in Keowee PRA (Open Issue #35)

In the Keowee PRA, failure rates are calculated by dividing the number of failures by the number of demands or hours in service

The above classical equation does not differentiate between a failure during an emergency actuation or normal actuation

System reliability is determined by evaluating the failure rates of the individual components and assigning appropriate exposure times 12

Keowee LOOP Failure Probability (Open Issue #36)

The frequency of the three LOOP initiators and the conditional failure probability of Keowee is provided below

The failure probability reported in the Keowee PRA is 0.0074

LOOP Event Event Freq.

Failure Prob.

Grid 0.0273/yr 0.0074 Switchyard 0.0486/yr 0.0086 Weather 0.0154/yr 0.0086 13 Keowee PRA Data Assumptions (Open Issue #37)

The Keowee units are started regularly and exposure times are established in accordance with the actual system operation (including grid generation and testing)

In a typical PRA, Technical Specifications provide the basis for exposure times on standby systems where testing provides the only demands 14

Keowee PRA Data Assumptions (continued)

(Open Issue #37)

Components not challenged during the normal start (e.g., time delay relays on ACBs 1 and 2) are assigned exposure times consistent with the Tech Spec surveillance

The framework in which the Keowee PRA was performed is consistent with the methodology used in other PRAs 15 Internal Flood Analysis of Oconee (Open Issue #39)

Random failures of the AC power system are included in the IPE analysis for internal floods

No equipment in the AC power system is directly impacted by the Turbine Building basement flood and the AC power system is quite reliable

Consequently, AC power failures have little impact on the CDF which results from internal floods 16

C~~'Consideration of Past Unavailability in the Keowee PRA (Open Issue #41)

The past events that resulted in LERs are included in the Keowee PRA if the event was the result of an actual failure

LERs that resulted from postulated scenarios where the design basis was not met are not considered because of the following reasons

- Keowee was still functional 1993 assessment indicates little core damage significance with past emergency power LERs

Past conditions in the LERs have been corrected and do not impact the current reliability of the emergency power system 17 Comparison of Oconee Emergency Power to Diesel Generators (Open Issue #46)

The IPE results in NUREG-0149 for CDF for SBO events for PWRs range from 1E-07/yr to 7E-05/yr with an average of 1E-05/yr

The Oconee CDF is calculated to be 1E-06/yr

Even with a 0.5 assumed failure probability of Lee, the Oconee SBO CDF is still in the low end of the above range 18 0

Comparison of Oconee Emergency Power to Diesel Generators (continued)

(Open Issue #46)

Using data from INEL-95/0035, the failure probability for a 2 train D/G system is 0.0225

Keowee failure probability is 0.0074 (0.0086)

Indicates good overall reliability of the total Oconee emergency power system as compared to the typical D/G system

Keowee sensitivity study results 0.01 base case data w/o operator recoveries 0.01 generic data with operator recoveries 0.013 generic data w/o operator recoveries 19 Demand Frequencies For The Keowee Data (Open Issue #47)

Duke agrees that for ideal and precise calculation of a component's reliability the failure rate data should come from identical components in identical conditions

Statistical data sources employ simplifying assumptions and consideration of equivalence

Generic data from the data sources is an aggregate of data from many plants with different but similar conditions

Therefore, use of data for like components with different demand frequencies is an acceptable PRA methodology 20 0

SGeneric Data Sensitivity Study For Undeveloped Events (Open Issue #48)

In the Keowee PRA, a sensitivity study was performed employing the generic failure rates

However, for approximately 5% of the basic events no generic failure rates were available

For this small fraction of the events the plant-specific data was used

Therefore, the generic study result is not a pure generic data sensitivity study 21 IPEEE

Uncertainty (01 #26)

Catastrophic versus more probable events (01 #27)

Seismic cutsets (01 #38)

Seismic fragility curves (01 #49) 22

Unavailability Decisions for the Emergency Power System (Open Issue #26)

Duke agrees that large uncertainties exist in the probabilities associated with the occurrence of catastrophic external events

Duke agrees that large uncertainties exist in the equipment response during catastrophic external events

Decisions on the unavailability of both Keowee units or the underground path consider the potential of severe weather 23 Unavailability Decisions for the Emergency Power System (continued)

(Open Issue #26)

Recovery actions for catastrophic external events are covered in abnormal procedures

Conservative assumptions are generally made with respect to the availability of Keowee overhead power and CT5 power during external events 24

Limitation of IPEEStudy (Open Issue #27)

External events from the industry standards and the IPEEE NRC guidance were reviewed for applicability

Specific events were selected in the IPEEE analysis because of their potential to affect many plant systems

