ML20128M328
| ML20128M328 | |
| Person / Time | |
|---|---|
| Site: | Brunswick  |
| Issue date: | 12/31/1984 |
| From: | Anoba R, Hammond J, Oliver R CAROLINA POWER & LIGHT CO. |
| To: | |
| Shared Package | |
| ML20128M312 | List: |
| References | |
| NUDOCS 8507110666 | |
| Download: ML20128M328 (92) | |
Text
{{#Wiki_filter:. _ . L J ce 'f CAROLINA POWER AND LIGHT CCMPANY BRUNSWICK STEAM ELECTRIC PLANT Diesel Generator Technical Specification Improvement Study December, 1984 _ .. Prepared By: R. C. Anaba (CIS - !!SR Unit) Assistance by: NUCO:1 I:lC. Reviewed By: ON Cc b\ '* s.-Ai R. E. Oliver (CIS) f Reviewed By: A,D . be.~~!
'J ',G Hac=ond (CIS)
Approved By: _
- O- I' J. D. :,j ~JettrLea (CIS) e 8507110666 850628 PDR ADOCK 05000324 pm P
1 .
o .i CORPORATE NUCLEAR SAFETY AND RESEARCH DEPARTMENT TABLE OF CONTENTS 1.0 Summary 2.0 Scope and Objectives 2.1 Introduction
2.2 Background
2.3 Objectives 2.4 Scope of Work
30 System Description
3 1 System History and Operational Experience 3.2 System Function 3 3 System configuration 3.4 Support Systems 3.5 Failure Criteria
' ~
4.0 Methodology - 4.1 Failure Modes and Effects Analysis and Success Criteria Evaluation 4.2 Fault Tree Development 4.3 Input Development 4.4 WAMCUT/ FRANTIC III Methodology 4.5 Risk Analysis Methodology 5.0 Results , 5.1 Risk Analysis 5.2 Sensitivity Analysis 5 3 Benefit Analysis 5.4 Diesel Generator Wearout Analysis 6.0 conclusions and Recommendations 7.0 References Apenndices A. Diesel Generator Data Analysis D. Emergency AC Power Reduced Fault Trees 9 9 w- O
i 's LIST OF TABLES , TABLE TITLE 31 Manpower Requirements for BSEP Diesel Generator Scheduled Maintenance Activities , 4.1 Decay Heat Removal Function Success Criteria immediately after Reactor Trip l 4.2 RHR System and Support System i Success Criteria , P 4.3 Case 1: Loss of Off-site Power in both Units - BSEP Diesel Generator Failure Criteria Evaluation 4.4 Case 2: Loss of Off-site Power In l Both Units And LOCA In One Unit - BSEP ! Diesel Generator Failure Criteria Evaluation 4.5 WAMCUT/ FRANTIC III Component Definitions - f i 4.6 WAMCUT Ccmponent Inputs For Reduced Fault Tree Models 4.7 FRANTIC III Test Configurations 4.3 FRANTIC III Input Summary For Case 1 - , Loss Of Off-site Power l 4,9 FRANTIC III Input Su==ary For Case 2 - l Loss Of Off-site Power With LOCA In One Unit ! 4.10 FRANTIC III Input Parameters For Periodically Tested C0mponents. 4.11 Basic Structure For Logic Model Development 5.1 Risk Analysis Results 1 5.2 Average Annual Risk Analysis Results - 5.3 Sensitivity Analysis Results 5.4 DG Dowel Pin Wearout Analysis Results A-1 Timo-Dependent Diesel Generator , Failure During Standby Period [ A-2 Not Used i ! s .
~ -.. . h
A-3 Diesel Generator Unscheduled Down time , A-4 Diesel Generator Test And Maintenance Unavailability, Scheduled and Unscheduled A-5 Average Unavailability, Excluding 1984 A-6 Diesel Generator Restoration Time, Scheduled And Unscheduled Test And Maintenance A-7 Mean Time To Repair Data For Brunswick Diesel Generators, Unscheduled Maintenance Only A-8 Quick-Start Failure Data For Brunswick A-9 Diesel Generator Maintenance Frequency t e e ___j ____ ______ _
3 1 Figur9 List Of Figures - Title 4-1 Diesel Generator Unavailability At Surveillance Tasting As A Function Of Test Interval 5-1 Correlation Between Test Frequency And Number of Failures Using NUREG/CR - 2989 Data A-1 'delbull Plot For The Standby Failure Data i 0 6 y l l l i l s .
.~. - .._.. _ ---
~6-t i
1.0 SUSARY l l The purpose of this study was to perform a probabilistic risk assessment i to revise the Technical Specification for the Diesel Generator System at the Brunswick Steam Electric Plant. In particular, it was the intent of this study to serve as the basis to revise the Technical Specification to , extend the Diesel Generator Allowed Outage Time (A0T) from three days to l l seven days, and to increase the surveillance test interval during a ! l Limiting Condition of Operation (LCO) from 12 hours to 72 hours. The current Technical Specification require that the Diesel Generator System l be placed in an LCO condition when a single diesel is removed from l service. In this study risk is defined as the prooability of not having l AC power available for decay heat removal or for a Loss of Coolant i Accident (LOCA) . ! l l The results of this study indicate that extending the Diesel Generator l ACT to seven days and increasing the LCO test interval to 72 hours woulc i slightly increase the level of risk over tne period of the LCO. However, l the resulting level of risk over the period of the LCO is still lower , than the average level of risk over the period of time when all diesels : are available in standby. There is also a slight increase in the average - annual risk level between current LCO and proposed LCO. However this , I difference is judged to be within the uncertainty and variation of the . . input data used for this study. It was concluded that the increase in risk during the LCO condition can i be justified in light of the following benefits: , o Reduc'ed wear on the diesel generators caused by increasing the . surveillance test interval during the LCO condition. e Operational flexibility and reduced chances of plant shutdown by extending the A0T. l o laproved maintenance quality through alleviation of the constraints of ! coroleting the work in a short time interval. [ I
& G
, ?
f 8 e
r l @ 2.0 SCOPE AND OBJECTIVES 2.1 Introduction This report presents the results of an evaluation to assess the benefits and risk impacts of proposed changes to the Brunswick Steam Electric Plant's Technical Specifications for the Emergency Onsite AC Power system. The proposed Technical Specification changes will increase the plant operating flexibility and reduce wearout effects on the diesel generator. The changes include an increase in the allowable outage time for a single out-of-service diesel generator and a reduction in the required surveillance frequency of the operable diesel generator systems when one diesel generator is out of service. The approach taken in this study has been to assess both the i benefits and risks of implementing these changes. The benefits l include, as noted above, i= proved operational flexibility for the plant, a reduction in the waarout rate of the diesel generators (because of a reduction in the number of starting tests that would be performed on each diesel), and potential improvements in diesel generator maintenance (because of longer allowed repair times). The ' increased risks resulting from those changes ' include an increa:s iri ' ' the amount of time that the emergency AC power system remains in a state in which a single diesel generator is out for repair, and a small reduction in the reliability of the remaining diesel generators during the period that the Limiting Condition for Operation (LCO) is in effect due to the reduction in surveillance test frequency. , The risk impacts noted above were quantified through the use of a timo-dependent unavailability analysis of the Brunswick Emergency Onsito AC Power system, using both the current Tecnnical l Specification requirements and the proposed changes to tns requirements. j 2.2 En e!< ?reund The current Technical Specifications for the diesel generaters system at the Drunswick Steam Electric Plant require that wnen ena of the four diccola is taken out of service the remaining diesels are to be tested within two hours and every 12 hours thereafter. Furthermore, the maximum allowable time that the diosol generater . - can bo out of servico before the plant (Unit 1 and Unit 2) must :o . . shut down is 72 hours. , Recently there has been a crowinc concern that frequent testing of l the diosol generators during the LCO condition can contribute to I premature failures of diesel components due to waarout and cyclic ratiguo mechanisms. Specifically, common modo failure of one or more accessory pump dowol pins and cap screws in the flex drive coupling drive plate (LER 1-82-78) on all four diesels are believed to be a result of cotal fatiguo of the dowel pin material due to the excessive number of engine starts. k O
It is felt that frequent fast starts degrade the engine / generator reliability and availability by increasing the probability for single component failure. Isolated failures of piston connection red bearing bolts, exhaust and inlet valve seats, and fuel injector parts are believed to be a result of extended unloaded operation during the LCO. Another area of concern is that the current A0T is not adequate to allow an unhurried on-line maintenance of the diesels (e.g., for repairs dowel pin require extensive engine disassembly that inspection / replacement). It is belioved that extending the ACT would reduce some of the concerns raised above. 23 Objectives The objectives of this study were to accomplish the following tasks using probabilistic methods and engineering judgment: ' o Evaluate the changes in risk associated with extending the ACT frem three days to seven days and by increasing the diesel - generator surveillance test interval during the LCO fres 12 hours to 72 hours. o Provide a qualitative discussion of benefits associated with ACT extension and surveillance test interval increase. o Quantify diesel generatar wearout phenomenon using plant data. 2.4 Scoce of Work The scope of the work that was required to accomplish the above objectives is outlined below: o Develcped fault tree models for the following cases:
- 1. Loss of offsite power Loss of cffsite power with a LOCA in one Unit 2.
o Determined the minimal cutsets for the above cases using the WAMCUT program. o Developed inputs for the FRANTIC-III computer code using WAMCUT results and plant specific testing and maintenance data. o Used FRANTIC-III to calculate the avera$e and peak riskstandby levels over the period of the normal diesel generator configuration and over the period of the LCO diesel generator configuration, o Calculated the average annual risk levels for the current LCO and the proposed LCO. s
F 8 o Reviewed plant data to determine any wearcut phenome.non associated with diesel generator testing. o Reviewed the qualitative benefits associated with A0T extension and surveillance test interval increase. I
30 SYSTEM DESCRIPTION i 31 System History and Operational Exnerienee The Brunswick Steam Electric Plant (BSEP) is a two unit BWR plant which has been in commercial operation since the mid-1970's (Unit 2 since 1975 Unit I since 1977). The SSEP Emergenay AC Power system consists of four diesel generators which supply 4160 V power to four emergency buses. Each emergency bus supplies 4160 V emergency power to both Units. During most of the study period, the Brunswick Technical Specifications allowed a single diesel generator to be inoperable for up to three day period, provided that augmented surveillance testing was performed on the remaining three diesel generators on a once per 12 hour period basis. For the period beginning from September 8,1976 to Nove=ocr 22, 1977, the Technical Specifications were written witn a DG allowed outage time of seven days. As of Novceber 22, 1977 the Technical Specifications were revised for a DG A0T of three days. Based upon satisfactory operating experience history, it .is ,
- proposed that the current Technical Specifications be amenced to allcw increased plant operatirnal flexibility and to recuce the overall wear en the diesel generators resulting from augmented surveillance testing during an LCO conditon.
Specifically, the follcwing two changes are proposed: o An increase in the allowable outage time for single diesel generator from three days to seven days o An increase in the required augmented surveillance test interval of the three operable diesel generators during above noted LCO condition frem once every 12 hours to once every 72 hours Experie nce gained frem testing, preventive maintenance and repair activities over a period of time indicate how well a plant's Technical Specifications have been optimi::ed. Any problems with the Technical Specifications may reveal themselves in terms of an increased number of forced outages due to such factors as having an allowable downti=e (inoperability time limit) that does not correspond with the .- actual times required for preventive maintenance. In Tabic 3.1 the msnpower requirements for BSEP DG preventive maintenance actions are listed. Table 3 1 indicates that the time requirec to perform routine diesel generator preventive maintenance actions do not provide adequate margin to allow for difficulties that may occur during an LCO condition. There needs to be a balance between testing and inspection requirements and equipment reliability. The evaluation of operating experience is one important task when considering proposed Technical Specification changes.
s . TABLE 3.1 MANPCWER REcUIREMENTS FOR BSEP DIESEL GENERATOR SCHEDULED MAINTENANCE ACTIVITIES No. of Hours No. of No. of 12 Procedure to Complete Maintenance Hour Shifts No. Description Frecuency Work Personnel Per Day PT-12.3 1 Emergency 18 mo. 60 10 2 Diesel Inspection MI-10-503C Dowel Pin Every 600 1 70 10 2 Inspection 100 Starts PT-12.3.5a DG. No. 1 54 me. 60 to 2 Inspection 54 me. 60 10 2 PT-12 3.5b DG. No. 2 Inspection
~
54 me. 60 10 2 PT-12 3.5c DG. No. 3 Inspection 54 me. 60 10 2 PT-12 3.5d DG. No. 4 Inspeccion As required 12 6 1 MI-10-503D DG Exhaust Inspection on schedule NOTES:
- 1) The information tabulated is based en discussions with BSEP plant personnel
- 2) The manpower requirements are based on the absence of any problems that are uncovered during the maintenance activity and can be considered minimum requirements.
- 3) The current A0T time for the BSEP DG's is 72 hours. .