Many hazards were eliminated due to inapplicability

Other hazards were screened out due to their extremely low probability of occurrence

Detailed analysis was performed for the events judged to be important contributors to the CDF

Duke believes all external events of significance for Oconee have been adequately considered 25 Seismic Cutset Frequencies (Open Issue #38)

The IPEEE submittal calculated the sampled mean seismically induced core melt frequency

The sampled mean seismic core melt frequency results in a value that is different from the summation of the seismic cutsets 26 0

Seismic Fragility Curves (Open Issue #49)

The fragilities for the Oconee components in the IPEEE have been enhanced

- Original Oconee equipment fragilities were determined in 1981 as part of the NSAC-60 study

- The fragilities of some equipment were updated in 1986 based oh block wall testing

- Incorporated plant walkdown and anchorage reviews into the fragilities 27 Seismic Fragility Curves (continued)

(Open Issue #49)

- Used information compiled in the earthquake experience database over the past several years

- Revised the fragilities to the currently recommended Uniform Hazard Spectra 28

FOR INTERNAL USE ONLY OCONEE EMERGENCY ELECTRICAL SYSTEMS

- DRAFT REPORT UNRESOLVED ISSUES_

ITEM PAGE DESCRIPTION RESOLVED NO.

NO.

1 ES-3 Review Duke commitment to modify permissives on emergency power path to delay loading of Keowee units until voltage and frequency are within approx. 90% of their nominal values.

The staff will review the particulars of the modification when they are available.

ES-S Ak e e taeb sep a G9"6 otu iesmm eaWAnceaeatni Stalumen togweomponent inthe QuanesV nagnyPowr *ser nsbigeeone 4

ES-3 Testing performed has not block-loaded the emergency power sources utilizing actual emergency loads in an integrated fashion, to the levels that could be seen during an actual event. Therefore, analytical work performed by Duke must be relied upon to a larger degree to provide the level of assurance necessary that the safety loads will start and operate properly.

5 ES-4 If the staff cannot establish to a reasonable degree of confidence from the analysis that the safety systems will perform properly without motor tripping or thermal damage, than additional more fully integrated loading tests may need to be performed to obtain the necessary level of confidence.

6 ES-4 As a minimum, it appears that overfrequency protection on the underground and overhead paths, and undervoltage protection on the underground path should be installed to supplement protection that is already in place.

7 ES-4 The staff does not know what protective trips are available when a Lee gas turbine is powering Oconee electrical equipment to protect against a failure of the gas turbinweltyre regulator or govemor that results in an out-of-tolerance voltage or frequency.

8 ES-4 The staff is not aware of any failure analysis performed on the Lee gas turbine govemor or generator voltage regulator that analyzes the potential for an out-of-tolerance voltage or frequency to occur.

9 ES-4 The staff believes that a system of voltage and frequency protection that does not rely on operator intervention is necessary to guard against Lee station voltage and frequency problems that may affect Oconee.

10 ES-4 The staff will determine whether more realistic loading tests should be performed on the Keowee and Lee gas turbine generators following its discussions with Duke on the CYME computer analyses.

11 ES-5 DUke plans to perform an integrated test of the SSF. The results could provide greater or lesser confidence in the ability of the SSF to perform its function, that in tum could affect the SBO CDF results. This issue may be revisited if the outcome of the SSF test is not satisfactory.

12 ES-5 The staff believes that the Kpowe units should be tested on a monthly frequency by using an emergency start signal to initiate the monthly start test.

13 ES-5 A bounding test should be provided unless the acceptance criteria for the proposed degraded grid and switchyard isolation functional test or Keowee load rejection test will provide the needed assurance that the single load rejection limits would not be violated if those tests were passed.

FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM PAGE DESCRIPTION RESOLVED NO.

NO.

14 ES-5 Duke should ensure that procedures are in place in each Oconee unit to prevent paralleling of the Unit Auxiliary Transformer and the Unit Startup Transformer except for brief power transfer evolutions.

15 ES-5 The staff believes that recovery of the switchyard battery chargers should be included in the emergency recovery procedures.

16 ES-6 The licensee is currently in the process of developing an integrated test for the SSF makeup pumps. The staff has concerns that the planned flow testing will not be adequate to provide sufficient data to verify adequate seal cooling flow to each pump during actual SSF makeup pump injection. The licensee should develop more detailed and meaningful acceptance criteria for these tests and submit the acceptance criteria for staff review.

17 ES-6 Because we have confidence that adequate seal cooling can be maintained (assuming acceptable integrated test results) if injection is restored within 10 minutes, the staff concluded that the licensee should place more emphasis on operator training and drills to assure this capability can be consistently and confidently maintained.

18 ES-6 Based on the staff's review of the Keowee PRA, the staff identified several actions that Duke should consider. The staff believes that implementation of these actions is needed to achieve the reported availability results in the Keowee PRA.

19 ES-6 Due to the importance of having a skilled technical specialist at the Keowee site during periods of severe weather and the unpredictable nature of traveling in these conditions, the staff believes that a Keowee technical specialist should be stationed at the site in advance, whenever severe weather is predicted to occur.

20 ES-T ftewwm t*is tda Salmin thntut ris*nant sstm Basedon tht e

tWthe tat s181 neede i a

i hstsee w be considaed tor 0eiabily monftoring.

21 ES-6 The staff believes the modifier for scheduling maintenance for mild weather only is reasonable provided that a program is referenced in the administrative controls section of the Oconee TS, and the program stipulates that pre-planned dual unit maintenance should be avoided and terminated (end equipment restored to operable status) when severe weather in imminent.

22 ES-7 The licensee should perform periodic testing of the air circuit breakers (particularly ACBs 3 and 4) and their new function and their associated logic, to ensure that these ACBs can achieve the reported high reliability values. Reliability goals should be established. ACB performance should be tracked and monitored against these goals under the maintenance rule.

23 ES-7 The licensee should perform periodic hot start tests.

24 ES-7 The licensee should evaluate potential mechanisms for a single common cause grid degradation failing both Keowee power paths when generating to the grid.

25 ES-7 Since the availability of the underground path from Keowee is risk significant, the staff recommends that Duke consider an additional TS to pre-align LEE during any CT4 maintenance activity.

26 ES-7 The staff recommends that Duke consider adding a TS to reduce the potential for clogging the Keowee generator cooling water strainers.

27 ES-8 The staff recommends additional documentation and modeling for certain areas to provide robustness. For example, the staff believes that the modeling of CTS power should be expanded from a 'black box' to explicitly account for all failure modes.

-2 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM PAGE DESCRIPTION RESOLVED NO.

NO.

28 ES-8 The external events section of this report identifies key assumptions and conclusions from the licensee's IPEEE concerning the ability of Oconee's Electrical Power System and the SSF to withstand external events. These assumptions and conclusions stem directly from the licensee's IPEEE and have not been verified by the staff.

29 6

The December 12, 1995, Duke letter describes recent initiatives that have been taken, or are planning to be taken, in order to improve the reliability and operation of the Oconee ac electrical system.

30 10 It was found that there are, or will be, features in the Keowee voltage regulator and governor that could preclude certain failures in those units from producing an out-of-tolerance voltage or frequency outside a certain range.

31 10 A maximum volts-per-hertz trip will be moved to the emergency lockout circuit.

32 10 There will be an overfrequency permissive on Keowee output breaker closure set at 110%,

separate from the govemor.

33 11 Alarms associated with the permissive will alarm an overfrequency condition during normal and emergency operation.

34 12 Duke committed in their January 31, 1996 letter, to implement modifications to monitor and alarm voltages and frequencies below the acceptable ranges.

35 12 It is believed that a system of voltage and frequency protection, permissives, and limiters that does not rely on operator intervention is necessary.

36 13 Overvoltage protection of the underground and overhead paths will be provided through the voltsihertz trip set at 110% that will be moved to the emergency lockout. Overfrequency protection on these lines will be provided only following a load rejection trip of the Keowee units for emergency generation purposes to Oconee. This will be implemented through overfrequency permissives set at 110% to be installed on the Keowee generator output breakers. Alarms will be initiated at 110% overfrequency condition. Additional overfrequency protection will be provided during startup of the Keowee underground and overhead units for a particular governor failure by installation of a governor flyball motor monitoring circuit. The staff believes that this alarm should be supplemented with a protective trip that will separate the Oconee loads from the Keowee generators on an unacceptable overfrequency condition. The trip should have suitable time delay, redundancy, and coincidence to ensure proper operation.

37 13 Duke has committed to modify permissives on the emergency power path to delay loading of the Keowee units until the voltage and frequency are within approximately 90% of their nominal values.

38 14 It is not clear how the connected Oconee loads and Keowee unit would respond to the frequency overshoot resulting from failure of a Keowee gate limit feature to runback to its low lirmit setpoint following its initial run up to 100%. The overfrequency protection should be designed to ensure satisfactory operation of equipment or separation of the equipment from the Keowee generator.

39 17 In their January 31, 1996 letter, Duke committed to modify permissives on the emergency power path to delay loading of the Keowee units until the voltage and frequency are within approximately 90% of their nominal values. This modification should resolve the issue. The staff will review the particulars of the modification when they are available.

-3 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEMPAGE DESCRIPTION RESOLVED NO.