- 4) The first dowel pin inspection exceeded 72 hours and required a special one ti=c exemption frem the NRC (LER 1-82-73) i i .
t.- e Of particular interest with respect to the Technical Specifications are the plant specific test and maintenance records. These aid in establishing equipment dsuntimes, restoration / repair times and testing frequencies. A summary of these factors is found in Appendix A. On average, the montnly maintenance frequency per diesel generator is 0.58, which implies that tne mean ti=e between diesel outages is approximately 1000 hours or approximately 40 days. , This value of the mean time between diesel outages (ie., 40 days) is used in this report as the baseline period for calculating the risk associated with the nor=al standby configuration of the Diesel Generator System. 3.2 System Function The purpose of the E=ergency Diesel Generator system is to provice a highly reliable on-site source of standby electrical power for the operation of E=ergency Systems and Engineered Safeguarcs Systems (ESS) when the normal off-site source of power is not available. s The E=ercency Diesel Generator system consists of four DG sets located in the DG building. Each DG set consists of a 10-cylinder, turbocharged, V-block, 4900 brake hp, 514 rpm horcbers ciesel engine driving a General Electric generator rated at 3850 kw at 0.o power ' factor, 4160V, 00 HZ, and 670 amps at 514 rps. - Each DG is supplied by its own starting air system, consisting of two compressors, two air receivers, and the necessary control valves and controls. The DG Fuel Oil System consists of a com=on seven-day, 225,000-gallon stcrage tank, and each DG is provided with a 23,300-gallon storage tank located outside the DG building below grouna anc a 5:aC-gallon dsy tank mounted on each engine base. Cooling water to the DCs is supplied by the Unit 1 and Unit 2 Nuclear Service Water Systems, which provide a redundant source to each DG. The diesels are designed to provide adequate emergency power to the plant on a loss of offsite power or on a loss of offsite power coincident with a LOCA to one Unit (FSAR Design Danis). The diesels also provide AC power to facilitate adequate heat re= oval capacility during loss of offsite power events as well as adequate AC power for emergency core cooling and/or makeup during a LOCA event. . . 3.3 Synten conrieruration During norr.al standby conditions each diesel is aligned to its designated emergency bus. The diesels will autcmatically start on the following signals: o Reactor low water level 3 Tech Snec: Greater than or equal to 2.5 inches; Actual: +45 inches
OR o High drywell pressure Tech Scect Less than or equal to 2 psig; Actual: +1.8 psig AND o Low reactor pressure Tech Seec: +410 115psis Actual: +410 psig E o Loss of off-site power 2 o Generator primary lockout S o Loss of emergency bus voltage - ' The following parameters or actions will trip the diesel generator when oper1 ting in the autc=stic mode: o Low lube oil pressure 27 psig o Overspeed 590 rpm > o Differential overcurrent 0.2 amps difference between phases o Reverse power 25% of phase o Loss of field 25% restraint o Generator phase overcurrent o Barring gear engaged . o Manus 1 (Operator Actien) The following parameters or actions will trip the diesel generator when operating in the manus 1 mode: o High lube oil temperature 1900F o fligh jacket water temperature 2000F o Low lube oil pressure 27 poig I
l 590 rps o Overspeed 12 psig o Low jacket water pressure o Loss of diesel generator control power 0.2 amps difference o Differential overcurrent between phases t 25% out of phase ; o Reverse power , 25% restraint o Less of field o Generator phase overcurrent , o Barring gear engagea o Manual (Operator Action) During an LCO condition, one diesel isaccelerated out of service and the schedule to on an diesels are tested
~
operacle '4hile in the test mode, the dissef demonstrato system availability. mode if an automatic generator start logic will override the test start signal is present. i 34 Suceert Syntats i There are two major support systems for the Diesel Generator System: ! service water and DC power. As currently ! Service water is required for each diesel generator. ! configured, two separate service water lines supply each diesel. One supply line originatos reem the Unit 1 service water system and the other line originates from the Unit 2 service water system. Loss of both sources of service water will result in the loss o diesel generator. DC power is required for each diesel generator Eachfordiesel engine control
,senerator circuits and output breaker control circuits.
is provided with a primary source of DC power and an aitornato . source of DC power (1. e., from the other Unit). sources of DC power requires Transfer frca primary to alternate Loss of both sources of DC power will result in operster action. the loss of a diosol generator. 0 e
{ i 4 3.5 Failura criteris The failure criteria used for this analysis are summarized below. Please see Section 4.1 of this report for a detailed Failure Modes and Effects Analysis (FMEA). Case 1 Loss of Offsite Power /No LOCA DC Fsilures Consequences I and 2 Loss of Heat Removal Capability in Unit 1 3 and 4 Loss of Heat Removal Capability in Unit 2 Case 2 Loss of Offsite Power with LOCA in ene Unit DO Fsilures Consequences 1 and 2 o Loss of Host Removal Capability. for Unit 1- .. . l OR o Power Unavailable for ECCO cooling / makeup if LOCA occurs in Unit 1 l l 3 and 4 o Loss of Heat Removal Capability for Unit 2 I l OR o Power Unavailable for ECCO cooling / makeup if LOCA occurs in Unit 2 l 1 and 4 o Power Unsva11able fer sC 0 cooling / makeup if LOCA occurs in Unit 1 or Unit 2 , 2 and 3 o Power Unavailable for ECCO cooling / makeup if LOCA occurs in Unit 1 or 2 I e
t . l t 4.0 METHODOLOGY 4.1 Failure Modes and Effects Analysis and Success Criteria Ev11ustien A Failure Modes and Effects (FMEA) Analysis was perfor=ed on the diesel Generator system to determine the failure criteria to be used for the fault tree development phase. Losses of equipment were evaluated on a system basis for the following cases:
- 1) Loss of offsite power
- 2) Loss of offsite power with LOCA in one Unit Systems that required AC power for Reactor Heat Removal and/or LOCA mitigation were considered for this analysis The success criteria for decay heat removal and LOCA mitigation are given in Table 4.1. The preferred mode of decay heat removal was assumed to be the RHR system in suppression pool cooling mode.
~
The success criteria for the various modes of' RiiR system sre , given in Table 4.2. The FSAR design basis for the plant requires two RHR pumps ano one core spray pump for the Unit with the LOCA event, and one RHR pump for safe shutdown of the other Unit. The FMEA for Case 1 is given in Table 4 3 and the FMEA for Case 2 is given in Table 4.4. 9 I e
- 1 I
l [
1 i L Table 4.1 ; r DECAY HEAT REMOVAL FUNCTION SUCC230 C3ITZRI A ; IMMEDIATELY AFTER REACTOR TRIP l
- 1. Main Condenser Available OR
- 2. RHR System in Suppression Pool Cooling Modo l
OR
- 3. Rosator Depressuri:ation to Cold Shutdown Entry Conditions AND i 3A. RHR System in Shutdown Cooling Modo I
OR 3D. RHR with Fuel Pool Cooling Assist and RWCU System i r LOCA MITICATION i SUCCE33 C3ITE3IA i
- 1. EllR System in LPCI mode with two R!!R pumps ava11scio
- 2. Coro Spray System with ono Core Spray Pump available l l
i l t t 1 l i 1 l ! l i t i
~. --.
-In-Table 4.2 RilR SYSTEM AND SUPPORT SYSTEM SUCCESS CRI'tERIA RitR - LPCI Mode SUCCESS CRITERIA Flow Full Flow from two R!!R pumps.
liest Transfer Not Required. Pump Seal Cooling Not Required. Bl!R Room Cooling One Cooler. RilR - Shutdown Cooling Flow Full Flow from one RilR pump. Ileat Transfer One heat exchanger in operation. rump Seal Cooling Not Required. ; Rild Room Cooling One Cooler. - RifR - with Fuel rool Cooling Annist l Flow Full Flow from one Riln pump. llent Transfer Doth fuel pool heat exchangers in i opearation. Pump Seal Cooling Not Required. RilR Room Cooling One Cooler. RitR - Suppresalon Pool Cooline, Flow Full Flow from one EllR pump. llent Transfer One heat exhanger in operation, rump teal Cooling Required. { RilR Room Cooling One Cooler. RilR forvica Water
- Flow Full Flow from one pump to an notive heat exchanger.
i I l l
. v) .
i Sarvien Watar , Flow Full Flow trem two pumps to one or both headers. Lubo Water Full riow from one lubo water pump.
' For Decay Heat Removal Function Only.
f O
O '. T.ible 4.3 C,ise 1: Loss of Orr-SLte Power in Both Units BSEP Diesel Generitor Fiilure Criteria Ev.11ustion DG Failure Cemeinstions Consequences 1&2 o Loss of Unit 1 AllR Loop A. & B. MOV's o Loss of Unit 1 Room Coolers A & B 3&4 o Loss of Unit 2 RHR Loop A&B MOV's o Loss of Unit 2 Room Coolers A & B 1&3 o Loss of Unit 1 RHR Loop A MOV's (Unit 1 RHR Loop B available with two RHR Pumps) o Loss of Unit 1 Room Cooler A (Unit 1 Room Cooler B hva11able> c Loss of Unit 2 Room Cooler A ^ (Unit 2 Room Cooler 8 availaolo) o Loss of Unit 2 BilR Loop A MOV's I (Unit 2 RHR Loop B available with two RilR pumps) 2&3 o Loss of Unit 1 Bl!R Loop B MOV's
- (Unit 1 HilR Loop A availablo with one AllR pump) o Loss of Unit 1 Room Cooler b (Unit 1 Room Cooler A availablo)
, o Loss of Unit 2 Room Cooler A '
(Unit 2 Room Cooler 8 available) o Loss of Unit 2 RHR Loop A MOV's ! (Unit 2 AliR Loop B available with ono Alla pu=p) 1&4 o Loss of Unit 1 AllR Loop A MOV's (Unit 1 RHR Loop D available ! with one RilR pump) o Loss of Unit 1 Room Cooler A (Unit 1 Room Cooler B availablo) . - o Loss of Unit 2 Rocs Coolor B (Unit 2 Room Coolor A availaolo) o Loss of Unit 2 RHR Loop B MOV's (Unit 2 AllR Loop A available with ono RilR pump) 2A4 o Loss of Unit 1 RllR Loop B MOV's (Unit 1 RHR Loop A available with two AllR pumps) o Loss of Unit 1 Room Cooler D a q
s s a l e
\37 t
, i' (Unit 1 Room Cooler A available) o Loss of Unit 2 Room Cooler B
' .J (Unit 2 Room Cooler A available)
' o Loss of Unit 2 RHR Loop B MOV's
, [ ,
t (Unit 2 RHR Loop A available with two RHR pumps) , ,, HOTES: This e evaluation is' based on a complete loss of off-site power to the pli.nt and successful starting and loading of the diesel generators. A d review of the decay heat removal system study for BSEP (Reference 13) indicates that for a complete loss of off-site power event, each BSEP Unit must have at least one RHR loop with one RHR pump available for suppression pool cooling and with RCIC or HPCI available for Reactor Vessel inv4ntory mak'eup immediately after a reactor trip. The availability of the HPCI and RCIC systems will depend on adequate DC i power supplies during the loss of off-site power condition. This can ce accomplished by successful diesel starting and loading , or adequate i station battery capacity. For this study it ens assumed that adequate ! DC power supplies were savailable for at least three hours after a loss of
'ofhite power event. The DG failure combinations that leac to a-loss of- -
l l decay heat removal capability in either Unit are given below: i s .( ' 1 and 2
*^
l \ or 3 and 4 f
' t
( ) I . s 1 (,* b
+- , >
\
/ (.'
\. ,
1 3 2t f g -'5 t (
. t ,
w R 4
' c 4
- .n' ;
) g
Table 4.4 Case 2: Loss of Off-Site Power in Both Units And LOCA in One Unit l BSEP Diesel Generator Failure Criteria Evaluation I DG Failure Co=bination(s) Consequences 1&2 o Loss of Unit 2 RHR Loop A & B LPCI MOV. , o Loss of Unit 1 RHR Loop A & B Cooling MOV's. o Loss of Unit 1 Room Coolers A
& B.
o Loss of Unit 1 Core Spray Loops A & B 3&4 o Loss of Unit 1 RHR Loop A & B LPCI MOV. o Loss of Unit 2 RHR Loop A & B Cooling MOV's. o Loss of Unit 2 Room Coolers A
& B.
o -Loss of Unit 2 Core Spray Loops A & B. 1&3 o Loss of Unit 1 RHR Loop A LPCI MOV. o Loss of Unit 1 RHR Loop A Cooling MOV's (Unit 1 RHR Loop B available with two RHR pu=ps for cooling or LPCI.) o Loss of Unit 1 Core Spray Loop A (Unit 1 Core Spray Loop B available.) o Loss of Unit 1 Room Cooler A (Unit 1 Room Cooler B available.) o Loss of Unit 2 Room Cooler A l (Unit 1 Room Cooler B ! available.) ~ o Loss of Unit 2 RHR Loop A LPCI MOV. o Loss of Unit 2 RHR Loop A cooling MOV's (Unit 2 RHR Loop B available with two RHR pumps for cooling LPCI.) o Loss of Unit 2 Core Spray Loop A (Unit 2 Core Spray Loop B available.)
DG Failure Consequences Combination (s) o Loss of Unit 1 RHR Loop A LPCI 2&3 MOV (Unit 1 RHR Loop B available with one pump for LPCI.) o Loss of Unit 1 RHR Loop B Cooling MOV's (Unit 1 RHR Loop A available with one RHR pump for cooling.) o Loss of Unit 1 Core Spray Loop B (Unit 1 Core Spray Loop A available.) o Loss of Unit 1 Room Cooler B (Unit 1 Room Cooler A available.) o Loss of Unit 2 Room Cooler A (Unit 2 Room Cooler B available.) o Loss of Unit 2 RHR Loop B LPCI MOV (Unit 2 RHR Loop A available with one pump for LPCI.) o Loss of Unit 2 RHR Loop A cooling MOV's (Unit 2 RHR Loop B available -with one RHR-- - pump.) o Loss of Unit 2 Core Spray Loop A (Unit 2 Core Spray Loop B available.) o Loss of Unit 1 RHR Loop B LPCI 1&4 MOV (Unit 1 RHR Loop A available with one pump for LPCI.) o Loss of Unit 1 RHR Loop A
' MOV's (Unit 1 RHR Loop B available with one RHR pump for cooling.)
o Loss of Unit 1 Core Spray Loop A (Unit 1 Core Spray Loop B available.) o Loss of Unit 1 Room Cooler A (Unit 1 Room Cooler B available.) o Loss of Unit 2 Room Cooler B Cooler (Unit 2 Room A available.) o Loss of Unit 2 RHR Loop A LPCI MOV (Unit 2 RHR Loop B availabic with one pump for LPCI.) o Loss of Unit 2 RHR Loop B MOV's (Unit 2 RHR Loop A available with one RHR pu=p for cooling.)
.!3a-DG Failure Consequences Combination (s) o Loss of Unit 2 Core Spray Loo!-
1 & 4 (Cont.) B (Unit 2 Core Spray Loop A available.) 2 & 4 o Loss of Unit 1 RHR Loop B LPCI. o Loss of Unit 1 RHR Loop B Cooling MOV's (Unit 1 RHR Loop A available with two RHR pumps for cooling or LPCI.) o Loss of Unit 1 Core Spray Loop B (Unit 1 Core Spray Loop A available.) o Loss of Unit 1 Room Cooler B (Unit 1 Room Cooler A available.) o Loss of Unit 2 Room Cooler B (Unit 2 Room Cooler A available.) o Loss of Unit 2 RHR Loop B LPCI MOV. o Loss of Unit 2 RHR Loop B Cooling MOV's (Unit 2 RHR Loop
~ -
A available with two RHR pumps' - for cooling or LPCI.) o Loss of Unit 2 Core Spray Loop
'B (Unit 2 Core Spray Loop A available.)