NO.

Snderground p

to accelerate up:naeaa Ia oi to-H c

o o

togir woud pqeonc orr t haoe motore (mulbple aswhoeu ld all.D uke Modification NSM ON-52966 will correct this problem by installing overrequency perrnissives on the Keowee generator output breakers set at 110%.

42 21 This report addresses some of the TS changes that are pertinent to the purposes of this report. The full body of the changes, including those that are addressed in this report, will be addressed in detail during the staff's review of the proposed TS changes.

43 23 Licensee to amend TS to include proposed Keowee hydro unit load rejection test (corresponding to an EDG test to reject load equal to 90 to 100 percent of its continuous rating...)

44 28 The annual test that includes differences in generator field flashing, use of auto synchronizer, and timing and application of gate limits in the governor system that are tested annually, should be tested monthly by using an emergency start signal to initiate the monthly start test. The staff will pursue this issue with Duke. The staff will also determine and evaluate the acceptance criteria for the test at that time.

45 28 Because the Keowee units periodically supply commercial power to the grid they likely satisfy the monthly loading surveillance requirement. The staff will verity with Duke that this test is similar to the ISTS monthly test.

46 29 The existing Oconee TS tests in this area (trip coil surveillance and operability, overhead and underground paths, etc) appear to be generally appropriate to the Oconee design. The proposed changes from the existing specifications also appear appropriate; however, the staff will provide its final evaluation of these changes following completion of its review of the changes. The staff will also evaluate the acceptance criteria to these surveillances during the review.

47 30 The staff believes that the Keowee Monthly Star Test -SR 4.6.1.a -should be a monthly start and load test performed for the Keowee units. The staff will pursue this issue with Duke.

48 32 It may be that the testing criteria Duke has referred to is in these other tests, although they don't identify that criteria in their discussion of those tests. The staff will pursue this issue with Duke.

49 32 A bounding test should be provided unless the acceptance criteria for the proposed degraded grid and switchyard isolation functional test or Keowee load rejection test will provide the needed assurance that the single load rejection limits would not be violated if those tests were passed.

50 33 The full load rejection test takes on added importance because the Keowee units are routinely paralleled to the grid for commercial generation purposes, the hydroelectric units undergo a large overfrequency excursion following a load rejection, and the Keowee generators are automatically connected to the Oconee loads following the load rejection.

The staff will review the details and acceptance criteria for this test during its review of the proposed Oconee TS revision.

-4 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM PAGE DESCRIPTION RESOLVED 51 33 The Keowee black start test is intended to demonstrate Keowee's ability to emergency start with only DC power available to its auxiliaries. Duke has stated that this feature is also periodically verified during the EPSL functional test when it is performed on Oconee Unit 1.

The test procedure did not substantiate the claim. The staff will be pursuing this with Duke.

52 34 The staff has not yet seen the data collected during the Keowee low power test that was performed in November 1995 to collect data on Keowee while the Keowee unit was loaded during its acceleration phase from a standby start. The EPSL logic will load the Keowee underground path unit with the Oconee unit accident loads during Keowee acceleration if a LOCA were to occur followed very quickly by a LOOP. If the overhead path were to fail, the Keowee underground path unit would also have to pick up the other two Oconee units' LOOP loads shortly after Keowee came up to full speed and voltage. Duke has provided a commitment to modify permissives on the emergency power path. The staff will review the particulars of the modification when they are available.

53 37 The staff will pursue its questions on the LOOP analysis regarding pump motor tripping with Duke in order to determine the accuracy of the results. If the results are found to be highly accurate but there are still small margins to motor tripping, then the staff will need to pursue the accuracy and reliability of the motor overload protective devices. If the staff cannot establish to a reasonable degree of confidence from the analysis, that the safety systems will perform properly without motor tripping or thermal damage, then additional more fully integrated loading tests will be needed to be performed to obtain the necessary level of confidence.

54 39 The staff believes a system of voltage and frequency protection that does not rely on operator intervention is necessary to fully guard against failures of the Lee gas turbine voltage regulator and govemor. The staff will pursue this issue with Duke.

55 41 The staff will pursue with Duke the status of the recommendations made in the analysis related to use of the Lee gas turbines and motor tripping. If the results are found to be highly accurate but there are still small margins to motor tripping, then the staff will need to pursue the accuracy and reliability of the motor overload protective devices. If the staff cannot establish to a reasonable degree of confidence from the analysis, that the safety systems will perform properly without motor tripping or thermal damage, then additional more fully integrated loading tests will be needed to be performed to obtain the necessary level of confidence.

56 41 Loading a Lee gas turbine by picking up system grid loads does not demonstrate the ability of the unit to instantaneously energize a foad block as would be required during an actual event. This issue will be further pursued during staff review of the proposed replacement Oconee Electrical TS.

57 43 In the case of an event that will automatically result in load shedding of the chargers and subsequent discharging of the batteries, the staff believes the recovery of the chargers should also be included in the emergency recovery procedures. The staff will verify with Duke that such recovery actions are included in the emergency procedures.

58 4

oo ti repot alo dte thatthe Keoue PRA severe weat rm dustunit waintnanceare raoah rvddt~ae duW untWaintenance avoled and also teminate *hene 5

48 The, eW V

w

'twomRerova orB"orn"A r4aTe stf alomend t* pusuewt Dmukan Oconse TSrqitarero N~eees e.lar em following the outcome at the 88F testand toe TS proposeidag._____

-5 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM PAGE DESCRIPTION RESOLVED 60 47 LER 287/80-0 stated that various means for assuring that the transformers are not used in parallel were under investigation. The staff could not determine if any follow-up corrective action had been implemented. The staff will pursue with Duke whether any action has been taken on this issue.

61 49 The lack of a Keowee degraded voltage/trequency model in the Oconee PRA does not appear to be a significant effect provided that Duke installs the additional voltage and frequency protective features the staff has called for in this report.

62 49 Duke has committed to modify the permissives on the emergency power path to delay loading of the Keowee units until the voltage and frequency are within approximately 90 percent of their nominal values. This modification should resolve the early loading issue relative to the Keowee underground path. The staff will review that particulars of the modification when they are available.

63 49 The occurrence of certain ground faults on the switchyard side of the Keowee main step-up transformer, together with a subsequent single failure of a safety-related switchyard circuit breaker to trip that would result in lockout in both Keowee units, might increase the common cause actuation of Keowee generator lockout probability to a degree that generating both units to the grid would be found to be slightly less reliable, rather than more reliable.

64 50 The occurrence of certain events involving tripping of generator loss-of-excitation relays might increase the common cause actuation of Keowee generator lockout probability to a degree that generating both units to the grid would be found to be slightly less reliable, rather than more reliable.

6 W

lt56 s not practicali theoon dieagn to prWide an integated tetfor eade pousibl continanon and permutation of mergency power sous ec. More mance therefera rewrded to~~~~w be placed on componettet, overtppingteele, parti kitgrated tets and componenta i theOconee emersgencyas wer syalem arebeintg std.

66 50 The staff does not have enough information on the CYME computer analyses performed for the Oconee emergency ac power systems to confirm at this time that they provide the level of assurance necessary that the safety loads will start and operate properly. The staff will further pursue its questions on these analyses with Duke.

67 50 The staff will determine whether more realistic loading tests are necessary on the Keowee and Lee gas turbine generators following its discussions with Duke on the CYME computer analyses performed with these units as power sources.

68 51 Duke plans to perform an integrated test of the SSF, the results of which could provide greater or lesser confidence in its ability to perform its function, that in tum could affect SBO results.

60 51 Titilr(tstepusu with Dukian OconSTS requirement relative to K(eowes dual Aernovedf~rm bipatteanuran svewe uathe. This iwu may be revisitd fobowng ithclne of Report 69:g ittlestand th TS poosed change.

70 51 As a minimum it appears that overfrequency protection on the underground and overhead paths, and undervoltage protection on the underground path should be installed to supplement protection that is already in place.

71 51 The voltshiertz limiter feature in the Keowee voltage regulator may be used together with an undervoltage trip to provide underfrequency protection if it can be shown that the comparable underfrequency setpoint will allow satisfactory operation of safety equipment.

0

-6 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM: PAGDESCRIPTION RESOLVED NO.-

NO.

72 51 A failure of the Lee gas turbine voltage regulator or governor that results in an out-of tolerance voltage or frequency, could expose the redundant safety-equipment of the three Oconee units to that voltage and frequency, unless protective trips separate the Oconee equipment. Thestaff does not know what protection of this kind is available when a Lee gas turbine in powering Oconee electrical equipment.

73 51 The staff is not aware of any failure analyses performed on the Lee gas turbine governor or generator voltage regulator that analyzes the potential for an out-of-tolerance voltage or frequency to occur.

74

51.

The staff believes a system of voltage and frequency protection that does not rely on operator intervention is necessary to fully guard against these failures. The staff will pursue this issue with Duke.

75 52 The staff believes the Keowee units should be tested on a monthly frequency by using an emergency start signal to initiate the monthly start test.

76 52 A bounding test should be provided unless the acceptance criteria for the iproposed degraded grid and switchyard isolation functional test or Keowee loadieection test will provide the needed assurance that the single load rejectionlimits would not be violated if those tests were passed.