NOTES This evaluation is base:d on a complete loss of offsite power to the plant and
- LOCA event in one Linit. This is the FSAR design basis for the plant and requires two RHR punys and one core spray pump for the Unit with the LOCA event, and one RHR pump for safe shutdown of the other Unit. Based on this evaluation, diesel generator failure combinations that did not meet the above success criteria were designated as failures for the case. These DG Failure Combinations are listed below.
1 and 2 3 and 4 1 and 4 2 and 3 6
4.2 Fault Tree Develoceent The fault trees for the cases previously defined were developed using the failure criteria developed in Section 4.1 and the Brunswick Diesel Generator fault tree model provided in Reference
- 1. The fault tree model from Reference 1 was modified to account for the plant specific configuration for the Brunswick . Diesel Generator System support systems and for the failure criteria developed in Section 4.1.
Initial fault tree models . developed for the Brunswick Diesel Generator System included failure to run contributions and testing and maintenance contributions. After initial cut set analyses using the WAMCUT computer code, it was determined that the fault trees could be reduced to include the most significant contributions while simplifying the input to the I FRANTIC III code by eliminating the "NOT" gates associated with the failure to run contributions. The reduced fault trees were developed with the following assumptions: o The failures to run contributions are negligible for loss of offsite power events not exceeding three hours in_ duration. , _. , o Testing and maintenance contributions are accounted for in the component inputs to the FRANTIC III code and need not be included in the reduced fault tree models. The reduced fault tree models are provided in Appendix B. 4.3 Inout Develooment The following section outlines the development of input data for the WAMCUT code and the FRANTIC-III code. t The components used for both codes are defined in Table 4.5. The inputs for the WAMCUT code are provided in Table 4.o. The FRANTIC-III test configurations are given in Table 4.7 The inputs for FRANTIC-III are given in Table 4.8 (Case 1) and Table 4.9 (Case 2). A detailed discussion of specific input categories is provided in section 4.3 4.3.1Initiatine Events .. l - L 4.3 1.1 Loss of Offsite Pcwer Events I Loss of offsite power event frequency data were obtained frc= Reference 10. For Case 1 (Loss of Offsite Power), an event frequency of 0.01/reacter year was used. This value corresponds to a less of offsite power event lasting for three hours. It was I assumed that there is at least 3 hours of DC power capacity subsequent to the loss of offsite power event. I
- , . , --, ..--__ , .~.,,-_,.,__ ~ ,.,__ -. . , _ , . , _ . . _ _ _ _ - , . , , , , . . - . - ..__.m _ , , - _ z. . . , -...... - -
r < For Case 2 (Loss of Offsite Power with LOCA in one Unit)This nn event frequency of 0.088/ reactor year was used. corresponds to the frequency of any loss of offsite power event which persists for = ore than a few seconds. It was assumed that a LOCA event requires successful starting cf the diesel generators in 10 seconds. The probability of offsite power events for exposure times less that one year was calculated using the following equation: P.= (Annual Initiating Event Frequency) x (T/365) (4.0) Where P = Probability of initiating event occuring over a time "T" T = Time of exposure, Days See Tables 4.6, 4.8 and 4.9 for specific values. 4.3.1.2 LOCA Events LOCA event frequency data were obtained form Reference 11. ' A LOCA event frequency of 1.4 E-3 per reactor year was tisec~ ~ in Case 2. LOCA event probabilities for exposure times less than one year were calculated using equation 4.0. F 4.3 2 sucocet systems and Common Cause Comconents The input date for support systems and common cause components were obtained frem Reference 1. These components have demand related inputs and remain unchanged by the different FRANTIC cases. Specific values are given in
, Tables 4.6, 4.8 and 4.9.
4.3.3 Diesel Geneester Cccoonents The inputs for these ecmponents were derived from Reference parameters, and proposed testing 1, current testing parameters. The current testing parameters and the proposed testing parameters are given in Table 4.7. Other specific values for these components are given in Tables 4.6, 4.8 and 4.9 _ L
Table 4.5 WAMCUT/ FRANTIC III Comconent Definitions Component Component Na=e Definition LOFP Loss of Offsite Power Event LOCA 1 LOCA Event in Unit 1 LOCA 2 LOCA Event in Unit 2 DG1UD DG No. 1 Unavailability DG2UD DG No. 2 Unavailability DG3UD DG No. 3 Unavailability DG4UD DG No. 4 Unavailability SWCCF SW System Common Cause Unavailability DGCCF DG System Common Cause Unavailability BACCF DC Power System Common Cause Unavailability DGHEC DG Human Error Com=on Cause Unavailability DC1UD Unit 1 DC Power System Unavailability DC2UD Unit 2 DC Power System Unavailability SWIUD Unit 1 SW System Unavailability SW2UD Unit 2 SW System Unavailability D e O
Table 4.6 WAMCUP Cercenent Incuts - For Reduced Fault Tree Models Component Source Name- Un2vailability Reference. (Note 1) 10 LOFP (Note 1) 11 LOCA1 (Note 1) 11 LOCA2 5.0E-2 14 DG1UD 5.0E-2 14 DG2UD 5 0E-2 14 DG3UD DG4UD 5.0E-2 1 SWCCF 8.0E-5 1 DGCCF 1.5E 4 1 BACCF 1.0E-5 1 DGHEC 2.0E-4 1 DC1UD 1.0E-4 1 DC2UD 1.0E 4 1 SW1UD 2.0E-3 1 SW2UD 2.0E-3 1 ... .. NOTES:
- 1. For intiating events such as Loss of Off-site Power or a Loss of Coolant Accident, the probability of the event occuring was determined using equation 4.0.
I
Table 4.7 FRANTIC-III Test Confieurations 40 Day Standbv (Note 1) Test Interval - 31 days staggered bases Test Duration 4 hours (DG assumed to be loaded during test) 3 Day LCO (Baseline Case) Test Interval - 12 hours Test Duration - 1/2 hour (Unloaded test) Action Time - Testing starts within 2 hours after LCO is initiated 7 Day LCO (Proposed Case) Test Interval - 3 days Test Duration - 4 hours (DG assumed to be loaded during test) Aotien time - Testing starts within 2 hours after LCO is initiated. NOTES:
- 1. The 40 day period was chosen on the basis of a =can time between DG outages of 1,000 hours (See Section 3 1) . During this period, all four diesels are assumed to be in a standby configuration.
- 2. During the LCO period, one diesel is assumed to be out of service.
.l
^ Table 4.8 FRANTIC III Inout Summarv for Case 1 - Loss of Offsite Power A. Initiating Event (s) Cceponent Residual Unavailability Name 365 Days 40 Days 7 Days 3 Days LOFP 1.0E-2 1.1E-3 1.9E-4 8.2E-5 B. Support Syste=s and Cc==en Cause Comoonents Component Residual Name Unavailability SWCCF 8.0E-5 DGCCF 1.5E-4 BACCF 1.0E-5 SW1UD 2.0E-3 SW2UD 2.CE-3 DC1UD 1.0E 4 ~~ ' ~ D C2UD 1.0E-4 DGHEC 2.0E-4 C. Diesel Generator Comoonents 40 Day Standby Weibull Test First Time For Component Scale Interval Interval Testing Residual Name Parameter (Days) (Days) (Hrs) Unavailability DG1UD 1.03E-4* 31 8.33E-2 4 5.0E-2 DG2UD 1.03E-4 31 7 4 5.0E-2 DG3UD 1.03E-4 31 14 4 5.0E-2 DG4UD 1.03E-4 31 21 4 5.0E-2 "Obtained from Reference 1 3 Dav LCO Weibull Test First Ti=e For - Cc=ponent Scale Interval Interval Testing Residual Name Parameter (Davs) (Days) (Hrs) Unavailability DG1UD 1.03E 4 5 8.3E-2 .5 5.0E-2 DG2UD 1.03E-4 5 8.3E-2 .5 5.0E-2 DG3UD 1.03E-4 5 8.3E-2 .5 5.0E-2
- 1. .. .
7 Day LCO Weibull Test First Ti=e For Cc=ponent Scale Interval Interval Testing Residual (Days) _ (Days) (Hrs) Unavailability Name Parameter OG1UD 1.03E-4 3 8.3E-2 4 5.0E-2 DG2UD 1.03E-4 3 8.3E-2 4 5.0E-2 1.03E-4 3 8.3E-2 4 5.0E-2 DG3UD t i i l
4 Table 4.9 FRANTIC III Input Summary for . Case 2 - Loss of Offsite Power With LOCA in One Unit . A. Initiating Event (s) Component Residual Unavailability J Name 365 Days 40 Days 7 Days 3 Days LOFP 8.dE-2 9.6E-3 1.7E-3 7.2E-4 LOCA 1 1.4E-3 1.5E-4 2.7E-5 1.2E-5 LOCA 2 1.4E-3 1.5E-4 2.7E-5 1.2E-5 B. Succort Systems and Co==en Cause Comoonents Same as in Case 1, Table 4.0. C. Diesel Generator Comoonents Same as in Case 1, Table 4.3 4 _ . I e T 4
- - - - e .- - ,, ,- , - , , , . _ _ _ . , , _ _ . , , _ ,;,_.,_
.4.4 WAMCUT/ FRANTIC III Methodolcev 4.4.1 Unavailability Modeling Several computational tools for unavailability modeling currently exist. The general structure for defining logic models and input parameters is co==en to the various codes. For this analysis, the FRANTIC III code (Reference
- 4) serves as the reference code. FRANTIC III is a time-dependent code that allows for unavailability calculations that relate to different test schemes, test intervals, repair ti=es, failure detection inefficencies, probabilities of test-caused failures, and preventive maintenance. Key ti=e-dependent parameters in the code are:
o Test interval; the first test interval (T)) and the periodic Test interval (T 2) o Average test pericd (TAU) o Average repair period (REPAIR) After a ti=e S = TAU + REPAIR, the component is ready to go_, _ into operational standby for the re=ainder of the test interval. If the remainder is called S 0, then: T2*3+30 = TAU + REPAIR + SO (4*l) For each one of the three time periods that make up the tes; interval a unique availability quantity is ce=puted: o Between-test unavailability contribution for time period S 0 o Test contribution for ti=e period TAU o Repair contribution for ti=e period REPAIR Co=putation of the test and repair contributiens requires as data input the average test period and the average repair in hours, respectively. A standby component is typically characterized by a constant unavailability per demand and ' running failure rate when it is in operation. For the between-test unavailability models the failure rate required as input for FRANTIC is the standby failure rate and not tne running failure rate. Empirical data show, however, that for a diesel generator set the standby failure rate and the running failure rate can differ by up to two orders of magnitude. A detailed data analysis must therefore precede the application of FRANTIC III. i , e
Empirical data indicate that the unavailability during the standby period consists of a constant (ti=e independent) part q, and time-dependent part LAMEDAX(T) (where LAMSDA is the standby failure rate). The constant part includes testing and maintenance-induced failures, as well as failures modes related to start-up and change of state. The standby failure rate represents contributions frem latent faults that accuculate during the standby period. A further discussion en standby failure rates is given in Appendix A. Assuming a linear unavailability model, then existing U.S. industry-wide data on diesel generaters indicate a 10% time-dependent part in the unavailability for a monthly test interval (Reference 5) In Figure 4.1 the ESEP specific data is compared with the industry-wide data. 4.4.2 Precaratien of FRANTIC III Incut The ec=putei- code input consists of two groups: (1) a system model in terms of an unavailability polync=ial, anc (2) cc=ponent data in terms of the failure, test anc maintenance characteristics. Preparation of the syste= model--i.e., the unavailability polyncmial--is done by using a qualitative fault tree code that generates a - boclean-expression in terms of minimal cut sets'(critical failure co=binations). The industry code WAMCUT (Reference ) is compatible with FRANTIC III and is a convenient preprocessor that generates the unavailability polync=ial for FRANTIC III. In general, the fault tree model used as FRANTIC III input needs to be echar2nt (i.e., inclusien of NOT gates in :n? fault tree structure should be avoided). This is simply a precaution again:t having a situation where the derivec unavailability polync=ill is incceplete or incorrect. The problem with inccherent fault trees is that there is no way of knowing in advance if a certain fault tree evaluation alger:.thm will deliver the right answer all the tire (Reference 7). In this case study, the system unavailability is evaluatec for operational standby assuming no unscheduled =aintenance, for LCO action statecent where ene diesel generator is cut of service and tht: others tested on an augumented basis and . finally for the case where the effects of extended allowable downtime are evaluated. The develop =ent of logic codels is discussed in more detail in paragraph 4.4.3 The FRANTIC III code handles four ccaponent types; constant unavailabilty cc=ponents, nonrepairable cc=ponents, monitored components, and periodically tested cocponents. The latter type is extensively used for this Technical Specificatien optici:stien analysis. Table 4.10 is a summary of the para.eters used to describe a periodically tested component in the FRANTIC !!! code. 1 5 . w
3 ,*. - 1 0 ,
~
~
/
/
a x r4 rH
.A 10-2 __
.a n _
5
- E Industry Wide Data (Reference 4-3)
BSEP Data Ucing ITUREG/CR-2909 Approach . l l l I 1 2 3 4 Test Interval in Weeks t Figure 4-1. Diesel Generator Unavailability at Surveillance Tecting as a function o f Test Interval.