77 52 Duke should ensure that procedures are in place in each Oconee unit to prevent paralleling of the Unit Auxiliary Transformer and the Unit Startup Transformer except for brief power transfer evolutions.

78 52 In the case of an event that will automatically result in load shedding of the Oconee switchyard battery chargers and subsequent discharging of the switchyard batteries, the staff believes the recovery of the chargers should be included in the emergency recovery procedures.

7u5 ummIyrtm

-ta mi 80 57 It is our understanding that integrated testing from the spent fuel pool via the SSF makeup flow path is planned in order to verify the system flow models. However, in the interim we have no technical justification to conclude (based on no identified flaws in the flow models beyond verification by test) that the flow disparities would be significant enough to reduce the flow to any one pump to a degree that would jeopardize the RCP seal (i.e., result in a flow rate less than the leak-off rate).

81 60 The licensee has committed to provide remote makeup capability to allow refilling the spent fuel pool without entering to fuel pool area.

Irsus d to th.aW to ammnt tesihtI dthe operat the SSs thre rewond

-~

4Waumed inhiuiW edehIgosicuisa n his our undustandng atte thvee sevice water pumheen ompt it 83 64 The only signaifcant existing concems relative to the SSF are the SSF makeup pump's marginal capacity (which results in a required initiation of 10 minutes) and the lack of integrated testing of the makeup pump's ability to supply balanced seal flow to all four reactor coolant pumps.

84 64 The licensee is currently in the process of developing an integrated test program for the makeup pumps. The actual conduct of the test would not occur until at least July 1996. The test is intended to verify the flow models used in the analysis by demonstrating approximately equal flow to each of the four reactor coolant pumps.

-7 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM PAGE DESCRIPTION RESOLVED NO.

NO.

85 64 Our understanding is that the present plans are to test flow up to the RCP injection point; this will verify the flows to each pump with the injection piping disconnected (does not include backpressure from pump and leakoff flow paths). Thus, their proposed integrated flow test will only test a portion of the entire flow path and will not include the RCPs. The staff is concerned that the planned flow test will not provide sufficient data to verify adequate seal cooling flow to each pump during actual SSF makeup injection.

86 64 Because of the marginal capacity of the makeup pumps, integrated testing that includes the seal return flow path should be performed to assure adequate seal cooling and to demonstrate adequate backpressure is maintained on the Number 1 seal during SSF events.

87 65 It would take only a small amount of seal degradation to result in excessive seal leakage during an SSF event. Therefore, adequate integrated testing should be performed.

88 65 The licensee should perform adequate integrated testing of the SSF makeup system and its capability to provide adequate seal cooling. Proposed acceptance criteria for these tests should be reviewed by the staff.

89 66 The licensee should put additional emphasis on operator training and drills associated with recognizing SSF events and the capability to initiate the SSF makeup system within 10 minutes from the loss of RCP seal cooling that accompanies all SSF events.

90 74 The staff believes that the Chi-Squared equation is not appropriate with respect to the number of degrees of freedom that it specifies for the Chi-squared distribution. Using the equation may be misleading since it produces a probability estimate that is about a factor of 3 lower than the 50 percent confidence value.

91 79 Due to the importance of having a skilled technical specialist at the Keowee site during periods of severe weather, and the unpredictable nature of traveling in these conditions, the staff believes that a Keowee technical specialist should be stationed at the site in advance whenever severe weather is predicted to occur.

92 83 The staff believes that the unreliability assigned to CTS power on a grid-related LOOP is potentially underestimated.

93 84 The Lee station is under the management of Duke's Fossil Generation Department.

Maintenance unavailability records are not kept with the Nuclear generation Department.

Verification of any past and future operational experience is therefore questionable.

94 84 The staff believes that due to the importance of power from CT5 to the Oconee ac power system, either the modeling of CT5 power should be expanded from a 'black box" to explicitly account for all the failure modes to determine more accurately its reliability and availability, or a more robust screening value should be used in the 'black box model.

95 84 Due to the importance in the emergency ac power system of all three Oconee units, the staff believes that the reliability of these components should be tracked under the maintenance rule implementation.

M JinalrST19, h licensesconwtd ticldheLeerenral PerStaleenr eyearsinwelaintenan asea a riiksagnateent sposn. ases on ti sleer. so ear sere Sdsinformaovliabn deladn?

al eelune WII be anadeed fr ellaidly~ mwang.

97 84 The second-most dominant cut set in the Oconee-Keowee ac power integrated model contains a modifier of 0.1 since Keowee maintenance is scheduled for mild weather periods.