~
28 6
The test override (bypass) unavailability q0 is a paraceter used to represent the probability of overriding a test and placing the cc=ponent in operation if demanded. When there is a test override, the unavailability will be lowered. A 90 = 0.1 =eans that there is a 90 percent value to the time-dependent unavailability q(t) . Additional parameters are ITYPE and f. The ITYPE is used to describe the renewal model that is used. Three renewal models are available in FRANTIC III: GOOD AS NEW test and repair (ITYPE = 1), GOOD AS OLD test and repair (ITYPE = 2), and GOOD AS OLD test and GOOD AS NEW repair (ITYPE = 3) . The test factor, f, is used to quantify the effects of wearout or burn-in en system unavailability. This factor allows for the investigation of changes in failure rate as a function of the nu=ber of tests perfermed. If f is set less than 1, then the reliability i=provement througn testing is studied, and f greater than 1 represents degradation. 4.4.3 Develoement of Loeic Models The basic philosophy behind the F'RANTIC III code-- and other - ~ ~ s1=ilar coces--is to relieve che analyst frc= the tedicus work involving the derivation of the often quite co= plex analytical expressions that reflect a particular renewal process. Given that a fault tree describing the basic functional structure of a system is available, the FRANTIC III code assigns the appropriate cc=ponent unavailability codels to the tree. This does not =ean, however, that any fault tree structure is applicable to a particular problem. The logic =cdel develop =ent needs to be depencent upon the particular test and maintenance strategies that are to be analysed. For the BSEP diesel generator system, the effects of changing test frequency and allowable downti=e are to se evaluated. The system can be in either operational stancoy with no diesel generater out of service for =aintenance, or in an LCO action statement with one diesel generator down for maintenance. In the forcer case there is a =enthly test interval and in the latter case the diesels in standby are tested every 12 hours. As a consequence, FRANTIC III requires one codel for each case; a " baseline" model and an "LC0" =cdel. This is chown in Figure 4.11 using a cause- . consequence diagrs= fer=at. The co=plete fault tree codels used for each case are presented in Appendix B. s .
- e. c.ula. p:i .:.1(
a n sge,. n
,. n ni s ; ,
n a y ,.
.in,.,M. F a
; i w m. cr sso u. rvn. t .se-O !.% .d ,i e c ;. n mG . in, e. l .
sit 5 ran, v i, e 1. n if u. v rut 6 i: n:e rva-e4 nneD9,nnn-e
- u ;
n .- ucuv a n ,yr, i i htm... s m# f eep ,,m-,14 a f4 a j. 1 . at
.. nni si{ m' f 3 .is 64 u van vi t.u.4 -
u rJ $n 7AMDBn 04.sn.H. n /en,1n s . . m s s.s a ., n ) unnuun u.vy. ea s u;.;9 1 a A,
. i.'csay4r. ] ,,w scI,', c. . O u c a e. :4 u r c :c e. tar ,
1.mn . s n,e- n ,- , .- g a /. G o,
.v"!Cf n ,i /Js.L$
v g a t1
. M d. O.w' l' m , nn nh sa -O,.ll 1O n a .J trnunle-01an kMe.a(
[:R f 4, Q. is e - , n i . .f us e = n
,n A
-; m , p e
.t .a 100v 111 VC1"/ C1 J ,, e
- J A ls uGMOa * -
S P.ts ,q, s.
'1.*. . . e . , - -. m i n - 4 w a . e r-
!. n!! MCbj O :
1 .:. I' a l'O b V C h 8.- 111 l' C a*V C3.1 Ti A!l g;L T*nn 1;ng a f a a .,,t=
- 4. = , 1 4 ,,. J , 'n sbn6 - gg a gg , gig, (:,g g ,
1, f. . , .s e
~L n>m m- - -
fs g_ yvn - -un 3is .- --s ,.a me =* .s a p a r-J)1'. /. :i A J2 1 ) iilt a O, p' a 1*".* P., C 4 l'O 1I! f:-
.Mf t n >= 1 . s e r >= m s i J .=
UY <1'1' 4 MC ,dn==,*,.A10dv.0, 1-%s : _9 E il 2.V M1 A C!J1 a 10];) e n n C O ,b = , - se n.se. s1 % , s - ~ . . nA-r j* n e e n===-a ."i 5 -
='
Iaam.~ v_ e. 1*1 1 1 N. 9l.![.122 1 1 91 I "3 e O ! . N NY. C U d b E E ' 'ad I N I E . D $ O a.I' s=t n A7.;.,;, = -
.l5 s*. pg 3 - a g g us ss s ,m .e= s= .a.. af. g is n g sg g n-srgap
[ eh b e'. kb~ w NN h
- b h .#. O I $ k 5a .h .de b N o. I i N [,,I, E l' T n e .- I'* .s n (. 3
. SCOD .. CtU V 'Lr 3" 2 I*AW ft 1.f r. 4 I. . g 1 1 e h. s .- a v: .4 .. . . c. s. 1 .
Albu e1 n e !. u u 1 1, m 22:alf c, i-w. Gi ic 4 C A* a 4'- mis m 4
? E.1 4 Q 3 94, ,n . . P *- 4 y g yg ,.,i , ; g i 1. 4 L 11#!1090.1 PsU u*.I a a a
I i i l l , 1 1 s 6
Diuel generator ' unavailabilitu nodel K : Fraction of time Systen in operacionali in LCO state standbu i K *i No l Yes N1-K i 1 LCO action statenent t Operator actica l All the DGs operational I uithin time I l . l l ! ! l No Yes No Yes
! l !
c-I Test and Repair Model l Systen Restoration ' i Na i fes ! I i _ . ... .
- 1 1 !
T. Plant Shutdowa 7 $ Continued 0ceration 7 FICURE 4-11. hsic structure for logic nodel development.
- . _ .L.,. . - . . . _ . . . . . , -
l 4.5 Risk Analysis Methodology There are a number of probablistic approaches to establishing LCOs and surveillance requirements for standby safety systems as described in NUREG/CR-3082 (Reference 9). Fce this study the relative risk method and the average annual risk comparison method were utilised. 4.5.1 Relative Risk Method Of particular interest is the relative risk method. The relative risk method li=its component outage time by constraining the exposure to risk during the outage to be no larger than the expcsure to risk during a baseline period in which the outage is assumed not to occur. A preliminary investigation of the risk during an LCO and during a normal standby period indicates that current Technical Specifications may be overly restrictive and yield significantly lower risk during LCOs than during the standby period. It was decided to adopt the relative risk method for the following reasons: o It constrains the outage duration by requiring that the ., , ' risk during the A0T be less than the risk curing a baseline period. o Insights are more clear when assessing risk on a relative basis rather than on an absolute magnitude basis. o The determination of the absolute limit of the system average unavailability is not required. The relative risk criterien is: R0 gg Rt ( .2) Where: Risk = Probability of not having AC power available for Decay Heat Re= oval or for a Loss of Coolant Accident (LOCA) . E0 = risk over the perioc of the outage of duration To RT = risk over the baseline period T, assuming no outage 80 and RT are in turn defined as: For Case 1 (Loss of Offsite Power).
i
~
R O* 1 (T0/365) 0 0 (4.3) - R 7 = 13 (T/365) QT (4.4) Where: 1 = Loss of offsite Power event frequency (events per reactor year) TO= Outage period, days T= Baseline period, days QO= DG system average unavailability over the pericd of the outage with one diesel out of service QT= DG system average unavailability over the period When all four diesels are in the nor=al standby configuration (1. e., baseline period). For Case 2 (Loss Of Offsite Power With LOCA In One Unit)
~ ~ ~~
R0* 1 (T0 /355) [2 (T /365) 0 QO B0* 1 2 (T 0/355)2 0 0 (4.5) RT* 132(T/365)2 o (4.6) Where: 2 = LOCA event fregency (events per Reactor year) 4.5.2 AVERAGE AN!!UAL RISE COMPARISC:1 METHOD ^ In order to evaluate the relative effect of the allowed outage ti=e ( ACT) on an annual basis, a =ethod wss required to calculate average annual risk using the results of paragraph 4.5.1. CASE 1 The average annual risk for Case 1 can be calculated using . . the following equation. . ( } Ra* 1 Where: Ra = Ave: e annual risk s ,
r-1 = Loss of offsite power event frequency (No. of events per (year) i Q= Average DG system unavailability Assu=ing that the status of the DG system alternates between the LCO state and the standby state, the average DG system unavailability can be determined over one cycle (1. e., TO+ T) by the following equations Q = (T0/TO + T) QO + (T/TO + T) QT (N 0) Using equations 4.3 and 4.4 CO 3 0 365/(To x1) 3 07=R7 365/(T x 33) Substituting these equations into equation 4.8 Q=R0 365/(T0 ' T) X , 1 +3 T 365/(T0 + T) x ) Substituting this expression.into equa; ion 4.7
~
1-
. Rg= R o 365/(T +T)j+R 7 365/(To + T)
R A=[365/(To+T))x(R0+3) T ("*9) Case 2 The ave-age annual risk for Case 2 can be calculated using the ft owing equation. R g=[1h2 (4.10) As deter =ined previously by equation 4.3 Q = (T0 /IO + I) CO + (I/IO + T) QT Using equatiens 4.5 and 4.6 00= R 0 (365/To }2 /}1 } 2 QT= R T (365/T)2 /b1 2 Substituting these equations into equations 4.8 and 4.10 R A = (365)2 80/Yo (T + To ) + R 7/T(T + TO) (4.11)
It is important to select an appropriate baseline period, T, for a fixed , T o such that the risk levels truly represents the actual values associated with the systems of interest. A reasonable definition of baseline period is based on the " natural" time periods that conform to the operational evolutions that the diesel generator systems go through as part of the normal plant operating procedures. For a yearly operations cycle, it is convenient to divide the total time into a normal standby period, unscheduled outage period, and a scheduled outage period (i.e., refueling outages). For this report the mean ti=e between DG outages (i.e. 40 days) was used as the baseline period for which all for diesels are assumed to be in the nor=al standby configuration. The FRANTIC code is used to compute R0 and R7 directly.
- O &
i 9
--r- -- - - - - -
~ - ' ~ - - . - __. __
- n,: -
a .. 5.0 Results 4 5.1 Risk Analysis As discussed _ in Section 4.5, the relative risk method and the average annual risk comparison method were used to evaluate A0T The extension and LCO surveillance test interval extension. results in Table 5.1 indicate that there is less than a half order .of magnitude increase in the level of risk between the baseline (3 day) LCO case and the proposed 7-day LCO case. Moreover, the risk levels for both LCO cases are lower than the 40-day standby risk level. The results in Table 5.2 indicate that there is also less than a half order of =agnitude increase in the average annual risk levels between the 3-day LCO case and the 7-day LCO case. Table 5.1 Relative Risk Analysis Results Avg / Max Avg / Max Avg / Max Plant Risk over Risk over Risk over Initiating the period the period Event Risk (Note 1) the period ,,_
~
of the of the of th'e 3-day LCO 7-day LCO 40-day LCO (Current) (Procosed) (Baseline) Loss of AC Power 4.38E-6/ 1.067E-5/ 1.37E-5/ Unavailable 4.432-6 1.19E-5 2.14E-5 Offsite Power for Decay (Case 1) Heat Removal In Unit 1 or Unit 2 . Loss of AC Power 1.84E-9/ 1.03E-8/ 7.07E-8/ Unavailable 1.86E-9 1.1E-8 1.086E-7 Offsite Power for Decay With LOCA Heat Re= oval in One or for Unit (Case LOCA Mitigation 2- FSAR Design Basis) NOTES:
- 1. Risk here means the probability of an event cccuring over the pericd of time considered.
? -
TABLE 5.2 AVERAGE ANNUAL RISK ANALYSIS RESULTS Average Annual Average annual Risk *4ith Risk With Initiating Plant 3-day LCO 7-day LCO Event Risk (Note 1) ( Current) (Procosed) Loss of AC Power 1.5E-4 1.9E-4 Offsite Power unavailable (Case 1) for Decay Heat Removal In Unit 1 or Unit 2 Loss of AC Power 7.4E-6 9.2E-6 Cffsite Power unavailable with LOCA for Decay in One Unit Heat Removal (Case 2 or for LOCA FSAR Design Mitigation Basis) . .- Notes
- 1. Risk here means the probability of an event occuring over a one year period.
- 2. The risk levels on this table do not include recovery factors for the RHR system (see reference 13) t
. st 4 -
5.2 Sensitivity Analysis A sensitivity analysis was performed on two parameters for the loss of offsite power case (Case 1); these are diesel generator unavailability on demand and diesel generator Weibull scale parameter. The diesel generator Weibull scale parameter is merely the hourly standby failure rate for an exponential distribution between testing. The results in Table 5.3 show that the risk levels for all configurations are approximately proportional to diesel generator unavailability on demand. The Weibull scale parameter has a higher sensitivity impact at larger exposure durations (i.e., 3 days, 7 days, 40 days, etc.) Table 5.3 Sensitivity Analysis Results Initiating Event: Loss of Offsite Power Sensitivity to DG Unavailability on Demand Avg / Max Avg / Max Avg / Max Risk over the Risk over the Risk over the DG period of the period of the period of the Unavailability 3-day LCO 7-day LCO 40-day Standby 2.0E-2 1.76E-6/1.81E-6 4.59E-6/5.31E-6 5.8E-6/1.11E-5 4.0E-2 3.49E-6/3.54E-6 8.61E-6/9.35E-6 1.07E-5/1.76E-5 5.0E-25 4.38E-6/4,43E-6 1.067E-5/1.14E-5 1.37E-5/2.14E-5 6.]E-2 5.28E-6/5.33E-6 1.28E-5/1.35E-5 1.72E-5/2.56E-5 8.0E-2 7.13E 6/7.18E-6 1.7E-5/1.78E-5 2.53E-5/3.51E-5 Sensitivity to DG Weibull Scale Parameter (Hourly Standby Failure Rate) DG Weibull Avg / Max Avg / Max Avg / Max Scale Risk over the Risk over the Risk over the Parameter 3- day LCO Period 7-day LCO Period 40-day Standby Period 1.03E-4' 4.38E-6/4.43E-6 1.067E-5/1.14E-5 1.37E-5/2.14E-5 2.0E-4 4.43E-6/4.53E-6 1.13E-5/1.27E-5 2. 39E-5 / 4. 37 E-5 4.0E-4 4.52E-6/4.72E-6 1.2SE-5/1.54E-5 5.17E-5/1.07E-4 6.0E 4 4.62E-6/4.92E-6 1.38E-5/1.81E-5 8.64E-5/1.85E-4
- Reference Case
l ! 7 L / ,, ,- I i , l 5.3 Benefit Analysis / 5.3 1 Benefits of the Procesed Technical Soecification Chances Testing frequencies and inoperability time limits defined in Technical Specificatiens have typically been chosen mainly on the basis of engineering judgment. Results frc= unavailability modeling can be used, however, as an aid t'o engineering judgment. Revisions to Technical Specifications can involve changes in testing frequencies and inoperability time li=its for standby equipment. These alterations can be shown to have negative effects on system unavailability anc also on plant risk. There are also benefits that compensate for a slight increase in risk. To arrive at a balanced perspective on the proposed Technical Specificatien revisions, it is essential that the theoretical asses ==ents are complemented by discussions that highlight the potential benefits in a reasonsole manner. This section provides an overview of correlations between testing frequencies and downtime, respectively, and diesel generator reliability. 5.3.2 Benerits Resulting from Increasing the Aurmented Surveillance Test Interval . ,. The purpose of periodic testing of standby cc=ponents is to demonstrate component availability. By detecting latent faults during testing, it is possible to maintain co=ponent operability. On the other hand, if the testing is too frequent, the compenent can be degraded througn wearcut effects. This in turn increases the component downti=e, ana reduces the reliability of the cc=ponent over a perica cf time. A diesel generator is typically tested once a conth, and in addition, it is tested on an accelerated basis if another diesel is down for maintenance and the power plant is in an LCO action statement. The current BSEP Technical Specifications prescribe testing every 12 hours while in an LCO. Of interest here is whether any correlatien xists between the number of failures and the testing frequency. Industry-wide diesel generator data as presented in .'iUREG/CR-1309 * (Reference 1) is here usedThe to result relato demand failures to shcwn in Figure 5-1 actual testing frequency. indicates a trend towards higher average numoer of failures as the test interval is reduced to below 400 hours. This is a general observation that includes experience from various types of diesel failures and is not restricted to LCO-induced tests only. . O
t o 9 f. 4G l l
.a l
I l n ! i l I t* I W +.
. r.
.kb I N D A I l I
s e no 4 e
.,,aseo.a. N..
. . . .4 6 l
4 l f f, I
=*
u i . I
* -)
- 4. /
*6 O
,1
-- 9 k =
g i'l AJ ~v J ik .1[ i.'l
" m .l. G'..,,....n.,.,
I, C D 0 i a ' .! '.! v.*.
- a '.t *)
L.r . f .;., .:. . 4 I V 4,d 1'd
. ., . .. . . r. . , , .. r . . .(I." */ A* '.' {.1W
*.? ?)C '.' .. i ! ' '*" !* g . . .'l Ct A
' thit l~
l il:*- * !:1 ! $ s' *ir21A A i14' jQQ /*t d.? G T, . 0 f} d b"s a M tsC
\i 8 I %'H t1 s .:. Q 1.5 .. .., .: J *28 *) f / U .? '
ne ..W,t :i T,' f1:5: d.n o u ,, 49OO O G4;; P. t tm a k
/.i e ./ .t u Mi t "..'
' 7. .i .V e 4 vM:.h{
V . .t jggg a g e *
- rp*
- a.* *i
- q v.=l * *1,8* =*** se . . see e-e e.r - .3 **.9t *.9 a Ia= . l .ru ft 9.1.g'
j a ,* C .'l .s v A d l)e is4; H is*t.'Ii
.t >,
* * * =e A C ;.: y :. L' c.' d '.h; i i tr . : t.ti a <,:,t
. '.'.':.1.;^s u u us .
hi. . . 1. q .s .. ,q,' L' , i . . . . , . 9 r. 4 '. c. ' ' i l '. 'P P ' " t .,. ',* '.} (M i .i.'.;'lJ.t u i. .i 4 ....6t...."" J. 4. i e e d I diJ a s.w bt v.i w.V/
% 4. .
.s 4.i, t c.
A 4
From an engineering judgment standpoint, this conclusion seems correct, since excessive amounts of testing results in faster wearcut of components due to friction effects and cyclic fatigue effects. For- example, the BSEP diesel Senerator manufacturer has recommended replacement of tne dowel pins that couple the engine shaft to the engine driven jacket water pump after every 600 starts due to observed cyclic fatigue failures in the pins (See Section 5.4). Similarly, the NRC staff has encouraged nuclear plants to reduce the number of unnecessary diesel generator fast starts as a result of observed industry trends in diesel generater wearout rates (Reference 14). A increase of the augmented surveillance test interval from once per 12 hours to once per 72 hours would result in a slightly increased standby unavailability of the diesel generator sets, since it would require a greater interval of time to discover any latent faults through testing. However, this test frequency reducticn co=cined with the extended ACT would reduce the number of required tests during an LCO condition by a facter of two. l 5.5 3 Benefits Resulting frem Extending Allowed Outage Time . .. " . (A0T) Li=1ts l The AOT limit allcwed for safety equi;=ent in Technical Specifications is typically 72 hours. Exceptions are the seven day li=it for systems such as the contain=ent enclosure cooling system supply and exhaust fans, and tha
> containment enclosure emergency air cleaning system filter and fan units in certain plants. Factors that deter =ine these time limits are past experience, equip =ent accessibility, unique equipmenc design features, etc., and i
of course the required system availability. The present ESEP Technical Specifications for the diesel generators require that both Units be placed in cold shutdown if the 12 hour A0T is exceeded. If a ti=e li=it does not reflect actual maintenance history, then this could potentially result in unnecessary forced reactor shutdowns. The industry-wide data on =aintenance curaticn for diesel ganerators indicate that with a 72-hour li=it, the mean duration is 21 hours, with an upper bound of 2d hours. For the cases with a seven day limit the mean caintenance duration is 40 hours, with an upper beund'of 60 hours (Reference 2). The BSEP data indicates that diesel generator maintenance activities at this plant have been r completed within the 72-hour period on an average basis. However, there have been some instances where the outage time exceeded 72 hours, ycr these cases, botn BSEP Unita
/ were shutdeun at the time or the plant continued to operate under special NRC exemption. To date diesel generator
m , y, ,
.t -43_
( si ~s yJ. (1 , A 7
/q,/ a f:k incperability has not resulted in forced shutdowns at the '
Brunswick Plant. However there is a possibility that a NfY i[( ~ forced shutdown may be required at some future date as a
> result of the present LCO. We have estimated that such a forced shutdown would result in a daily revenue loss of i
$492,000.00.
f
' The\ extension of the Technical Specification Allowed Outage i' '
./ ,_
,Ti=e would also provide the plant staff with increased flexibility in the performance of scheduled maintenance
- , activities. For example, the dowel pin replacement j ,.
evolution mentioned previously could potentially result in an equip =ent outage of somewhat greater than 72 hours but less than 7 days.- Under the current Technical Specification requirements, this maintenance can only be scheduled during periods' of extended plant shutdowns (e.g., refueling outages). Under the proposed change, the above evaluation l could be performed during non-shutdown times. l
- Additonally an increase in the allowable outage ti=e will
.o reduce the stress on maintenance personnel to complete diesel generator repairs to meet the. current 72 hour ~
^
~
limit. Maintenance personnel would be given more flexibility to perform their repairs. This would in turn increase the thoroughness -and quality of the maintenance evolution. No quantitative data exist to allow an accurate
' assessment of the benefits of this effect. However itdue seems reasonable that positive benefits will ensue to
(. /
> reouction in time stress on personnel.
,js
#~ Diesel Generator Weareut Analysis 5.4 f
- In triis 'section the potential wearcut characteristics of the f
BSEP diesel generators are evaluated using the Weibull methodology (Reference 12) and the DG dowel pin failure data.
- 5.4.1 Plant Data Evaluatien A review of plant records indicate the following data on the failure of DG dowel pins.
p .
\
f? DG Number of DG No. Starte to Failure l' 1230 2 1659
- 3 1492 4 1494 1
3
~ .
I
These data can be translated in terms of time to ' failure ! using the following equation. . TTF = (Nu=ber of starts to failure) T (5-1) Where: TTF = Ti=e to failure
-T = Test interval.
The'following data were developed using equation 5-1 above: T = 12 hours DG Number of ; No. Days to Failure 1 615 2 829.5
~- -
3 746 4 747 T = 3 days DG Number of . No. Days to Failure l l 1 3690 2 4977 3 4476 4 4482 T = 31 days , DG Nu=ber of No Days to Failure 1 38130 . 2 51429 3 46252 4 46314 5.4.2 Weibull Methodoloey The WEIPLOT computer code (Reference 12) was used to calculate the Weibull parameters for the dowel pins. The Weibull methodology is based en the following equations:
- )
i l l R(5) = exp (-((t- (0))/n) *"B) (5-2) andMTTF=n['(1+1/B) (5-3) Where: B = Weibull slope or shape parameter n = characteristic life t = time to failure MTTF = mean time to failure P = Gamma function 5.4 3 Results The WEIPLOT cc=puter calculations produced the results shown in Table 5.4. Table 5.4 DG Devel ?in
- dear Out Analysis Result Standard MTTF Deviation !! Correlation j[ B (Days) (Days) (Davs Coefficient 12 hrs 8.6 681 94 720 .90 3 days 7.7 4382 670 4661 .9o 31 days 7.7 45286 6226 48164 .96 The results above clearly show the detrimental effects of accelerated testing on the BSEP DG dowel pins. The MTTF is significantly reduced as the test frequency increaseo.
6.0 conclusions and Reco==endations . The results of this study indicate that for all cases analy:ec, the risk level during the LCO conditions considered is less than the risk level during the normal standby configuration of the diesel generater system (i.e., all four diesels in standby). Although there is a slight increase in the LCO risk level when the ACT is extended to seven days and the diesel generator surveillance test interval is increased to 72 hours, the increase is within the uncertainty and variation of the input data used in this study. It is further concluded that the slight increase in the average annual risk level is warranted in light of the benefits associated with reducing diesel generator wearout, reduced chances of forced plant shutdown, and reduced time stress en
=aintenance pe,rsonnel.
The following reco==endations to revise the Diesel Generator System Technical Specifications have been formulated as a result of this study: o The A0T can be increased from three _ days to_seven cays _ , o The surveillance test interval during the LCO can be increased from 12 hours to 72 hours. TA , .. o During( LCT- surveillance testing, the diesels should be 1ca 'ed ;to at least 50i, of rated load for at least two houra. This is similar to a monthly load test requirement. I
m 7.0 References- ,
- 1. Battle, R. E. , and D. J. Campbell, " Reliability of Emergency AC Power Systems at Nuclear Power Plants", NUREG/CR-2969, July, 1983.
- 2. Public Service Company of New Hampshire and Yankee Atc=ic Electric Company, "Seabrook Station Probabilistic Safety Assessment', December, 1983
- 3. Vesely W. E. , and F. F. Goldberg, " FRANTIC III - A Computer Code for Ti=e Dependent Unavailability Anslysis", NUREG-0193, October, 1977
- 4. Ginsburg, T., and J. T. Powers, " FRANTIC III - A Computer Code for Time Dependent Reliability Analysis (User's Manual)", U. S. Nuclear Regulatory Commission, April, 19o4 (draft).
- 5. Manka=o, T., and U. Pulkkinen, " Dependent Failures of Diesel Generators", Nuclear Safety, Vol. 23, No. 1. 1962.
- 6. Erdmann, R. C., F. L. Leverens, and H. Kirch, "WAMCUT, A ^
computer Code for Fault Tree Evaluation", EPRI NP-bO3,* Jime7 1978.
- 7. Worrell, R. B., D. W. St:ick , and B. L. Hul=e, "Pri=e Implicants of Noncoherent Fault Trees", IEEE Transactions en Reliability, Vol. R-30, No. 2. 1981.
'. Porn, K., " Implementation and Use of FRANTIC for Time Dependent Unavailability Anslysis", Studsvik/SD-51/62, July, 1981.
- 9. Lofgren, E. V., and F. Varcolik, "Probabilistic Approaches to LCOs and Surveillance Requirements for Standby Safety Syste=s", NUREG/CR-3082.
- 10. H. Wyckoff, " Losses of Offsite Power at Nuclear Power Plants' , May, 1984, NSAC/EPRI.
- 11. Reactor Safety Study, WASH-1400.
- 12. Wiebull Failure Distribution Analysis and ? letting Syste=; . -
EPRI EM-3658-CCM, October, 1984.
- 13. Gaertner, J. P., J. H. Holderness, K. D. Kimball, NSAC/03,
" Brunswick Decay Heat Removal Probablistic Safety Study",
December, 1984
- 14. NRC Generic Letter 84-15, " Proposed Staff Actions to Improve and Maintain Diesel Generator Reliability", dated July 2, 1984.
s .
A-1 Appendix A , BRUNSWICK STEAM ELECRIC PLANT DIESEL GENERATOR DATA ANALYSIS z This data analysis is based on the CP&L in-house data sources and LERs for the . time period 1976 to 1983 When assessing failure parameters, only those events that are classired as true failures in the CP&L data bank for the diesel generator system are considered. When assessing equipment unavailability parameters, all the events that cause diesel generator downtime are considered. A.1 Unavailability During Standbv For periodically tested standby components, empirical data indicate that the unavailability during the standby period consists of a constant (time-independent) part q, and a time-depencent part At (where h is the standby failure rate). The constant part includes testing and maintenance errors, and failure modes related to start-up and change of state. The standby failure rate represents centributions fr0= latent faults that accu =ulate during stancby. They include effects frc= varicus environ = ental factors, fatigue cue . , , I to dynamic loads, and degradation of gaskets, hoses, etc. There are not =uch data published on standby failure rates. Manka=o (Reference A-1) has studied U. S. diesel generator operating experience. By using linear and non-linear unavailability cocels some valuable insights have been gained on how to realistically model standby co=ponents. Battle and Campbell (Reference A-2) also acknrwiedge the time-dependent and ti=e-independent unavailability contributions. They employ a very simple, but conservative, a;;rcach to ec=puting the rates fro = available operating experience data sources. Data on standby equipment operability orginate frc= observations mace during periodic testing. It is therefore difficult to determine whether a particular failure should be considered as ! having a time-dependent or ti=c-independent source. Further= ore, the two unavailability contributions need not originate frca } completely different failure modes. There are examples of maintenance errors where impurities are left in the equipment causing failures at a later time. Mankamo (Reference A-3) presents an overview of failure mechanisms and contributions to diesel generator unavailability. The impcrtance of the unavailability concept discussed above is that it aids in establishing realistic models for how the unavailability benaves as a function of the time spent in standby. By omitting the time-indee2ndent part. then it is implied that the component unavailability always is sero after successful testing. The
A-2 practical i= plication of this on technical specifications is that very short test intervals are advantageous. On the other hand, by omitting the ti=e-dependent part, then the practical i= plication is that surveillance testing does not provide us with any benefits. A.2 Weibull Shane Parameter for the Standby Failure Rate This section provides a brief discussion on some of the difficulties encountered when analyzing the time-dependent unavailability contributien using the available BSEP raw data--in this particular case restricted to the LERs only. The failure classification is subjective but aids in displaying the dependence between the subjectiveness (assumptions) and the chape factor. A total of 10 time-dependent failures are identified, see Table A-
- 1. This is to be ce= pared with the results presented in NUREG/CR-2989 (Reference A-2). The problem is now to determine the ti=e to failure. In addition, it is si= ply assumed that eacn time-depencent event is unique and each failure represent.s the culmination of a degradation process starting on day one of operational stancby. Our study period is 1976 to 1983, and the ti=e to failure is countec rec = January 1, 1976.