The staff believes this is reasonable modeling assumption, provided that weather considerations is referenced in the administrative controls section of the Oconee TS, and the program stipulates that pre-planned dual unit maintenance should be avoided and terminated (and equipment restored to operable status) when severe weather is imminent.

-8 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM PAGE DESCRIPTION RESOLVED NO.

NO, 98 84 Because the staff believes that the unreliability of CT5 power is potentially underestimated, the validity of the assumption that the total Oconee emergency power system to two 100 percent EDGs is questionable.

99 84 When Duke integrated the Keowee PRA with the IPE, the CDF at Oconee Unit 3 for loss of as power events was calculated to be 1 E-6/yr. Because the same concerns with power from CT5 are carried through to this value, the staff believes that this value could be potentially underestimated.

100 85 In light of its importance, the staff believes that the SSF failure probabilities used for the IPE update (which is now being performed by Duke) should be examined thoroughly. In addition, since the validity of the data is dependent on the adequacy and scope of tests, any testing issues associated with the SSF should be resolved.

101 89 To ensure high reliability of automatic operation of ACBs is maintained, periodic testing (especially ACBs 3 and 4), their new functions, and the associated logic should be performed.

102 89 Periodic hot start testing of the Keowee units should be performed. Statistical failure probabilities should be determined from the test results data.

103 89 The sensitivity analysis assumed that dual Keowee unit generation occurs 3.4 percent of the time. Should this percentage increase, the risk impact of this common cause failure increases. Consequently, the staff feels that the licensee should determine how this event can occur and its likelihood. The staff needs this information to conclude there is no significant reduction in Keowee's reliability if both units are generating to the grid, particularly if the percentage of time that both Keowee units are generating to the grid is increased.

14 9 0

The licensee should consider taking opportunities to compare actual data for components that are tested/demanded infrequently but demanded on an emergency start with the data of similar components with more frequent test/demand intervals.

105 97 Due to the importance of the strainers on the common intake header of the Keowee generator cooling water system, Duke should consider adding a TS item to assure periodic surveillance of these strainers.

106 97 The staff believes that a Keowee technical specialist should be stationed at the site in advance whenever severe weather is predicted to occur.

107 98 The staff suggests that use and interpretation of the results of the Keowee PRA shruld be applied with the generic data.

108 98 Most of the plant specific failure data are lower than their generic counterparts. A large amount of the plant specific data was obtained from the experience of Keowee during grid generation. Although it is reasonable to assume that the majority of this experience data would be applicable to emergency demand situations, there are some subtle differences in operation, such as the voltage adjust, for which grid generation experience applicability may be questionable. The Keowee PRA does not provide justification for the applicability of these cases.

109 98 Demand frequencies of components such as relays, for which plant specific data was calculated by combining all components of the same type into a type code, rang from daily to quarterly to every refueling outage. Ideally, all components in the same group or type code should have the same or similar periodic testing frequencies. At Keowee, some components that are only challenged every refueling outage are applied the same failure probability as components that are demanded daily. Therefore, the staff believes the results of the generic data sensitivity study are more robust that the base case.

-9 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM PAGE DESCRIPTION RESOLVED NO.

NO.

110 98 Duke classified certain basic events as "undeveloped events." For these basic events, Duke could not find applicable generic data. Therefore, the "undeveloped event' failure rates were calculated using Keowee plant specific data only and remained at their plant specific values for the base case and the generic data sensitivity study. Because of this, the generic data result is not a completely generic data assessment.

111 99 The failure probabilities were not discussed in the text or the tables of the Keowee PRA.

112 99 For the integration with the IPE, Duke calculated the probability of losing all ac power for Oconee Unit 3 to be 6.35E-5/yr for internal events including severe weather. The staff believes that this probability could potentially be underestimated.

113 100 The Lee station is under the management of Duke's Fossil Generation Department.

Maintenance unavailability records are not kept with the Nuclear Generation Department.

Verification of any past and future operational experience is therefore questionable.

114 100 The staff believes that the unreliability assigned to CT5 power on a grid-related LOOP is potentially underestimated.

115 100 Due to the importance of the Lee station in the emergency ac power system of all three Oconee units, the staff believes that the licensee should establish reliability goals for the CT5 power components, and performance should be tracked and monitored under the maintenance rule implementation.

116 100 In a letter dated January 31, 1996, the licensee committed to include the Lee/Central Power system in the maintenance rule as a risk significant system. It is expected, therefore, that specific failure modes will be considered for reliability monitoring under that rule.

117 100 The probability of a Severe Weather LOOP with both Keowee units in maintenance was reduced by a factor of 10 in the Keowee PRA to account for maintenance being scheduled for mild weather periods. This assumption is reasonable, provided the TS stipulate that pre planned dual unit maintenance should be avoided and also terminated (and equipment restored to operable status) when severe weather is imminent.