A Weibull plot is presented in Figure A-1, and it shows the da'ta ' analysis results. Note that the sa=ple size in this plot is 20, and remember that the shape parameter is a function of the sa=cle size. It has been assumed that we are observing 20 similar components. In this context it is assumed that the components have equal reliability i=portance values. There are four diesel generator sets, each consisting of several subsystr=s 6d can cause a " failure of emergency AC power". Assuming there are five i=portant subsystems in each set, then the total sa=ple size is i equal to four-times-five. Our Weibull plot includes only 8 of the 10 events; two events are considered as burn-in failures and have been discarded. There are two data clusters in the plot and if a straight line is fitted to all the data points, then the shape factor is equal to 1.5. Note that it has been assumed that the ti=e to failure follows a Weibull distribution. Ignoring the two data points centered around a ti=e to failure of approximately 25,000 hours, then the shape facter is equal to 2.5. For a shape factor of 1.5 the Weibull plot gives a scale para =eter - of 100,000 hours. This implies a mean-ti=c-to- failure (MTTF) of 2,022 hours. Assuming exponentiality, tnen this IETF-value corresponds to a standby failure rate of 5.0 x 10-4/ hour which is a factor of 5 higher than the value given in NUREG/CR-2909 (see Table 9.15.13). This sensitivity of this parameter has been assessed in section 5.2 of this report.
A-3 A.3 Test and Maintenance Unavailability Data , The data for BSEP show that few. maintenance activities are perfor=ed on the diesels while the reactor is shut down. There is a reason behind this; during power operation there are two non-emergency sources available for the onsite power system. On Unit shutdown one source--the normal station service transformer--is lost and the operability of the diesel generator system therefore continues to be essential. Tables A-3 through A-9 display the actual test and maintenance downti=es and restoration times. Distinction is made between scheduled and unscheduled caintenance. The decision rule applied to the raw data is that those maintenance activities directly triggered by a reactor scram or a test are classified as unscheduled. The - data show that on the average all maintenance activities have been completed well within the ti=e limits defined in the Tecnnical Specifications. However, there were some isolated cases where the outage time exceeded 72 hours. For these cases both BSEP units were shutdown at the time, or continued plant operation was allowed as a result of special NRC exemption. REFERENCES A-1. Mankamo, T. V., and U. Pulkkinen, " Dependent Failures of Diesel Generators", Nuclear Safetv, Vol. 23, 1982, pp. 32-40. A-2. Battle, R. E. , and D. J. Campbell, " Reliability of E=ergency AC Power Syste=s at Nuclear Power Plants", NUREG/CR-2989, July, 1963 A-3 Mankamo, T. V., optimising Test Intervals of Standby Diesel Generators", Reliability in Electrical and Electronic Cc=conents , and Systens, North-Holland Publishing Co=pany, 1962.
a! *3,L.% n., , , h ,. ,I, ,. un 4 .s . 1 1 a'sc. des dC nn n. j w. *.l'as.y Vtesn nr ., + a m n, v- .r. a 1 M 1112 r!
- s i;'Olikp!!i uiCDCi i:'li C- e' f. M b u t/ 1 G1 l ui C.e.s P,... e ,, .. e !. .' 1, n ag;a 13 g la .,y y o u , t s N. n-.,; " .m tj.1, n
u m. a, e . .2 g.gn rp un p *-3,.,- 3. . . .s n n- q. m. .r, n, . c .- n , s ,. n 4 .
.C.v C li b 41 0 11 :sc tu .i d '. i tAl c .av e:I b !n" 0 01 " .:.# y s n nv e. u t t i
. sess.
. .( - 7 . , , ,.. ,...4
. , . i. ,, -
.i. .:. . 1. ,,i a.r t fc uniin :
d Iticl U11 V, c m. l ,, lG n.
.i u usu i
- .0. .e us
.t. ~ . . . , . , , p n - ., m .
~
1iu aticcu .r u s t s.1 u s 9/1 9CA U. . ,,1 yn.u w sP-
.... 4., 1.....
u, w :: a :a. a,, n - s c. ,.:
~,u .
u,;11 u1%, 2 43 o - {.- 5, M =) a. w iee.w - j.5 =c= es . . , e.g
.a' ~uav
- f. 93
.-.7.-.
.- .s.e te. 4 -
g 1 .a. w. h. . i ,5
- 1.' Q ips .e I .p=-qe.g sa. se, ..
s y T..
,a . .,- ,, .-..=3 .a 1 ,-().
- y.s .w a. .er
.i.s pa m .1
=
.J.
e., -a. d. .,1 g,
, .s
. b~ r 2 C '
if N . i O ta, .:.vi**q .s _ . '4 p *. ! . . . . . .. n , 4j
, r.
tl "" ! 0 .i t.: S M11 M, a*C . b, e. t.*u: 8
-r2 - a I e, %
gI , 'I 1 Y C . 3 . D, U .i t."l 2 91 M r
/ ~ n#
;4 ,-. n
= .7 8d .N q= s j .e et as ye i a a 4 pf p 2, .i t p
_ . .i- _
..e 3 av ,P a.
O st r.: 2 t;C [V11 1111 C .11 uul!C lJ 10 b, hC' 9: f O d f2i.~1 b6 .s*
.% 4 3 m ; l' d. - - 1**L U D s b' m1 1/ C V . M, U l l a v M1' 1 1. O ill A, ttl"
. . . , .3 - . , w m p .2 ...4 1.11 A, to/I; .'-: .'} t'.A'O O W11' 0 V
,,. n g -.
.c :;a. ..gg .
0 .c g , . n,. . .
.. n , 4
- 4. .
.r - s
.i. g m 34 .i, .i
_rt n , i.,V C U U,, U, .Fr , .: U. .i. i.t .
,i,181 1 i V 1A1' 1 1 '.' . lii t/Y; J. - .
M q . .f g A,. in - , ,n ,.n>. . n ro e a <, < U - L.1 -n sCu.u v-- . U.,,L, :.- g ,, i t J Cil:21 bi b 1 15 t C
'l ih > .,
1,. v - t
,. . e,..,, ,-.y. ..s ,_. n . - . r .,nsn, ob h N$ 8 h #g n 451 h l I e [in e fterea.g4Ap
#...9 t M*t.
C;; Y 11'u si:5h7:8 , "e e e
- w. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ . _ _ . _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ __
g.4
. .g e. .,, , p. 4. q m.
s . n c .,1 /s - -. - , -..
- l l r. c .n _ a2 , . I x a. -93 .;.n -i .m il kl !.la il V //1 C O C '. M C f ! ~'i'G U U .:i. h!!28.iilC M M A'J'd. i/u 5!ilI b A sGC
% 2 . .
gg b'OE.70 IIIIS 1 O 180 03 [
- 3. u i :"i r* ( ,
j 0 9 8' d fli"' O i r- 1 QQt3 1 p q .: i409 1500 j qQ/! u s.nt .:. .r s u .t < i ...:u :. q. "i .- : a u r.: Liu- uu .:. j G .; .. .- u ~ t 90 ._ ii 3
;iv CA -9 di- --
94 a ui -. - .:. d u :"I u v .:. u .1. j .~..M . b a ? g .3.. .' ( i
*e
~*4 (J!
. .a .=
M (t 'El ./ $ Il .[ [J U I.) es , ST .
;) q <5 .,,. &*O 4< s ti". - -: d =.: -
sib u ln .:. : =.' b Y b . *
~ :' '
f[ #', ~$ Ib
....,"dt.;
c Y2
+ >< , 7 .. .. .%.. ,... .!. ;t i U till n G A'y) .. fih*;uu V l
4 .
16DLE 6-4, Diesel Generator Test and liaintenance linavailability, Scheduled and linscheduled. Ilnavailabilitu DGH 1976 1977 1978 1979 1900 1981 1982 1983 1984 1 3,1-3 5.7-2 5.0-3 0,1-3 2,9-2 1,2-2 9,7-3 1,7-2 9,1-3 l 2 2.3-3 1,7-2 2,3-4 9,7-3 1,9-2 9.7-3 0.2-3 2,3-2 9.6-3 l, 3 1,9-3 1,5-2 9,5-3 5.4-3 8.8-2 1,7-2 1,2-2 3.9-2 3.3-3 4 1,7-2 1,5-2 1,1-2 1,5-2 5,1-3 1.9-2 1,5-2 3,1-2 1,1-1 l
-a Exponentia: notation is given in abhpeviated forn; e.g., 3,1-3 : 3.1 x 10
e , (=
- a qp r g .
Q
".?
b M"I'JL 9'i
'4 (%%. %
2
$'$ t t .'% N% "% rp3" s d */ C !.T.'=1 t.
!59 %= 951 U"!4G{1Gi)s.lj.1)9 a pkas he* 4
- IlM2ln4l i'.LtA 3 7 .] 3i%4 1 g=% rf.ll. 'l h. h /
/ O --
a iiMA i. *i. a - 1 V .ad a .n ,. _- n s 4 . , = , P= % 5,1 51 *** t e ilf } V J f (Ii'd., M "n "" 5 i M;sO (j."b ' =l/ V l. f../.?,li.J..t s y if0 e
=.n. .c T.
-4
=.
- 7. u e
.3 mn , m 3
I r , f.' i% * *.1 ' ?1 %m 5 .=% ~
,4
- 5) . '.1 ' ' .~. ~ I.i tt 1
- y .1 .-
, 4 *ta s= .'= E ~J J .
O '~ e i. e. <
=J .'. 4 4.
s 1 es ,mee as 7 if' 5.3 I h IJ
.s -
n ~ d 956 W A a ts a l
~l ,
l
tri ,. , "':" " l* CJ C ris CJ Its c 5"'-: - e ns . -
'"":: ' (Ci ""r5 =! w 5 :ll" @
e l T * ::: I C - . c;E-* A
--e I CD P- Oc l -- 63 i rs:: c in c 2 -
= CN G3 O: LI":: I X A lll -@
"1"l5 ==-4 1. -CJ4 C T X
@ f .- 1/t ' CJ C M i Its dJ P-= '--:
C A A os N =: f.,J:: I/I C"> 4 r5 T C e--a C3 C CO T AC O: 0: v r::: c=
== CN -4 =--i T 04 Ac ^ X
-4 v 8/I E %Q --g
""C: AG^ (C C C A
- try cJ . .-4 r5 c Iy AAc Oc l ] ..-e O ==;,0ll 5 CJ
= c= i os o T -- -
- c- ui A Q CN -1 -! -i G3 "C:: st tt r T . . ifi
.--.e I ce c v _- c, : In
=* ""!: -e e--a 0 4 A A
":: = % A !3e% :: O CJ i ! --e --e dJ A ""'-
i W i c '""!"* ""Tll :":: 204C c i cac::: l l X c = : : cJ r - =: f.', r0 A CN 4 CC T-= It": CC C $ = s-* *v C3 -( I -- d - A 131 @
~ !* i I/i* ::n c; =1 A C CJ Eft -
==-e C C P 15l @ CG
% OJ j GJ rd & v--*
- 3:::::C 2 O 1 '"I" :*C a 3 C ca . A. , _
E--e i CC i O.- CC --i 84 C A:*l: cS - C CN -( OO 6:1 rd . - CJ..-4 CCC C * -( =: A C & e-* CJ O c its 1:::* c -* c: -! ::.:: c
*-* C A C % :: r5 C rd "" :::
-&' 2 ,G A c2 O*" "
fd * +=-e si'* > =. C 212 A* A E--* O-- } CO P= Oc l- & ^ c tyg A o 1:J# c CN W - .i --i 00 -i t#3. w A => ty: C n-4 .-4 A & C C at I/I CJ et ::; C C c'-""l4 CJ cu o c2 .a:: c.: A A A= = ce ::: e ca T: c ca - at CC CT C C C O* 3 AC P= !.r : ca --( o': -& v -e C C "'- CW CN ~ - - Oa O ~: (G c6 c ,, Lcs
-4 -&* i - C C -e C C c ru e C:: C -M C C 11 CJ cJ "Z" 20 --e S. e 1/3 . - . ,--e d -&- @ =: C
- -: e & % C C C Ifi :=:
c.: et C '"~ :" -- :='I C c -e e-e c III CJ P- It": Li- I O: Cc =* C.: cJ ..::: it rt it " Cll!'~5 . CN IJ":: O3 .-4 -i -C: A '3r3 2 E als %- C j -4 ty; tyi. -e c C; c
.--4 et 1 C i= O N '"C: CA4 C =. g*s-* C1.1 C.: rd rC: -
1/3 & :""; "O e--* e--* C C C IJ1 ; Ac , C: l A -* - G = ":: CJ t#i 11J
. - r1J <
gN- T ll - CT'ai CN ' st it rt" --* 75 " t: & o 4
;"::i E--* C's -4 63 ; = = = . ' CJ dJ CCO
--I I c = C c 4 .:':: .-4 ris
- rT: C := A c c rt -: A sc *""l -C: C: M O its Z w c I !
-.:=
.c;: - . ..___ __
C:' -4 00 0". T r::' A c " :: E=.3 , (=i u
~
e 1st
": CJ 5 .e 1 &-
me c h ,
- - TABLE A-7. Mean Tine to Repair 1)ata for Brunswick Diesel Generators, IInscheditled liaintenance Only.
t
- lean Tine to Repair a a a a L a a 1978 1979 1980 1981 1982 1983 1984 DGH 1976 1977 _ _ _ _ _ _ _ __
1 6 1 17 12 1 10 16 1 14 4 2 4 8 2 7 25 5 32 14 26 J 12 0 21 12 47
- . 4 1
l ilotes: a: The CP&L data for unscheditled naintenance incoinlete (thisno refers to the CP&L-prepared The Table given flTTRs 3 for above years are 1976-198l), base don
- DG identifications gitan.
LCO ,ata, Lils and latle 3. . L: Janttang - htigtist.
1 I i = i.n
! W %
- D- *:ll
..-e c CJ I~'t ftt 4 III 2 % lyn % Its
=m CJ % ;"'-= % % %
H S C .- a g c ;;;;;; l
, i - 4 o ::: o
= * + 4 4 r$ co -> l l CJ c -
O- tr :: - O: 2 C G3 O"3 OG --4 04 o
. g-me e- 4 I/I .N D c -++ %
- 'llI CJ C5 W %
C % c: Ifl % - . 5=I fr' ::'::' It! ::::: I cJ Irv - cJ 1.n cJ Q % Itr Ift % I/I %
- _
i o c ce cA als C 4e % o % O
.3 % %
- e-4 --* A rd I c.: tri . - c
- -e r::3 o ty --e *%
sTI c! ..=.4 C .T =M fit ECs C C C C r!J C "% CJ 1
% C c * --e C 4J A l
@ % cJ
- 4 CJ
- e--* Iti i
% 4
- C cJ C 4 * -*
5 l ::' ..-. "C: : :1% 5 -* fa"* C
.--* l #- l ICi E--* 't % 4 iCi EE.* * '
f& 9 P.Ti.e , l
=
% C i T . --i T C4 %" 4 T e4 :C; f.nr 7 l 1
4 o ' D ' I c:- uo --4 co c . T !J ;r C2 cc v e
- C:: ---4 --4 --i n. c --4 04 CD I Orl -- 3 G"I O '! C'*3 m N S
I i- ; i cc ::C cc I I I
.e= '~.,. I I I T Oc --I
("> T ft"> 1.t"I CO CO cc i f.=J G 04 04 0 *I I I I A 1 A ' co co co , -A ,--i s's i=3
*=
I.--+ l s .
M C
' s/t GJ
..e M
N
- f' cc =t in C .
CD C c> 11 U l e g cJ GJ
% c3 m b =
r :i -- CJ CC C C fri -e C l *CJ . 1 d { C 1 -C c*4tnC mO3H ccWWcoW . C l M C CJ :: C*'"> ---i -I G3J ' "J IN:T
- E *"-'* * !
C ,
- c::b i o,--
e -e 1 %c f9 CC s- . - = i u a 4 W a, A - C A C IZ % N I i I e A c C } . C ! l 5 C 0 ' t = CLs
-.e- -
~ -== L13
.e i/1 l m CJ l w = r--* l -
a .; o - I I _e
- 1 1
=;
m ' A me W co ch cc -iou co w I ' M W W W W cc cC C.:cc c CN rr icm ch ch: ch ch ch C C.I
-, - H O C# --i - s H H e T_.,e "I ~~
~
A l l e
- =: '
m C n; H - seen.s 4
W.e-O N* nn...
;s -
WEIBULL SLOPE / o.,,_,.,.
- /
/
.. /
.-i..
/
~
/
t.. /
/-
V. (7 BETA = 2 5 eg, , BETA = 1 5
- s i e
# l
/ l
; .1 -
/
. i t
/
- t
.n. . . r. I ^ /
a e
.: a e l',i o '
~ _
- h. _E0.0 ,.,
v , .i 3 C I I c r- e,, o 1f'.O ci _ 5 S . '.' - 40 _
, [> . C ~
R . '. ) - l I 1 I II i i f i r ij 1.0 I f I e l 10 19 5 j. Time to 1ailur'c, T FICURE A-1. Weibull l' lot for the Standly 1:ailure Date s , , , . j
\
B-l i l APPENDIX B - EMERGENCY AC PO' DER REDUCED SYSTEM FAULT TREES O O
#g o e
I I 1
.. - ,'a.
,, n. m.. . .. .-i. d e <i u de****
s ...!. i i f I a, ik :s % e L u a Li n .G.J t i u u ue:i i 1 i ! i p, . c. - i .
\ .::i't.u. a .r s, l a
l !
~
T A CS O t '. i' #il7#"" 7 Tild5I\
~ '
l ' I l 1.n if c/ V..t V i .A' il .+. .. L ' " = ' ' ' ' l I I i
! 7 [k ')('$@kj i t SI 'l 'i' f.b l .V M i.* E i . ' i l 1,' t il
! i, f ,g w h !s
,2 hh .,hy . ' . - ,
i , . . . . - , .eda' ~hs .... g d mEmma ehd e se sees tse N, $ $ " $ M d ,$ .5 e k M sd e"nsd 3 I I i *ssemass a - e enemmase -- .r t k.
r
?0'AER !!N4UAILABLE !
. f6R EX HEAT REM 0'J Ab
! ! G1
, , . .t 0FFSITE AC -
- F0'AER IS UN.4Nilt3LE (LOFF) v DIESEE i rurocrev
.-H-Lwe - %,;w.0'1 cx :41_L l'y? 't'
~
GI. F:"nIno MODn tu ~ G2 /4_
;. rur.e.cevieu
. 4. ati vs m.314u-r s i
! 4C P0'4ER Tall'JRE - _~4
. , nn db
- a i , n? ,glql.,.rs
- t, .e.rt l.
. m.; e:: r. *
,wt c .:. __
ChSI!E AC DG HUMN ? W.R 00MW:N ERROR C.;'JSE CONfRIB,. CONTRIE!!IION (DO:1EC) G3 Lj_ n,g,-[d59eN Nd. ,,
-I
. c4m. . ; cc..q1e. . t. o, , -
LJ
.< e,v. enr ..
r u. CC. mom C;uit <-
- s
- .4 Lpi e.r..
, .. e2 . t. m. .I :.u uc.3 (SkCCF) DO.
C,2MN 0090t.i C;nE
% e..m. r
.rts n.a,r
. l ud _
FAliiE, - ti4CCF) g[G v'V[l s
- e
M INDEPEiDU!I TAILDEES
! DG INDEfB!DDiI NC,-
F4ILURE
,-w
< ~,I G14 M 1 AND 2 DG 1 AND 4
^ i UMUAllASLE I :UMVAILABLE ! rm.
G7 i , t ! G7 i I DG NO 2 ' !DGNO.3 i } DG NO. 4 !
! DG NO.1
! UN4UAflAELE ' !!!N4U41l4 ELE i !UN4UAll4 ELE! ! Uh4UAll4PLE I GI$ /E ? Cig /4 G15 - ?> C17
! DG NO .1 L
! UN4UnllkBLE _! ~
~,
' GIS
'i'
- 1 ISW SYSIDI ! ! DC PO!!ER !
- UM'J AllAELE , iUwv4114FtE i
i C26 /R G22 'h
- r. ~~s -
30 NS. 1 F4llU?.E TO ST(51 GM DEM%D (D11'J0)
- DG N0. 2 ~~- E
' !!HU:Ila?lE
- 10
' S'A S'lS!EM >
! .DC F M 3 i U!?cea!LsELE UM'ntl:9'_E I I
G23 9
'C07 J! .
k.j DG N0, 2 . FAlluRE TO SI,SI ON Ce@D .
DG NO. 2 -I
- UMU AILABLE ! ---
c ,, C1' t
!SW SYSIEM i ' iDCF04ER !
' iUN404ILABLEI
' UN4 VAIL 4BLE I fn;1 '-
w _
.s 2 6
a q4. e
'J e w FAILU2E 20 S!4RI ON DE.9ND G03UD) l S
~
!UN4V0l[dflE 5
g*-
*18 i
t t l '
.SW SYSTEM : : 00 F0'C TUN 4 VAIL 4BLE1 l t UNWAIL98LE '
QQ. s53 ,- (a W ; p Qel
) e iA!!UEE !0 s!aRI ON DEMM Q3MD) 6 0 0
% 6
; M No,1 -'!!'
~
tSEEVICE!!0TER . I UitM4IL4BLE m
!G26
.R r"%
si s.s' UNIT 1 S4 UNIT 2 SW S'MIE4 SYSIEM UNAU41LABLE UN9V4lLs3tE (SW1'JDJ t SW.Us)
~~
i $$b). li 'Ea TE? p V4!!aSLE _'de 8 e .,. .
- k. / s..,i UNI! I 'i'J ::f!! ; IW SYSfCM tulEM UMV9119tf ? 91'>ai LAELE (Sil1DD) ($14!?)
~ . . . -,
;:l.)I u..J W- . ..,,I.e3 , 4 LN JU L1ELI y
,244 St5IDt s'.S TIi UM'JAILABLE G M I'.CE (ill10D) t SCD)
h 8 \)
' SEEUIEF t. 'les m l! Lg y ER l ~
n.
# 6 U e,e d
*% f*
UN!I 1 S'A . Utili 2 S'! ,
$'lSIDt !YSTE.M liMVAIRELr "NwalLAELE (S'410D) (Sk200) 4 8 % 9
i 1 [- .. . ( f ! r M NO,1 7h
~
D0 ?0.'3ERl:
,'UMVA!!4ELI r
- n. )
' G20 ;
1 i 1 i f n t & i w- v , t yN!T 1 M UNIT
- DC 1 I
OWE 3 POER fJNU14114BLE UMY4 LULE GC13) GC2d ) , ( 4 i I i
' ' 6: $ ? it
~
I OC 5 iL. ' UM' Jai LGLI ' , t i l , t 23 i i 8. (
. l(
PSEP, F0P:R ! UM'1Ai!.ABL' 'M Valt,E ! ; l- (DC10D) (0CW1 l T l l l
' j$ NO, 3 P I) l r
1LC POWIR f~ i m.'1,11!.2:fvue J I l . . . r i e ' ' ( e t p *% >= *s (
%? .. . .
( I UNii ; l0 unit 2 N ; 109t2 l NJtR ! l Din!L12LI um4JLCI
- .601u0) GCL) [
I , i i . i
1 t l i I e l l
- M P0. 4 -"'- ?
LCFM3 l 1 UNAU41t4ELI' t l
, e
..e
[ I
\ .s ......
ete. , ce ,npI +A *e ** e ,- . t T A A i r0JE3 19'Wi3 1
@U..a,L,;DL? , s4 ',.w,,,4n.sstuBLE 4 6 'we e r 5 ve ds el e /
l S s e
E l 4 e o f p l' l l I 1 1 I i , [ f e., 9% h M # I nb +
\n,ut rivs!.a.a 11e i n e; a. Nu nuai
- 3) >
} \n1I? .L4 .a .J i
h i r/we e,r m?revern t.i/ e n. , Luca uC ' r <! s si . .
.; 1 Ynbss I l
- e l
l. yl/) D / W tt f I Nii Y ?,5 t t( slV !, . J in s ? !. 2. O
/' h U ' '* / '
di fd J/ 41 i j it!I'fl / t ?
,1 ta Y,0?)M j T
- a' p Vla A V A' 51 i .4 4 - A $. t.u i
I
~z e
t J
% e
[' . P%ER LF!A9AII4 ELE TOR H HEAT RDt099.L !
.m
! ! G1 .
r e OTTSIIE AC .
<f POWERIS ' .
,,-~
t!?!avaltaELE (LOTP) v MESIL C'! P
' tEMISOKl.7..
oc < r d [?.IOR
%. t.{..nII '
G2 h
/
D!EF.CIN0'l ONSITE -i
- M PauER Dil'1RE ' ~
.n
.,r v:-r, a f.
FAl.'JFE I
^/,!?!!g'!NS l
,.. o
.. wu a . .
.- esitt e.e l M ... r.:'i
. X "IR c.90s 0;USE C't:f RIB. ,
F??.03 DON!?!?JI!;N
.r' G3 i .
(M3EC)
,f W[iI AC ~
?
NWERCgreti O U! E 0.1rRi!.'
,b e
/ ,e-
- s. .
t,ty; cit 1 -
, p w ,;;33- ,...
fijljst t.
- STAT 10101T (Acl) M c' rMi' C troit GL[ Elf
'A 3
- 2 ibllJ.h i Q.RCI) 9,/;u un )l gE
~ ~
.* 6 DG INDEPDDINI FAIUJRES
! MP sTN . .)" r7 'f<
24 s17L.d.u,61
, L'? t
._ FAIL.u I
44 , h [ *)&i M 1 AtiD 2
- e. im UAllqPLE; G7
- i t
- ,lBa a of rA
) . 9 m ua, lv)
<3 a ne s9 ya s!p), .)
a i ljN304 {[Q[l[ !' y a"N000][03(( Uf@UQ {[^((,[ ! J - . ! ,-
..A -*
L. 4 "d2 3 ' -}- . e . [ E
- w e e
- 4 g* m e, .
"Y l
,en J rv eJe*1.rw o0 .i I wa yb .cours. i i
. f riu.
t ',1 ) * * ? .
- T " f i g; s. , iaeys*
a .i4' " } s N 0 L w d F* *1 s N 1 a..i f 1.,1
' yw s
FNO8 . (1* *%O O r
*kD _
,.*., I.!5 ~ b-s.-
f
*'l'J
^ U.} li've A FAILURE TO
. SIAP.I ON DE$hD (D^1UD) an"' 9 r
g.,j/
.n
- ot A [,t 7,. ycl,. :
1 1 C a ') (v')s.'. Ts r--W, . p,* s ., O - A.m'irQ ., li s.* *, iw?i__
*. .
- f *
- Tr ? 7 eni ai
' ' .11.I M
.* Y f * ?:Lu? 7 i L.1.
1 ^)) g "d as a asa
's. ,,/
np o,) , ; j .J , T s PAL T T itJ.7 TA s '] t-TAty '.a s s9
/ J & !* in 4 ,\ tI h $
17%,Nb lf s d'alis.jii)
% s
r- -
- r
!DG NO. 3 m i UN4UAI!4BLE t
,-~,
, ,, -n*7 i
1 l
. Si.
nac,lcT a L.a. ,
,,w..-,<a
.v .
rUN4U41LABLEI SNeua!LaSLE ceo m , I, ,_ m-i:: .,: , t.
,__.s DC NO. 3 7A11EE io c?
21 dant
.~.1 , .s..l.
7:TMANis ' b w. , L . .* ifQ e 9 k s
.. 'DG N0, 1 7C
; SEMICE 1:ATEP. I -
! UN4UAILABLE l
l
~
.m.