118 100 Because of the importance of CT4 (underground transformer), the staff recommends that Duke consider an additional TS to pre-align Lee during any CT4 maintenance activity, as is stipulated for dual Keowee maintenance activities.

119 101 A periodic hot start test should be performed to more accurately determine failure probabilities.

120 102 To ensu' s high reliability is maintained, periodic testing of ACBs (especially ACBs 3 an' 4),

their new functions and the associated logic, should be performed. Reliability goals should be established, and performance should be tracked and monitored against these goals under the maintenance rule.

121 102 Periodic hot start testing of the Keowee units should be performed. Statistical failure probabilities should be determined from the test results data. Reliability goals should be established at the values used in the Keowee PRA and performance should be tracked and monitored against these goals under the maintenance rule.

122 102 The licensee should evaluate potential mechanisms for a single common cause grid degradation failing both Keowee power paths when generating to the grid.

123 103 The sensitivity of the results to changing the failure probability of infrequently tested components was somewhat significant, at 15 percent. The licensee should consider the potential for differences in the failure probabilities of components with widely varying test/demand frequencies by taking opportunities to compare actual data for components that are tested/demanded infrequently but demanded on an emergency start with the data of similar components with more frequent test/demand intervals.

-10 FOR INTERNAL USE ONLY

FOR INTERNAL USE ONLY ITEM PAGE DESCRIPTION RESOLVED NO.

NO.

124 103 The Keowee PRA was performed under a less stringent framework than typical plant PRAs.

For this reason, the staff and licensee should be sensitive to changes in any operating practices, test frequencies, time in configurations, etc., and their impacts should be addressed accordingly, 125 103 The IPEEE assumptions and conclusions presented in the report come directly from the licensee's IPEEE submittal and have not been verified by the staff.

126 107 The seismic fragility curves relating equipment failure probability versus ground acceleration will need to be reviewed further to assess the reasonableness of their current estimates.

127 107 The seismic cut set frequencies in Table 3-4 will need to be explored further; the cutsets appear to total 1.7E-04, not 3.6E-05.

1us 101 Pub. equ... i d

th Ocss sbmatt Dcebr 1996 8ased a Gosed I~cneetprpoed eahdJag

-sscpee 129 110 Failure of equipment due to internal floods is dominated by submerged equipment.

Examination of the internal fault trees indicates the potential for crediting the Keowee or Lee power sources through the link to the internal events electric power fault trees. It is assumed that these power supplies were included but unclear what, if any, impact they had on the results.

130 ES-8 The contribution of less significant but more likely external event scenarios may be misrepresented in the licensee's IPEEE submittal.

131 51 The SBO results are also sensitive to the assumed availability and reliability of the LEE gas turbines. This issue, therefore, may be revisited if the outcome of the SSF test is not satisfactory.

132 94 Since the uncertainty analysis was performed with plant specific bayesian-updated data, this analysis does not encompass the concerns the staff has with this data. The staff believes that interpretation of the Keowee PRA results should be based on generic data rather than plant specific bayesian-updated data.

133 94 The demand failure rate of components in similar groupings had varying testing frequencies.

This grouping could bias the results of components that are tested infrequently.

134 94 Duke did not justify the applicability of using grid generation experience to calculate certain basic event probabilities that are specific to a Keowee emergency demand (i.e., the voltage adjust).

135 113 Since external event and severe weather event uncertainties can be large, the staff believes that there uncertainties should be taken into consideration when making decisions regarding the unavailability of the Oconee emergency AC power system.

136 ES-8 Since external event and severe weather event uncertainties can be large, the staff believes that these uncertainties should be taken into consideration when making decisions regarding the unavailability of the Oconee emergency AC power system.

137 ES-8 The staff intends to perform a thorough review of the Oconee IPEEE.

1. The items appearing in this table were taken from the NRC staff report on the Oconee Emergency Electrical Distribution System. Each is not necessarily met to stand alone, reference should be made to the Report itself for a more complete explanation. Duplicate items may appear in the chart if they are duplicated in the Report.

- 11 FOR INTERNAL USE ONLY

v t e Letter Sequence '
TAC:M93550, (Open)
Initiation Request, Request Acceptance... Administration Questions 22 January 1998 26 January 1998 22 April 1997 15 January 1998 22 December 1995 Answers Results Approval... Other: ML15117A350, ML15117A352, ML15238A368
MONTHYEAR1997-02-18 [Table View]

ML15118A448

Text

SUBJECT:

SUMMARY

Enclosures:

Enclosures:

Navigation menu

Search