N g
, y
_./ %_.* UNIT 1 SW UNIT 2 SW SYSTIM SYSTEM UNAURILAELI !!N9?AILAELE (SWi!D) (SW2UD) , y .: y ~) -- ': i ~g tij i[ '" . Al l"O
.g:14I'riec u . ,, 4.%-" c
~.,
T 's 6 O
#e e*ep
%. **e,ed
!!*J{ P
- s'i i 11 { } S rii YS b ib(( }
"9FJAIIAELE -
!!NAV.4ILADT P renuwele:rs
.-JHlu w n.v/
~ ' '
p pn 3 h30!EU4I[3
; UN40aiLELI
^
i $ua {n e*e p e**q
% ,$ 'e, aw' UN!! 1 'W UNii 2 SW en Jida er r.a i r.aei co tsr.u i:
UbADbIbAE[5 S.CUS[bi b .,5 (SW1uD) 9:(205) n ..
d e . i DC F0MEP. 9.9.w.
. . n puqu. e. ,
- Vuw
/??
w
/* s /* ,
\ . s b e,n.d .%,.#
li.N,II .1 uv Ll3'11 uo ww ne
.a i-m me erg T Vii" M IVril.h U"!UlallAELE !!NA'JAILAELE T^ tly/
(uka? /T Mil *p uk suvaunt
- t. u's nn M:
i . .) . 2s ~*2..- i r gf- M AI!") hh$ kELE
~
l'S S . . * .
%w f..,_ , , . . ,
=e _a l8MIY 1,'
I OP'
*i'i{ T 7 EP **
.n- = a . . A r,-
;'. a t.h : .' 4 L 5 NN'D$I[Abb5 ', Y,T)S [ b'-i[b5 mi . 5' (s J ua t = >dl ~ .i r.
~T
^ $(,s1 'e'hNi' .4
. s!% TV?LG
** ST e t,!*M
' ; vnaw ib i J a D; ?.
.-~.
.sa 4
. I2M p~~~s. f~~ *
%.. ~
egr= f na we 11 a ,v qtIl79l 9 n/* a yu
)[t . .-i .I.( .). s%^[it(D.
i!.):3 6;.7 [ ? .i1.!.?.?. 0 tti* .'* M u l * *i e-d.. f f
- 04 ?.T.
IT TL! *f v4!!'i's w< f. N'7 wuund
??9 I *
.* .s 9 e ag I. > l }
! . *P 3 O .i? aIIY 6UM // d a'iOJiO k.4.ii.Je/ il *aI.*/ did I
i 74 + e's \M. {' # ,4-
-- (1 s
J'f d t t l
- i E '.$ h,A M k'$ e i
1 I' I I
.' ik e . ',.-1 M f. f til',*1 **1.EgP*'d %
3.*E,t, *
-* L,a,1 . . J. o,
=
.I !J B .< * , e ,4, . ', ,!-
M E * .a 're *
=
m es
/,/ f. .J.
- de 4 il e,n as mIIe e s"of. dE'.2 mad a l' t
E -i l a 5s' f $i ! I a d$
$ t 'b3 1# # 1 "! I .f -tia- .a y =3 l
V .u 'd .1, s i f m i f A w" i i e u 1 a ,p e,= v, i. e
. n%an s I &
ea .u + s =E rei y,t** d i
.d Y b.id I d d '. VI Uni 1.3 8 'd .". .L m = ~ n n r g1, s r ; y1
.f " .'? .
m
<L f ,i* ; Q T- .3-1
^t
- a i t 4'm Ns d.. f 'J
- P...p,t. g /1
, .j_.
o . d d,t g .. s d 5 em o um em as d44e4 cae mem b sma d e cae J an . 4.M$ - - ' " -" - - - M_ _- _ e
iACPOWERUNAVAILAELEI iICR LOCA IN EIIHER UNIT [-w'i
, G1 t
l i ??WER INAVAILAELE '
' i ?0!!ER !!MAVAILAELE i
' FOR LOCA IN inR LOCA IN !
i
! ijtilf 1 UN!! i n ,,
' G51 G52
,- s ,~,
I AC POWER !!NAVAILABLE j lAC POWER UNAVAILABLE ! ' '
if0R LOCA UR RX HEAI . LOCA1 LOCA2 FOR LOCR UR EX HEAI ! G53 i REMOV4L IN ONE !!!i!I : G53 i REMOURI, itt OsE Uti[I I A ,k L "-- 2 e
- g
- 6
TAC P0'ER A !]!cVAllABLE F 4 ---
~
-' iFOR10CAcRRYHEAT ED:0'JAL IN ONE UNII !
3 m! G53
- p. 1!
\__.
s < OFTSITEAC rw,i
- v. .a. .: .-
litiAVAILEE i (LOFP) ! i 4 DIESEL
! EMERTiCY ONSITE I GENERAIOR
' oC PM'IR FAILURES i MODEL r#
J4 en
- I*2'*d 6
ENEk(W
.% (OUL.3 2-ily
() f.I9: N, -f na ' _.dn:.p
-h f* *s .ssa .
$4 e="*,
i . , 4 e v.e ud 4t 11_ ento. s i ;
; FAILUPE i 6
. b'h.M c tl4 P D.4 TJa Dit v.f T T.\AU '.
f4 4 i* i
~J a i ONSIII AC !
e t? Uf!M..b.t
.r:
I r a nIIr.s
)[r eva s dtm I h_ N 01)M.IE*1 u.
n_.m .n, n, , re.!t. .. tw..it .o.La., C0!i!RI?JI!ON . G2EC) G3 LO_ f -I4 )[ i b C MON k . - cvi+-!.m.
. .e, c.V>Jp2,i i . n. i.
I
!"~~s pn ud
,,~~c s
1> ,4^ SW SYSIEN ' ' '
~~~'
C."' ' ""'On,c"'
' "#* "= rp ...e.i g
_#' p
-1 Ey'7g,i E v,g2.M$Ib.U,
[ _A1 _3 Hr (s$W ':,-f s war/
?,.F O v-o i WS.ij i<
(',v':n e .ui-t"?,.4 ('.*i s d a .f. Cahl EnlluRE
.m. . . m.. . ..r.
. m-ec.-)
u ct o v
% .j?u??
S $ wwa / f 4
V O ^F e DG INDI? INDENT FAILUEES iDGIGEPENDENTN&, ~~
~
Fell 0RE i i -. ,~. u~e C14 1 4 - i g. l'
- i. i
=
!. i.
!hkAhAffkBfE hh4hAffkEfE hkAhAf kBfE f f hhAhAffkBfE' G51 /L G52 /F4 G53/h G54 /!L r.c no a ., . ., :,r iUNAUAILABLE! -
i i
, ,- m . . . . . -
! I G15
- i
'M S"SIEM ! : DC P0',iER !
i i !!h4VallAELr - UNAUAIL4BLE'
.s ! x i;26 _/ R- f ~,, G23 _/ 4
.r a 1
DG NO, 1 FAILURE IO SIARI ON DEMAND (DGiUD) i-l e A 6 ?A #1 -ts ' !I$nUnib8LE I t I l
, ~m o '. e e
(=4 ? JAG
,=~9
'1 h 0 *
- at -, . : .%u bOI 8 i > '=*Oh 9 h. .r s h f. .N 1 .$ el ,
- - h I w--
~
8 t G27 .S\ ,e ~3 G23lh_ s ! P /5 er $2 Ed' e u*) 1.'s
! FA!Li!?.E IO SIAR! 03
- DE"M4D '
, (DG200)
=, ,
- --. -, -,--,-,..-,.---,._,--,.,,_,-r,-,--,..,,----- , ,r.,, y,,r,, , , , _ . . , . - , - , - . , , ~ve-
l M NO. 3
!!N4UAILABLE l FS-i i
rm avn
' IJAI
. 4
! e.11'tva.1a eva*r.u. : n.e w a ne'urp v f1**s
! !_!N.61J 3 [ -T -g Qi T.
' i ei .,tj A [ f.^.27 IINo . .me Y ! ,
<*' ^
,94 ; D '.
^00
,J s.V
.54 -
e t b e,e d'-/ e d _' J I'. IIllOT I.f)
. . a +s u
# I -- -
ars-um 5 tk Gu*uo)
.' ~J gr).
s 'r - : s M .*r k-I!MU41L4BLE - g* *%a r i j
, o I e t .. .Iw .. ..
- I!b^!)h
.M f I.""J u~
. !NE N,Nk fIk f I^k e
% i-I J
Ed [- t e,, - e.e* J alm b E5.5.:_ ~?.:) SIARI 03 IIMND (DG4UD) . .
. 3 ,
- - 9 . . . _ .
y-.. q s
- I
's -
- AND 2
! t!M4V4ILABLE !
I r% i
!G51 w
f i
$$ II).1 i D$ II), 2 !
. Ut!4 VAIL 9 ELE !
! UtiAYAILABLE !
ct e, . . clb' o I .-J
. d'J. yt ,a ,iv . n4 i r
' ???d A 5 U-lA00I[.i$l5
~ _ _
7
.so tj'rg S * !6 2 R.
; J.A 3.J, J .- 3 kl } e
;. 'I atS EI A w. p
*II l% {WII"I*I1_.i7a.'a[s e l'd.'
u 'V . ' N & I{ 3*s, a a.s. $.7 [I A e 3.1 1
-J 1 -h' JAW a t
n [;3"Jov iib . I 'O A - 43 4
.nND,1^f'I r J.P.;..*1 - r iot A UT_7
,e".,
8 O uea . . Y I ' e , 3 e,!w: ,'.-ei_ngic
. e sr m-u.m,9 ai <$1 s1a..t.:.
J . a t , I
d - +d - _ .. w _ ww-_ -
- w. _, w S
e + 4 h d
?$
IVJa k '% = 2 ~.2 *~ i $)ji 9 l i f[Nkt$1LAE.LEI W "4
** I h l
cJ v.,. a a i l[Nn u At
=j ntii f ',H u!-T A
auC s',s
. . .M4s'ild jl)8 a eib i un T usuJ 6pi y .- - .
r
%47 sa g 623 3trA f
_? ____ 1 4 e
+
i s o
-~--v-- , - < , - - . - , - , , , , - - , .,_._, . _ . , ._.,_,,, , , , _ _ , , , _ , _ , _ , _ , _ _ _ _ _ , _ , _ _ , _ , _ _ ,__ _ _ , , , _ _ _ _ _ , _ _ _ , __ _ _ _ _
- Ir3 M.1 r,~.e nriv,. ira..
m
"/3- o
- UMVAILABl.E
..m .
1: i C)? - t
,s i 7-~,. ~ _s % .)
UM!I 1 DC UNIT 2 DO POWER POWER dNAVAILAELE UtiAUAILABLE (D01UD) (DCEUD) A i ne yo 9- r-- i.d'.t r,0k..R r. ! =9
; tN. 0.!.!f.i[Y cui.r i e***e c., ,
R.4
\* %j !!>yI.I,t nr=0L. - -
r a re. p "- 9 t-r UN..c!.? .
-w .
m..,i.1v.11.r
. A.p; .r. .s e in.tAiJrillA: pr v.
(DC10D) (DC20D) m .% m -s
.d "P?M3 I-
; TJ#.'ti"! s.n, hut.I at a ;
3;i'vn l l
. ~ , a p ~ ,.
be b._,# ,e GMf!. .! "P 64 sv f!UTT. wnA u9 ow he NA,.'[
, 2 u f"f,y n
'.C.h.
b I' I'p;.'ilt
;]()^ *{ I- $~~$ ? ?
ysv;., ..s -I!*2Aff (g* JK . gs*4 ' "-{g. EI r en..w.. i 1 l i I l l D
46 Y i ' i
. r' .ps.,tR. 2 v.C. w4l
.,v.
r.
,m_m_
t
. %d
~~
f~%
'%_~ -
P,^ ? h..v-a T. .t
!!PT. ,o U_
*. N.1 *1
^!
"Urk{E, T . A4 { Qp v n.,,
L
.. . in GCitiM GC2'M J e
O b g
o
.. t DC N0. 1 -I:
i SEEUICE WATER I-UNAUAILABLE 3
! G26
,a ,
v s..-
"T 1 SW !!NII O SW 6?EM SYSIEM UNAVAILASLE UtGVAILABLE (SW1UD) (SW20D) b,,J 13
" ,,3s e [4,
.s SEEVICE WAIE3
!EinVAILAELE i
.s
_ '2Y i
*s,,: '.,_'
UNIT 1 SW UNII 2 SW S?SIEM S'!SIEM UNAVAILABLE UNAVAILABLE (SW123) (SW200)
, s r 4 --
tHJ
! StadlC .e' J,.. . -Tu.,
I I .*- 0 5 I }l*8]N m i ,.. v A a LA.7 == 7
,.=,
OOD hJ
,o*o, ,.*o,
"/
*. 5..-
voerrw
& I 4 A a.s t eve::a J A J s s.. s i i i a f c. ,Vi v al .. . ,u ,.1 ,,
,,araus o 1. - )
} %
!)$ IN. 4 -
acre.iibt; m V - n Nsio r.s.h l. a Ht*^H A TL' <, _E! 7 uni.1vr.1 n ~ _.a,
, , =, _ e 6 6 _@' O
.F 3
~
%. ,.A * ,pd I $
t ill i . w r. i
.i.L. . wd o
** V. b' lul ? TW C3/C?fM J.;Aui
..N(iO A n f Yn.au ANI . I'*sljh.
- s .I .I.O D$
nusu , is!!10D) (SR23D) t O
7
., . 3.
S m k *
.g at -
g .w' .} _. e ,4 3 a n s J - I' i "t./
-r4 we' v4 e r g a na$ asA . s'shJs sind z.m s mN
. n ag- -
'L 5-
- t f h. "
3w( 4. - 444 E'1m~p
? j'9pk a ,,p -
.r'-- PT r-
'"1** M; ai t.1.
- W"-'. a 4 4
.m J As 15 9. 8 , a a -a a i ~a se L.J g-. -:? .. *ar a '-
, - . ~. -- -
Ek *c')- -- k I l
! 4 .= !. - p (t ) .-, 1 2', s
, =+ r ,t ,
, . . i. 4 . a. a.
. i:
[
/
y d M j. J $ t -
" [4Ifg -5 f 4,g .. e . . '
m.4 .g ._a a b
.$ I ""1,
+*
,4 % *4 4 1 ' ' ~
F i i. .- J" u.da,7% s. E-tis
.s > s n
'te s &
s As 9 m .
- a.
g g ws-airigw. - -- - - -~ - - awe's ama hwakte n. .at rtT=='*- g tw *s w_ :-t - :&_.-h~ , . h
.}}